Graduate Research Award Program Application: 2OtO-2OIl Page 2 of 2

Application Checklist and Application Packaqe Gover Sheet

Applicant's name: Yaze Vikrant Suhas Date'.0511712010Last First Middle

I x ] A completed application checklist and application package cover sheet

I x ] Personal information of applicant

I x ] Qualifications of applicant

I x ] Research project proposal

I x ] Reference letter #1

I x j Reference letter #2

I x ] Research advisor form

tx I Official transcripts from all colleges and universities attended

Please list all institutions from which official transcripts have been requested. Transcripts may be

included in the submission packet if properly sealed from the registrar.

lf official transcripts will be sent separately, please indicate that they have been ordered and

include an unofficial copy of each of your transcripts in the application package.

Ix ] Writing Sample

(The writing sample should be submitted with the application. An appropriate writingsample might be a previous publication, a paper written for a class assignment, a

research project report, or similar example of professionalwriting. lf papers were co-

authored, the role of the applicant must be clearly described. Writing samples may not

be longer than 25 pages.)

Graduate Research Award Program Application : 2QLO-20L1 Page 1 of 3

APPLICATION FORM

GRADUATE RESEARCH PROGRAM ON PUBLIC-SECTOR AVIATION ISSUES

Sponsored By: Administered By:FederalAviation Administration Airport Cooperative Research ProgramU.S. Department of Transportation Transportation Research Board, NationalAcademies

PART I_ PERSONAL INFORMATION OF APPLICANT

/Plcacc Tvncl

1. Full legal name:Vaze Vikrant Suhas

2.

3.

4.

5.

6.

7.

8.

o

'10.

11.

Last First

Date of birth: 06/28l'1983

College or University currentlyMajor Field: Civil EngineeringDegree objective: [ ] Master's

Middle Former name (if any)

Place of birth: Mumbai (lndia)

enrolled at: Massachusetts lnstitute of Technology-

Ix ] Doctorate

Citizenship: lndian

Gender: [x]Male [ ] Female

Ethnicity (optional) :

I American lndian or Alaskan Native: origin in any of the original peoples of North America

I Black: origin in any of the black racial groups

I Hispanic: Mexican, Puerto Rican, Central or South America, or other Spanish culture or origin,regardless of race

x I Asian or Pacific lslander: origin in any of the original peoples of the Far East, Southeast Asia,

orthe pacific lslands. lncludes China, Japan, Korea, the Philippine lslands, Samoa, and thelndian Subcontinent

I White: orig in in any of the original peoples of Europe, North Africa, or the Middle East

Mailing address: 88, Columbia Street, Apt #1, Cambridge MA 02139

Permanent address: 88, Columbia Street, Apt #1, Cambridge MA 02139

Telephone numbers - Mailing: (617)777 4031- Permanent: Mailing: (617)777 4031-

Email address: vikrantv@mit.edu

Expected month and year of graduation: August 201 1

Names of two people from whom you are requesting reference letters.

a. Prof. Cynthia Barnhad

b. Prof. Amedeo Odoni

12. Name and title of faculty research advisor for this project:

Prof. Cynthia Barnhart, Associate Dean of Engineering for Academic Affairs, School ofEngineering, Massachusetts lnstitute of Technology.

Graduate Research Award Program Applicat¡on: 2010-2011 Page 2 of 3

PART II - QUALIFICATIONS OF APPLICANT

College / University Location Major FieldDates

AttendedGPA Degree

Date degreeawarded/expected

Vlassach usetts I nstitute¡f Technoloqv

)ambridge,VIA

livil=noineerino

Sept 2008 -)resent i.00/5.00 )hD \ug 201 I

Vlassachusetts I nstituterf Technoloqv

)ambridge,VIA

lransportati)n

Sept 2005 -June 2007

i.00/5.00 VIS )une 2007

ndian lnstitute offechnoloov. Bombav

Vlumbai,ndia

livi!inqineerinq

l¡rhr ?ôôl -Mav 2005

1.83/10.00 3S Sept 2005

(Please Type)

13. Education: ln reverse chronological order, list colleges or universities attended

Please explain any interruption(s) of schooling, i.e., military training, illness, etc.)

14. Professional Experience. ln reverse chronologicalorder, list professionalexperience, including

summer and term-time work.

15. Awards, honors, and publications: List fellowships, scholarships, and other academic and/orprofessional positions, held since entering college or university.

Name of Emplover Location Dates Nature of Work

Lehman Brothers lnc New York, NY 3o'n July 2oo7 -22nd Aug2008

Quantitative andAlgorithmic Trading:Strategy Development,Mathematical Modelinq

Sabre Airline Solutions Southlake, TX 30'n May 2006 -18th Aug 2006

Operations ResearchAnalyst lntern: StatisticalModelinq, Optimization

Swiss Federal lnstitute ofiechnology (ETH)

Zurich, Switzerland 5"' May 2005 -l oth July 2oo5

Summer Intern: TravelBehavior Modeling,Nonlinear Optimization

Budapest University ofTechnology and Economics(BUrE)

Budapest, Hungary 10'n May 2004 -gfh July 2oo4

Summer lntern: GlobalPositioning SystemPerformanceMeasurement, SoftwareDevelopment

Award, Honor, or publication Date(s) Description

)residential Fellowship, Massachusettsnstitute of Technology

2005-2006

\dministered by the MIT office of the Provost,)residential Fellowships fund the tuition andiving stipend of awardees for their first academicrear at MlT.

)resident of lndia Gold Medal iept 2005\warded for best academic performance amongall graduating students of all disciplines in lndiannstitute of Technoloqy, Bombav

3est Undergraduate Thesis Award iept 2005 ndian lnstitute of Technology, Bombay

nstitute Medal2,2003,4.2005

est academic performance in the Department ofivil Enqineerinq for all 4 years of college

rlational Talent Search Scholarship 1 999-2005\warded inl999, by the National Council oflducational Research and Training (lndia).ncludes a monthly stiPend.

raper Publication in the lnternationalJournal of Revenue ManagementAccepted)

\pril 2011,laze,V. and C. Barnhart (2011). An Assessmentrf the lmpact of Demand Management Strategiesbr Efficient Allocation of Airport Capacity.

raper Publication in the Transportationì,esearch Records: Journal of theIransportation Research Board

July 2009

,laze,Y., C. Antoniou, Y. Wen and M. Ben-Akiva2009). Calibration of Dynamic Traffic\ssignment Models with Point-to-Point Trafficiurveillance.

raper Publication in the Transportationìesearch Records: Journal oftheIransportation Research Board

Jan 2008Rai, R., M. Balmer, M. Rieser, V. Vaze, S.

Schonfelder and K. Axhausen (2008). Capturingluman Activitv Spaces: New Geometries.

Graduate Research Award Program Appl¡cation: 2010-2011 Page 3 of 3

.,r ña¡nrihô \,^,,r ..,rêêr nnale anr-{ hor¡¡ this research will contribute to achieving those goals (youlv. ugùvllw9 yvul vqlvv¡

may add Page(s) if necessarY):

My long term career goal is to make a meaningful contribution to the field of transportation and

*ôoititf towards solviñg the pressing problems of societal interest. Working as a faculty member

at a toþ university will ñot only prov¡de a high quality research environment but also present an

opportun¡ty to traln the future transportation professionals. Therefore, in the intermediate term,

my objective is to obtain a faculty position at one of the leading universities in transportation

r"r."r"n. My doctoral research, !o far, has focused on air transportation. I believe that this

project will bê a critical step towards attaining my intermediate and long term objectives.

Over the next several decades, enabling the transportation of the growing number of travelers

around the globe will constitute a grand challenge to the mankind. New paradigms leading to

the developñrent of new and innovãtive modes, vehicles and infrastructure will be required to

meet this challenge. At the same time, efficient utilization of available resources and newly built

infrastructure wii be of paramount importance. Several insightful developments in the

disciplines such as micro-economics, operations research, statistics, game theory, algorithms

and computer science offer a set of useful tools for making informed decisions. My current

research and future research interests revolve around the applications of these theoretical

developments in order to address practical problems in the field of transportation and mobility.

I believe that this project will be helpful to me in multiple ways. The project will allow me to

interact with imporlani stakeholders including the airlines, airport operators and the Federal

Aviation Administration (FAA). Such interactions and discussions will provide a strong

foundation for understandìng the practical aspects of the problem. These will serve as the first

steps in the process of being able to transfer academic research into practice. I also hope that

the research, along with ihe associated industry interactions, will create future funding

opportunitier îor fuñier projects on mobility research. On the academic front, I expect that the

project will result in two públications, a technically focused paper aimed at publication in ajouinat such as Transportation Science or Operations Research; and a more application-'oriented

paper targeteä at a journal such as Air Traffic Control Quarterly or the recently

instituted AGIFORS Journal. overall, the project will serve as an important stepping stone for

beginning my career as a transportation faculty member'

I af¡rm that the information provided in this application is true and complete to the best of my knowledge.

[NOTE: This application is not complete without a signature.]

Date.o..ã11.8 /Z-a\o

Graduate Research Award Program Application: 2010-2011

APPLICATION FORM

GRADUATE RESEARCH AWARD PROGRAM ON PUBLIC-SECTOR AVIATION ISSUES

Page 1 of 2

Sponsored By:Federal Aviation AdministrationU.S. Department of Transportation

Administered By:Airport Cooperative Research ProgramTransportation Research Board, National Academies

PART III -RESEARCH PROJECT PROPOSAL

(Please Type)

Name: VazeLast

Title of Research Project:

Vikrant SuhasFirst Middle

Dale'. 0511712Q10

The Role of Airline Competition in Airport Congestion Mitigation using Demand Management

ln 500 words or less, describe the proposed research project. lnclude project objectives,

methodology, and expected outcomes. Also indicate how this research work could benefit the

aviation community, and contribute to your career goals.

Airport congestion is imposing a tremendous cost on the world economy. Due to increasing

demand for the limited number of airport slots, demand often exceeds the available capacity at

the congested airports. Demand management measures such as administrative slot-controls,

market mechanisms or any combinations thereof, have the potential to restore the demand-

capacity balance over a medium-to-short time horizon with relatively small investment. For a

t.,...trfu| implementation of a demand management scheme, the preferences of the various

concerned stákeholders need to be taken into account. Several past proposals for the

implementation of market-based schemes have been fiercely opposed by a variety of players,

most notably by the airlines. A key ingredient missing in the discussion of these proposals has

been the role of frequency competition among the airlines. Recent research by Yaze and

Barnhart (2010, 20llt indicates that such competition is an important reason for the inefficient

utilization of airport resources. The main objective of the proposed research is to investigate the

role of airline còmpetition under demand management and to evaluate the impact of different

demand management schemes on the various stakeholders including airlines, airport operators,

passengers, as well as on the society as a whole.

prior research has often hinted at the possibility that all the stakeholders, including the airlines,

can be better off if the demand management schemes are implemented efficiently. This project

aims to substantiate these claims through game-theoretic analysis of the frequency competition

between airlines. The scheduling problem for an individual airline is traditionally modeled as a

profit maximiza1ion problem. Hòwever, because the prof,rtability of an airline depends not only

ãn its own schedule ùut also on the competitors' schedules, I plan to model the set of competing

airlines as a system of profit-maximizing agents and solve it using the Nash equilibrium solution

Graduate Research Award Program Application: 2O7O-2OL| Page 2 of 2

concept. I plan to validate the model estimates by comparing with actual frequency data' Using

these modéls, I plan to evaluate the airlines' responses to a few candidate demand management

strategies, incluãing more efficient administrative slot controls and some simple forms of

conge-stion-based lãnding fees. The expected outcome of this research is an evaluation

framework for estimating and comparing the impacts and effectiveness of different demand

management strategier. ih. evaluation will include quantifying the impacts in terms of delay

costsãnd profitabiñty to the airlines; in terms of the delay costs, fares and level-of-service to the

purr.rrg.rr; and in teims of the queuing delays and emissions at the airports' This research can be

Ùeneficial for testing and evalùating different demand management schemes before actually

being implemented, and thus, will serve as a tool for aviation policy decisions.

So far in my doctoral research, I have focused my attention on understanding the effects of

airline competition on the aviation resource utilization. This project will serve as a natural

follow-up in tirat ciirection. in the iong rerm, i wouid iike to be at tlie forefront of multi-

disciplinãry research in the f,reld of transportation, especially geared towards applying the

concãpts and ideas from varied fields including economics, operations research, game theory,

statistics and computer science, to solve the major problems being faced by the transportation

industry. I believè that working on this project will be an invaluable contribution towards

achieving my long-term career goals.

References:o Vaze, V and C. Barnhart (2011). An Assessment of the lmpact of Demand Management

Strategies for Efficient Allocation of Airport Capacity (Accepted). lnternational Journal of

Revenue Management.¡ Vaze, V and C. Barnhart (2010). Price of Airline Frequency Competition. Technical Paper.

Department of Civil and Environmental Engineering, MIT'

IPlease attach additional sheets as needed']

C"pÐl

Graduate Research Award Program Application: 2010-2011 Page 1 of 2

APPLICATION FORM

GRADUATE RESEARCH AWARD PROGRAM ON PUBLIC.SECTOR AVIATION ISSUES

Sponsored By.Federal Aviation AdministrationU.S. Department of Transportation

Administered By.Airport Cooperative Research ProgramTransportation Research Board, National Academies

PART V - FACULTY RESEARCH ADVISOR

To be completed by the apolicant:

NOTE: ln order to enrich the educational experience gained from your proposed research proiect, it isnecessary for you to request a facutty member from your university who is familiar with your researchproject to act as a research advisor to you during the course of the proiect. Please provide the following'information

and ask the facutty member to complete the form. You should submit it with your application.

The research advisor may also be a reference respondent.

Applicant's Name: Vaze Vikrant Suhas-Date: 05/10/2010Last

Title of Proposed Research Project:

First Middle

The Role of Airline Competition in Airport Congestion Mitigation using Demand Management

To be completed by the faculty research advisorNOTE: The research product of research award recipients can be considerably enhanced if a facultymember at the appticant's university acts as an advisor to the applicant during the conduct of the

research. Therefore, each applicant is required to designate such an advisor who will be available to

him/her throughout the course of the research project to provide advice as it progresses. When research

papers by award winners are pubtished by the Transportation Research Board, the faculty member will be

identified as the research advisor.

Faculty Research Advisor's Name:

-Prof. Cynthia Barnhart

Title: Associate Dean for Academic Affairs, School of EngineeringDepartment: _School of EngineeringUniversity:

-Massachusetts

lnstitute of TechnologyMailing Address:_Room 1-206,77 Massachusetts Avenue, Cambridge, MA 02139-4307Email: cbarnhar@mit.edu Phone:_(617) 253-381 5

1. Have you examined the applicant's proposed research plan? Yes -X-

No

2. Do you consider the applicant's research plan reasonable? Yes -X- No

lf no, please comment.

Page 2 of 2

Graduate Research Award Program Application: 2010-2011

PARTV: Page2

3.Doyoubelievethatthisapplicantcancompletetheproposedresearchwithinthetimeframeindicated? Yes

-x- t'lo

-'

lf no' please comment'

4. wiil the appricant receive academic credit for this work? Yes

-

No -X. .-

lf yes' please

:...¡i¡ala thê nât!!!.e of this aeademic "'"äù.1üiá: ceceiving õ'-mic cledit in no way counts

lllulrJcltç !r rv I ru¡vr

against the aPPlicant'l

5. Please indicate briefly how you plan to monitor and advise on the work of the applicant on this

project.

I will meet at least weekly with Vikrant to review progress and discuss next steps' Additionally' I

wiil ask Vikrant to document his work i. tnã'i"irïiinesis ch;pi;rs and journai paper drafts' and I

wi* provide him feedback on his *orx. 'Èinäir},"i *it-nérp viränt to Rno ine right meetings' etc' for

n¡äi" *óótt his findings to the community'

6'lamwillingtobetheresearchadvisortotheapplicantiftheapplicantreceivesthisresearchaward. Yes

-X- No

- oate, zbU /4 tutÒSignature:

Graduate Research Award Program Application: 2010-2011 Page 1 of 2

APPLICATION FORM

GRADUATE RESEARCH PROGRAM ON PUBLIC-SECTOR AVIATION ISSUES

Sponsored By: Administered BY:

FederalAviaiion Administration Airport Cooperative Research Program

U.S. Department of Transportation Transportation Research Board, NationalAcademies

PART IV. REFERENCE LETTER ON APPLICANT

(Please Type)

This section to be completed bv Applicant:

Name of Applicant'. YazeLast

Vikrant Suhas- Date 05/05/2010First Middle

Applicant's major field of studY Givil Engineer¡ng-

Title of Proposed Research Project: j

The Role of Airline Competition in Airport Congestion Mitigation using Demand Management

This section to be completed bv reference respondent:

NOTE: This appticant has named you as one of two people who know his/her academic and professional

experience and ability. Your views witt help us evaluate this applicant's qualifications for receiving an

award for conducting research on the above proiect.

ptease complete this form, make another copy, and place each copy in a separate envelope. Seal

both envelopes, sign each across úhe sea/, and return BOTH envelopes to the applicantforinclusion witn tne apptication to be submitted to the Transportatign Research Board.

Name of Reference Respondent: Prof. Cvnthia Barnhart

Tifle: Associate Dean for Academic Affairsr school q[ Enqineerinq,/Organization: Massachusetts lnstitute of Tech4oloqv

Mailing address: Rm 1-206 '- 77 Massachusetts Avenue. Cambridqe. MA 021394307

Emailaddress: cbarnhar@mit.edu Phone: (617) 253-3815

ln what capacity do you know the applicant? lserve as his research supervisor. doctoralthesis

supervisor, teachino assistant supervisor.

How long have you known the applicant? Since Fall 2005

Graduate Research Award Program Applicationl. ZOLO-ZOLI Page 2 of 2

PART lV: Page 2

1. please evaluate the applicant in the following areas as compared with other individuals of

comparable training, age and experience.

Z. please comment on the ability of the applicant to carry out the proposed research in a timely

manner. I have extreme -confiflence

that Vikrant can carry out this work in a timely

manner. Vikrant is perhaps the most productive graduate student I have encountered in

my career.

3. please add any other comments that you consider to be pertinent to the evaluation of the

applicant and that are not covered adequately by your other answers. Attach additional sheet(s) if

nää"rsary. Vikrant is one of the best, if not the best, graduate research assistant with

whom I have worked tn my 22 years as a faculty member'

CONFIDENTIALITY: The information contained in this reference letter shall not be available to the

applicant or otherwise publicly disclosed except as required by law.

Signature of Reference Respondent: Date: May 6, 2010

Ì rfefendinr Aboveaverage

AverageBelow

average

lnsufficientopportunity to

observeJU tÞlal lu ll lL

(nowledqe of major field xìesearch skills x)roblem solving skills X\---¿:.,:¿,,rlEiál,lVll,! Y

-eadership xlVritten communication x)ral communication x

Graduate Research Award Program Appiication: 2010-2011 Fage 1 of 2

APP¡-[GÁ,T!Oh¿ FORTW

GRADUATE RËSEARTI-I PROGRAM ON'¡ FUBLIC-SECTOR AVIATIOF{ ISSUES

Sponsored By: Administered By:Federal Aviation Administration Airport Cooperative Research Frogramu.S. Department of Transporiation Transportation Research Board, Nlational Acadernies

FART IV - REFËRENCE LËTTER ON APPLICANT

(Please Type)

This section to be çompleted by Applicant:

Narne of Applicant: VazeI ^^¿

,Viårrant SuhasFirst h¡liddle

Date t5/û5/2t10

Applicant's nnajor field of str,rdy Civil Ëngineering

Titte of Proposed Research Projeet:

The Role of Airline Competition in Airport Congesiion Mitigation using Demand lVlanagement

This section to be comnleted bWeference respondent:

Ê'lOï8. This appticant has narned yau as one af flrua peaple who knaw hislher academic and professianalexperience and abitity. Yaur views wiil help us evaluate this applicant's qualifications far receiving anaward for conductingr research an the above praject.

Pfease cormpfefe fftrs fonra, rmake *mofÉ¡ er cøpy, amd pface eaafi aopy ím a seBarafe emvefope. Seafbotfu enve!øpes, sfgtn eacf¡ æcross ffie seæd, amd return ÆÕf"Ff enwefopes fo úfie appßi*awt farinclusiøn witk tke appiicatian fo he s¿¡þm itted tø ffte lranspo rtatian Researcf¡ &oard.

Name of Reference Respondent. Amedeo Odoni

Title: Professor of Aeronautics and Astronautics and Professor of Civil and Ënvironmental Ëng'g

Organizaiion: Massachusetts Institute of Technology

ft/ïailing address: Room 33-219, MlT, Cambrídge, MA 02139

Email address: _arodoni@mit.edu Phone: (617)-253-7439

In what capacity do you know the applicant? _Teacher, research supervisor, my teaching assistant_[-low long have you known the applicant? 0 years

Graduate Research Award Frogranr Application: 2010-2011 Page 2 of 2

PART lV: Page 2

1. Please evaluate the applicant in the following areas as compared with other individuals ofcomparable training, age and eNperience.

)utstandinç Aboveaverage

Average Belowaverage

lnsufficientopportunity to

observe(nowledge of major field X

ìesearch skills X)roblem solving skills X

reativity X

-eadership X

/üritten communication X

Jral communication X

2. Flease comment on the ability of the applicant to carry out the proposed research in a timelymanner.

The evaiuation ìn the above Table (ltem '1) may seem exceptionally strong. This is because Vikrant is atruly exceptional student. l-'le is probably the best graduate student (FhD or Master's level) currently atMIT.

3. Please add any other cornments that you consider to be pertinent to the evaluation of theapplicant and that are not covered adequately by your other answers. Attach additional sheet(s) ifnecessary.

As indicated under ltem 2 above Vikrant is truly exceptional. l-{e is working on a very interesting anddifficult PhD thesis topic (also the subject of this application for an ACRF fellowship) in which he isstudying (a) how airline competition with respect io flight frequency has affected airport congestion and(b) what impact alternative dernand management strategies might have in this respect. This is a subjectof critical importance to national policy-makers in the air transportation area. l-le has been able to puttogether a terrific analysis, which combines many diverse methodologies (optirnization, game theory,queuing theory, economics) all of which he can handle in expert fashion. He is also in the process ofputting together three excellent working papers based on his research to date - all of which will, I amsure, become first-rate journal publications. After some polishing, these papers will contain more thanenough material for a first-rate PhD thesis -- and he still has a year to go untll he graduates!

Of the numerous PhD students (more than 100) ihat I have supervised or worked closely with over 41years on the MIT faculty, only a handful had impressed me as nruch as Vikrant has at this stage of theirdevelopment. I cannot think of a stronger case for an ACRP fellowship than the one presented by thisstudent and this research topic.

CONFIDEhiTIA[-lTY: The information contained in this reference letter shall not be available io iheapnlicant or otherwise publicly disclosed

Signature of Reference Respondent: Date:

S_qtlfl€-

Draft Paper fVikrant Vaze, Cynthia Barnhart]

An Assessment of Impact of Demand Management Strategies forEfficient Allocation of Airport Capacity

Vikrant Vazel

Cynthia BarnhartllMassachusetts

lnstitute of Technology, Department of Civil and Environmental Engineering

Abstract

In the recent yeais, airline delay costs have become a signifìcaüt component of operating costs.

Airport demand management strategies have the potential to mitigate congestion and delays.

However, the extent to which the delays can be reduced using demand management is not clear.

In this paper, we develop a bound on the minimum possible level of delays that can be achieved

while satisfying all the passenger demand. The problem is formulated as a large-scale integer

linear programming model and solved using linear programming relaxation and heuristics. A

network delay simulator is used for calculating the delays under different capacity scenarios. The

results provide an estimate of the inefficiencies in the usage of airport infrastructure in the

domestic US due to competitive scheduling decisions by the airlines.

1. Introduction

Growing congestion at major US airports translates into several billion dollars of delay costs

each year. Total flight delays rose sharply during much of the current decade. Although the

current economic recession has led to airline schedule reductions and consequently resulted in

delay reduction over the last couple of years, large delays are expected to return as soon as the

economic crisis subsides (Tomer and Puentes,2009). Aircraft delays result in passenger delays

and discomfort, as well as additional fuel consumption and green house gas emissions. Various

studies have estimated the total economic impacts of delays and the numbers differ considerably

depending on the assumptions and scope of each study. In the report prepared for the Senate

Joint Economic Committee (JEC), the total economic cost of air traffìc delays to airlines,

passengers and the rest of the economy was estimated to be $4i billion for year 2007, including

$31 billion in direct costs (Schumer and Maloney,2003). The Air Transport Association (2008)

has estimated the direct costs of air traffic delay in 2007 to be $14.5 billion, which is less than

half of the JEC estimate. Notwithstanding these differences in actual delay cost estimates, it is

widely accepted that the delays are imposing a tremendous cost on the overall economy.

According to Bureau of Transportation Statistics (2008), delays to approximately 50Yo of the

delayed flights in2007 were due to the National Aviation System (NAS). The leading cause of

NAS delays was weather, which resulted in 65.60Yo of such delays, while volume was the second

most important cause responsible for 18.91% of the NAS delays. Volume related delays are due

Draft Paper ikrant Vaze, C thia Barnhartl

to scheduling more operations than available capacity, while weather related delays are due to

capacity reduction in adverse weather conditions causing the realized capacity to drop below

demand. Both these types of delays are due to scheduling more operations than the realized

capacity and hence can be termed as delays due to demand-capacity mismatch. Demand-capacity

mismatch thus contributed to 84.51% of the NAS delays in 2007, making it a primary cause of

flight delays in United States.

Capacity enhancement and demand reduction are the two natural ways of alleviating this

demand-capacity mismatch. Capacity enhancement measures require a longer time horizon and

more investment compared to demand management strategies. Demand management can reduce

delays by reducing the demand for airport resources without affecting the number of passengers

being transported. However, the extent to which the delays can be reduced using demand

management is not clear. ln this research, we assess the maximum possible impact of these

demand management strategies. All the analysis is based on extensive amounts of publicly

available data on airline schedules, passenger flows and airport capacities. Detailed flight and

passenger flow information was obtained from the Bureau of Transportation Statistics website

(BTS, 2008). Actual realized airport capacity values for one entire year were used for calculation

of flight delays.

Before proceeding further, let us define some terminology that will be used frequently in this

paper. In all our models, a market is defined as a passenger origin and destination pair. A

segment is defined as an origin and destination pair for non-stop flights. A path is defined as a

sequence of segments along which a passenger is transporled from origin to destination. A flight

leg is defined as a combination of origin, destination, departure time and arrival tirne of a non-

stop flight. An itinerary is a sequence of flight legs along which a passenger is transported f¡om

origin to destination. in the United States, a ground delay program (GDP) is a primary tool for

rnanaging the airport capacity in real-time wherein the arrival rates into reduced capacity airports

are controlled by coordinating the departure times of impacted flights. The duration for which

the impact of a GDP persists at an airport is called the GDP duration. We will refer to the actual

network of flights operated by multiple airlines in the US domestic markets in 2007 as the

existing network. Also we will refer to our delay minimizing network as the single airline (or SA)

network.

The rest of this paper is organized as follows. Section 2 provides a review of demand

rnanagement strategies and explains the motivation behind solving the single airline scheduling

problem. Section 3 briefly reviews the literature on airline scheduling and highlights the

important differences between the previous research and the problem at hand. Section 4 provides

a detailed problem statement for this study. Section 5 describes the modelling framework.

Section 6 outlines the set of algorithms used to solve the problem. Section 7 provides details of

data sources and implementation. A summary of results is provided in section 8.

Draft Paper [Vikrant Vaze, Cynthia Barnhart]

2. Airport Demand Management

The passenger demand for air travel in the US domestic markets has been increasing rapidly over

the last couple of decades and is expected to double within the next l0 to 15 years (Ball et al.,

2007b). Delays are expected to outpace the demand considerably with each lo/o increase in

airport traffic expected to bring about a 5oZ increase in delays (Schaefer et al., 2005). In spite ofthe increase in passenger demand, the average aircraft size in terms of the seating capacity in the

US domestic flights is shrinking rapidly. According to the analysis by Bonnefoy and Hansman

(2008), the average aircraft size has shrunk by more than 30o/o over the 17 year period from 1991.^ a^^1 'T.L:^ -^,,-Ll +La+ *a-a â:^l^+. ¡¡¡ fn froncnnrf fha

"o-o -tt-het nfLQ LVV /. ItllJ tULlBllty llIç¿IIIJ Lll4r tlrulv rrrórrLJ 4lw ¡rvwwùù4rJ Lv Lr4r¡oHv¡L Lrrv rq¡r¡v ¡rq¡¡rvv¡ vr

passengers. This phenomenon of proliferation of smaller aircraft is closely connected with

frequency competition among competing airlines. Higher frequency shares are believed to be

associated with disproportionately higher market shares because the two follow what is

commonly known as an S-shaped relationship (Belobaba, 2009). Therefore, there is a tendency

among the competing airlines to match non-stop frequencies in key markets to retain market

share. As a result, airlines tend to favour large numbers of flights using small aircraft rather than

a smaller number of flights using larger aircraft. Runway capacity is usually the most restrictive

element at major airports and it is also the predominant cause of the most extreme instances ofdelays (Ball et al., 2007b). From the runway capacity perspective, takeoff (and landing) of a

small aircraft requires almost the same time as that of a large aircraft. Hence frequency

cornpetition between airlines is responsible for aggravating airport congestion and delays.

Demand management strategies refer to any adrninistrative or economic regulation that restricts

airport access to users. Few of the most congested US airports, such as Kennedy (JFK), Newark

(EWR) and Laguardia (LGA) Airports in New York City area, O'Hare (ORD) Airport at

Chicago and Reagan (DCA) Airport at Washington DC, have been slot controlled in one way or

the other for a long time. Current slot allocation strategies are inefficient because of large

barriers to market entry (Dot Econ Ltd.,2001) and use-it-or-lose-it rules that encourage over-

scheduling practices (Harsha, 2008). Market-based mechanisms such as slot auctions and

congestion pricing have been proposed as efficient means of reducing congestion. There is

extensive literature (Ball et a1.,2007a; Ball, Donohue and Hoffman, 2006; Daniel, 1992; DoI"

Econ Ltd.,2001; Fan and Odoni,200l; Grether, Isaac and Plott, 1979;1989) suggesting that

market-based approaches, if designed properly, allocate scarce resources efficiently and promote

fair competition. Harsha (2008) shows that market-based mechanisms can lead to airline

schedule changes that reduce the demand for runway capacity, without reducing the number ofpassengers transported. This is achieved by better utilization of capacity in off-peak hours and by

greater usage of larger aircraft.

In theory, these pricing and auction mechanisms should bring the demand and supply into

balance by removing the inefficiencies in the system. However, the extent to which the system-

wide delays can be reduced by these mechanisms is still unclear. On one hand, restricting airport

Draft Paper ikrant Vaze, CYnthia Barnhart

utilization to a very low level can practically ensure the absence ofcongestion related delays, but

this could mean that the airport is highly underutilized and all the passenger demand might not

be satisfied. on the other hand, scheduling a very large number of operations can satisfy all the

passenger demand but the delays could reach unacceptable levels. An important question is what

minimum level of airport utilization and delays needs to be permitted in order to satisfy all the

passenger demand.

In this research, we measure the extent to which airport capacity in the us domestic air

transportation nefwork is being inefficiently utilized. The aim is to build a schedule that would

minimize the delavs in the absence of frequency competition. In order to obviate the effects of

competition, we assume a single airline that will satisfi all the passenger demand without

compromising the level-oÊservice for passengers' A network delay simulator (Odoni and

pyrgiotis, 2009) is used to estimate the delays for the resulting network' The delay values for the

single airline network are compared with those for the existing network under various realistic

capacity reduction scenarios. All the days in an entire year are divided into 5 categories based on

the total GDp duration on that day across all the busy airports. Each category contains equal

numbers of days and hence, can be assumed to be equally likely. one representative day from

each category is chosen for delay calculations. These delay estimates serve as theoretical lower

bounds on the system delays when airport capacity is allocated most efficiently. The maximum

possible delay reduction will indicate the maximum potential impact of implementing efficient

demand management techniques. if insignif,rcant, then passenger demand has already reached a

Ievel where targe delays are inevitable and capacity enhancetnent is the only realistic means of

delay reduction. On the other hand, if the results suggest substantial delay reduction under the

single airline case, then the existing level of passenger demand can be efficiently served using

the existing infi.astructure with much lower delays and there is ample opportunity for congestion

rnitigation using demand management strategies'

3. Airline Schedule DeveloPment

The airline schedule development process includes decisions regarding daily frequency'

deparlure times, aircraft sizes and crew schedules. Due to the enormous size and complexity of

the airline schedule development process, the problem is typically broken down into four stages:

l) timetable development; 2) fleet assignment; 3) maintenance routing; and 4) crew scheduling

(Barnhart, 2009). The task of deciding the set of flight legs to be operated along with the

corresponding origin, destination and departure time for each leg is called timetable

developrnent. Although the entire timetable development problem can be modelled as an

optimization problem, practitioners typically focus on incremental changes to existing schedules'

The fleet assignment problem involves a profit maximizing assignment of fleet types to flight

legs. Hane et al. (1995) proposed a leg-based fleet assignment model which assumes independent

leg demand and average fares. Jacobs, Johnson and Smith (1999) and Barnhart, Knicker and

Lohatepanont (2002) proposed itinerary-based fleet assignment models that produce significant

Draft Paper ikrant Vaze, Cynthia Barnhart]

profrt improvement over the leg-based models. Maintenance routing is the assignment of specifìc

aircraftto individual flight legs while satisffing the periodic aircraft maintenance requirements.

The maintenance routing problem is typically solved as a feasibility problem or as a through-

revenue maximization problem (Barnhart, 2009). The problem of assigning a cost minimizing

combination of pilot and cabin crews to each flight leg is called crew scheduling. Because of the

complicated duty rules and pay structure for airline crews, crew pairing has long been regarded

as a challenging problem. Crew pairing is modelled as a set partitioning problem and solved

using techniques such as column generation and branch-and-price (Barnhart et al., 1998b;

Desrosiers et al., f995; Klabjan, Johnson and Nemhauser, 2001).

The aforementioned sequential solution process may result in suboptimal solutions. Many

researchers have tried to integrate some of the stages into simultaneous optimization problems.

Rexing et al. (2000) proposed joint models for flight retiming within time windows and fleet

assignment. Lohatepanont and Barnhart (2004) present an integrated model for incremental

schedule development and itinerary-based fleet assignment. Clarke et al. (1996) and Barnhart et

al. (1998a) have proposed models to incorporate the effect of maintenance routing while making

the fleet assignment decisions. There is a large body of literature on the integration of aircraft

routing and crew scheduling problems including Cohn and Barnhart (2003), Cordeau et al.

(2001) and Klabjan ef al. (2002).

All the models mentioned above aim to produce precise schedules that maximize planned profit.

Models developed for the purpose of this study are different in several important ways. Rather

than producing an operable schedule, the main purpose is to obtain a bound on delays. Because

of the complex and stochastic relationship between schedule and delays, any such bound will

have to be approximate. Therefore there is no point in developing a very precise schedule.

Moreover, the problem of single airline schedule development is even larger in size than the

schedule development problem for any existing airline, which itself is solved sequentially due to

tractability issues. Therefore in this study, we use aggregale models that are suffìcient for our

purposes while maintaining tractability. Instead of profìt maximization, the objective is delay

minimization subject to satisfaction of demand and level-of-service requirements. Therefore,

only the relevant decisions such as timetable development and aircraft sizes are included in the

problem. We do not solve the aircraft routing and crew scheduling problems in this research.

Harsha (2008) has proposed an aggregated, integrated airline scheduling and fleet assignment

model (AIASFAM) to help airlines place a bid in a slot auction. The itinerary-based version of

this model is an extension of the Barnhart, Knicker and Lohatepanont model, with more

aggregate tirne-line discretization for computational tractability. The models presented in this

study share some characteristics with the AIASFAM model.

Draft Pa ikrant Vaze, Cynthia Barnhart]

4. Problem Statement

In order to obtain a lower bound on airport congestion, we assume the existence of a single

monopolistic airline. The problem at hand is to design a schedule for this single airline with the

objective of minimizing airport congestion, while satisffing all the demand and maintaining a

comparable level-of-service. The utilization ratio is defined as the ratio of demand to capacity of

a server, which in this case is an airport. Queuing theory suggests that the average flight delay is

an increasing and convex function of the utilization ratio (Larsen and Odoni, 2007). Considering

the tremendous size of the problem at hand, using a non-linear objective function would make

rhe nrohlem intractable. Total delay is a nonlinear and stochastic function of the number of-__- t----

scheduled flights. Therefore, we aim to minimize the maximum utilization ratio as a surrogate

objective function for the scheduling problem. Due to the convex relationship between the

utilization ratio and delays, the effect of the maximum utilization ratio on total delay in a

queuing network is disproportionately high. Further, the maximum utilization ratio is a

deterministic function of the airline schedule and can be modelled as a linear function of the

decision variables.

An important modelling consideration is how to capture the demand satisfaction requirements for

every market and every time period of the day. The single carrier must be able to transport all

passengers who are currently transported by existing airlines, from their respective origins to

their respective destinations. We model this by dividing the day into multiple time intervals and

ensuring that all the passengers who are currently transported during a particular interval

continue to be transported during the same interval in the new schedule. We define the level-of-

service (as perceived by air passengers) as the number of stops in an itinerary. Almost 97 '6% of

all US domestic passengers travelled on non-stop and one-stop itineraries in2007 (8TS,2008)'

Hence, in the single airline scheduling model, we assume that all the passengers must be

transported on itineraries with at most one stop.

Service frequency is another irnportant criterion of level-of-service in the current competitive

environment. Therefore, in our model, we require that the single airline provide at least the same

daily frequency on each non-stop segment as the ffictive frequency provided by the existing

carriers. Cohas, Belobaba and Simpson (2005) propose amodel of effective frequency available

to air passengers faced with a choice between multiple competing carriers' When more than one

airline operates in a market, the effective frequency depends on how closely the schedules are

matched. For example, consider two competing carriers, each offering n flights per day' If one

airline schedules flights at a time when the other airline does not offer service, then the effective

frequency increases. However, if one airtine schedules aU its flights close to the departure times

of flights by the other airlines, then the number of different options to the passengers does not

increase above n. Thus, the important criterion in deciding the effective frequency is the

closeness of cornpeting airline schedules. We calculate the effective non-stop frequency for a

segment as the total number of non-stop flights offered by all carriers as long as the flight

Draft Paper ikrant Vaze, CYnthia Barnhart

departure times are not too close to each other. In particular, if the departure times of two flights

are separated by less than one hour, then we assume the two flights to be equivalent to a single

flight. The minimum frequency to be provided on each non-stop segment by the single airline

must be greater than or equal to the effective frequency currently provided by all the existing

carriers on that segment.

Before designing a schedule, a decision must be taken regarding the network structure. Network

design involves decisions about nefwork type i.e. hub-and-spoke or point-to-point, choice of

hubs, choice of non-stop segments, choice of allowable airports for passenger connections. One

possible approach would be to include all these decision into our single-airline optimization

problem. For the problem size under consideration, that would lead to an integer optimization

problem involving over one-hundred million variables. Instead, we solve the problem

sequentially in three stages.

5. Modelling Framework

Figure 1 provides a schematic description of the overall modelling framework. The first stage is

the Network Design Q..lD) stage, which involves decisions about the number and location of

hubs, candidates for non-stop routes and allowable airports for passenger connections' The

network structures of existing airlines were used as a guideline for our network design model'

Many of the major airlines in US domestic market today have a set of 4 or 5 major hubs. The

direct flights are allowed to bypass the hub for a few important markets with large demand' Our

selection of hubs was made based on qualitative criteria including the number of operations in

the existing network, available capacity, geographic location and weather. Our choices of non-

stop segrnents bypassing the hub were made based on the market demand corresponding to the

non-stop segments. We allow passengers to connect only at the hubs.

The second stage involves the daily Frequency Planning and Fleet Assignment (FPFA) problem'

A delay minimizing schedule should have fewer flights per day and better distribution of flight

time to avoiding clustering of demand near peak hours. Obtaining a good feasible solution is the

main aim of our FPFA stage. A good solution will keep the number of flights to a minimum, so

that airport usage is rninimized. However there are multiple optimal solutions that minimize the

total number of flights. [n order to produce an efficient schedule it is important to choose the

most appropriate aircraft.sizefor each segment so as to avoid excessive seating capacity. This is

achieved by choosing cost coefficients (denoted by cs,¡ in the formulation) for each fleet type

such that the cost increases with increasing seating capacity and cost per seat decreases with

increasing seating capacity. These costs are consistent with those that the airlines report through

Form 4l financial reports (8TS,2008). Satisfaction of the daily demand and the minimum daily

frequency requirement are the two main constraints. Output of this second stage includes the

daily frequency of service on each segment and fleet types assigned to each segment. Constraints

(l) through (5) provide integer programming fonnulation for the FPFA problem.

Draft Paper ikrant Vaze, C nthia Barnhart

FPFA Formulation:

Notation:

K: Set of fleet types

,S: Set of segments

P = Set of paths

M: Sef of markets

cr,¡r: OperaTing cost of fleet type k on segment s, s e 'S and ft é- K

C¿ : Seating capacity of fleet type k, k e K

D*:Dally demand in market iø, m e M

I : Minimum daily frequency to be provided for segment s, s e ^S

P(*): Set of paths associated with market m, m e M

uf :[å:"pathp contains :if;ü|: pepandses

Decision variables:

x,,¿:No. of flights of fleet type kon segments per day, s e s andke K

yr: No. of passengers on path p per day, p € P

Formulation:

Mtntmize II cs,k*xs,k

seS keK

Subject to:

Lper@)!p:D¡n'YmeM

Zt rrcxr,t * Ct 2 Zp re 6P, x Yo' V s e "S

Luuxxr,n >f,,Vse S

x5,¡re V.+,Vse S qndkeK

loeZ+,VPeP

(1)

(2)

(3)

(4)

(5)

Draft Paper [Vikrant Vaze, Cynthia Barnhart]

Constraint (1) ensures that the total daily demand for each market is satisfìed. Constraint (2)

ensures that the total number of seats on each segment is sufficient for carrying all the passengers

whose paths contain that segment. Constraint (3) enforces that the daily frequency of service in

each market is at least equal to the effective frequency currently provided by the existing carriers

in that market. Constraints (4) and (5) restrict the possible values for the number of passengers

and number of flights to non-negative integers.

The third stage involves actual Timetable Development (TD). Similar to the approach adopted by

Harsha (2008), the departure and arrival times are aggregated to the nearest hour to keep the

number of decisions variables low. Given the daily frequencies and fleet assignment for each

segment, output of this stage produces the scheduled set of flight legs. Constraints (6) through

(l l) provide integer programming formulation of the Timetable Development model.

The objective function is to minimize the maximum utilization ratio across all busy airports

across all airport-time period (ATP) pairs. The duration of each ATP is I hour in this case.

Constraint (6) enforces the satisfaction of demand for each market-time period (MTP) pair.

Constraint (7) ensures that the total number of seats for each flight leg is at least equal to the total

number of passengers whose itineraries contain that flight leg. Constraint (8) ensures that the

minimum daily frequency requirement is satisfied. Constraint (9) relates the maximum utilization

ratio to the operations at each airport during each hour. Constraints (10) and (11) restrict the

possible values for the number of passengers and the frequencies to non-negative integers.

TD Formulation:

Notation:

I : Set of airports

1: Set of itineraries

Z: Set of MTPs

F: Set of flight legs

Z: Set of ATPs

L(*): Set of MTPs associated with market m, m e M

Dt: Demand in MTP l, I e L(m) and m e M

C¡: Seating capacity for fleet type assigned to flight legl f e F

/ : Minimum daily frequency to be provided for segment s, s e S

HC,: Hourly capacity (i.e. maximum total number of operations) for ATP t, t € T(a) and a e A

Draft Paper [Vikrant Vaze, Cynthia Barnhart

T(a): Set of ATPs associated with airport a, a e A

(f : Set of itineraries associated with MTP I, I e L(m) andm e M

F(s) : Set of flight legs associated with segment s, s e 'S

Ãr. :{I,if itinerary ¿ contains flight leg / i e I andf e F"/ -tO, otherwise

.,,r. :íI,ifATPt is utilizedbyflightleg/ t eT and,fe Fr f t0, otherwise

Decision variables:

x¡: Frequency for flight legl f , F

/¡: No. of passengers on itinerary i, i e I

rn,o.,: Maximum utilization ratio for airport hourly capacities

Formulation:Minímize r*ot

Subject to:

Ii e rlr¡ lt : DI,Y I e L(m)'m e M

xf * Cf >- L¡.rt6i * !t,v f e F

I¡.r1r¡xf 2f,VsES

Z¡reTi x xf 3r^ax * HCtv t eT(a)'ae A

x¡eZ+,VfeF

lfZ+,Yiel

6. Solution Algorithm

The combination of problem size and our objectives in addressing this problern allow us to use

an approximate solution algorithm. We solve the FPFA LP relaxation and round up the resulting

solution to nearest integer values greater than or equal to the LP optimal solutior, value' Due to

the nature of the constraints in FpFA formulation, none of the constraints is violated if segment

frequencies are increased.

(6)

(7)

(8)

(s)

(10)

(11)

Draft Paper [Vikrant Vaze, Cynthia Barnhart]

Solution to the FPFA problem involves determining daily frequency values, which are relatively

large integers. The impact of rounding up is comparatively small. But for the TD problem, the

solutions are highly fractional because the hourly frequencies are much smaller than daily

frequency values. Therefore, the rounding up procedure worsens the objective function

dramatically. Much of the solution's non-integrality stems from the markets in which demand is

extremely low per day. Therefore, the LP solution has very small fractions of flight legs serving

small markets, and the solution is not of sufficient quality for our purposes. Therefore, we solved

the TD problem in two steps. The TD solution procedure is described schematically in figure 2.

In the first step, the TD LP relaxation was solved for a smaller sub-problem involving all the

mrr!¿atc.r,ith ¡ d¡i!.-,rlem¡n¡l of .ât leacf ?-50 nassense!"s- These constituted ove-r 609/o of the totalIII4I r\vLJ YY rtrr s s@¡¡J uvrrre¡

demand. These markets include all the candidates for non-stop service bypassing the hub. This

LP solution was rounded upward to the nearest integers. Due to the nature of constraints in the

TD formulation, an increase in value of any .r variable in a feasible solution does not affect

feasibility. Moreover, most of flights in these important markets serve as connecting flights for

smaller markets. Therefore the additional seating capacity made available due to flight rounding

is very likely to be utilized to carry passengers in remaining smaller markets. The remaining

problem was solved using a greedy heuristic, as depicted in figures 3 and 4.

In the first step of the heuristic, as described in figure 3, additionalnon-stop flights are scheduled

to satisfy the demand of markets with a daily demand of less than 250 passengers and with at

least one endpoint at a hub (the Hub Markets). The markets are processed one after the other in

decreasing order of demand. Additional flights are scheduled such that the maximum utilization

ratio across all the affected airport-time period (ATP) combinations is minimized al each step.

Scheduling an additional non-stop flight increases the utilization ratio for the origin airport

during the departure hour and also increases the utilization ratio of the destination airport during

the arrival hour. Among all the departure time choices for a market-time period (MTP)

combination, the one which minimizes this maximum utilization ratio is chosen.

ln the second step of the heuristic, as described in figure 4, additional flight legs are scheduled to

satisfu the demand for markets with a daily demand of less than 250 passengers, where neither

endpoint is a hub (the Non-hub Markets). These are the markets with low daily demand with

neither origin nor destination being a hub. For these markets, the demand has to be satisfied by

one-stop itineraries. The algorithm processes the markets in decreasing order of demand and

schedules additional flights on hrst or second or both legs of an itinerary so as to minimize the

maximum utilization ratio among allthe affected ATPs.

The optimization algorithm ignores the aircraft flow balance constraints, which is an important

component of the airline schedr,rling procedure. The purpose of this study is not to come up with

a schedule that can be operated using actual aírcraft fleets- our goal is not to design a

rnonopolistic airline schedule for the US- but rather to fìnd a lower bound on the levels of delays.

However, to assess the potential impact of aircraft balance requirements, after completing the

Draft Pa r [VikrantUaze[. thia Barnhartl

optimization plocess, we performed a post-processing step wherein additional flights were added

to the optimal unbalanced schedule in order to balance it. Fortunately, because the passenger

demands typically are balanced in both directions of a market, the resulting schedule is not too

far from a balanced schedule. In our algorithm' we add one flight at a time from an airport with a

surplus of a particular aircraft type to an airport with a deficit of the same aircraft type' The

departure time of each additional flight is chosen to greedily minimize the maximum utilization

ratio of the affected ATPs at each step. The results presented in the next section provide the

statistics on both the balanced and unbalanced single airline (SA) networks, and compare these

statistics with the existing network'

Flight delays are estimated using the Approximate Network Delay (AND) mo<iei <ieveiopetì 'oy

odoni and pyrgiotis (2009). Each airporr is modelled as a M(t)lE¡(t)lI queuing system with

infìnite queuing capacity. The model tracks each flight in the network and updates the airport

demand profiles based on revised flight leg arrival and departure times' The same model is used

to estimate the delays in the existing network and the balanced single airline network' Airport

capacity data for one full year is used to classify the days into 5 categories' The delay values fpr

both networks are compared for one representative day belonging to each category'

7.Data Sources and Implementation Details

schedules of major us domestic airlines were obtained from the Airline on-Time Performance

Database provided by the Bureau of Transportation Statistics (200s)' A l0% sample of the

passengers carried by each airline per quarter is provided in the Airline origin and Destination

Survey(DBIB)availableonthesamewebsite.AnestimateofthetotalpassengerdemandperOD market is obtained by rnLrltiplying the DBlB passenger number by 10' The top 20 airlines in

the uS, which constitute over 950/o of the total traffic, are considered for the analysis' Demand

for each market per day is divided into 4 time periods that roughly correspond to morning peak

period, off-peak period, evening peak period and another off-peak period' Given the daily

demands, the Decision window Model by the Boeing Airplane company (1997) is used to

calculate the demand in each time period of the day' Apart from these 20 airlines' there are

several other flights such as cargo, general aviation and international flights which are not

included in our analysis. We assume that these remaining flights continue to be operated as they

currently are. We simply add those operations to the total operations at each airport for

calculating the utilization ratios'

The Federat Aviation Administration has published benchmark capacity values for the 35 most

congested US airports (FAA, Z0O4)'For each one of those airports' the report provides 3 values

of capacity along with the corresponding probabilities of their realization, based on weather'

wind and other conditions. The capacity is measured as the maximum number of operations

(takeoffs and landings) possible at an airport per hour' Based on the capacity distributions for

these airports, we use capacity values for each airport that correspond to the maximum capacity

Draft Paper ikrant Vaze, Cvnthia Barnhart

that is available for at least 95% of the time at that airport. We use these capacities only for

calculating the utilization ratios. The delay calculation is based on the actual realized capacity

values under each weather scenario.

The analysis is carried out for Tuesday, l6th October 2007. Because this analysis is aggregate in

nature, we use only three generic fleet types. We call them Wide Body (WB), Narrow Body

(NB) and Regional Jets (RJ). The seating capacity values for these three fleet types are chosen to

be the average seating capacities of the corresponding aircraft in the 2008 fleet of the seven

Iargest US airlines.

Actual data on GDp durations and hourly capacities for the 35 busiest US airports was compiieci

fbr the entire year tiom Apr 2007 to Mar 2008 in order to generate realistic capacity reduction

scenarios. Based on the total duration of GDP across these 35 airports, 366 days in the year were

divided into 5 weather categories, namely, very good, good, normal' bad and very bad, each

containing approximately the same number of days. Therefore each of the categories can be

considered to be equally likely. For delay calculation purposes, a typical day belonging to each

category was chosen such that the total GDP duration for this day is close to the median of total

GDP duration for all days belonging to the same category.

8. Results and Discussion

For our single airline network, the maximum utilization ratio over all airports over all ATPs was

found to be 90.77%0. In comparison, the maximum utilization ratio in the existing network was

1600/0. Table I provides the cumulative distribution of utilization ratios across the 35 busiest

airports, across 24 hours of the day. The first column contains the percent utilization ratios and

the next three columns show the number of times that value was exceeded in the existing

network, unbalanced single airline (SA) network and balanced single airline (SA) network

respectively. Each combination of airport and hour is counted as one observation.

Utilization RatioExistingNetwork

UnbalancedSA Network

Balanced SA

Nefwork

>150o/o I 0 0

> l40o/o J 0 0

> l30o/o 8 0 0

> l20o/o l9 0 0

> lI0Yo 28 0 0

> l00o/o 55 0 0

> 90o/o 76 I 1

> 80Yo r33 60 83

Draft Paper [Vikrant Vaze, C thia Barnhart

> 70o/o 196 l3l tsl

> 60Yo 275 196 2r0

> 50Yo 350 268 292

Table 1: Comparison of Airport Utilization Ratios

Table 2 compares some important metrics for the unbalanced and balanced single airline

networks to those for the existing network. US domestic flights in October 2007 had aî average

load factor of 7ï.45yo (BTS, 2O0g). The load factor for the single airline network is lower,

reflecting that the single airline network design problem had iess ftexibiiity in seiecting the mosi

appropriate aircraft for each flight leg due to aggregation ofaircraft sizes into 3 broad categories'

We define block hours as the difference between the scheduled arrival time and scheduled

departure time of a flight. The unbalanced single airline network requires 260/o rewer flights and

golo lower total block hours than the existing network. For the year 2007, the DB1B Market

database (8TS,2008) shows that approximately 31.98% of the passengers travelling on major

US airlines were connecting passengers, which is very close to that for the single airline network'

Figure 5 shows the distribution of the utilization ratio over a day aT John F. Kennedy

International Airport, New York (JFK), one of the busiest airports in US' The single airline

network spreads operations more equally over the day at the congested airports. This efficient

utilization of off-peak hourly capacity is one of the factors that contribute to lowering the

congestion levels in the single airline network.

Table 2: Network Performance Metrics

The other imporlant factor that contributes to lower congestion is an increase in average aircraft

sizes. Table 3 illustrates the number of flights being operated by each aircraft size. The strong

shift towards wide body aircraft can be observed in both the unbalanced and balanced networks'

Most of the increase in the number of wide bodies comes from a decrease in usage of narrow

bodies.

Metric

Existing

Network

Unbalanced SA

Network

Balanced SA

Network

No. of Flights 20,539 t5,231 16,872

Fraction of Connecting Passengers 31".98% 33.77% 33.77%

Load Factor 78.4s% 71.77% 64.51%

Total No. of Block Hours 44,103 40,756 45,739

Draft Pa r lVikrantVaze, Cynthia Barnhart

Fleet Type WB NB RJ

Existing Network 363 (1 .77%) 12788 (62.26%) 7388 (3s.e7%)

Unbalanced Network 328s (2t.s7%) 60s8 (39.77%) 5888 (38.66%)

Balanced Network 3s6r (2r.rt%) 677s (40.16%) 6s36 (38.74%)

Table 3: Aircraft Size Distributions

Table 4 compares the total aircraft delay under 5 different capacity scenarios. The delay values

are computed excluding any propagated delay due to late arriving aircraft. in each capacity

scenario, the single airline network produces substantially lower delays, with the delay reduction

ranging from 534oto 88%. As the congestion worsens, the absolute, as well as percentage, delay

reduction increases. Across different scenarios an average delay reduction of 81.72% can be

achieved in the absence of competition. This provides an estimate of the inefficiencies that arise

in US domestic air transportation due to the competitive scheduling practices of carriers using

the existing airport infrastructure. The results suggest that congestion related delays could be

reduced to less than one fifth of the existing level if there was no competition' ln other words,

competition is responsible for more than a 4}}%worsening of congestion related delays' Most of

the improvement is due to efficient utilization of airport capacity during off-peak hours and a

strong shift towards larger aircraft. Hence, there is significant room for improvement in

congestion even with the existing airport infrastructure. Passenger demand has not yet reached a

level where delays as high as present are unavoidable. Given the available capacity, efficient

administrative controls and/or market-based mechanisms can potentially lead to substantial

reductions in airport congestion and delays.

Scenario

Existing Network

Delay (aircraft-min)

Single Airline Network

Delay (aircraft-min) Reduction

Very Good

Good

Normal

Bad

Very Bad

7495.05

L4682.30

27998.76

3508L.44

64026.52

3552.97

4090.06

5940.40

6289.88

742t.76

52.60%

72.14%

78.78%

82.07%

88.4r%

Average 29856.82 5459.02 81..72%

Table 4: Total Flight Delay under various weather Scenarios

This research has identified two extreme possibilities. At one end is the current situation where

inefficient capacity allocation coupled with profit maximizing scheduling practices under

cornpetition has lead to large delays and disruptions. At the other end of the spectrum is the

situation where delay minimizing scheduling in the absence of competition can lead to very low

er [Vikrant Va Barnhart

delays. The critical next step is to devise intelligent mechanisms and incentives that will force

the airlines to migrate their schedules from those in place today towards delay-minimizing

schedules, in the presence of market competition. The impact of these mechanisms on airline

profitability and passenger level-of-service needs to be assessed carefully'

References

Air Transport Association (ATA) (2008). Cost of Delays' Obtained through the Internet:

:llv'tww.ai ialtonics/ATC* faccessed I0l 1212008].

Ball, M.O., Ausubel, L.M., Bernardino, F., Donohue, G., Fiansen, Ivi' and Fioifüiari' K' (2007a)

,Market based alternatives for managing congestion at New York's LaGuardia airport'' Paper

Presented at the AirNeth Annual Conference. April I I -l3, 2007 ' The Hague, Netherlands'

Ball, M.o., Barnhart, c., Nemhauser, G. and odoni, A. (2007b)'Air transportation: irregular

operations and control' , Handbook in OR & MS, Vol' 14, No' 1, pp'23-38'

Ball, M.o., Donohue, G. and Hoffman, K. (2006) 'Auctions for the safe, efficient and equitable

allocation of airspace resources" In; Cramtofl, P., Shoham, Y. and Steinberg, R' (eds'),

C o mb inato ri al Auct io ns, (pp. 5 07-5 3 8), Cambridge : MIT Press'

Barnhart, C. (2009) 'Airline schedule optimizatio n' , In: Belobaba, P', Odoni, A' and Barnhart' C'

(eds.), The Gtobal Airline Industry, (pp.183-212), Wiley'

Barnhart, C.,

'FIight string

220.

Barnhaft, c., Johnson,E.L.,Nemhauser, G., Savelsbergh, M. and vance, P. (1998b)'Branch-

and-price: column generation for solving huge integer programs' , Operations Research' Vol' 46'

pp.316-329.

Barnhart, c., Knicker, T., and Lohatepanont, M. (2002) 'Itinerary-based airline fleet

assignment' , Transportation Science. Vol. 36, pp'199-217 '

Belobaba, P. (2009) 'Overview of airline economic, markets and demand', In: Belobaba' P''

odoni, A. and Barnhart, c. (eds.), The Global Airline Industry, (pp.47-72), Wiley.

Boeing Airplane Company (Igg7) Decision Window Path Preference Methodologlt Description,

Seattle, WA.

Bonnefoy, p.A. and Hansman, R. J. (2008) 'Overview of recent forces and trends in the airline

industry and challenges for future operations'. Paper Presented at the Lufthansa-European

Btts ine s s S cho o I C onference, 2008.

Boland, N., Clarke, L., Johnson, E.L , Nemhauser, G', and Shenoi' R' (1998a)

models for aircraft fleeting and routing' , Transportation Science, Y ol- 32, pp'208-

Draft Paper lVikrant Vaze, Cynthia Barnhart

Bureau of Transportation Statistics (BTS) (2008). Airline On-Time Statistics and Delay Causes.

Obtained through the Internet : http : //www. bts. sov, [accessed 15 I 12120081'

Clarke, L., Hane, C., Johnson, E.L. and Nemhauser' G. (1996) 'Maintenance and crew

considerations in fleet assignment' , Transportation Science' Vol' 30, pp'249-260'

Cohas, F.J., Belobaba, p.P. and Simpson R.W. (1995). 'Competitive fare and frequency effects

in airport market share modelling', Journal of Air Transport Managemenf, Vol' 2, No' 1' pp'33-

45.

Cohn, A. a.nd Barnhart, C. (2003). 'Improving crew scheduling by incorporating key

maintenance routing decisions', Operations Research, Vol. 51. pp.387-396. .

Cordeau, J., Stojkovic, G., Soumis, F. and Desrosiers, J. (2001)' 'Benders decomposition for

simultaneous aircraft routing and crew scheduling' , Transportation Science, Yol' 35, pp'375-

3 88.

Daniel, J. (lgg2) Peak\oad-congestion pricing and optimal capacity of large hub airports" with

application to the Minneapolis St.PauI airport. PhD thesis. Massachusetts Institute of

Technology, Cambridge MA.

Desrosiers, J., Dumas, Y., Solomon, M.M. and Soumis, F. (1995) 'Time constrained routing and

scheduling' , Handbook in OR & MS Network Routing, pp'35-139'

Dor Econ Lrd. (2001). Auctioning Airport Slots. obtained through the Internet:

hup;//www.dotecon.com/publications/slotaucto.pdf[accessed 05/01/2010]'

Fan, T.P. and Odoni, A.R. (2001)'The potential of demand management as a short-term means

of relieving airport congestion' . Paper Presented at the EUROCONTROL-FAA Air Traffic

Management R&D Review Seminar, 2001'

Federal Aviation Administration (FAA) (2004). Airport capacity benchmark report' Obtained

through the Internet: httP: .faa.eovlabout/office orsl@bench/, [accessed t5 I 12/2008].

Grether, D.M., Isaac, R.M. and Plott, c.R. (1979) Alternative methods of allocating airport slots"

perþrntance and evaluation.Tecì'tnical repoft, Polinomics Research Laboratories'

Grether, D.M., Isaac, R.M.

Experimental Economics and

Hane, C.4., Barnhart, C., Johnson, E.L.,

(1995) 'The fleet assignment problem:

P ro gramnting, Y ol. 7 0, PP.2l I -232.

and Plott, c.R. (1989) The Allocation of Scarce Resources:

the Problem of Allocating Airport Slots, Boulder: Westview Press.

Marsten, R.8., Nemhauser, G.L., and Sigismondi, G'

solving a large-scale integer program', Mathematical

Draft Pa r [Vikrant Vaze, Cynthia Barnhart

Harsha, P. (2008). Mitigating airport congestion: market mechanisms and airline response

model..PhD thesis. Massachusetts Institute of Technology, Cambridge MA'

Jacobs, T.L., Johnson, 8.L., and Smith, B'C. (1999) 'O&D FAM: Incorporating passenger flows

into the fleeting process' . Paper Presented at the 39th Annual AGIFORS Symposium' October 3-

8,1999. New Orleans LA.

Klabjan, D., Johnson, E.L. and Nemhauser, G.L. (2001) 'solving large airline crew scheduling

problems: random pairing generation and strong branching', Computational Optimization and

Applications, Vol. 20, PP.7 3 -91.

Klabjan, D., Johnson,E.L.,Nemhauser, G.L., Gelman, E., and Ramaswamy, S. (2002)' 'Airline

crew scheduling with time windows and plane count constraints' , Transportation Scienc¿, Vol'

36, pp.337-348.

Larsen, R.C. and Odoni, A.R. (2007) (Jrban operations research,2nd Ed., Dynamic ldeas'

Lohatepanont, M. and Barnhart, C. (2004) 'Airline schedule planning: integrated models and

algorithms for schedule design and fleet assignment' , Transportation Science, Vol' 38, pp'19-32'

Odoni, A. and Pyrgiotis, N. (2009) 'A Stochastic and dynamic queuing model of a network of

major airpofts'. Paper Presented at the INFORMS annual meeting. october l1-14,2009' San

Diego CA.

Rexing, 8., Barnhart, c., Knicker, T., Jarrah, 4., and Krishnamurthy, N. (2000) 'Airline fleet

assignment with time windows" Transportation science, vol. 34, pp.1-20'

Schaefer, 4.J., Johnson, 8.L., Kle¡"wegt A'J' and Nemhauser, G'L' (2005) 'Airline crew

scheduling under uncertainty', Transportation Science, Vol. 39, pp.340-348.

Schumer, c.E. and Maloney, c.B. (2008) Yottr Jlight has been delayed again: Jlight delays cost

passengers, airlines, and the (JS econonty billions. The US Senate Joint Economic Committee'

May 2008.

Tomer, A. and puentes, R. (2009) Expect delays. an analysis of air travel trends in the United

States. Brookings Metropolitan infrastructure Initiative Series.

Draft Paper ikrant Vaze, Cynthia Barnhart]

Figure 1: Modelling Framework

Solve LP Relaxation ofTD 5ub-

problem

Round Fractional OPtimal Values

UPward

Greedy Heuristic For Hub Markets (i.e. A

Market with at least One Endpoint at a

Hub) With Small Demand

Greedy Heuristic For Remaining Non-Hu

Markets (i.e. A Market with Ne¡ther

Endpoint at a Hub) with small Demand

Frequency Planningand Fleet

Assignment (FPFA)

TirnetableDevelopment

(TD)

Figure 2: Timetable Development Algorithm

Draft Paper I C thia Barnhart

Figure 3: Greedy Heuristic for Hub Markets with SmallDemand

Vikrant Vaze,

Sort Hub Markets in Decreasing Order of Demand

For Each Market, Sort Time-Periods in lncreasing Order of Time

Select First MTP from the List

Select Next MTP from the List

ls CurrentMTP Dema

ls anv MTP

Remaining in

the List?

Direct Flightwith Empty

Seats?.

Schedule an Additional Flight

to Minimize the Maximum

Utilization Ratio Among the

Affected AirPorts

Allocate as ManY Passengers

as Possible to EmPtY Seats

on Direct Flights

Draft Paper [Vikrant Vaze, nthia Barnhart]

Sort Non-Hub Markets in Decreasing Order of Demand

For Each Market. Sort Time-Periods in lncreasing Order of Time

Select First MTP from the L¡st

Select Next MTP from the Lìst

ls CurrentMTP Deman

an ltinerarVwith Vacancy

Allocate as ManY Passenger

as Possìble to ltineraries withVacancy on Both Legs

an ltìnerarYw¡th Vacancy

Only on 15t

Schedule Additional 2^o Leg Flights

to Minimize the MaximumUtilìzation Rat¡o Among tl'ìe

Affected AirPorts

an ltinerarywith VacancyOnly on 2"d

Schedule Additional 1sr Leg Fl¡ghts

to Minimize the MaximumUtil¡zat¡on Ratio Among the

Affected Airports

Schedule an Additional FIight on

Each Leg of an ltìnerarY toMinimize the Maximum UtilizationRatio Anrong the Affected Airports

Figure 4: Greedy Heuristic for Non-hub Markets with Small Demand

Draft Paper [Vikrant Vaze, Cynthia Barnhart]

r,80%

L60%

L40%

L20%

a.P Loo%(oú.c ao%o.F(!.! 60%'Ff

40%

20%

oo/.

& Unbalanced

{ü Bäräìiëèì**

Hour

Figure 5: Distribution of Utilization Ratio over aDay at JFK Airporl

AcådeniE r eoo1-e00ä

Code Crurse Nåne

CHCHcs

HSMAItENOPHPH

101 Chemi strg I115 cheñi5tFg Lab I101 ComputÈr Prtgranning ånd

UtiIizationl0l Econonics103 Hat-henatics I1 1 1 t¡ork shóF Pract ice I101 NS0101 Phgsics I1 15 PhUsicE Låh I

sPI : 9.31 CPI : 9.31

Acðderqic I e00e-e003

Code Course N¡rqe

CE e01 Solid MechåniËs ICE e05 Fluid Mechanics ICE e11 Sol id I'lÊch¿nics Lab tCE 413 FluÍd I'lechanicË Lab ICE ¿15 Civil Engineering DråuingES Ê81 Environnehtal Enqg' and

Sc i enceHS e05 InLroduÊt ítrn to SociologgHA. e03 Håthènåtics I I I

sPI : ?.é3 CPr ! 9,47

AÉademic ¡ e003-2o04

Code Counse Nane

CE 30l Fluid l'lechanics IIICE 3O3 Soil HechënicE Iig gOz Design of Structures ICE 31 I Design in open Channel Flot¡CE 3lE Desigrr Lâbot.atoFg ICE 317 Stl^ucturå1 Mechênics IICE 319 Soi I I'lechanics Låb ¡GS 307 GeologgGS 317 Geologg Lab

sPI I 9,93 CPI : ?.ó?

5ÉnÊsLÊr*:1

credit Gråde

ó,0

ó.0

6.06.O4.O0.06.0

ABAEAA

AAAEAAPPBBAA

CDde

CECELñ

CHtfAME

HE

NO

PHFH

AcadÉfiic r ¿001-e0ou

CöuFSe Nðr'le

EngineÉr-ing Hach¿rricsConstructiün EnginÊeringCheñistrg IIChenistng Lab IIlf at-henåtics I I1..!ork shop PråcL i cÊ I IEr'gineering Gr'aphics andDr" åu i ngNS0PhgsiES IIFhgs iEE Låh I I

Senester:1

Credit Grade

8.0 AA6.0 AA3,0 AA ..,,.

1 ,5 AA., "5. O AA..6.0 AA': .

.l

6.0 ÌAA"8.0 :rEB

10e1se10e11ó104LL?119

ra?1 0-c11ó

SPI ¡ 9.44 CPI:9.39

C0dÉ

CE ê04cE e0ócE a 1.6

cE el8CE Ê80

CE ???EE OOEEE OO3

HS EOE

SPmËsteF : Ê

CPedit GrèdÉ

6.0 AA6,0 AA6.0 AB1,5 ABS.O AE4,0 AE5.0 AE

O.O PP6.0 AA1,5 AA

ÐE¡tts5Lr:f ! ¿

CrÉdi¿ GrádÈ

6.0 AA6.0 AA3.0 AA1.5 AAÉ.0 AA

6.0 AA6:,0 AA1,5 .'AAé.O AA

SenesLPr' I Z

CFedit Ërade

É,0 AA6.0 AA6.0 AA6.0 AA1.5 AA

3.0 AAS.O AAO,O PP3.O AA

AcådËilic e e00U-Ê0O3

CouFse Nåne

GÈüdesu IFluid MÊchåniEs I IGÉúdësg Låb IFluid Mechånics Lab I I[,,]åLeF and [,JåsÈe [^lat-eFEhgineeringSLnuctunal MeEhaniEs IPrinciptÈs of ElectriÈå1 EnggFrinciples c,f ElecLr'icð1Eng i neer^ ing LabInt-roducLi*n ta FhilÕsüPhU

9enÊsteP 3 1

credit llrade

ó.0 AAó.0 AAó.0 AA3.0 AA3.0 AAó.0 AA1.5 AAó.0 AA3.O AB

sPI:10.0 cFI;9.6

Çode

cE 30ecE 304cE 306cE 30&cE 31ecE 314cE 318CE 3EOcE 396cE 398

AcådÉnÍc ¡ eO03-20O4

C.JursÉ Na{¡e

Geüdes9 I ISsil, llechanÍcs IISLnucturål Mechånics IIIDesign of SLruEturÉs IIËÉodÉs9 Lab IISDi I l"lÉchånics LËb I IDÉEign Lèboråtorg IITrðìlsportåLicn Engg ItloFks VisitsSErqinaF

sPI ¡ 10.0 CFl | 9.76

CodÉ Course Nane

çE 3SB Prècticöl TråiningCE 403 Design of StFuctul^es IIICE 4o7 Foundðtion Engg.CE 411 Design LáboFatorg IIICE 415 TransPortation EngineerÍng IICE 417 tlater Resources EngineeringCE 4ó3 Probåbilistic and St¿tistical

l'lethodE i11 Civi I EngCE 497 B'Tech. PFoiecL (Gråde Poirl¿5

uill be included ihES 403 Envi.FonñentðI StudiesHS 487 Plahnihg and Econúnic

Policies in India

SpI ¡ 10.0 CpI ¡ 9.8

TBTAL CREDITS EARNEDTOTAL CUHULATIVE POINTS

credit Gt'ad*

0.oó.0é,03.0é,oó.o6.O

8.0

0.0é.0

PPAAAAAAAAAAAA

| 4?2.?5:415e.0

Co de

cE 46?

cE 4S9cE 499CE óEOcE 754

HS 496

FPAA

CoutrsÉ NanÉ

Elenents of SLruct uFå1DgnåmicsPåvenenL SgstÈms EngineEringB. TÉch. Proiect- I IFinite Elenent HethodsEEonúr¡ic Evalu¡ticn andAnålUËiE of Tr'ånsPortåtiûnFrÐ .j eE t,Ë1ûhå1izétion: A Súciè1PerspeELivÉ

sPI : tO.0

Cr-edit GFådÈ

6,0 AA

6.0 AA15.0 AAé.0 AAó.0 AA

6.0 AA

OF ADHiSSIONYEAF & I'IONTHOF ÊRADUATION

cFI | 9.93

JUL EOOl

HAY EOOSDEFUTY REGISTRAF/AS

islaceneuIcr\X .lrstt [icgistrar' (Aca<leuric)L l.'L I-lor;:l¡av --lùù ù76.

I I ld¡'lvtassacfruseN lnst¡tute of Technology

Acader¡ticTranscript

@ffEce øf ËE"oe RegE N*u, 77 Massachusetts AvenueCambridge, Massachusettso2139-4307

Vikrant Suhas Vaze

HIT ID: 92L 844 962 Birthdate: 28'JUN-1983

Adm'itted as a Regular Student for Fall Term 2005'2006

from INDIAN INST 0F TECHNOL0GY

HUMBAI, INDIA

Program/Degree Obiective as of Current Term:

civil and Environmental Engineering (Transportat'ion

Studì es ) /Doctorate

Operat'ions Research/l'laster' s

Subject Subiect Ñane Lvi Creri Grade

SPRING TERM 2008'2009 COURSE: 1 D

1.982 Research: Civìl & Envìron Eng

6.855 Network 0ptim'ization***

SUHI4ER TERI4 2009 COURSE: 1 D

1.982 Research: C'ivil & Envìron Eng

***FALL TERM 2009-2010 COURSE: 1 D

L.23L Plannìng & Design:Aìrport Sys

L.982 Research: Civ'il & Env'iron Eng

***JANUARY TÊRl'l 2009'2010 COURSE: 1 D

1.THG Thesis

Subject Subject Name Lvl Cred Grade

GRADUATE STUDENT

N24PHL?A

GRADUATE STUDENT

N24P

GRADUATE STUDENT

HI2AN24P

GRADUATE STUDENT

HL2JFALL TERl'l 2005-2006 C0URSE: 1 CT

L.221 Transportat'ion Systems H

L.222 Transport Demand & Economics' -H

-*.223 Transp Po'l'icy, Strategy, Hgmt H

L.224 Carrier Systems , ,', 'H

1.225 Transportatìon Flow Systems H

1.264 Data Internet & Sys Integ Tech H

1.982 Research: Civil & Envirón ,Eng N,

**rrSPRING TERl,l 2005-2006 COURSE: L CT

L.202 Demand }4ode]ing H

L.982 Research: C'ivil & Env'iron Eng N'

6.431 Applìed Probability G

* * * t.. .,,. ,.

FALL TERM.2006-2007 COURSE: 1 CT '..L.L24 iFoundatìons of Software Eng 'tl'L.g78 Sp Grad Stud: Civ & Env Eng G

I.982 Research: Civil & Environ Eng N

I.THG Thesis H

2.098 ' Optimìzat'ion Hethods H

***JANUARY TERM 2006'2007 C0URSE: 1 CT

1.982 Research: Civil & Envìron Eng N

1.THG Thesìs H

***SPRING TERI4 2006'2007 COURSE: 1 CT

t.234 Airline Management H

L.982 Research: Civ'il & Environ Eng N

1.THG Thesis H

***FALL TERl'l 2008'2009 C0URSE: 1 D

1.203 Logistical &Transport Planning H

L.982 Research: Cjvil & Env'iron Eng N

1.983 Teaching: Civil & Env'iron Eng N

***-- ContÍnued in Next Column -'

GRADUATE STUDENT

6¡' A I'6 .A ,,

.6' A

,,6 A '

:6 A,t, '

LzA ILzP.

GRADUATE STUDENT

:r 12 A .._.-:: . l,,.15 P::.....

't? A..lr !,\ :., ::

GRADUATE STUDENT

L2 .A '..:

l-P'20 "P.

20 illA .::

t2A

GRADUATE STUDENT

1PL JlA

GRADUATE STUDENT

L2A20P204

GRADUATE STUDENÎ

L2A33P15P

Jr**********************x****************************************

08-JUN-2007' Awarded the Degree of Master of Science in' ., TransPortatìon .

05.FEB.2009,Doçtoial'General Exam'inat'ion completed

********'rJr**.*****tr******rr*********rkr(***rr*********************

Graduate Cumulative GPA: 5.0 (on a 5.0 scale)*******i***-***tr**********************************************

-. END OF RECORD -.-- No Entries Valid Below This Line "

OFFICIAL TRANSCRIPT: ISSUED 10-HAY'2010

0rder #: 000942i18

Issued to

LAHdENCE D. GOLDSTEIN

il,:i å'tä[;:: :Hfflï:tu{ ffiM #^. {:mÅ{aÅ*.*.This transcript is printed on security påper and does not requìre a raised seal.

To confirm authenticity, see reverse side. lnformation must not be disclosed

to other parties without prior wrìtten consent of the student.A black and white document is not an original