Research ArticleCalculating the Contribution Rate of Intelligent TransportationSystem in Improving Urban Traffic Smooth Based on AdvancedDID Model

Ming-wei Li1 Jun Yun1 and Na Liu2

1Management School Wuhan University of Technology Wuhan 430070 China2The Ministry of Transport Research Institute Beijing 100029 China

Correspondence should be addressed to Ming-wei Li 467827811qqcom

Received 23 May 2015 Revised 21 August 2015 Accepted 25 August 2015

Academic Editor Julien Bruchon

Copyright copy 2015 Ming-wei Li et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Recent years have witnessed the rapid development of intelligent transportation system around the world which helps to relieveurban traffic congestion problems For instance manymega-cities in China have devoted a large amount of money and resources tothe development of intelligent transportation systemThis poses an intriguing and important issue how tomeasure and quantify thecontribution of intelligent transportation system to the urban city which is still a puzzleThis paper proposes amatching difference-in-difference model to calculate the contribution rate of intelligent transportation system on traffic smoothness Within the modelthemain effect indicators of traffic smoothness are first identified and then the evaluation index system is built and finally the ideasof the matching pool are introducedThe proposed model is illustrated in Guangzhou China (capital city of Guangdong province)The results show that introduction of ITS contributes 925 to the improvement of traffic smooth in Guangzhou Also the researchexplains the working mechanism of how ITS improves urban traffic smooth Eventually some strategy recommendations are putforward to improve urban traffic smooth

1 Introduction

In August 2009 Premier Wen Jiabao clearly indicated theimportance of establishing the ldquoExperience Chinardquo centerand seizing the initiative of the new round global networktechnology development Since then a new wave of Internetof Things (IOT) researching is rising in China NowadaysChina is on the rapid development and construction of theInternet ofThings which has brought unprecedented oppor-tunities for themanagement and service of the transportationindustry In 2011 the Ministry of Transport focused onpromoting the development of transportation through IOTon the basis of the present development industry situationDuring the research aimed at guaranteeing the better devel-opment of transport industry Guangzhou was chosen as thedemonstration area and a ldquocity intelligent transportationapplication demonstration project based on IOTrdquo was orga-nized Moreover as a new IOT product technology Intelli-gence Transportation System (ITS) was applied in the traf-fic industry Compared with the traditional transportation

projects the system mentioned above shows dynamic novelcomplex extensive characteristics and so forth In recentyears ITS has been constructed and applied in China while atthe same time it hasmade certain achievements on improvingthe urban trafficoperation and service so as tomake the trafficsmooth safe green and convenient

However urban intelligent traffic is a complex integratedsystem whose development and construction are affected bymany factors such as the urban traffic environment popu-lation structure geographical location and economic devel-opment level With the rapid development of urban intelli-gent transportation the transportation situation has gainedcertain improvement However considerable attention hasbeen currently paid to the evaluation index system andmethods regarding the application effects of ITS Besidesthere are relatively rare researches on the application effectsin the micro level of ITS For example researches about thecontribution of ITS in proving urban traffic environment andtravel quality traffic smoothness and security are relativelyblank Moreover European Union issued the Guidebook for

Hindawi Publishing CorporationMathematical Problems in EngineeringVolume 2015 Article ID 564230 12 pageshttpdxdoiorg1011552015564230

2 Mathematical Problems in Engineering

Assessment of Transport Telemetric Applications UpdatedVersion in September 1998 which systematically expoundedthe evaluation thought of ITS project in EU Besides it hadbeen used to evaluate the ITS project in London Parisand other cities However the evaluation is just related tothe development and construction effects rather than theinvolvement with the application effects In 2000 the UnitedStates issued ldquoNational ITS Architecturerdquo (version 30) andcomprehensively expounded the evaluation contents andmethods of ITS The evaluation content stipulated includesevaluation design performance cost and profitability analy-sis risk analysis and evaluation results However the micro-level research of ITS acting on urban traffic is not involved

Although in the later research Underwood and Gehring[1] evaluated the ITS application effects from social individ-ual and enterprise benefits the research content and resultare still in themacro level Furthermore Novak andMcDnald[2] studied the potential macroeconomic effects of ITS in theUnited States in which direct employment economic multi-plier productivity technology diffusion and competitivenessare included Moreover Lee [3 4] Zavergiu [5] Brand [6]and Beimborn et al [7] established user efficiency analysismodels cost benefit analysis related models and some otherrelevant models so as to research the economic value of ITSprojects However these researches are more from the per-spective of ITS economic value instead of the improvementof traffic environment besides their evaluation indicatorsare different Then Xiao [8] who used the analytic hierarchyprocess and fuzzy mathematics theory to evaluate the effectsof ITS in improving urban comprehensive transportationsystem once argued that the effects of ITS in improving trafficenvironment cannot be ignored In addition Zhao et al [9]researched from public transport facilities service benefitand social responsibility four aspects in total so as to establishthe comprehensive evaluation index system of urban publictraffic system Additionally Chen and Liao [10] evaluated ITSapplication effects from six aspects namely economic effectsocial effect resource consumption environmental impactand development ability In all these researches ITS applica-tion effects have been analyzed to a certain extent Howevermore researches are still focusing on the economic and overallsocial effects Besides there has not yet been in-depth analysison the specific effects of ITS to urban traffic Until 2011Cao et al [11] defined and quantified the evaluation indexsystem of ITS application effects from five aspects comfortsafety convenience timeliness and coordination which hasachieved a certain innovation and laid the foundation for thefurther studyHowever it appeared that there have been somedeficiencies with the evaluation method The limitations ofthe methods led to the complexity of the evaluation processand affected the authenticity of the resultsTherefore in orderto ensure the authenticity of the evaluation results more suit-able innovationmethods should be developed and employed

However during this period certain results in terms ofthe researches of evaluation methods have been obtainedFor instance Leviakangas and Lahesmaa [12] discussed theapplication of AHP to analyze the cost benefits of ITS It is toosubjective to determine the weights of the efficiency indica-tors and this might widen the gap between evaluation results

and actual situation Moreover Mattingly [13] constructed amodel to show the value of ITS project in the monetary formwhich was integrated through the analytic hierarchy processand multiple attribute value function method After that themodel was used to evaluate ITS effects of Anaheim in Califor-nia thus confirming the feasibility of themodel However themodel cannot evaluate the application effects of the micro-level object effectively Wang and Walton [14] discussed thegrey correlation analysis (GRA) in evaluating ITS projectZhang et al [15] combined analytic hierarchy process withgrey correlation method to evaluate the effects of the sub-systems of ITS Nevertheless all these researches still cannotavoid the influence by the model characteristics namely thesubjective evaluation results are stronger which can seriouslyaffect the veracity of the practical evaluation results

In order to get a better evaluation model some scholarsproposed DID model due to the mature theoretical basis andpractical verification of the model Moreover DID is moresuitable for the evaluation of the benefits of new policy andthe results are alsomore robust For exampleWooldridge [16]established correspondingDIDmodels according to differentdata types while Ai and Norton [17] and Puhani [18] studiedthe parameter estimation computing method of generalizedlinear equation in DID model Moreover Drykker [19] usedWooldridgersquos test to verify the correlation between randomerrors and then corrected the model by generalized leastsquares method By studying the minimum wage system inBritain from 1991 to 2000 Galiani et al [20] evaluated theeffects of minimum wage system on employment BesidesConti [21] calculated the contribution rate of euro in improv-ing per capita GDP and labor productivity level by DIDmodel Lewis and Reiley [22] researched the contribution ofoffline sales products advertising for profit change In addi-tion Liang and Zhou [23] Wu [24] Wang [25] and Dengand Wang [26] innovatively selected DID model to analyzethe contribution of soil and water conservation engineer-ing new agriculture insurance policy and foreign strategicinvestment and so forth However most researches focus onthe fields of medical equity and economic reform whichcould be inflicted with great short-time changes and easinessto control group However there are rare researches in thetransportation field while at the same time the applicationsare lacking innovation Furthermore researches on effects ofITS in improving urban traffic congestion are relatively blankWith more difficulties in choosing the control group therewould be more limitation in terms of the application of DIDBesides the complexity of ITS also determines the difficultiesin selecting control groupThus it is necessary to improve theDID model and the innovative introduction of MDIDmodeis very important

Despite the achievements acquired by the researches inthe traffic smooth field [27ndash31] more researches are focusingon the cause and influence factors of traffic congestion orother fields Besides there are relatively scarce quantitativeresearches between transportation operation service andITS Although some scholars have studied the relationshipbetween traffic smooth and ITS there are still limited qual-itative researches All of these may lead to the fact thatspecific functions of ITS in improving traffic smooth cannot

Mathematical Problems in Engineering 3

be expressed clearly Therefore strengthening quantitativeresearch of ITS in improving urban traffic congestion isparticularly important and necessary [32]

In order to acquire more scientific ITS investment andtechnology-integrated application with higher efficiency thispaper takes Guangzhou (capital city of Guangdong provinceChina) as an example (Guangzhou city is the ITS demonstra-tion city in China) based on comprehensive analysis Afterthat urban traffic smooth is chosen as the research objectBased on the literature analysis from domestic and foreignresearch results characteristics of ITS and its influencefactors have been refined Then relevant analysis has beenconducted on the function of ITS to urban traffic smooth Inaddition a set of evaluation index system has been built soas to comprehensively objectively and accurately reflect thebenefits of ITS in improving urban traffic smooth Addi-tionally the indexes related to urban traffic smooth areextracted firstly then the indexeswhich are greatly influencedby ITS are stripped out and finally the influencing factorsof these indexes are figured out Finally with advancedDID model (MDID matching difference-in-difference) andother related econometric models the contribution rate iscalculatedThrough the research the contribution rate of ITSin improving Guangzhoursquos traffic smooth has been acquiredand the mechanism of ITS in perfecting urban traffic smoothhas been proposed

2 Model Formulation

21 Introduction to DID Model

211 DID Model General DID estimation model is

119910119894= 120573 + 120572

119894119889119894+ 119906119894 (1)

In the formula 119889119894= 1 means experimental object and

119889119894= 0 means control object Average treatment effect on

the treated ATTmeasures the average return of experimentalgroup and control group

120572ATT

= 119864 (120572119894| 119889119894= 1) (2)

Investigating dynamic change and introducing time vari-able to the estimation model

119910119894119905= 120573 + 120572

119894119889119894119905+ 119906119894119905 (3)

The virtual variable 119905 = 1199050is time section before policy

implemented and 119905 = 1199051is the time section after policy

implemented Moreover policy implementation effect is onlyacting on the experimental group Observation time meets1199050lt 119896 lt 119905

1 and the formula is acquired as follows

119864 [119910119894119905| 119889119894 119905]

=

120573 + 119864 [120572 | 119889119894= 1] + 120576

119905if 119889119894= 1 119905 = 119905

1

120573 + 120576119905

others

(4)

Control group

Experimental group

Experimental group

Historicaldata

Control group

Currentt = 0

t = 0

t = 1

t = 1

dataΔy1

Δy2

Δy = Δy2 minus Δy1

Figure 1 Application ideas of DID model

Here the value of ATT is shown

120572ATT

= 119864 [120572 | 119889119894= 1]

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

minus 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

(5)

Because 119864[ATTDID] = 120572ATT the estimator of DID is

ATTDID = [1199101

1199051

minus 1199101

1199050

] minus [1199100

1199051

minus 1199100

1199050

] (6)

DID is the unbiased estimators of ATT The estimatorcan not only measure the effects before and after the policyimplemented but alsomeasure the policy differences betweenexperimental group and control group After that the policyimplementation effects are obtained

Here setting the area that has built the ITS as ldquoexperi-mental grouprdquo the area without ITS is developed as ldquocontrolgrouprdquo DIDmodel can effectively control the effects of otherpossible factors Besides it can not only distinguish the indi-vidual difference in experimental group after using ITS butalso separate time differences of different areasThen it comesto the identification of the application effect of ITS (Figure 1)

From the perspective of the DID model research andapplication the selection of control group mainly relies onsubjective judgment and lack of objective and quantitativeevaluation standard When control group is a single individ-ual the results will depend on the individual data to a greatextent and result in poor robustness model Therefore it isnecessary to improve the DID model

22 Introduction to Matching Algorithm Due to the strongsubjectivity of DID model in selecting control group theauthenticity of evaluation results will be affected Thusmatching algorithm is introduced in the paper to solvethe problem In terms of matching algorithm adopting anapproximation experimental algorithm is implemented todistinguish experimental group and control group effectivelyMatching algorithm believes that individuals share the samecharacteristics due towhich they also have the same responseto the same policy variables That is to say the decision-making behaviors of experimental group and control groupare not affected by unobservable variables Moreover match-ing algorithm can enlarge the select range of control groupwhich can avoid the subjective randomness in selectingcontrol group and eliminate the estimation deviation caused

4 Mathematical Problems in Engineering

by nonmutual support domain and nonidentical distribution[33]

Matching model assumes that the selection biases areentirely determined by observable variables and it also rulesout the selection biases caused by unobservable factorsIt means that when the control variables are determinedwhether affected by policy or not they will have nothing todowith the value of index119884 Call it conditional independenceassumption (CIA) and mathematical formula is

119864 (119884 | 119883 119889 = 1) = 119864 (119884 | 119883 119889 = 0) 119884 perp 119889 | 119883 (7)

The CIA assumption in matching model would limitsome variables selected into control group but the restrictiondoes not cause interference to the precision of results It is forthe reason thatDIDmodel allows the existence of observationfactors

Here setting120572ATT0

is theATTof control variable and120572ATT1

is the ATT of experimental group When CIA assumption issatisfied formulas are acquired as

120572ATT0

= 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

= 0

(8)

That is to say the change of control group is 0 before andafter the policy implemented Therefore 120572ATT is

120572ATT

= 120572ATT1

minus 120572ATT0

= 120572ATT1

minus 0

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

(9)

However there will bemultidimensional problems due tothe increasing matching strategy with the rise of dimension119883 In the early method covariate (CV) is used for thecalculation which would result in ldquodimension disaster (whenthe dimension increased the volume of space will increasequickly which lead to the fact that the available data willbecome very sparse)rdquo and the calculation amount is too largeIn order to solve the multidimensional problem Rosenbaumand Rubin [33] put forward the propensity score (PS) asa covariate function If CV meets CIA hypothesis then PScould also meet the requirement Besides the dimensionreduction can be completed successfully by transferring theinformation contained in CV to PS Therefore PS(119883) isselected to replace the CVrsquos119883 and formula119884 perp 119889 | 119883 rArr 119884 perp

119889 | 119901119904(119883) is employed to achieve the dimension reduction

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883)) = 119864 (119863 = 1 | 119884 (0)

119884 (1) PS (119883))

= 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) PS (119883) 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) 119883) |

119884 (0) 119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [PS (119883) | 119884 (0) 119884 (1) PS (119883)]

= PS (119883)

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883))

= Pr (119863 = 1 | PS (119883)) = PS (119883)

(10)

minus4 minus3 minus2 minus1 0 1 2 3 40

00501

01502

02503

03504

Kdensity treatmentKdensity control

Figure 2 Common support domains

When dealing with PS Heckman believed that Logist(Logist model uses cumulative Logist distribution functionto describe the relationship between selection probabilityand explanatory variables and limited the value of explainedvariables ranging from 0 to 1 119901

119894= 119864(119884 = 1 | 119883

119894) =

1198901205730+1205731119883119894(1+119890

1205730+1205731119883119894) the odds ratio selection or not selection

is 119901119894(1 minus 119901

119894) taking logarithm of the odds ratio to obtain the

measurementmodel 119871119894= ln(119901

119894(1minus119901

119894)) = 120573

0+1205731119883119894+120576119894) has

more advantages than Probit (Probit model uses the standardnormal distribution function to describe the relationshipbetween the selection probabilities and the explanatory vari-ables and limited the value of explained variables rangingfrom 0 to 1 119901(119884

119894= 1) = Φ(119911

119894) = int119911119894

minusinfin(1radic2120587) exp(minus1199042

1198942)119889119904

then it implemented simple inverse function transformationand got the following 119911

119894= Φminus1(119901119894) = 120573

0+ 1205731119883119894+ 120576119894)

Therefore the estimation values of matching can be obtainedwith the following formula

ATTMatching =1

1198731

1198731

sum

119894=1

(1199101119894minus

1198730

sum

119895=1

1199081198941198951199100119895) (11)

1198730and 119873

1refer to the total number of individuals

119908119894119895means the differences between experimental group and

control group and it meets 119908119894119895isin [0 1] sum1198730

119895=1119908119894119895= 1

Meanwhile in order to select optimal control group com-mon support hypothesis is introduced here as 0 lt 119875(119889 = 1 |119883) lt 1 The cross region between experimental group andcontrol group is shown in Figure 2

23 Induction to MDID Model MDID model aims to findoptimal control group on the basis of DID model throughbringing inmatchingmodule andmatching algorithmThenvirtual processing individuals have been developed before theintervention implemented Last the virtual processing indi-vidual comes to the implementation of the double differencebased on the matching selection of control group

MDIDmodel has the following advantages when it comesto calculating the contribution rate (1) MDID could solvethe human factors generated by DID model (2) Match-ing algorithm could not solve time fixed effects but DIDcould (3) MDID allows the existence of invisible factors

Mathematical Problems in Engineering 5

Table 1 Basic principle of MDID

Before the policy change After the policy change Differential valueExperimental group 120573

0+ 1205732+ 120572 + 120575 120573

0+ 1205731+ 1205732+ 120574 + 120572 + 120575 Δ119910

119905= 1205731+ 120574

Control group 1205730+ 1205732+ 120575 120573

0+ 1205731+ 1205732+ 120575 Δ119910

119888= 1205731

Differential value mdash mdash ΔΔ119910 = 120574

(4) Matching process does not need lots of data whichmakes it possible to build larger optional libraries (5) MDIDmodel is relatively complete and easy for the inspection andinference

According to the DID model and matching estimatorMDID estimator is established as followsATTMDID

=1

1198731

1198731

sum

119894=1

(11991011198941199051

minus 11991011198941199050

) minus

1198730

sum

119895=1

119908119894119895(11991001198951199051

minus 11991001198951199050

)

(12)

In view of the above estimators considering the otherinfluencing factors in control variables the following panelmodel is constructed here to estimate ATT

119908119894119895119910119894119905= 1205730+ 1205731119905119894119905+ 1205732119909119894119905+ 120572119894119889119894119905+ 120574119905119894119905119889119894119905+ 120575119894+ 120583119894119905 (13)

Among them 119910119894119905is random variable119908

119894119895is weight and 119909

119894119905

is control variable 1205730 1205731 and 120573

2are coefficient vectors and

1205731means how the economic behaviors of the experimental

group and control group change with time under the policyintervention 120575

119894and 119906

119894119905are random disturbance items which

are not changing with time 120572119894is the difference between

control group and experiment group and it does not changewith time 119905

119894119905119889119894119905is the cross term of virtual variable 120574 is the

target variable and it is the change degree of the experimentgroup after accepting the policy

The original hypothesis of DID is 1198670 120574 = 0 when

there is no policy intervention 120574 = 0 and DID effects ofexperimental group and control group are 0

Table 1 summarizes the basic principle of MDIDmethod

3 Numerical Study

31 Analysis of Smooth Indicators

311 Qualitative Analysis Traffic smooth can be explainedfrom two aspects smooth state and smooth efficiencySmooth state can be understood as road unimpeded whichcould be represented by urban average traffic running speedUrban traffic smooth depends on traffic conditions of eachsection or road intersection in thewhole trafficnetworkGen-erally speaking lots of congestion reasons resulted in poortraffic smooth with the main reasons being the mismatchingand incompatibility of traffic demands and supplies

Trafficdemands aremainly influenced by urban economytraffic composition motorized travelling and other factorsAmong them urbanization process affects travel demandsfor the reason that more urban population and advancedeconomy mean greater travel demands Besides traffic formscould directly determine the motorized travel demands and

Table 2 Main impact indicators of traffic smooth

Category Indicator contents Relevance to the smoothThe socialeconomic

PopulationGDP Negative correlation

Trafficdemand

Motor vehicle ownershipCivilian car ownership Negative correlation

Traffic supplyRoad mileage road area Positive correlationPublic vehicle number Positive correlationPublic transportation

passenger Positive correlation

the smooth state of urban traffic Basic operation directlydetermines urban traffic intensity such as travel consumptionand distance

Traffic supply mainly contains urban traffic managementand planning road infrastructure related transportationservice facilities and so forth Urban trafficmanagement andplanning determine the spatial layout of urban transporta-tion such as road network structure and travel directionMoreover road infrastructure includes road mileage roadarea road network structure public transportation railtransportation and other facilities while the transportationservice facilities mainly include vehicle parking and vehicletransfer

Based on the above analysis and modeling necessity fac-tors whose data are quantifiable and accessible are selected toform the indicators as shown in Table 2

(1) Population and GDP Economic development will bringsubstantial increase in urban population and will also leadto a great travel demand Moreover economic developmentcould lead to the production of more purchasing power ofmotor vehicles Therefore the more development the urbaneconomy has gained the greater the traffic travel demandwillbe which also means higher requirement of transportationsupply and carrying capacity

(2) Number of Motor Vehicles As the visual expression ofroad traffic and parking demands vehicle amount largelydetermines the requirements ofmotor vehicle travel In termsof the slowly growing urban roads and tight urban landthe rapid growth of motor vehicles will provide the roadwith more carrying capacity which would quickly reachsaturation Besides the insufficient parking supply capacitywill directly result in severe urban traffic congestion

(3) Road Mileage and Road Area Urban road is the carrierof urban traffic and road mileage as well as road areas candirectly determine the carrying capacity of the road network

6 Mathematical Problems in Engineering

Table 3 Related data of traffic smooth in Beijing

Year ASPD GDP POP CVECH RODM RODA PVECH PPER2000 189 3161 1357 1041 2471 3502 10353 3482001 190 3711 1383 1145 2493 3701 12945 3922002 198 4330 1423 1339 2504 3857 15046 4352003 201 5024 1456 1631 3347 4315 16753 3712004 200 6033 1493 1824 4064 7287 18451 4392005 234 6970 1538 2097 4073 7437 18503 4502006 225 8118 1581 2391 4419 7286 19522 3982007 205 9847 1633 2734 4460 7632 19395 4232008 231 11115 1695 3137 6186 8940 21507 4712009 231 12153 1755 3681 6247 9179 21716 5172010 222 14114 1961 4497 6355 9395 21548 505Data sources Beijing Transportation Development Annual Report

Longer road and larger areas mean more driving smooth ofurban traffic However due to the restriction of urban plan-ning and land construction funds and many other factorsthe growth rate of urban roads is lower than that of motorvehicle Taking Beijing as an example the average annualgrowth of roadmileage is 1012 from2005 to 2010 Excludingthe large-scale construction during the 2008OlympicGamesthe average annual growth rate is only 3 which is less than13 of the average annual growth of motor vehicle

(4) Number of Public Vehicles After years of explorationoptimizing the traffic travel structure has become more andmore important Besides improving the carrying capac-ity and service level of public transport also becomes animportant way to solve the urban traffic congestion problemThe number of public vehicles which plays a positive rolein easing urban traffic congestion could reflect the supplycapacity of urban public transport

(5) Number of Public Transport Passengers Public transportpassenger can describe supply capacity of public transportfrom the perspective of transportation production Generallypeople believed that with greater amount of public transportpassenger urban transport operation would turn better Atpresent the main public transport travel modes are publicelectric vehicle (regular bus) mass transit and taxi

312 Quantitative Analysis The purpose of the quantitativeanalysis is to clarify the correlation among indicators andthe relationship between indicators and traffic smooth Asshowed in foregoing analysis the paper defines average trafficspeed of road network as the characterizing indicator ofurban traffic smooth Considering the regularity betweenindicators and data availability the paper collects the data ofBeijing from 2000 to 2010 for further analysis as shown inTable 3

Here the following symbols are used to represent thename of each indicator

ASPD average traffic speed of road network (kmh)GDP GDP (hundred million Yuan)

POP resident population (ten thousand people)CVECH number of civilian vehicles (ten thousandvehicles)RODM road mileage (km)RODA road area (ten thousand m2)PVECH number of public vehiclesPPER number of public transport passengers

Utilizing EViews 60 to implement correlation test thespecific results are shown in Table 4

The results showed that the correlation coefficient ofGDPPOP other factors and ASPD is all above 07 With highcorrelation these factors are main factors affecting averagetraffic speed of road network

Indicators are screened by referring to the existing evalu-ation achievements about traffic smooth both at home andabroad and following the metering modeling principles ofldquofrom general to simplerdquo In terms of screening the indepen-dent variables the CIA assumption in MDID model couldprovide some supporting functions independent variablesare influenced by ldquointerventionrdquo but it cannot influence theldquointerventionrdquo According to MDID assumption variableswhich cannot meet CIA assumptions will be eliminatedFinally Granger test is used as the specific implementationmethod

Step 1 Invoke the analysis results of speedrsquos factors andestablish timing database

Step 2 Conduct Granger test between influencing factorsand ASPD one by one

Step 3 Strip out the factors of ASPD Granger reason

The results show that Granger test result of traffic speedis significant (119875 value lt 005) Hence the number of publictransport passengersrsquo indicator is not suitable for the inde-pendent variable of speed

32 Building theMatching Pool DIDmodel needs a compari-son city while at the same timeGuangzhou city is the only ITSdemonstration project city in China Therefore the controlgroup cannot be the ITS demonstration project cityHoweverthere are two major drawbacks for the manual selectionof the control group cities First how much difference thecontrol group cities can tolerate It may be 10 or 30There is no objective standard Second in case of one singleselected city the results will heavily depend on the data ofthe chosen city Under the circumstances the introductionof matching process can solve these two issues effectivelyand provide theoretical support for selecting optimal controlgroup Before the execution of the matching algorithms thefirst thing is to select the cities that can be allowed to enterinto the matching pool

According to the model theory it should be assumedthat policy is exogenous in terms of selecting control groupThat is to say ITS only affects Guangzhou and has no effecton control group cities or the influences can be ignored

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

2 Mathematical Problems in Engineering

Assessment of Transport Telemetric Applications UpdatedVersion in September 1998 which systematically expoundedthe evaluation thought of ITS project in EU Besides it hadbeen used to evaluate the ITS project in London Parisand other cities However the evaluation is just related tothe development and construction effects rather than theinvolvement with the application effects In 2000 the UnitedStates issued ldquoNational ITS Architecturerdquo (version 30) andcomprehensively expounded the evaluation contents andmethods of ITS The evaluation content stipulated includesevaluation design performance cost and profitability analy-sis risk analysis and evaluation results However the micro-level research of ITS acting on urban traffic is not involved

Although in the later research Underwood and Gehring[1] evaluated the ITS application effects from social individ-ual and enterprise benefits the research content and resultare still in themacro level Furthermore Novak andMcDnald[2] studied the potential macroeconomic effects of ITS in theUnited States in which direct employment economic multi-plier productivity technology diffusion and competitivenessare included Moreover Lee [3 4] Zavergiu [5] Brand [6]and Beimborn et al [7] established user efficiency analysismodels cost benefit analysis related models and some otherrelevant models so as to research the economic value of ITSprojects However these researches are more from the per-spective of ITS economic value instead of the improvementof traffic environment besides their evaluation indicatorsare different Then Xiao [8] who used the analytic hierarchyprocess and fuzzy mathematics theory to evaluate the effectsof ITS in improving urban comprehensive transportationsystem once argued that the effects of ITS in improving trafficenvironment cannot be ignored In addition Zhao et al [9]researched from public transport facilities service benefitand social responsibility four aspects in total so as to establishthe comprehensive evaluation index system of urban publictraffic system Additionally Chen and Liao [10] evaluated ITSapplication effects from six aspects namely economic effectsocial effect resource consumption environmental impactand development ability In all these researches ITS applica-tion effects have been analyzed to a certain extent Howevermore researches are still focusing on the economic and overallsocial effects Besides there has not yet been in-depth analysison the specific effects of ITS to urban traffic Until 2011Cao et al [11] defined and quantified the evaluation indexsystem of ITS application effects from five aspects comfortsafety convenience timeliness and coordination which hasachieved a certain innovation and laid the foundation for thefurther studyHowever it appeared that there have been somedeficiencies with the evaluation method The limitations ofthe methods led to the complexity of the evaluation processand affected the authenticity of the resultsTherefore in orderto ensure the authenticity of the evaluation results more suit-able innovationmethods should be developed and employed

However during this period certain results in terms ofthe researches of evaluation methods have been obtainedFor instance Leviakangas and Lahesmaa [12] discussed theapplication of AHP to analyze the cost benefits of ITS It is toosubjective to determine the weights of the efficiency indica-tors and this might widen the gap between evaluation results

and actual situation Moreover Mattingly [13] constructed amodel to show the value of ITS project in the monetary formwhich was integrated through the analytic hierarchy processand multiple attribute value function method After that themodel was used to evaluate ITS effects of Anaheim in Califor-nia thus confirming the feasibility of themodel However themodel cannot evaluate the application effects of the micro-level object effectively Wang and Walton [14] discussed thegrey correlation analysis (GRA) in evaluating ITS projectZhang et al [15] combined analytic hierarchy process withgrey correlation method to evaluate the effects of the sub-systems of ITS Nevertheless all these researches still cannotavoid the influence by the model characteristics namely thesubjective evaluation results are stronger which can seriouslyaffect the veracity of the practical evaluation results

In order to get a better evaluation model some scholarsproposed DID model due to the mature theoretical basis andpractical verification of the model Moreover DID is moresuitable for the evaluation of the benefits of new policy andthe results are alsomore robust For exampleWooldridge [16]established correspondingDIDmodels according to differentdata types while Ai and Norton [17] and Puhani [18] studiedthe parameter estimation computing method of generalizedlinear equation in DID model Moreover Drykker [19] usedWooldridgersquos test to verify the correlation between randomerrors and then corrected the model by generalized leastsquares method By studying the minimum wage system inBritain from 1991 to 2000 Galiani et al [20] evaluated theeffects of minimum wage system on employment BesidesConti [21] calculated the contribution rate of euro in improv-ing per capita GDP and labor productivity level by DIDmodel Lewis and Reiley [22] researched the contribution ofoffline sales products advertising for profit change In addi-tion Liang and Zhou [23] Wu [24] Wang [25] and Dengand Wang [26] innovatively selected DID model to analyzethe contribution of soil and water conservation engineer-ing new agriculture insurance policy and foreign strategicinvestment and so forth However most researches focus onthe fields of medical equity and economic reform whichcould be inflicted with great short-time changes and easinessto control group However there are rare researches in thetransportation field while at the same time the applicationsare lacking innovation Furthermore researches on effects ofITS in improving urban traffic congestion are relatively blankWith more difficulties in choosing the control group therewould be more limitation in terms of the application of DIDBesides the complexity of ITS also determines the difficultiesin selecting control groupThus it is necessary to improve theDID model and the innovative introduction of MDIDmodeis very important

Despite the achievements acquired by the researches inthe traffic smooth field [27ndash31] more researches are focusingon the cause and influence factors of traffic congestion orother fields Besides there are relatively scarce quantitativeresearches between transportation operation service andITS Although some scholars have studied the relationshipbetween traffic smooth and ITS there are still limited qual-itative researches All of these may lead to the fact thatspecific functions of ITS in improving traffic smooth cannot

Mathematical Problems in Engineering 3

be expressed clearly Therefore strengthening quantitativeresearch of ITS in improving urban traffic congestion isparticularly important and necessary [32]

In order to acquire more scientific ITS investment andtechnology-integrated application with higher efficiency thispaper takes Guangzhou (capital city of Guangdong provinceChina) as an example (Guangzhou city is the ITS demonstra-tion city in China) based on comprehensive analysis Afterthat urban traffic smooth is chosen as the research objectBased on the literature analysis from domestic and foreignresearch results characteristics of ITS and its influencefactors have been refined Then relevant analysis has beenconducted on the function of ITS to urban traffic smooth Inaddition a set of evaluation index system has been built soas to comprehensively objectively and accurately reflect thebenefits of ITS in improving urban traffic smooth Addi-tionally the indexes related to urban traffic smooth areextracted firstly then the indexeswhich are greatly influencedby ITS are stripped out and finally the influencing factorsof these indexes are figured out Finally with advancedDID model (MDID matching difference-in-difference) andother related econometric models the contribution rate iscalculatedThrough the research the contribution rate of ITSin improving Guangzhoursquos traffic smooth has been acquiredand the mechanism of ITS in perfecting urban traffic smoothhas been proposed

2 Model Formulation

21 Introduction to DID Model

211 DID Model General DID estimation model is

119910119894= 120573 + 120572

119894119889119894+ 119906119894 (1)

In the formula 119889119894= 1 means experimental object and

119889119894= 0 means control object Average treatment effect on

the treated ATTmeasures the average return of experimentalgroup and control group

120572ATT

= 119864 (120572119894| 119889119894= 1) (2)

Investigating dynamic change and introducing time vari-able to the estimation model

119910119894119905= 120573 + 120572

119894119889119894119905+ 119906119894119905 (3)

The virtual variable 119905 = 1199050is time section before policy

implemented and 119905 = 1199051is the time section after policy

implemented Moreover policy implementation effect is onlyacting on the experimental group Observation time meets1199050lt 119896 lt 119905

1 and the formula is acquired as follows

119864 [119910119894119905| 119889119894 119905]

=

120573 + 119864 [120572 | 119889119894= 1] + 120576

119905if 119889119894= 1 119905 = 119905

1

120573 + 120576119905

others

(4)

Control group

Experimental group

Experimental group

Historicaldata

Control group

Currentt = 0

t = 0

t = 1

t = 1

dataΔy1

Δy2

Δy = Δy2 minus Δy1

Figure 1 Application ideas of DID model

Here the value of ATT is shown

120572ATT

= 119864 [120572 | 119889119894= 1]

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

minus 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

(5)

Because 119864[ATTDID] = 120572ATT the estimator of DID is

ATTDID = [1199101

1199051

minus 1199101

1199050

] minus [1199100

1199051

minus 1199100

1199050

] (6)

DID is the unbiased estimators of ATT The estimatorcan not only measure the effects before and after the policyimplemented but alsomeasure the policy differences betweenexperimental group and control group After that the policyimplementation effects are obtained

Here setting the area that has built the ITS as ldquoexperi-mental grouprdquo the area without ITS is developed as ldquocontrolgrouprdquo DIDmodel can effectively control the effects of otherpossible factors Besides it can not only distinguish the indi-vidual difference in experimental group after using ITS butalso separate time differences of different areasThen it comesto the identification of the application effect of ITS (Figure 1)

From the perspective of the DID model research andapplication the selection of control group mainly relies onsubjective judgment and lack of objective and quantitativeevaluation standard When control group is a single individ-ual the results will depend on the individual data to a greatextent and result in poor robustness model Therefore it isnecessary to improve the DID model

22 Introduction to Matching Algorithm Due to the strongsubjectivity of DID model in selecting control group theauthenticity of evaluation results will be affected Thusmatching algorithm is introduced in the paper to solvethe problem In terms of matching algorithm adopting anapproximation experimental algorithm is implemented todistinguish experimental group and control group effectivelyMatching algorithm believes that individuals share the samecharacteristics due towhich they also have the same responseto the same policy variables That is to say the decision-making behaviors of experimental group and control groupare not affected by unobservable variables Moreover match-ing algorithm can enlarge the select range of control groupwhich can avoid the subjective randomness in selectingcontrol group and eliminate the estimation deviation caused

4 Mathematical Problems in Engineering

by nonmutual support domain and nonidentical distribution[33]

Matching model assumes that the selection biases areentirely determined by observable variables and it also rulesout the selection biases caused by unobservable factorsIt means that when the control variables are determinedwhether affected by policy or not they will have nothing todowith the value of index119884 Call it conditional independenceassumption (CIA) and mathematical formula is

119864 (119884 | 119883 119889 = 1) = 119864 (119884 | 119883 119889 = 0) 119884 perp 119889 | 119883 (7)

The CIA assumption in matching model would limitsome variables selected into control group but the restrictiondoes not cause interference to the precision of results It is forthe reason thatDIDmodel allows the existence of observationfactors

Here setting120572ATT0

is theATTof control variable and120572ATT1

is the ATT of experimental group When CIA assumption issatisfied formulas are acquired as

120572ATT0

= 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

= 0

(8)

That is to say the change of control group is 0 before andafter the policy implemented Therefore 120572ATT is

120572ATT

= 120572ATT1

minus 120572ATT0

= 120572ATT1

minus 0

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

(9)

However there will bemultidimensional problems due tothe increasing matching strategy with the rise of dimension119883 In the early method covariate (CV) is used for thecalculation which would result in ldquodimension disaster (whenthe dimension increased the volume of space will increasequickly which lead to the fact that the available data willbecome very sparse)rdquo and the calculation amount is too largeIn order to solve the multidimensional problem Rosenbaumand Rubin [33] put forward the propensity score (PS) asa covariate function If CV meets CIA hypothesis then PScould also meet the requirement Besides the dimensionreduction can be completed successfully by transferring theinformation contained in CV to PS Therefore PS(119883) isselected to replace the CVrsquos119883 and formula119884 perp 119889 | 119883 rArr 119884 perp

119889 | 119901119904(119883) is employed to achieve the dimension reduction

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883)) = 119864 (119863 = 1 | 119884 (0)

119884 (1) PS (119883))

= 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) PS (119883) 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) 119883) |

119884 (0) 119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [PS (119883) | 119884 (0) 119884 (1) PS (119883)]

= PS (119883)

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883))

= Pr (119863 = 1 | PS (119883)) = PS (119883)

(10)

minus4 minus3 minus2 minus1 0 1 2 3 40

00501

01502

02503

03504

Kdensity treatmentKdensity control

Figure 2 Common support domains

When dealing with PS Heckman believed that Logist(Logist model uses cumulative Logist distribution functionto describe the relationship between selection probabilityand explanatory variables and limited the value of explainedvariables ranging from 0 to 1 119901

119894= 119864(119884 = 1 | 119883

119894) =

1198901205730+1205731119883119894(1+119890

1205730+1205731119883119894) the odds ratio selection or not selection

is 119901119894(1 minus 119901

119894) taking logarithm of the odds ratio to obtain the

measurementmodel 119871119894= ln(119901

119894(1minus119901

119894)) = 120573

0+1205731119883119894+120576119894) has

more advantages than Probit (Probit model uses the standardnormal distribution function to describe the relationshipbetween the selection probabilities and the explanatory vari-ables and limited the value of explained variables rangingfrom 0 to 1 119901(119884

119894= 1) = Φ(119911

119894) = int119911119894

minusinfin(1radic2120587) exp(minus1199042

1198942)119889119904

then it implemented simple inverse function transformationand got the following 119911

119894= Φminus1(119901119894) = 120573

0+ 1205731119883119894+ 120576119894)

Therefore the estimation values of matching can be obtainedwith the following formula

ATTMatching =1

1198731

1198731

sum

119894=1

(1199101119894minus

1198730

sum

119895=1

1199081198941198951199100119895) (11)

1198730and 119873

1refer to the total number of individuals

119908119894119895means the differences between experimental group and

control group and it meets 119908119894119895isin [0 1] sum1198730

119895=1119908119894119895= 1

Meanwhile in order to select optimal control group com-mon support hypothesis is introduced here as 0 lt 119875(119889 = 1 |119883) lt 1 The cross region between experimental group andcontrol group is shown in Figure 2

23 Induction to MDID Model MDID model aims to findoptimal control group on the basis of DID model throughbringing inmatchingmodule andmatching algorithmThenvirtual processing individuals have been developed before theintervention implemented Last the virtual processing indi-vidual comes to the implementation of the double differencebased on the matching selection of control group

MDIDmodel has the following advantages when it comesto calculating the contribution rate (1) MDID could solvethe human factors generated by DID model (2) Match-ing algorithm could not solve time fixed effects but DIDcould (3) MDID allows the existence of invisible factors

Mathematical Problems in Engineering 5

Table 1 Basic principle of MDID

Before the policy change After the policy change Differential valueExperimental group 120573

0+ 1205732+ 120572 + 120575 120573

0+ 1205731+ 1205732+ 120574 + 120572 + 120575 Δ119910

119905= 1205731+ 120574

Control group 1205730+ 1205732+ 120575 120573

0+ 1205731+ 1205732+ 120575 Δ119910

119888= 1205731

Differential value mdash mdash ΔΔ119910 = 120574

(4) Matching process does not need lots of data whichmakes it possible to build larger optional libraries (5) MDIDmodel is relatively complete and easy for the inspection andinference

According to the DID model and matching estimatorMDID estimator is established as followsATTMDID

=1

1198731

1198731

sum

119894=1

(11991011198941199051

minus 11991011198941199050

) minus

1198730

sum

119895=1

119908119894119895(11991001198951199051

minus 11991001198951199050

)

(12)

In view of the above estimators considering the otherinfluencing factors in control variables the following panelmodel is constructed here to estimate ATT

119908119894119895119910119894119905= 1205730+ 1205731119905119894119905+ 1205732119909119894119905+ 120572119894119889119894119905+ 120574119905119894119905119889119894119905+ 120575119894+ 120583119894119905 (13)

Among them 119910119894119905is random variable119908

119894119895is weight and 119909

119894119905

is control variable 1205730 1205731 and 120573

2are coefficient vectors and

1205731means how the economic behaviors of the experimental

group and control group change with time under the policyintervention 120575

119894and 119906

119894119905are random disturbance items which

are not changing with time 120572119894is the difference between

control group and experiment group and it does not changewith time 119905

119894119905119889119894119905is the cross term of virtual variable 120574 is the

target variable and it is the change degree of the experimentgroup after accepting the policy

The original hypothesis of DID is 1198670 120574 = 0 when

there is no policy intervention 120574 = 0 and DID effects ofexperimental group and control group are 0

Table 1 summarizes the basic principle of MDIDmethod

3 Numerical Study

31 Analysis of Smooth Indicators

311 Qualitative Analysis Traffic smooth can be explainedfrom two aspects smooth state and smooth efficiencySmooth state can be understood as road unimpeded whichcould be represented by urban average traffic running speedUrban traffic smooth depends on traffic conditions of eachsection or road intersection in thewhole trafficnetworkGen-erally speaking lots of congestion reasons resulted in poortraffic smooth with the main reasons being the mismatchingand incompatibility of traffic demands and supplies

Trafficdemands aremainly influenced by urban economytraffic composition motorized travelling and other factorsAmong them urbanization process affects travel demandsfor the reason that more urban population and advancedeconomy mean greater travel demands Besides traffic formscould directly determine the motorized travel demands and

Table 2 Main impact indicators of traffic smooth

Category Indicator contents Relevance to the smoothThe socialeconomic

PopulationGDP Negative correlation

Trafficdemand

Motor vehicle ownershipCivilian car ownership Negative correlation

Traffic supplyRoad mileage road area Positive correlationPublic vehicle number Positive correlationPublic transportation

passenger Positive correlation

the smooth state of urban traffic Basic operation directlydetermines urban traffic intensity such as travel consumptionand distance

Traffic supply mainly contains urban traffic managementand planning road infrastructure related transportationservice facilities and so forth Urban trafficmanagement andplanning determine the spatial layout of urban transporta-tion such as road network structure and travel directionMoreover road infrastructure includes road mileage roadarea road network structure public transportation railtransportation and other facilities while the transportationservice facilities mainly include vehicle parking and vehicletransfer

Based on the above analysis and modeling necessity fac-tors whose data are quantifiable and accessible are selected toform the indicators as shown in Table 2

(1) Population and GDP Economic development will bringsubstantial increase in urban population and will also leadto a great travel demand Moreover economic developmentcould lead to the production of more purchasing power ofmotor vehicles Therefore the more development the urbaneconomy has gained the greater the traffic travel demandwillbe which also means higher requirement of transportationsupply and carrying capacity

(2) Number of Motor Vehicles As the visual expression ofroad traffic and parking demands vehicle amount largelydetermines the requirements ofmotor vehicle travel In termsof the slowly growing urban roads and tight urban landthe rapid growth of motor vehicles will provide the roadwith more carrying capacity which would quickly reachsaturation Besides the insufficient parking supply capacitywill directly result in severe urban traffic congestion

(3) Road Mileage and Road Area Urban road is the carrierof urban traffic and road mileage as well as road areas candirectly determine the carrying capacity of the road network

6 Mathematical Problems in Engineering

Table 3 Related data of traffic smooth in Beijing

Year ASPD GDP POP CVECH RODM RODA PVECH PPER2000 189 3161 1357 1041 2471 3502 10353 3482001 190 3711 1383 1145 2493 3701 12945 3922002 198 4330 1423 1339 2504 3857 15046 4352003 201 5024 1456 1631 3347 4315 16753 3712004 200 6033 1493 1824 4064 7287 18451 4392005 234 6970 1538 2097 4073 7437 18503 4502006 225 8118 1581 2391 4419 7286 19522 3982007 205 9847 1633 2734 4460 7632 19395 4232008 231 11115 1695 3137 6186 8940 21507 4712009 231 12153 1755 3681 6247 9179 21716 5172010 222 14114 1961 4497 6355 9395 21548 505Data sources Beijing Transportation Development Annual Report

Longer road and larger areas mean more driving smooth ofurban traffic However due to the restriction of urban plan-ning and land construction funds and many other factorsthe growth rate of urban roads is lower than that of motorvehicle Taking Beijing as an example the average annualgrowth of roadmileage is 1012 from2005 to 2010 Excludingthe large-scale construction during the 2008OlympicGamesthe average annual growth rate is only 3 which is less than13 of the average annual growth of motor vehicle

(4) Number of Public Vehicles After years of explorationoptimizing the traffic travel structure has become more andmore important Besides improving the carrying capac-ity and service level of public transport also becomes animportant way to solve the urban traffic congestion problemThe number of public vehicles which plays a positive rolein easing urban traffic congestion could reflect the supplycapacity of urban public transport

(5) Number of Public Transport Passengers Public transportpassenger can describe supply capacity of public transportfrom the perspective of transportation production Generallypeople believed that with greater amount of public transportpassenger urban transport operation would turn better Atpresent the main public transport travel modes are publicelectric vehicle (regular bus) mass transit and taxi

312 Quantitative Analysis The purpose of the quantitativeanalysis is to clarify the correlation among indicators andthe relationship between indicators and traffic smooth Asshowed in foregoing analysis the paper defines average trafficspeed of road network as the characterizing indicator ofurban traffic smooth Considering the regularity betweenindicators and data availability the paper collects the data ofBeijing from 2000 to 2010 for further analysis as shown inTable 3

Here the following symbols are used to represent thename of each indicator

ASPD average traffic speed of road network (kmh)GDP GDP (hundred million Yuan)

POP resident population (ten thousand people)CVECH number of civilian vehicles (ten thousandvehicles)RODM road mileage (km)RODA road area (ten thousand m2)PVECH number of public vehiclesPPER number of public transport passengers

Utilizing EViews 60 to implement correlation test thespecific results are shown in Table 4

The results showed that the correlation coefficient ofGDPPOP other factors and ASPD is all above 07 With highcorrelation these factors are main factors affecting averagetraffic speed of road network

Indicators are screened by referring to the existing evalu-ation achievements about traffic smooth both at home andabroad and following the metering modeling principles ofldquofrom general to simplerdquo In terms of screening the indepen-dent variables the CIA assumption in MDID model couldprovide some supporting functions independent variablesare influenced by ldquointerventionrdquo but it cannot influence theldquointerventionrdquo According to MDID assumption variableswhich cannot meet CIA assumptions will be eliminatedFinally Granger test is used as the specific implementationmethod

Step 1 Invoke the analysis results of speedrsquos factors andestablish timing database

Step 2 Conduct Granger test between influencing factorsand ASPD one by one

Step 3 Strip out the factors of ASPD Granger reason

The results show that Granger test result of traffic speedis significant (119875 value lt 005) Hence the number of publictransport passengersrsquo indicator is not suitable for the inde-pendent variable of speed

32 Building theMatching Pool DIDmodel needs a compari-son city while at the same timeGuangzhou city is the only ITSdemonstration project city in China Therefore the controlgroup cannot be the ITS demonstration project cityHoweverthere are two major drawbacks for the manual selectionof the control group cities First how much difference thecontrol group cities can tolerate It may be 10 or 30There is no objective standard Second in case of one singleselected city the results will heavily depend on the data ofthe chosen city Under the circumstances the introductionof matching process can solve these two issues effectivelyand provide theoretical support for selecting optimal controlgroup Before the execution of the matching algorithms thefirst thing is to select the cities that can be allowed to enterinto the matching pool

According to the model theory it should be assumedthat policy is exogenous in terms of selecting control groupThat is to say ITS only affects Guangzhou and has no effecton control group cities or the influences can be ignored

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Mathematical Problems in Engineering 3

be expressed clearly Therefore strengthening quantitativeresearch of ITS in improving urban traffic congestion isparticularly important and necessary [32]

In order to acquire more scientific ITS investment andtechnology-integrated application with higher efficiency thispaper takes Guangzhou (capital city of Guangdong provinceChina) as an example (Guangzhou city is the ITS demonstra-tion city in China) based on comprehensive analysis Afterthat urban traffic smooth is chosen as the research objectBased on the literature analysis from domestic and foreignresearch results characteristics of ITS and its influencefactors have been refined Then relevant analysis has beenconducted on the function of ITS to urban traffic smooth Inaddition a set of evaluation index system has been built soas to comprehensively objectively and accurately reflect thebenefits of ITS in improving urban traffic smooth Addi-tionally the indexes related to urban traffic smooth areextracted firstly then the indexeswhich are greatly influencedby ITS are stripped out and finally the influencing factorsof these indexes are figured out Finally with advancedDID model (MDID matching difference-in-difference) andother related econometric models the contribution rate iscalculatedThrough the research the contribution rate of ITSin improving Guangzhoursquos traffic smooth has been acquiredand the mechanism of ITS in perfecting urban traffic smoothhas been proposed

2 Model Formulation

21 Introduction to DID Model

211 DID Model General DID estimation model is

119910119894= 120573 + 120572

119894119889119894+ 119906119894 (1)

In the formula 119889119894= 1 means experimental object and

119889119894= 0 means control object Average treatment effect on

the treated ATTmeasures the average return of experimentalgroup and control group

120572ATT

= 119864 (120572119894| 119889119894= 1) (2)

Investigating dynamic change and introducing time vari-able to the estimation model

119910119894119905= 120573 + 120572

119894119889119894119905+ 119906119894119905 (3)

The virtual variable 119905 = 1199050is time section before policy

implemented and 119905 = 1199051is the time section after policy

implemented Moreover policy implementation effect is onlyacting on the experimental group Observation time meets1199050lt 119896 lt 119905

1 and the formula is acquired as follows

119864 [119910119894119905| 119889119894 119905]

=

120573 + 119864 [120572 | 119889119894= 1] + 120576

119905if 119889119894= 1 119905 = 119905

1

120573 + 120576119905

others

(4)

Control group

Experimental group

Experimental group

Historicaldata

Control group

Currentt = 0

t = 0

t = 1

t = 1

dataΔy1

Δy2

Δy = Δy2 minus Δy1

Figure 1 Application ideas of DID model

Here the value of ATT is shown

120572ATT

= 119864 [120572 | 119889119894= 1]

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

minus 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

(5)

Because 119864[ATTDID] = 120572ATT the estimator of DID is

ATTDID = [1199101

1199051

minus 1199101

1199050

] minus [1199100

1199051

minus 1199100

1199050

] (6)

DID is the unbiased estimators of ATT The estimatorcan not only measure the effects before and after the policyimplemented but alsomeasure the policy differences betweenexperimental group and control group After that the policyimplementation effects are obtained

Here setting the area that has built the ITS as ldquoexperi-mental grouprdquo the area without ITS is developed as ldquocontrolgrouprdquo DIDmodel can effectively control the effects of otherpossible factors Besides it can not only distinguish the indi-vidual difference in experimental group after using ITS butalso separate time differences of different areasThen it comesto the identification of the application effect of ITS (Figure 1)

From the perspective of the DID model research andapplication the selection of control group mainly relies onsubjective judgment and lack of objective and quantitativeevaluation standard When control group is a single individ-ual the results will depend on the individual data to a greatextent and result in poor robustness model Therefore it isnecessary to improve the DID model

22 Introduction to Matching Algorithm Due to the strongsubjectivity of DID model in selecting control group theauthenticity of evaluation results will be affected Thusmatching algorithm is introduced in the paper to solvethe problem In terms of matching algorithm adopting anapproximation experimental algorithm is implemented todistinguish experimental group and control group effectivelyMatching algorithm believes that individuals share the samecharacteristics due towhich they also have the same responseto the same policy variables That is to say the decision-making behaviors of experimental group and control groupare not affected by unobservable variables Moreover match-ing algorithm can enlarge the select range of control groupwhich can avoid the subjective randomness in selectingcontrol group and eliminate the estimation deviation caused

4 Mathematical Problems in Engineering

by nonmutual support domain and nonidentical distribution[33]

Matching model assumes that the selection biases areentirely determined by observable variables and it also rulesout the selection biases caused by unobservable factorsIt means that when the control variables are determinedwhether affected by policy or not they will have nothing todowith the value of index119884 Call it conditional independenceassumption (CIA) and mathematical formula is

119864 (119884 | 119883 119889 = 1) = 119864 (119884 | 119883 119889 = 0) 119884 perp 119889 | 119883 (7)

The CIA assumption in matching model would limitsome variables selected into control group but the restrictiondoes not cause interference to the precision of results It is forthe reason thatDIDmodel allows the existence of observationfactors

Here setting120572ATT0

is theATTof control variable and120572ATT1

is the ATT of experimental group When CIA assumption issatisfied formulas are acquired as

120572ATT0

= 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

= 0

(8)

That is to say the change of control group is 0 before andafter the policy implemented Therefore 120572ATT is

120572ATT

= 120572ATT1

minus 120572ATT0

= 120572ATT1

minus 0

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

(9)

However there will bemultidimensional problems due tothe increasing matching strategy with the rise of dimension119883 In the early method covariate (CV) is used for thecalculation which would result in ldquodimension disaster (whenthe dimension increased the volume of space will increasequickly which lead to the fact that the available data willbecome very sparse)rdquo and the calculation amount is too largeIn order to solve the multidimensional problem Rosenbaumand Rubin [33] put forward the propensity score (PS) asa covariate function If CV meets CIA hypothesis then PScould also meet the requirement Besides the dimensionreduction can be completed successfully by transferring theinformation contained in CV to PS Therefore PS(119883) isselected to replace the CVrsquos119883 and formula119884 perp 119889 | 119883 rArr 119884 perp

119889 | 119901119904(119883) is employed to achieve the dimension reduction

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883)) = 119864 (119863 = 1 | 119884 (0)

119884 (1) PS (119883))

= 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) PS (119883) 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) 119883) |

119884 (0) 119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [PS (119883) | 119884 (0) 119884 (1) PS (119883)]

= PS (119883)

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883))

= Pr (119863 = 1 | PS (119883)) = PS (119883)

(10)

minus4 minus3 minus2 minus1 0 1 2 3 40

00501

01502

02503

03504

Kdensity treatmentKdensity control

Figure 2 Common support domains

When dealing with PS Heckman believed that Logist(Logist model uses cumulative Logist distribution functionto describe the relationship between selection probabilityand explanatory variables and limited the value of explainedvariables ranging from 0 to 1 119901

119894= 119864(119884 = 1 | 119883

119894) =

1198901205730+1205731119883119894(1+119890

1205730+1205731119883119894) the odds ratio selection or not selection

is 119901119894(1 minus 119901

119894) taking logarithm of the odds ratio to obtain the

measurementmodel 119871119894= ln(119901

119894(1minus119901

119894)) = 120573

0+1205731119883119894+120576119894) has

more advantages than Probit (Probit model uses the standardnormal distribution function to describe the relationshipbetween the selection probabilities and the explanatory vari-ables and limited the value of explained variables rangingfrom 0 to 1 119901(119884

119894= 1) = Φ(119911

119894) = int119911119894

minusinfin(1radic2120587) exp(minus1199042

1198942)119889119904

then it implemented simple inverse function transformationand got the following 119911

119894= Φminus1(119901119894) = 120573

0+ 1205731119883119894+ 120576119894)

Therefore the estimation values of matching can be obtainedwith the following formula

ATTMatching =1

1198731

1198731

sum

119894=1

(1199101119894minus

1198730

sum

119895=1

1199081198941198951199100119895) (11)

1198730and 119873

1refer to the total number of individuals

119908119894119895means the differences between experimental group and

control group and it meets 119908119894119895isin [0 1] sum1198730

119895=1119908119894119895= 1

Meanwhile in order to select optimal control group com-mon support hypothesis is introduced here as 0 lt 119875(119889 = 1 |119883) lt 1 The cross region between experimental group andcontrol group is shown in Figure 2

23 Induction to MDID Model MDID model aims to findoptimal control group on the basis of DID model throughbringing inmatchingmodule andmatching algorithmThenvirtual processing individuals have been developed before theintervention implemented Last the virtual processing indi-vidual comes to the implementation of the double differencebased on the matching selection of control group

MDIDmodel has the following advantages when it comesto calculating the contribution rate (1) MDID could solvethe human factors generated by DID model (2) Match-ing algorithm could not solve time fixed effects but DIDcould (3) MDID allows the existence of invisible factors

Mathematical Problems in Engineering 5

Table 1 Basic principle of MDID

Before the policy change After the policy change Differential valueExperimental group 120573

0+ 1205732+ 120572 + 120575 120573

0+ 1205731+ 1205732+ 120574 + 120572 + 120575 Δ119910

119905= 1205731+ 120574

Control group 1205730+ 1205732+ 120575 120573

0+ 1205731+ 1205732+ 120575 Δ119910

119888= 1205731

Differential value mdash mdash ΔΔ119910 = 120574

(4) Matching process does not need lots of data whichmakes it possible to build larger optional libraries (5) MDIDmodel is relatively complete and easy for the inspection andinference

According to the DID model and matching estimatorMDID estimator is established as followsATTMDID

=1

1198731

1198731

sum

119894=1

(11991011198941199051

minus 11991011198941199050

) minus

1198730

sum

119895=1

119908119894119895(11991001198951199051

minus 11991001198951199050

)

(12)

In view of the above estimators considering the otherinfluencing factors in control variables the following panelmodel is constructed here to estimate ATT

119908119894119895119910119894119905= 1205730+ 1205731119905119894119905+ 1205732119909119894119905+ 120572119894119889119894119905+ 120574119905119894119905119889119894119905+ 120575119894+ 120583119894119905 (13)

Among them 119910119894119905is random variable119908

119894119895is weight and 119909

119894119905

is control variable 1205730 1205731 and 120573

2are coefficient vectors and

1205731means how the economic behaviors of the experimental

group and control group change with time under the policyintervention 120575

119894and 119906

119894119905are random disturbance items which

are not changing with time 120572119894is the difference between

control group and experiment group and it does not changewith time 119905

119894119905119889119894119905is the cross term of virtual variable 120574 is the

target variable and it is the change degree of the experimentgroup after accepting the policy

The original hypothesis of DID is 1198670 120574 = 0 when

there is no policy intervention 120574 = 0 and DID effects ofexperimental group and control group are 0

Table 1 summarizes the basic principle of MDIDmethod

3 Numerical Study

31 Analysis of Smooth Indicators

311 Qualitative Analysis Traffic smooth can be explainedfrom two aspects smooth state and smooth efficiencySmooth state can be understood as road unimpeded whichcould be represented by urban average traffic running speedUrban traffic smooth depends on traffic conditions of eachsection or road intersection in thewhole trafficnetworkGen-erally speaking lots of congestion reasons resulted in poortraffic smooth with the main reasons being the mismatchingand incompatibility of traffic demands and supplies

Trafficdemands aremainly influenced by urban economytraffic composition motorized travelling and other factorsAmong them urbanization process affects travel demandsfor the reason that more urban population and advancedeconomy mean greater travel demands Besides traffic formscould directly determine the motorized travel demands and

Table 2 Main impact indicators of traffic smooth

Category Indicator contents Relevance to the smoothThe socialeconomic

PopulationGDP Negative correlation

Trafficdemand

Motor vehicle ownershipCivilian car ownership Negative correlation

Traffic supplyRoad mileage road area Positive correlationPublic vehicle number Positive correlationPublic transportation

passenger Positive correlation

the smooth state of urban traffic Basic operation directlydetermines urban traffic intensity such as travel consumptionand distance

Traffic supply mainly contains urban traffic managementand planning road infrastructure related transportationservice facilities and so forth Urban trafficmanagement andplanning determine the spatial layout of urban transporta-tion such as road network structure and travel directionMoreover road infrastructure includes road mileage roadarea road network structure public transportation railtransportation and other facilities while the transportationservice facilities mainly include vehicle parking and vehicletransfer

Based on the above analysis and modeling necessity fac-tors whose data are quantifiable and accessible are selected toform the indicators as shown in Table 2

(1) Population and GDP Economic development will bringsubstantial increase in urban population and will also leadto a great travel demand Moreover economic developmentcould lead to the production of more purchasing power ofmotor vehicles Therefore the more development the urbaneconomy has gained the greater the traffic travel demandwillbe which also means higher requirement of transportationsupply and carrying capacity

(2) Number of Motor Vehicles As the visual expression ofroad traffic and parking demands vehicle amount largelydetermines the requirements ofmotor vehicle travel In termsof the slowly growing urban roads and tight urban landthe rapid growth of motor vehicles will provide the roadwith more carrying capacity which would quickly reachsaturation Besides the insufficient parking supply capacitywill directly result in severe urban traffic congestion

(3) Road Mileage and Road Area Urban road is the carrierof urban traffic and road mileage as well as road areas candirectly determine the carrying capacity of the road network

6 Mathematical Problems in Engineering

Table 3 Related data of traffic smooth in Beijing

Year ASPD GDP POP CVECH RODM RODA PVECH PPER2000 189 3161 1357 1041 2471 3502 10353 3482001 190 3711 1383 1145 2493 3701 12945 3922002 198 4330 1423 1339 2504 3857 15046 4352003 201 5024 1456 1631 3347 4315 16753 3712004 200 6033 1493 1824 4064 7287 18451 4392005 234 6970 1538 2097 4073 7437 18503 4502006 225 8118 1581 2391 4419 7286 19522 3982007 205 9847 1633 2734 4460 7632 19395 4232008 231 11115 1695 3137 6186 8940 21507 4712009 231 12153 1755 3681 6247 9179 21716 5172010 222 14114 1961 4497 6355 9395 21548 505Data sources Beijing Transportation Development Annual Report

Longer road and larger areas mean more driving smooth ofurban traffic However due to the restriction of urban plan-ning and land construction funds and many other factorsthe growth rate of urban roads is lower than that of motorvehicle Taking Beijing as an example the average annualgrowth of roadmileage is 1012 from2005 to 2010 Excludingthe large-scale construction during the 2008OlympicGamesthe average annual growth rate is only 3 which is less than13 of the average annual growth of motor vehicle

(4) Number of Public Vehicles After years of explorationoptimizing the traffic travel structure has become more andmore important Besides improving the carrying capac-ity and service level of public transport also becomes animportant way to solve the urban traffic congestion problemThe number of public vehicles which plays a positive rolein easing urban traffic congestion could reflect the supplycapacity of urban public transport

(5) Number of Public Transport Passengers Public transportpassenger can describe supply capacity of public transportfrom the perspective of transportation production Generallypeople believed that with greater amount of public transportpassenger urban transport operation would turn better Atpresent the main public transport travel modes are publicelectric vehicle (regular bus) mass transit and taxi

312 Quantitative Analysis The purpose of the quantitativeanalysis is to clarify the correlation among indicators andthe relationship between indicators and traffic smooth Asshowed in foregoing analysis the paper defines average trafficspeed of road network as the characterizing indicator ofurban traffic smooth Considering the regularity betweenindicators and data availability the paper collects the data ofBeijing from 2000 to 2010 for further analysis as shown inTable 3

Here the following symbols are used to represent thename of each indicator

ASPD average traffic speed of road network (kmh)GDP GDP (hundred million Yuan)

POP resident population (ten thousand people)CVECH number of civilian vehicles (ten thousandvehicles)RODM road mileage (km)RODA road area (ten thousand m2)PVECH number of public vehiclesPPER number of public transport passengers

Utilizing EViews 60 to implement correlation test thespecific results are shown in Table 4

The results showed that the correlation coefficient ofGDPPOP other factors and ASPD is all above 07 With highcorrelation these factors are main factors affecting averagetraffic speed of road network

Indicators are screened by referring to the existing evalu-ation achievements about traffic smooth both at home andabroad and following the metering modeling principles ofldquofrom general to simplerdquo In terms of screening the indepen-dent variables the CIA assumption in MDID model couldprovide some supporting functions independent variablesare influenced by ldquointerventionrdquo but it cannot influence theldquointerventionrdquo According to MDID assumption variableswhich cannot meet CIA assumptions will be eliminatedFinally Granger test is used as the specific implementationmethod

Step 1 Invoke the analysis results of speedrsquos factors andestablish timing database

Step 2 Conduct Granger test between influencing factorsand ASPD one by one

Step 3 Strip out the factors of ASPD Granger reason

The results show that Granger test result of traffic speedis significant (119875 value lt 005) Hence the number of publictransport passengersrsquo indicator is not suitable for the inde-pendent variable of speed

32 Building theMatching Pool DIDmodel needs a compari-son city while at the same timeGuangzhou city is the only ITSdemonstration project city in China Therefore the controlgroup cannot be the ITS demonstration project cityHoweverthere are two major drawbacks for the manual selectionof the control group cities First how much difference thecontrol group cities can tolerate It may be 10 or 30There is no objective standard Second in case of one singleselected city the results will heavily depend on the data ofthe chosen city Under the circumstances the introductionof matching process can solve these two issues effectivelyand provide theoretical support for selecting optimal controlgroup Before the execution of the matching algorithms thefirst thing is to select the cities that can be allowed to enterinto the matching pool

According to the model theory it should be assumedthat policy is exogenous in terms of selecting control groupThat is to say ITS only affects Guangzhou and has no effecton control group cities or the influences can be ignored

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

4 Mathematical Problems in Engineering

by nonmutual support domain and nonidentical distribution[33]

Matching model assumes that the selection biases areentirely determined by observable variables and it also rulesout the selection biases caused by unobservable factorsIt means that when the control variables are determinedwhether affected by policy or not they will have nothing todowith the value of index119884 Call it conditional independenceassumption (CIA) and mathematical formula is

119864 (119884 | 119883 119889 = 1) = 119864 (119884 | 119883 119889 = 0) 119884 perp 119889 | 119883 (7)

The CIA assumption in matching model would limitsome variables selected into control group but the restrictiondoes not cause interference to the precision of results It is forthe reason thatDIDmodel allows the existence of observationfactors

Here setting120572ATT0

is theATTof control variable and120572ATT1

is the ATT of experimental group When CIA assumption issatisfied formulas are acquired as

120572ATT0

= 119864 [119910119894119905| 119889119894= 0 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 0 119905 = 119905

0]

= 0

(8)

That is to say the change of control group is 0 before andafter the policy implemented Therefore 120572ATT is

120572ATT

= 120572ATT1

minus 120572ATT0

= 120572ATT1

minus 0

= 119864 [119910119894119905| 119889119894= 1 119905 = 119905

1] minus 119864 [119910

119894119905| 119889119894= 1 119905 = 119905

0]

(9)

However there will bemultidimensional problems due tothe increasing matching strategy with the rise of dimension119883 In the early method covariate (CV) is used for thecalculation which would result in ldquodimension disaster (whenthe dimension increased the volume of space will increasequickly which lead to the fact that the available data willbecome very sparse)rdquo and the calculation amount is too largeIn order to solve the multidimensional problem Rosenbaumand Rubin [33] put forward the propensity score (PS) asa covariate function If CV meets CIA hypothesis then PScould also meet the requirement Besides the dimensionreduction can be completed successfully by transferring theinformation contained in CV to PS Therefore PS(119883) isselected to replace the CVrsquos119883 and formula119884 perp 119889 | 119883 rArr 119884 perp

119889 | 119901119904(119883) is employed to achieve the dimension reduction

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883)) = 119864 (119863 = 1 | 119884 (0)

119884 (1) PS (119883))

= 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) PS (119883) 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119884 (0) 119884 (1) 119883) |

119884 (0) 119884 (1) PS (119883)] = 119864 [119864 (119863 = 1 | 119883) | 119884 (0)

119884 (1) PS (119883)] = 119864 [PS (119883) | 119884 (0) 119884 (1) PS (119883)]

= PS (119883)

Pr (119863 = 1 | 119884 (0) 119884 (1) PS (119883))

= Pr (119863 = 1 | PS (119883)) = PS (119883)

(10)

minus4 minus3 minus2 minus1 0 1 2 3 40

00501

01502

02503

03504

Kdensity treatmentKdensity control

Figure 2 Common support domains

When dealing with PS Heckman believed that Logist(Logist model uses cumulative Logist distribution functionto describe the relationship between selection probabilityand explanatory variables and limited the value of explainedvariables ranging from 0 to 1 119901

119894= 119864(119884 = 1 | 119883

119894) =

1198901205730+1205731119883119894(1+119890

1205730+1205731119883119894) the odds ratio selection or not selection

is 119901119894(1 minus 119901

119894) taking logarithm of the odds ratio to obtain the

measurementmodel 119871119894= ln(119901

119894(1minus119901

119894)) = 120573

0+1205731119883119894+120576119894) has

more advantages than Probit (Probit model uses the standardnormal distribution function to describe the relationshipbetween the selection probabilities and the explanatory vari-ables and limited the value of explained variables rangingfrom 0 to 1 119901(119884

119894= 1) = Φ(119911

119894) = int119911119894

minusinfin(1radic2120587) exp(minus1199042

1198942)119889119904

then it implemented simple inverse function transformationand got the following 119911

119894= Φminus1(119901119894) = 120573

0+ 1205731119883119894+ 120576119894)

Therefore the estimation values of matching can be obtainedwith the following formula

ATTMatching =1

1198731

1198731

sum

119894=1

(1199101119894minus

1198730

sum

119895=1

1199081198941198951199100119895) (11)

1198730and 119873

1refer to the total number of individuals

119908119894119895means the differences between experimental group and

control group and it meets 119908119894119895isin [0 1] sum1198730

119895=1119908119894119895= 1

Meanwhile in order to select optimal control group com-mon support hypothesis is introduced here as 0 lt 119875(119889 = 1 |119883) lt 1 The cross region between experimental group andcontrol group is shown in Figure 2

23 Induction to MDID Model MDID model aims to findoptimal control group on the basis of DID model throughbringing inmatchingmodule andmatching algorithmThenvirtual processing individuals have been developed before theintervention implemented Last the virtual processing indi-vidual comes to the implementation of the double differencebased on the matching selection of control group

MDIDmodel has the following advantages when it comesto calculating the contribution rate (1) MDID could solvethe human factors generated by DID model (2) Match-ing algorithm could not solve time fixed effects but DIDcould (3) MDID allows the existence of invisible factors

Mathematical Problems in Engineering 5

Table 1 Basic principle of MDID

Before the policy change After the policy change Differential valueExperimental group 120573

0+ 1205732+ 120572 + 120575 120573

0+ 1205731+ 1205732+ 120574 + 120572 + 120575 Δ119910

119905= 1205731+ 120574

Control group 1205730+ 1205732+ 120575 120573

0+ 1205731+ 1205732+ 120575 Δ119910

119888= 1205731

Differential value mdash mdash ΔΔ119910 = 120574

(4) Matching process does not need lots of data whichmakes it possible to build larger optional libraries (5) MDIDmodel is relatively complete and easy for the inspection andinference

According to the DID model and matching estimatorMDID estimator is established as followsATTMDID

=1

1198731

1198731

sum

119894=1

(11991011198941199051

minus 11991011198941199050

) minus

1198730

sum

119895=1

119908119894119895(11991001198951199051

minus 11991001198951199050

)

(12)

In view of the above estimators considering the otherinfluencing factors in control variables the following panelmodel is constructed here to estimate ATT

119908119894119895119910119894119905= 1205730+ 1205731119905119894119905+ 1205732119909119894119905+ 120572119894119889119894119905+ 120574119905119894119905119889119894119905+ 120575119894+ 120583119894119905 (13)

Among them 119910119894119905is random variable119908

119894119895is weight and 119909

119894119905

is control variable 1205730 1205731 and 120573

2are coefficient vectors and

1205731means how the economic behaviors of the experimental

group and control group change with time under the policyintervention 120575

119894and 119906

119894119905are random disturbance items which

are not changing with time 120572119894is the difference between

control group and experiment group and it does not changewith time 119905

119894119905119889119894119905is the cross term of virtual variable 120574 is the

target variable and it is the change degree of the experimentgroup after accepting the policy

The original hypothesis of DID is 1198670 120574 = 0 when

there is no policy intervention 120574 = 0 and DID effects ofexperimental group and control group are 0

Table 1 summarizes the basic principle of MDIDmethod

3 Numerical Study

31 Analysis of Smooth Indicators

311 Qualitative Analysis Traffic smooth can be explainedfrom two aspects smooth state and smooth efficiencySmooth state can be understood as road unimpeded whichcould be represented by urban average traffic running speedUrban traffic smooth depends on traffic conditions of eachsection or road intersection in thewhole trafficnetworkGen-erally speaking lots of congestion reasons resulted in poortraffic smooth with the main reasons being the mismatchingand incompatibility of traffic demands and supplies

Trafficdemands aremainly influenced by urban economytraffic composition motorized travelling and other factorsAmong them urbanization process affects travel demandsfor the reason that more urban population and advancedeconomy mean greater travel demands Besides traffic formscould directly determine the motorized travel demands and

Table 2 Main impact indicators of traffic smooth

Category Indicator contents Relevance to the smoothThe socialeconomic

PopulationGDP Negative correlation

Trafficdemand

Motor vehicle ownershipCivilian car ownership Negative correlation

Traffic supplyRoad mileage road area Positive correlationPublic vehicle number Positive correlationPublic transportation

passenger Positive correlation

the smooth state of urban traffic Basic operation directlydetermines urban traffic intensity such as travel consumptionand distance

Traffic supply mainly contains urban traffic managementand planning road infrastructure related transportationservice facilities and so forth Urban trafficmanagement andplanning determine the spatial layout of urban transporta-tion such as road network structure and travel directionMoreover road infrastructure includes road mileage roadarea road network structure public transportation railtransportation and other facilities while the transportationservice facilities mainly include vehicle parking and vehicletransfer

Based on the above analysis and modeling necessity fac-tors whose data are quantifiable and accessible are selected toform the indicators as shown in Table 2

(1) Population and GDP Economic development will bringsubstantial increase in urban population and will also leadto a great travel demand Moreover economic developmentcould lead to the production of more purchasing power ofmotor vehicles Therefore the more development the urbaneconomy has gained the greater the traffic travel demandwillbe which also means higher requirement of transportationsupply and carrying capacity

(2) Number of Motor Vehicles As the visual expression ofroad traffic and parking demands vehicle amount largelydetermines the requirements ofmotor vehicle travel In termsof the slowly growing urban roads and tight urban landthe rapid growth of motor vehicles will provide the roadwith more carrying capacity which would quickly reachsaturation Besides the insufficient parking supply capacitywill directly result in severe urban traffic congestion

(3) Road Mileage and Road Area Urban road is the carrierof urban traffic and road mileage as well as road areas candirectly determine the carrying capacity of the road network

6 Mathematical Problems in Engineering

Table 3 Related data of traffic smooth in Beijing

Year ASPD GDP POP CVECH RODM RODA PVECH PPER2000 189 3161 1357 1041 2471 3502 10353 3482001 190 3711 1383 1145 2493 3701 12945 3922002 198 4330 1423 1339 2504 3857 15046 4352003 201 5024 1456 1631 3347 4315 16753 3712004 200 6033 1493 1824 4064 7287 18451 4392005 234 6970 1538 2097 4073 7437 18503 4502006 225 8118 1581 2391 4419 7286 19522 3982007 205 9847 1633 2734 4460 7632 19395 4232008 231 11115 1695 3137 6186 8940 21507 4712009 231 12153 1755 3681 6247 9179 21716 5172010 222 14114 1961 4497 6355 9395 21548 505Data sources Beijing Transportation Development Annual Report

Longer road and larger areas mean more driving smooth ofurban traffic However due to the restriction of urban plan-ning and land construction funds and many other factorsthe growth rate of urban roads is lower than that of motorvehicle Taking Beijing as an example the average annualgrowth of roadmileage is 1012 from2005 to 2010 Excludingthe large-scale construction during the 2008OlympicGamesthe average annual growth rate is only 3 which is less than13 of the average annual growth of motor vehicle

(4) Number of Public Vehicles After years of explorationoptimizing the traffic travel structure has become more andmore important Besides improving the carrying capac-ity and service level of public transport also becomes animportant way to solve the urban traffic congestion problemThe number of public vehicles which plays a positive rolein easing urban traffic congestion could reflect the supplycapacity of urban public transport

(5) Number of Public Transport Passengers Public transportpassenger can describe supply capacity of public transportfrom the perspective of transportation production Generallypeople believed that with greater amount of public transportpassenger urban transport operation would turn better Atpresent the main public transport travel modes are publicelectric vehicle (regular bus) mass transit and taxi

312 Quantitative Analysis The purpose of the quantitativeanalysis is to clarify the correlation among indicators andthe relationship between indicators and traffic smooth Asshowed in foregoing analysis the paper defines average trafficspeed of road network as the characterizing indicator ofurban traffic smooth Considering the regularity betweenindicators and data availability the paper collects the data ofBeijing from 2000 to 2010 for further analysis as shown inTable 3

Here the following symbols are used to represent thename of each indicator

ASPD average traffic speed of road network (kmh)GDP GDP (hundred million Yuan)

POP resident population (ten thousand people)CVECH number of civilian vehicles (ten thousandvehicles)RODM road mileage (km)RODA road area (ten thousand m2)PVECH number of public vehiclesPPER number of public transport passengers

Utilizing EViews 60 to implement correlation test thespecific results are shown in Table 4

The results showed that the correlation coefficient ofGDPPOP other factors and ASPD is all above 07 With highcorrelation these factors are main factors affecting averagetraffic speed of road network

Indicators are screened by referring to the existing evalu-ation achievements about traffic smooth both at home andabroad and following the metering modeling principles ofldquofrom general to simplerdquo In terms of screening the indepen-dent variables the CIA assumption in MDID model couldprovide some supporting functions independent variablesare influenced by ldquointerventionrdquo but it cannot influence theldquointerventionrdquo According to MDID assumption variableswhich cannot meet CIA assumptions will be eliminatedFinally Granger test is used as the specific implementationmethod

Step 1 Invoke the analysis results of speedrsquos factors andestablish timing database

Step 2 Conduct Granger test between influencing factorsand ASPD one by one

Step 3 Strip out the factors of ASPD Granger reason

The results show that Granger test result of traffic speedis significant (119875 value lt 005) Hence the number of publictransport passengersrsquo indicator is not suitable for the inde-pendent variable of speed

32 Building theMatching Pool DIDmodel needs a compari-son city while at the same timeGuangzhou city is the only ITSdemonstration project city in China Therefore the controlgroup cannot be the ITS demonstration project cityHoweverthere are two major drawbacks for the manual selectionof the control group cities First how much difference thecontrol group cities can tolerate It may be 10 or 30There is no objective standard Second in case of one singleselected city the results will heavily depend on the data ofthe chosen city Under the circumstances the introductionof matching process can solve these two issues effectivelyand provide theoretical support for selecting optimal controlgroup Before the execution of the matching algorithms thefirst thing is to select the cities that can be allowed to enterinto the matching pool

According to the model theory it should be assumedthat policy is exogenous in terms of selecting control groupThat is to say ITS only affects Guangzhou and has no effecton control group cities or the influences can be ignored

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Mathematical Problems in Engineering 5

Table 1 Basic principle of MDID

Before the policy change After the policy change Differential valueExperimental group 120573

0+ 1205732+ 120572 + 120575 120573

0+ 1205731+ 1205732+ 120574 + 120572 + 120575 Δ119910

119905= 1205731+ 120574

Control group 1205730+ 1205732+ 120575 120573

0+ 1205731+ 1205732+ 120575 Δ119910

119888= 1205731

Differential value mdash mdash ΔΔ119910 = 120574

(4) Matching process does not need lots of data whichmakes it possible to build larger optional libraries (5) MDIDmodel is relatively complete and easy for the inspection andinference

According to the DID model and matching estimatorMDID estimator is established as followsATTMDID

=1

1198731

1198731

sum

119894=1

(11991011198941199051

minus 11991011198941199050

) minus

1198730

sum

119895=1

119908119894119895(11991001198951199051

minus 11991001198951199050

)

(12)

In view of the above estimators considering the otherinfluencing factors in control variables the following panelmodel is constructed here to estimate ATT

119908119894119895119910119894119905= 1205730+ 1205731119905119894119905+ 1205732119909119894119905+ 120572119894119889119894119905+ 120574119905119894119905119889119894119905+ 120575119894+ 120583119894119905 (13)

Among them 119910119894119905is random variable119908

119894119895is weight and 119909

119894119905

is control variable 1205730 1205731 and 120573

2are coefficient vectors and

1205731means how the economic behaviors of the experimental

group and control group change with time under the policyintervention 120575

119894and 119906

119894119905are random disturbance items which

are not changing with time 120572119894is the difference between

control group and experiment group and it does not changewith time 119905

119894119905119889119894119905is the cross term of virtual variable 120574 is the

target variable and it is the change degree of the experimentgroup after accepting the policy

The original hypothesis of DID is 1198670 120574 = 0 when

there is no policy intervention 120574 = 0 and DID effects ofexperimental group and control group are 0

Table 1 summarizes the basic principle of MDIDmethod

3 Numerical Study

31 Analysis of Smooth Indicators

311 Qualitative Analysis Traffic smooth can be explainedfrom two aspects smooth state and smooth efficiencySmooth state can be understood as road unimpeded whichcould be represented by urban average traffic running speedUrban traffic smooth depends on traffic conditions of eachsection or road intersection in thewhole trafficnetworkGen-erally speaking lots of congestion reasons resulted in poortraffic smooth with the main reasons being the mismatchingand incompatibility of traffic demands and supplies

Trafficdemands aremainly influenced by urban economytraffic composition motorized travelling and other factorsAmong them urbanization process affects travel demandsfor the reason that more urban population and advancedeconomy mean greater travel demands Besides traffic formscould directly determine the motorized travel demands and

Table 2 Main impact indicators of traffic smooth

Category Indicator contents Relevance to the smoothThe socialeconomic

PopulationGDP Negative correlation

Trafficdemand

Motor vehicle ownershipCivilian car ownership Negative correlation

Traffic supplyRoad mileage road area Positive correlationPublic vehicle number Positive correlationPublic transportation

passenger Positive correlation

the smooth state of urban traffic Basic operation directlydetermines urban traffic intensity such as travel consumptionand distance

Traffic supply mainly contains urban traffic managementand planning road infrastructure related transportationservice facilities and so forth Urban trafficmanagement andplanning determine the spatial layout of urban transporta-tion such as road network structure and travel directionMoreover road infrastructure includes road mileage roadarea road network structure public transportation railtransportation and other facilities while the transportationservice facilities mainly include vehicle parking and vehicletransfer

Based on the above analysis and modeling necessity fac-tors whose data are quantifiable and accessible are selected toform the indicators as shown in Table 2

(1) Population and GDP Economic development will bringsubstantial increase in urban population and will also leadto a great travel demand Moreover economic developmentcould lead to the production of more purchasing power ofmotor vehicles Therefore the more development the urbaneconomy has gained the greater the traffic travel demandwillbe which also means higher requirement of transportationsupply and carrying capacity

(2) Number of Motor Vehicles As the visual expression ofroad traffic and parking demands vehicle amount largelydetermines the requirements ofmotor vehicle travel In termsof the slowly growing urban roads and tight urban landthe rapid growth of motor vehicles will provide the roadwith more carrying capacity which would quickly reachsaturation Besides the insufficient parking supply capacitywill directly result in severe urban traffic congestion

(3) Road Mileage and Road Area Urban road is the carrierof urban traffic and road mileage as well as road areas candirectly determine the carrying capacity of the road network

6 Mathematical Problems in Engineering

Table 3 Related data of traffic smooth in Beijing

Year ASPD GDP POP CVECH RODM RODA PVECH PPER2000 189 3161 1357 1041 2471 3502 10353 3482001 190 3711 1383 1145 2493 3701 12945 3922002 198 4330 1423 1339 2504 3857 15046 4352003 201 5024 1456 1631 3347 4315 16753 3712004 200 6033 1493 1824 4064 7287 18451 4392005 234 6970 1538 2097 4073 7437 18503 4502006 225 8118 1581 2391 4419 7286 19522 3982007 205 9847 1633 2734 4460 7632 19395 4232008 231 11115 1695 3137 6186 8940 21507 4712009 231 12153 1755 3681 6247 9179 21716 5172010 222 14114 1961 4497 6355 9395 21548 505Data sources Beijing Transportation Development Annual Report

Longer road and larger areas mean more driving smooth ofurban traffic However due to the restriction of urban plan-ning and land construction funds and many other factorsthe growth rate of urban roads is lower than that of motorvehicle Taking Beijing as an example the average annualgrowth of roadmileage is 1012 from2005 to 2010 Excludingthe large-scale construction during the 2008OlympicGamesthe average annual growth rate is only 3 which is less than13 of the average annual growth of motor vehicle

(4) Number of Public Vehicles After years of explorationoptimizing the traffic travel structure has become more andmore important Besides improving the carrying capac-ity and service level of public transport also becomes animportant way to solve the urban traffic congestion problemThe number of public vehicles which plays a positive rolein easing urban traffic congestion could reflect the supplycapacity of urban public transport

(5) Number of Public Transport Passengers Public transportpassenger can describe supply capacity of public transportfrom the perspective of transportation production Generallypeople believed that with greater amount of public transportpassenger urban transport operation would turn better Atpresent the main public transport travel modes are publicelectric vehicle (regular bus) mass transit and taxi

312 Quantitative Analysis The purpose of the quantitativeanalysis is to clarify the correlation among indicators andthe relationship between indicators and traffic smooth Asshowed in foregoing analysis the paper defines average trafficspeed of road network as the characterizing indicator ofurban traffic smooth Considering the regularity betweenindicators and data availability the paper collects the data ofBeijing from 2000 to 2010 for further analysis as shown inTable 3

Here the following symbols are used to represent thename of each indicator

ASPD average traffic speed of road network (kmh)GDP GDP (hundred million Yuan)

POP resident population (ten thousand people)CVECH number of civilian vehicles (ten thousandvehicles)RODM road mileage (km)RODA road area (ten thousand m2)PVECH number of public vehiclesPPER number of public transport passengers

Utilizing EViews 60 to implement correlation test thespecific results are shown in Table 4

The results showed that the correlation coefficient ofGDPPOP other factors and ASPD is all above 07 With highcorrelation these factors are main factors affecting averagetraffic speed of road network

Indicators are screened by referring to the existing evalu-ation achievements about traffic smooth both at home andabroad and following the metering modeling principles ofldquofrom general to simplerdquo In terms of screening the indepen-dent variables the CIA assumption in MDID model couldprovide some supporting functions independent variablesare influenced by ldquointerventionrdquo but it cannot influence theldquointerventionrdquo According to MDID assumption variableswhich cannot meet CIA assumptions will be eliminatedFinally Granger test is used as the specific implementationmethod

Step 1 Invoke the analysis results of speedrsquos factors andestablish timing database

Step 2 Conduct Granger test between influencing factorsand ASPD one by one

Step 3 Strip out the factors of ASPD Granger reason

The results show that Granger test result of traffic speedis significant (119875 value lt 005) Hence the number of publictransport passengersrsquo indicator is not suitable for the inde-pendent variable of speed

32 Building theMatching Pool DIDmodel needs a compari-son city while at the same timeGuangzhou city is the only ITSdemonstration project city in China Therefore the controlgroup cannot be the ITS demonstration project cityHoweverthere are two major drawbacks for the manual selectionof the control group cities First how much difference thecontrol group cities can tolerate It may be 10 or 30There is no objective standard Second in case of one singleselected city the results will heavily depend on the data ofthe chosen city Under the circumstances the introductionof matching process can solve these two issues effectivelyand provide theoretical support for selecting optimal controlgroup Before the execution of the matching algorithms thefirst thing is to select the cities that can be allowed to enterinto the matching pool

According to the model theory it should be assumedthat policy is exogenous in terms of selecting control groupThat is to say ITS only affects Guangzhou and has no effecton control group cities or the influences can be ignored

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

6 Mathematical Problems in Engineering

Table 3 Related data of traffic smooth in Beijing

Year ASPD GDP POP CVECH RODM RODA PVECH PPER2000 189 3161 1357 1041 2471 3502 10353 3482001 190 3711 1383 1145 2493 3701 12945 3922002 198 4330 1423 1339 2504 3857 15046 4352003 201 5024 1456 1631 3347 4315 16753 3712004 200 6033 1493 1824 4064 7287 18451 4392005 234 6970 1538 2097 4073 7437 18503 4502006 225 8118 1581 2391 4419 7286 19522 3982007 205 9847 1633 2734 4460 7632 19395 4232008 231 11115 1695 3137 6186 8940 21507 4712009 231 12153 1755 3681 6247 9179 21716 5172010 222 14114 1961 4497 6355 9395 21548 505Data sources Beijing Transportation Development Annual Report

Longer road and larger areas mean more driving smooth ofurban traffic However due to the restriction of urban plan-ning and land construction funds and many other factorsthe growth rate of urban roads is lower than that of motorvehicle Taking Beijing as an example the average annualgrowth of roadmileage is 1012 from2005 to 2010 Excludingthe large-scale construction during the 2008OlympicGamesthe average annual growth rate is only 3 which is less than13 of the average annual growth of motor vehicle

(4) Number of Public Vehicles After years of explorationoptimizing the traffic travel structure has become more andmore important Besides improving the carrying capac-ity and service level of public transport also becomes animportant way to solve the urban traffic congestion problemThe number of public vehicles which plays a positive rolein easing urban traffic congestion could reflect the supplycapacity of urban public transport

(5) Number of Public Transport Passengers Public transportpassenger can describe supply capacity of public transportfrom the perspective of transportation production Generallypeople believed that with greater amount of public transportpassenger urban transport operation would turn better Atpresent the main public transport travel modes are publicelectric vehicle (regular bus) mass transit and taxi

312 Quantitative Analysis The purpose of the quantitativeanalysis is to clarify the correlation among indicators andthe relationship between indicators and traffic smooth Asshowed in foregoing analysis the paper defines average trafficspeed of road network as the characterizing indicator ofurban traffic smooth Considering the regularity betweenindicators and data availability the paper collects the data ofBeijing from 2000 to 2010 for further analysis as shown inTable 3

Here the following symbols are used to represent thename of each indicator

ASPD average traffic speed of road network (kmh)GDP GDP (hundred million Yuan)

POP resident population (ten thousand people)CVECH number of civilian vehicles (ten thousandvehicles)RODM road mileage (km)RODA road area (ten thousand m2)PVECH number of public vehiclesPPER number of public transport passengers

Utilizing EViews 60 to implement correlation test thespecific results are shown in Table 4

The results showed that the correlation coefficient ofGDPPOP other factors and ASPD is all above 07 With highcorrelation these factors are main factors affecting averagetraffic speed of road network

Indicators are screened by referring to the existing evalu-ation achievements about traffic smooth both at home andabroad and following the metering modeling principles ofldquofrom general to simplerdquo In terms of screening the indepen-dent variables the CIA assumption in MDID model couldprovide some supporting functions independent variablesare influenced by ldquointerventionrdquo but it cannot influence theldquointerventionrdquo According to MDID assumption variableswhich cannot meet CIA assumptions will be eliminatedFinally Granger test is used as the specific implementationmethod

Step 1 Invoke the analysis results of speedrsquos factors andestablish timing database

Step 2 Conduct Granger test between influencing factorsand ASPD one by one

Step 3 Strip out the factors of ASPD Granger reason

The results show that Granger test result of traffic speedis significant (119875 value lt 005) Hence the number of publictransport passengersrsquo indicator is not suitable for the inde-pendent variable of speed

32 Building theMatching Pool DIDmodel needs a compari-son city while at the same timeGuangzhou city is the only ITSdemonstration project city in China Therefore the controlgroup cannot be the ITS demonstration project cityHoweverthere are two major drawbacks for the manual selectionof the control group cities First how much difference thecontrol group cities can tolerate It may be 10 or 30There is no objective standard Second in case of one singleselected city the results will heavily depend on the data ofthe chosen city Under the circumstances the introductionof matching process can solve these two issues effectivelyand provide theoretical support for selecting optimal controlgroup Before the execution of the matching algorithms thefirst thing is to select the cities that can be allowed to enterinto the matching pool

According to the model theory it should be assumedthat policy is exogenous in terms of selecting control groupThat is to say ITS only affects Guangzhou and has no effecton control group cities or the influences can be ignored

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Mathematical Problems in Engineering 7

Table 4 Correlation between indicators

Probability CorrelationASPD GDP POP CVECH RODM RODA PVECH PPER

ASPD 1000 0748 0703 0727 0805 0830 0826 0704mdash 0008 0016 0011 0003 0002 0002 0016

GDP 0748 1000 0982 0992 0967 0929 0893 08280008 mdash 0000 0000 0000 0000 0000 0002

POP 0703 0982 1000 0995 0934 0888 0851 08220016 0000 mdash 0000 0000 0000 0001 0002

CVECH 0727 0992 0995 1000 0955 0905 0867 08320011 0000 0000 mdash 0000 0000 0001 0002

RODM 0805 0967 0934 0955 1000 0956 0916 08270003 0000 0000 0000 mdash 0000 0000 0002

RODA 0830 0929 0888 0905 0956 1000 0937 08240002 0000 0000 0000 0000 mdash 0000 0002

PVECH 0826 0893 0851 0867 0916 0937 1000 08000002 0000 0001 0001 0000 0000 mdash 0003

PPER 0704 0828 0822 0832 0827 0824 0800 10000016 0002 0002 0002 0002 0002 0003 mdash

This hypothesis implied a few requirements (1) the modelallows observation factors (2) in addition to the policythe influences of remaining factors are the same (3) thecharacteristics of control and experimental group are steady

According to the above requirements and combined withthe situation of Chinarsquos administrative division 36 centercities are selected to make up the matching pool Howeverconsidering the outlier features of Tibetrsquos data the investiga-tion only uses Suzhou as the replacement of Lhasa

MDIDmodel requires that the cross-section datamust beimplemented before the policy Therefore we can choose thedata of the 36 cities in 2010 as alternative library cross-sectiondata as shown in Table 5

33 Analysis of Matching Process Based on the above dataProbit model is employed to do the robust regression Theresults are shown in Table 6

The results show that in matching Probit regression pro-cess the robust test of PVECHrsquos coefficient is nonsignificantand there is higher colinearity between PVECH and CVEHAfter the removal of PVECH the values of 1198772 and AICSICare experiencing obvious change

The colinearity problems of RODMand RODA cannot beignored and it is unsuitable for them to be put into the samemodelThus onlyRODMis retainedTherefore the statisticalmagnitude of the 4 variancesrsquo Wald 1205942 is 1989 the value of119875 is 00005 the likelihood of log pseudo is minus22760393 andpseudo 1198772 is 05019 with the fitting degree being acceptable

Then the matching process comes to the implementa-tion of Probit regression and calculating the value of PSUnder common support domain radius matching algorithmmethod (radius value is 01) is used to select controlgroup cities and calculate the weights of each city Thesignificant control group cities are Shenzhen Shanghai andSuzhou Then the Gaussian kernel density function 119870(119901

119894minus

119901119895119895isin119889=0

) = 119890minus(119901119894minus119901119895)221205902

is employed (setting Gaussianwidth parameter as 01) Last the weight of control groupcities is calculated according to the following formula

120596119894=

119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902

sum119870119894(10038171003817100381710038171003817119901119894minus 119901119895

10038171003817100381710038171003817119895isin119889=0) = 119890minus(119901119894minus119901119895)221205902 (14)

The results are shown in Table 7PS value of the remaining 32 cities is less than 0001 so it

is not listedFor guaranteeing the robustness of model doing loga-

rithm process to data and recalculating PS value 02265(Guangzhou) 04052 (Shenzhen) 01697 (Shanghai) and01371 (Suzhou) the results show that the model is providedwith good stability

Therefore a virtual city composed of 0403 Shenzhen0139 Shanghai and 0458 Suzhou has been set city A =

0403 lowast Shenzhen + 0139 lowast Shanghai + 0458 lowast SuzhouIt is supposed that each indicator of city A is the same toGuangzhou Then city A is set as the control group to cal-culate the contribution rate

After that contribution rate is calculated by collectingrelevant data of the experimental group (Guangzhou) andcontrol group (Shenzhen Shanghai and Suzhou) from 2008to 2012 as shown in Table 8

Contribution rate can be calculated by the changes of119910 stemming from the changes of 119909 divided by the totalchanges of 119910 According to Solow Growthmodel andmakinga difference of logarithm the coefficient of virtual variable isthe contribution rate Meanings of variables and indicatorsare shown in Table 9 Therefore GRP is set as the groupvirtual variable If GRP = 0 it means that the variable shouldbe in control groupAnd ifGRP = 1 itmeans that the variableshould be in the experimental group Set ITS as policy virtual

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

8 Mathematical Problems in Engineering

Table 5 Cities and corresponding data in matching pool

City GDP CVECH POP RODM RODA PVECHShijiazhuang 3401018600 90165300 100281 1475 4147 9065700Taiyuan 1778053900 60504800 35402 1780 2432 3935300Hohhot 1865711600 32155100 27876 720 1609 1018900Shenyang 5017540000 98431200 80465 2895 5706 1027500Dalian 5158160000 94488500 66382 2899 4135 714500Changchun 3329032900 66484500 75770 3659 6260 2393500Harbin 3664853800 65243500 99181 1427 3296 2161800Nanjing 5130649182 83052400 78603 5599 9576 1953900Hangzhou 5949168700 124805600 85197 2194 4754 7030400Ningbo 5163001700 87743400 74429 1439 2379 1668900Hefei 2961670000 38606000 54049 2013 4323 1670700Fuzhou 3123409200 43373400 69927 1101 2327 1110800Xiamen 2060073700 38539100 35313 1213 2994 798700Nanchang 2200105900 36242900 50266 965 1806 630300Jinan 3910527100 79735900 67462 4498 5907 1229700Qingdao 5666190000 97557100 86077 3409 5893 1554400Zhengzhou 4040892600 96301000 80904 1338 3158 2157200Wuhan 5565930000 104650000 94400 2682 7273 983800Changsha 4547057300 100867700 68435 1781 3618 2850200Guangzhou 10748282792 159893400 122896 6986 9731 739600Shenzhen 9581510100 166967400 101610 12613 8941 309900Nanning 1800261300 43219900 70262 1307 3205 2168100Haikou 617194800 23803800 15935 111 316 1880800Chengdu 5551333600 259930000 114437 2610 6460 581900Guiyang 1121817400 60442500 43052 872 1348 1141600Kunming 2120370000 132421500 63200 1420 2815 206700Xirsquoan 3241690000 95716200 84545 2428 5342 1898600Lanzhou 1100389800 24610000 32356 906 2162 914600Xining 628280000 30990000 22101 433 737 744800Yinchuan 769422700 22559700 18531 506 1652 620200Beijing 14113580000 449720000 191096 6355 9395 1158400Urumqi 1338517200 127140000 31100 1632 2005 1508500Chongqing 7925580000 114300000 287200 5130 9931 1430300Tianjin 9224460000 158240000 126373 5439 9159 729700Shanghai 17165980000 175510000 225648 4713 9299 2711400Suzhou 9228910000 126100100 99190 2904 5928 793700Data source urban passenger statistical reports of the Ministry of Transport

variables and GRP lowast ITS as cross terms and the final modelis shown as119908 lowast ASPD = 120573

0+ 1205731lowast RODM + 120573

2lowast CVECH + 120573

3

lowast GDP + 1205721lowast GRP + 120572

2lowast ITS + 120574

lowast GRP lowast ITS + 120583

(15)

34 Analyzing the Calculated Results The data above arenot in the symmetrical distribution so it is difficult for theconditional expectation of least squares model to accuratelyreflect the real situation Therefore the paper establishes thefollowing fixed effects estimation model and used quantileregression to achieve fitting

The estimation and test results of coefficient and varianceare shown in Table 10

The paper can obtain the following conclusions based onthe estimation results

(1) The coefficient of GRP lowast ITS could reflect the netcontribution rate of ITS project to GuangzhoursquosASPD and the value is about 00925 Hence thenet contribution rate of ITS project to the ASPDincreasing of Guangzhou is 925 Moreover 925of the effects are contributed by ITS project in termsof improving the traffic smooth

(2) Variance 1198772 = 07041 which means that 7041 ofthe total variation can be explained by the model

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Mathematical Problems in Engineering 9

Table 6 Regression results

Variable Coef Robust std err 119911 119875 gt 119911 [95 conf interval]GDP 137119864 minus 07 507119864 minus 08 27 0007 374119864 minus 08 236119864 minus 07

CVECH 701119864 minus 08 297119864 minus 07 024 0814 minus513119864 minus 07 653119864 minus 07

RODM minus00000411 00001591 minus026 0796 minus00003529 00002707

POP minus00091995 00039181 minus235 0019 minus00168789 minus00015201

cons minus3895322 06351935 minus613 0 minus5140279 minus2650366

Table 7 PS values and the weights

Significant city PS WeightGuangzhou 02465 1Shenzhen 02917 4028Shanghai 01034 1389Suzhou 03310 4583

with four variables and two free dummy variableswhile the other 30 is caused by factors that are notincluded in the model

(3) Table 10 shows that the intercept coefficient cons =0054003 Due to the fact that the control group can-not completely lose connection with ITS project thevalue indicating the effect proportion of ITS project tocontrol group is 54 and this value can be ignoredTherefore cons = 0054003 also fully illustrates thescientificity in selecting control group and supportingthe authenticity of the contribution results

(4) The coefficient of policy virtual variables is minus0031374which means that traffic smooth of control group isalso affected by time factor and other related factorsBesides traffic smooth level decreased by 314from the year 2008 to 2012 which shows that trafficsmooth of control group is relatively stable during theperiod and it is beneficial to use MDID contributionassessment model

(5) The regression coefficient of group virtual variable isminus0035766 It means that before the implementationof ITS project the difference of traffic smoothbetween control group and experimental group isonly 358 It also shows that the selection of controlgroup and experimental group in this paper is quitereasonable

(6) The coefficients of RODM CVECH and GDP are1085374 minus1867530 and 1865249 respectively prov-ing that the RODM and GDP have a positive corre-lation with traffic smooth and the correlation valueof GDP is higher than RODM In contrast CVECHhas a negative correlation with traffic smooth

The applications of ITS will inevitably exert a positiveinfluence on the urban transportation development How-ever with the rapid development of the city and coupled withthe effects of many other factors on ITS the urban trafficsmooth has not been significantly improvedTherefore based

on these conclusions the paper puts forward policy recom-mendations to improve urban traffic smooth from severalaspects

(1) The paper enhances the management level of urbantraffic and improve the existing road network struc-ture City should figure out the requirements of itsmarket economy take into consideration the currentsituation of its own traffic and promote reform ofits management system For example the city couldactively promote the separation of enterprise and gov-ernment so as to preferentially develop public trans-port strengthen the using of information technologyand develop multicore urban spatial structure More-over in terms of doing road network planning by thegovernment they should predict the traffic environ-ment in a long period in the future The situationwhere the occurrence of the later construction woulddeny the preoutcomes will be avoided Last perfectthe static traffic facilities such as establishing three-dimensional garage

(2) It improves traffic smooth under the condition ofensuring effective urban development process Thedevelopment of economy and increasing populationis bound to increase the number of motor vehiclesThus the traffic smooth state will be affected Thentraffic demand will be guided reasonably and thetraffic congestion will be eased by adopting theprice mechanism such as improving vehicle taxesand parking fees charging road congestion fees andimplementing public transport travel concessions

(3) It optimizes urban transportation infrastructureoperational state At present Chinarsquos motor vehicleownership has increased by 15 while the urbanroad has just grown by 3 This ratio is extremelyunreasonable Due to the limitation of urban landresources it is necessary tomaximize the limited landresources and road utilization rate For example com-mon cross overpass will be established and intelligenttraffic lights will be used to control the intersectionFirstly set intersection spacing of different roadreasonable levels and then strengthen intercon-nection rates between roads After that shortenthe travel distance furthermore install intelligenttraffic display and allocate the traffic flow to be morereasonable Finally it comes to reducing travel time

(4) It reasonably configures traffic composition modeFirstly improve facilities and services of public

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

10 Mathematical Problems in Engineering

Table 8 Data of experimental group and control group

City Year ASPD RODM POP CVECH GDP

Shenzhen

2008 2990 5849 9333250 1252747 77867920002009 2740 6035 9746450 1419005 82013176002010 2640 6184 10161050 1669674 95815101002011 2800 5977 10419718 1939653 115055298002012 2650 6015 10507432 2210821 12950100000

Suzhou

2008 3130 5013 8973807 840431 70780900002009 2570 5178 9247993 1027058 77402000002010 2580 5348 9918973 1261001 92289100002011 2590 5524 10493582 1518866 107169900002012 2600 5672 10533855 1778977 12011650000

Shanghai

2008 1698 15844 21021141 1321229 140698700002009 1770 16071 21754742 1471052 150464500002010 1664 16687 22564845 1755100 171659800002011 1564 16792 23250594 1947515 191956900002012 1470 17316 23639464 2126637 20181720000

Guangzhou

2008 2100 5434 10841750 1171731 82873816002009 1870 5519 11511550 1340540 91382135002010 2270 6986 12289649 1598934 107482827922011 2290 7072 12730500 1857740 124234390002012 2530 7100 12795143 2041592 13551200000

Data source statistical yearbook of the cities and traffic statistics yearbook

Table 9 Meanings of variables and indicators

Main variable Variable declarationDecision variable ASPDExplanatory variable CVECHControl variable RODM GDP

Policy variable119905119894119905= 1means after the implementation of

ITS project (2011-2012)119905119894119905= 0means before the implementation of

ITS project (2008sim2010)

Category variable119889119894119905= 1means the objects of experimental

group119889119894119905= 0means the objects of control group

transport which then can attract more people totake public transportation For example the layout ofpublic transport developing various forms of rapidtransit buses implementing different discount buscards at peak period and improving public transportshare ratio may be optimizing

4 Conclusions and Further Study

This paper reveals the influence factors and evaluationindexes of ITS on urban traffic smooth based on the analysisof ITS in terms of improvement of urban traffic smoothCombining the improved DID model we calculate thecontribution rate of ITS in improving urban traffic smoothThrough innovative study of the DID model the researchovercomes the application limitations of the DID model

which can serve as a useful tool for policy making of trans-portation department and evaluation of demonstrationproject Also the study provides new insights for quantita-tive research of urban transportation industry in terms ofinvestment and decision The result reveals that ITS worksto improve urban traffic smooth to some extent SpecificallyITS helps to relieve the traffic congestion through taking fulladvantage of the urban road area reduce vehicle travel timeand distance and improve the urban traffic smooth In anutshell the proposed quantitative analysis methodology inthis paper can provide some new insights for ITS decisionmakers for strategic planning purposes Moreover this studycan provide theoretical reference for the performance analy-sis of cityrsquos large demonstration projects and innovative ideasfor the technical processing of DID model

Although ITS has improved the traffic smooth to someextent the improvement of average travel speed is still notobvious as a result of the increase of motor vehicles In viewof the above the contributions of ITS cannot be ignored interms of improving the average travel speed Accordingly itis necessary to accelerate the construction of ITS and make apractical action in improving urban traffic smooth Althoughthe modeling methodology proposed in this study provides anew avenue for evaluating the effects of ITS on urban trafficsmooth further extensions are necessary

(1) The paper identifies the main indicators that affecttraffic smooth However it is difficult to quantifythe related degree between the indicators and ITSsince there are plenty of factors that influence trafficsmooth As such some high correlation indicators

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Mathematical Problems in Engineering 11

Table 10 Estimation results of the model

Variable Coef Bootstrapstd err 119905 119875 gt |119905| [95 conf interval]

RODM 1085374 0201156 5400000lowastlowast 0000 06303271 154042

CVECH minus1867530 0890793 minus2100000lowast 0065 minus3882644 01475842

GDP 1865249 0568868 3280000lowast 0010 05783794 3152118

GRP minus0035766 0050119 minus0710000 0494 minus01491443 00776114ITS minus0031374 0049013 minus0640000 0538 minus01422493 00795012GRP lowast ITS 0092490 0063868 1450000 0182 minus00519902 02369696cons 0054003 0094830 0570000 0583 minus01605178 02685238119905 statistics in parentheses lowast119875 lt 005 and lowastlowast119875 lt 001

may be ignored and this may further cause a signif-icant bias in the evaluation capability of the modelAccordingly it is worthy of more attention in a fur-ther study

(2) Because of the special characteristics of ITS such ascomplexity dynamism and the hysteresis of appli-cation effects one key problem is to determine thestarting time-point of evaluation firstly An appro-priate starting time-point can guarantee that theevaluation results reflect the application benefits ofITS accurately Therefore in the future research it isnecessary to study the evaluation time-point of ITSapplication effects Also it is necessary to choose thepoint data to evaluate the contribution rate and tostrengthen the scientific rationality of the results

(3) Although the MDID model works well in the selec-tion of control group it has high requirements of theindividuals that enter the matching poolTherefore itneeds to strengthen the discussion about the selectionof control group in further research As such DIDmodel can be preferably used in traffic fields

Conflict of Interests

All the authors have declared that they have no conflict ofinterests to this work

References

[1] S E Underwood and S G Gehring ldquoFramework for evaluatingintelligent vehicle-highway systemsrdquo Transportation ResearchRecord vol 1453 pp 16ndash22 1994

[2] D C Novak and M McDnald ldquoA general overview of thepotential macroeconomic impacts of ITS investment in theUnited Statesrdquo inProceedings of the AnnualMeeting of the Trans-portation Research Board Washington DC USA January 1998

[3] D B Lee Benefit-Cost Evaluation of ITS Projects SeattleWSDOT Web Site (SE-26) FHWAFTA Joint Program OfficeUSDOT 1999

[4] D B Lee and L R Klein Benefit-Cost Evaluation of ITS TravelerInformation Kiosks Volpe Center USDOT 1997

[5] R Zavergiu ldquoIntelligent transportation systemsmdashan approachto benefit studiesrdquo Tech Rep TP 12695E TransportationDevel-opment Centre Montreal Canada 1996

[6] D Brand ldquoApplying benefitcost analysis to identify and mea-sure the benefits of intelligent transportation systemsrdquo Trans-portation Research Record no 1651 pp 23ndash29 1998

[7] E A Beimborn Z Ren-peng and M Neluheni ldquoA frameworkandmethods for evaluation of benefits of intelligent transporta-tion systemsrdquo Transportation Research Board Record 17772001

[8] H-B Xiao ldquoEvaluation index system and evaluation methodson city integrated transport systemrdquo Transportation Science ampTechnology vol 234 no 3 pp 87ndash90 2009

[9] Q Zhao S-M Huang S-X Di and G-S Lu ldquoComprehensiveevaluation index system of urban public traffic systemrdquo Trans-port Standardization vol 208 no 11 pp 83ndash86 2009

[10] B-W Chen and X-F Liao ldquoResearch on evaluation indexsystem of highway transportation sustainable developmentrdquoChina Journal of Highway and Transport vol 22 no 5 pp 112ndash116 2009

[11] L Cao W-X Zha W-T Zhao and S-B Wang ldquoEstablishingtransfer evaluation index system of BRTrdquo Transport Engineeringand Safety vol 246 no 11 pp 126ndash128 2011

[12] P Leviakangas and J Lahesmaa ldquoProfitability evaluation ofintelligent transport system investmentsrdquo Journal of Transporta-tion Engineering vol 128 no 3 pp 276ndash286 2002

[13] S P Mattingly Decision Theory for Performance Evaluation ofNew Technologies Incorporating Institutional Issues APP Lin-eation to Traffic Control Implementation University of Califor-nia Berkeley Calif USA 2000

[14] Z Wang and C M Walton ldquoA multicriteria approach forevaluation and alternative selection of ITS deploymentsrdquo inProceedings of the 11th Annual World Congress on IntelligentTransportation Systems (ITS rsquo04) Nagoya Japan 2004

[15] Q Zhang J-L Shen and C Li ldquoBidding appraisal mechanismand synthetic appraisal method in ITS projectsrdquo Journal ofTransportation Systems Engineering and Information Technol-ogy vol 5 no 4 pp 18ndash29 2005

[16] J M Wooldridge Introductory Econometrics A Modern Ap-proach South-Western College Publishing 2nd edition 2002

[17] C Ai and E C Norton ldquoInteraction terms in logit and probitmodelsrdquo Economics Letters vol 80 no 1 pp 123ndash129 2003

[18] P A Puhani ldquoThe treatment effect cross difference and interac-tion term in nonlinear lsquodifference-in-differencesrsquo modelsrdquo IZADiscussion Paper 2 2008

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

12 Mathematical Problems in Engineering

[19] D M Drykker ldquoTesting for serial correlation in linear panel-data modelsrdquoThe Stata Journal vol 3 no 2 pp 168ndash177 2003

[20] S Galiani P Gertler and E Schargrodsky ldquoWater for life theimpact of the privatization of water services on child mortalityrdquoJournal of Political Economy vol 113 no 1 pp 83ndash120 2005

[21] M Conti ldquoThe introduction of the euro and economic growthsome panel data evidencerdquo Journal of Applied Economics vol 17no 2 pp 199ndash211 2014

[22] R A Lewis and D H Reiley ldquoOnline ads and offline salesmeasuring the effect of retail advertising via a controlledexperiment on YahoordquoQuantitative Marketing and Economicsvol 12 no 3 pp 235ndash266 2014

[23] J Liang and Y Zhou ldquoAn empirical study on GEM and enter-prises innovationrdquo Science Research Management vol 34 no 2pp 89ndash92 2013

[24] Y-L Wu ldquoSoil and water conservation projects and nationalfood safetymdashevidence from the panel data of northeast blacksoil areasrdquo China Population Resources and Environment vol24 no 5 pp 304ndash308 2014

[25] H Wang ldquoNew farming maintain to evaluate the contributionof the rural residentsrsquo consumption growth in Chinamdashbased onDID modelrdquoModern Business no 2 2014

[26] M Deng and Q-H Wang ldquoRelationship between Chinarsquosstrategic introduction of overseas investment and X-efficiencyof Chinese-funded banksrdquo International Business no 2 pp 21ndash30 2014

[27] A Tzedakis ldquoDifferent vehicle speeds and congestion costsrdquoJournal of Transport Economics and Policy vol 14 no 1 pp 81ndash103 2005

[28] C Wright and P Roberg ldquoThe conceptual structure of trafficjamsrdquo Transport Policy vol 5 no 1 pp 23ndash35 1998

[29] H Rehborn and M Koller ldquoA study of the influence of severeenvironmental conditions on common traffic congestion fea-turesrdquo Journal of Advanced Transportation vol 48 no 8 pp1107ndash1120 2014

[30] AWaadt ldquoTraffic congestion estimation servicemobile assistedpositioning schemes in GSM networksrdquo Procedia EarthandPlaetary Science vol 1 no 1 pp 1385ndash1392 2009

[31] Y-L Li and CWang ldquoUrban planning and population densityfactors affecting urban traffic development researchrdquo XinjiangSocial Science no 3 pp 42ndash46 2013 (Chinese)

[32] M Anisimova ldquoDarwin and Fisher meet at biotech on thepotential of computational molecular evolution in industryrdquoBMC Evolutionary Biology vol 15 article 76 2015

[33] P R Rosenbaum and D B Rubin ldquoThe central role of thepropensity score in observational studies for causal effectsrdquoBiometrika vol 70 no 1 pp 41ndash55 1983

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of