paper 127[1]

8/6/2019 Paper 127[1]

1/13

ThirdInternationalConferenceonProductionResearchThirdInternationalConferenceonProductionResearchThirdInternationalConferenceonProductionResearchThirdInternationalConferenceonProductionResearch

AmericasAmericasAmericasAmericasRegion2006Region2006Region2006Region2006(ICPR(ICPR(ICPR(ICPR----AM06)AM06)AM06)AM06)

IFPRIFPRIFPRIFPRABEPROABEPROABEPROABEPRO----PUCPRPUCPRPUCPRPUCPR----PPGEPSPPGEPSPPGEPSPPGEPS

INTEGRATINGPROJECTMANAGEMENT,CONCURRENTINTEGRATINGPROJECTMANAGEMENT,CONCURRENTINTEGRATINGPROJECTMANAGEMENT,CONCURRENTINTEGRATINGPROJECTMANAGEMENT,CONCURRENT

ENGINEERING,ANDENGINEERINGDESIGNTOIMPROVESHIPENGINEERING,ANDENGINEERINGDESIGNTOIMPROVESHIPENGINEERING,ANDENGINEERINGDESIGNTOIMPROVESHIPENGINEERING,ANDENGINEERINGDESIGNTOIMPROVESHIP

DESIGNDESIGNDESIGNDESIGN

EmersonPieroni

GraduateStudent-COPPE/[email protected]

RicardoManfrediNaveiro

ProfessorESCOLAPOLITCNICAandCOPPE/[email protected]

Abstract:Abstract:Abstract:Abstract:

There are different perspectives among the design and development research

community. Past research shows that there are at least four common perspectives

addressing this issue. Furthermore, there are significant differences among papers

withineachof the perspectives,not only in the methodology usedandassumptions

made, but also in the conceptualization of how product development is executed.

Project Management, Concurrent Engineering and methodologies of EngineeringDesign are examples of those perspectives and are usually treated as isolated

disciplines.However,therearemanyinteractionsamongthetasksrelatedtothem.We

investigate these interactions,common tasks andconcepts.Our research integrates

these viewpoints with the main concepts, tasks and decisions providing a

comprehensiveunderstandingofshipdesign.

Keywords:Keywords:Keywords:Keywords:ProductDevelopmentProcess,DesignMethodology,ShipDesign.

1111 IntroductionIntroductionIntroductionIntroduction

There is no single theory to the product development process. The differences of

products developed by firms result in particular methodologies with emphasis on

certainaspectsoftheproduct.KrishnanandUlrich(2001)showthatthereareatleast

four common perspectives addressing the product development process:marketing,

organizations,engineeringdesign,andoperationsmanagement.

8/6/2019 Paper 127[1]

2/13




However,theauthorsobservedthatthedecisionsmadeduringtheprocessseemto

remainconsistentatcertainlevelofabstractionforallfirms.Thisisacommonpointof

allproductdevelopmentprocesses.Moreover,therearefamiliarcharacteristicsoftheproductdevelopmentprocessindependentoftheproducttobedesigned.

ProjectManagement,ConcurrentEngineeringandEngineeringDesignareexamplesof

thoseperspectives.However,therearemanyinteractionsamongthetasksrelatedto

them.Someofthesetasksarerepeatedovertheparticularmethodologies.Wemust

knowtheseinteractionsinordertoimproveunderstandingofShipDesign.

ThispaperpresentsastudyofShipDesign.WeusetheDesignStructureMatrix(DSM)to thispurpose.Wemap the interactionsamongthemainactivitiesacceptedbythe

literaturerelatedtotheProjectManagement,totheConcurrentEngineering,andtothe

ShipDesign,andwithDSMestablishasequenceofrealizationofthem.Itresultedina

singlewaytovisualizetheShipDesignandthemaintasksinvolvedinit.Weshowthat

this representation helps the Project Manager to take strategic decisions about

CommunicationalandOrganizationalPlanningandProgramManagementTasks.

2222 ProductDevelopmentPerspectivesProductDevelopmentPerspectivesProductDevelopmentPerspectivesProductDevelopmentPerspectives

Krishnan and Ulrich (2001) present four different perspectives of the Academic

Communities about the Product Development Process (PDP). The perspective on

product,typicalperformancemetrics,dominantrepresentationalparadigmanddecision

variablesaredifferentforeachfield.Despitethedifferencespresentedbyeachgroup,

there is a certain consistence about the decisions made during the design. Those

decisionsseemtoremainregularatcertainlevelofabstractionforallproductdesign.

However,thefirstpointtobeobservedisthatthedecisionsaretimedependent.They

aremadeduringseveralpointsofthePDPandarearesultfromthedifferentlevelof

participationof each field during the PDP. For instance, the Marketing Teamhas a

greatlevelofdecisionsmadeduringthefirststepofthePDP,despiteitscontributionto

thewhole designcycle.Other point tobeaddressed is thatmanyperspectives and

decisions are inserted in methodologies and theories of Project Management,

EngineeringDesign,andConcurrentEngineering.Thosedisciplinesareusuallytreated

inanisolatedway.However,thosedisciplinesarecomplementary.

8/6/2019 Paper 127[1]

3/13




AccordingtoPMBOK(2000),projectsarecompoundbyprocesses,andaprocessisa

successionofactionsthatconductstoaresult.Therearetwotypesofprocess:project

managerial process and product oriented process (PMBOK, 2000). The first one isrelated to the description, organization and conclusion of the work and is the main

focusofprojectmanagement.Thesecondisrelatedtothespecificationandcreationof

theproductandiscalledEngineeringDesign.Inthecaseofshipdesign,theproduct

orientedprocessisusuallyaddressedbytechnicalshipdesignprocess.Despitethis

differentiation,thereareinteractionsbetweentheseprocesses.Theyinteractduringthe

wholedesigncycle.Forinstance,thetimemanagementcannotbedefinedwithoutthe

knowledgeoftheactivitiesoftheshipdesign.

Concurrent Engineering incorporates that group of processes by means of new

requirements (that insert new activities to address them) and parallel planning of

activitiesexecution.ConcurrentEngineeringstrivestodotherightjobrightthefirsttime

and isbasedon two fundamentalobservations (ZANGWILL, 1992). The first is that

changesbecomemoreandmorecostly,thelatertheyaredoneinaproject,andthe

secondisthatdoinginparallelthedifferentstepsofaprojectmaketheprojectdone

more quickly then doing the steps sequentially. The consequences are that new

requirementsofproduction,maintenance,operations,etc.mustbeaddressed duringtheearlierstagesofdesignandthedependenciesamongthemmustbeanalyzedin

ordertoexecutetheactivitiesinparallel.

Todaytherepresentationoftheshipdesignprocessismadeinamacroscopiclevel,

generally in a system level. Itmasks some particularities of the process.Whenwe

analyze the process in the activity level we obtain a better understanding of the

process.

3333 ShipDesignShipDesignShipDesignShipDesign

To illustrate an engineering design methodology we present the ship design spiral

(figure2).Theconceptofthedesignspiralwasintroducedin1959byProfessorJ.H.

EvansofMIT(Evans,1959),andisthemostwidelyacceptedmethodologyoftheship

design(LAVERGHETTA,1998).

Thespiraldescribestheprocessasasequenceofspecificdesigndisciplines,bothof

synthesis(e.g.hullgeometry,arrangement)andanalysis(e.g.stability,seakeeping)in

8/6/2019 Paper 127[1]

4/13




order toachievea balanceddesign thatmeets the requirements. The spiral depicts

important characteristics of the design like the iterativeness and the progressive

elaborationofthedesign.However,thespiralrepresentstheprocessinamacroscopiclevel.Someofthosedisciplinesincorporatehundredsofactivities.

FigureFigureFigureFigure1111::::GenericShipDesignSpiralGenericShipDesignSpiralGenericShipDesignSpiralGenericShipDesignSpiral

When we group the activities into disciplines we mask the interactions among the

activitiesportrayingthedesignprocessaslinear.Thisisnotthecase.Theprocessis

better described as quasi-linear (LAVERGHETTA, 1998). As point Laverghetta,

designactivitieswithinthespiralrelyoninputfromandprovideoutputtoalmostevery

node of the process. As such, engineers and designers must have access to

information(whetheractualorestimated)fromeachdesigndisciplineandtheymustbe

acutely aware of the potential feedbackeffects caused by the changing output from

their own products. To control it, managerial activities play a capital role into the

process.However,itisnecessarytomaptheactivitiesinteractionsoftheprocess.

4444 IntegratingmethodologiesandtheoriesIntegratingmethodologiesandtheoriesIntegratingmethodologiesandtheoriesIntegratingmethodologiesandtheories

4.1 Concepts of Design Structure Matrix

Toanalyzetheinteractionsweusethedesignstructurematrix(CHO,2001).Weuse

DSMbecausethetoolsindicatedbyPMBOKarebasedondiagramsand,forcomplex

systems,isdifficulttoanalyzetheinteractionsasitwillbeshownlater.Browning(2001)

Requirements

PrincipleDimensions

HullGeometry

Arrangement

Resistance

PropulsionSystemStructure

AuxiliarySystems

Weights

Stability

Seakeeping

Cost

8/6/2019 Paper 127[1]

5/13




providealiteraturereviewoftheuseofDSMandEppingeretal(2004)providethe

conceptsofDSM.

TheDSMisasquarematrixinwhichisassignedthetasksidenticallyorderedatrow

andcolumns.The initialprocess istomarkthecellsto identifywhena row receives

informationfromacolumn.Figure4showsthattheoriginalDSM(leftmatrix)presentsa

seriesoftasksorderedaccordingexecutionsequencebothatrowsandcolumns.Note

that the row A has the number 1 marked at columns C and E. It means that the

executionoftaskAneedsinformationfromtasksCandE.Ontheotherhand,looking

atcolumnA,onecannotethenumber1markedatrowsBandC.Itmeansthatthe

task A gives information to tasks B and C. Above diagonal marks has a relevantsignificance:itindicatesthatanearliertaskisdependentonalatertask.Theoriginal

DSM(leftmatrix)showsthattaskAdependsoninformationoftasksCandEthatwere

notexecutedyet.

Figure2Figure2Figure2Figure2ExampleofoptimizationprocesswiththeDSM.ExampleofoptimizationprocesswiththeDSM.ExampleofoptimizationprocesswiththeDSM.ExampleofoptimizationprocesswiththeDSM.

After concluding the establishingofmarksof the originalmatrix itmustbeexecuted

partitioningalgorithmstoreorderthesequenceoftaskstoreducethenumberofabovediagonalmarks.Itpermitsthatallnon-necessaryiterationswereretiredoftheprocess.

Itcanbeverifiedbylesseningmarksequalnumber1abovethediagonal.Ontheother

hand,taskswithnecessaryiterationweregrouped,asnotedintherightmatrixoffigure

4(tasksAandC;tasksF,HandI).Thistypeofgroupinghasspecialimportancefor

team organization grouping persons who necessities an intense changing of

information.Also,notethepresenceoftasksthatcanbeexecutedinaparallelway,as

tasksGandL.TaskLdoesnotdependoninformationfromtaskGandvice-versa.

8/6/2019 Paper 127[1]

6/13




Finally, task D depend on information from task B to be executed and it is called

sequentialtasks.

4.2 Planning Activities interaction analysis

Firstofall,weanalyzetheplanningactivitiesinteractionsofaproject.Itwilldefinethe

bestorderofexecutionofplanningactivities.Toanalyzetheinteractionamongthose

activitieswebaseonPMBOK(2000).ThePMBOKliststhemainactivitiesrelatedto

theprojectmanagementandindicatesinformationinputandoutputforeachactivity.

Basedon the interactions,we insert the dataon the DSM (figure3)and process it

(figure 4). Activities of Quality Management, Risk Management and Integration

Managementareretiredof theprocessbecausetheyprovideandreceiveinformationofalmostallactivities.Consequently,theyplayaroleofintegrationactivitiesandmust

be considered along the process.Other information input related to historical data,

premisesandrestrictionsdependontheofficeandtheproducttobedevelopedand

mustbeconsideredaccordinglytotheoffice.

Aftermappingtheinteractionsamongtheactivities(figure3)weprocesstheDSMand

it result in thebest order of executionof the activities (figure 4). The neworder of

executiondiminishesiterationprocessesascanbeobservedlookingabovediagonalmarks.Theonlyiterationprocessisthatnecessarytotheexecutionoftheplanning,

where there are mutually dependencies. Indeed, resource planning and activity

durationestimatingaremutuallydependent.For instance, the activity ofshipweight

estimatingcan beperformedbyone ormoreengineers. The greater the number of

engineersallocatedtotheactivitythelesserthetimeofexecution.

Figure3Figure3Figure3Figure3DSMbeforeprocessing.DSMbeforeprocessing.DSMbeforeprocessing.DSMbeforeprocessing.

8/6/2019 Paper 127[1]

7/13

8/6/2019 Paper 127[1]

8/13




specification of the product, like structural detailing, electrical system detailing and

others. If the scope detailing changes, new activities will be addressed and is

necessary to map the interactions among those activities. Consequently,communications planning and schedule development will change. Despite

organizationalplanningmightbeaffected,itmaybenotimportanttochangeit.Itmust

be analyzed. Consequently, at each step of development we must reevaluate the

planningoftheproject.

The third point to beobserved is that the principles ofConcurrent Engineering are

contained inside of the Scope Detailing and Definition, of the ActivityDefinitionand

Sequencing, and of Schedule Development. Consequently, changes in the scopedetailinganddefinitionandintheactivitysequencingaffectpracticallyalltheactivities,

including Communications Planning, Organizational Planning, Resources Planning,

Schedule Development and others. With the implementation of the concepts of

ConcurrentEngineeringnewactivitieswillbedefinedanditisnecessarytomapthose

interactions.Itwillaffectallplanningactivities.

4.3 Engineering activities interaction analysis

Followingtheorderofactivitiesdefinedbyfigure4 toexecutetheengineeringdesign,thefirststepistoplanandtodefinethescopeofthedesign.Tothepurposeofthis

paper the scope is to designa generic ship. The requirementsof this ship are that

commontoalmostallships,likevelocity,stability,seakeepingandothers.Thereare

specificactivitiestoaddressthoserequirements.Todefinetheactivitiesandtomapthe

interactions among them, we base on the Principals of Naval Architecture (Lewis,

1989),traditionalbookofNavalEngineering,onLaverghetta(1998),andonBogosian

Neto(2005).Atleastallactivitieslistedonthispapermustbeexecutedtodesignships.

However,eachshiphasitsownrequirementsanditmustbetranslatedintoactivitiestoaddress them.Themanagermust knowhow theseactivities interactwith the others

activitieslistedonthispaper.

Aftertheactivitydefinition,lookingatfigure4,thenexttwostepsareresourceplanning

andactivitydurationestimating.However,weobservethatactivitysequencingdoesnot

depend on thoseactivities andmust beperformed ina parallelway. Therefore, we

analyzetheactivitysequencing,mappingtheinteractionsoftheactivitiesbecauseof

thescopeofthispaper.

8/6/2019 Paper 127[1]

9/13




Figure6showstheDSMbeforetheprocessinglistingtheactivitiesneededtoexecute

theshipdesignandmappingtheinteractionsamongthemasexplainedearlier.Figure

7showstheDSMaftertheprocessing.Figure8showstheprecedencediagramoftheactivitiesbasedonfigure6.

Figure6Figure6Figure6Figure6DSMbeforeprocessing.DSMbeforeprocessing.DSMbeforeprocessing.DSMbeforeprocessing.

Figure7Figure7Figure7Figure7DSMafterprocessing.DSMafterprocessing.DSMafterprocessing.DSMafterprocessing.

8/6/2019 Paper 127[1]

10/13




Figure8Figure8Figure8Figure8PrecedenceDiagrambasedonactivitiesoffigure6.PrecedenceDiagrambasedonactivitiesoffigure6.PrecedenceDiagrambasedonactivitiesoffigure6.PrecedenceDiagrambasedonactivitiesoffigure6.

WeobservethatDSMaftertheprocessingreorderedthesequenceofaccomplishment

of activities that diminished the number of iterations of the project, what it can be

visualized by the number of markings above of the main diagonal line in the

comparisonoffigures6and7.Also,observingtheDSMaftertheprocessing,wenotice

thesproutingoftwoblocksthatneedaniterativeprocessofresolution.Thesearethe

calledplannediterationsoftheprocessandarenecessarytotheproject.

When identified those coupled activities the manager must be aware of the time

schedulingtoexecutemajorloopsontheprocess(e.g.activitiesfrom13to21offigure

7).Sometimes,despitemeetingtherequirements,withotherloopsonesystemcanbe

optimized.However,thereisatimelimitation.Thisisonepointofmanagerialdecision

(Management Control figure 7 line 22). Also, it can be discovered errors, new

information, and other unplanned troubles (e.g. bad estimative) that necessity to

reevaluate some technical activities creating major loops. Quality and Risk

Managementmust beaware of these facts creatingactivities of control.The others

8/6/2019 Paper 127[1]

11/13




managerslikeTime,Scopeandsoonmustbeinformedofthedecisionsmadeinorder

toreevaluatetheirplansandmakedecisionsabouttheexecutionofthedesign.

Identifyingiterationloopsplaysastrategicfunctiontotheoffice.Somekindofplanned

iteration canbeperformedbyspecificsoftware. Itcanbepurchasedby theofficeor

even developed. It will depend on the objectives of the office. Sometimes a low

investment can better the quality and can accelerate the design, diminishing risks.

Onceidentifiedtheiterationgroupitmustbeanalyzed.

Also,analyzing the interactionswe can plan the communicationsof thedesign. It is

easytoidentifywhoneedsinformationandwhoprovidesinformationaswellitiseasytoidentifywhentheinformationmustbeavailable.Observingthecolumnnamedlevel

figure7weidentifycoloredgroupsofactivities.Ifthosegroupsarenotassociatedwith

planned iteration, it means that those groups can be performed in a parallel way,

becausetheydonotdependoninformationofeachother.Forinstance,itcanbeseen

lookingatactivities10,11and12.Theycanbeperformedinaparallelway.However,

activities13,14,15and16areassociatedwithaplannediterationthatcanbeseen

lookingtheblueblockinsidethematrix.Theymustbeexecutedbymeansofiteration.

Otherpointtobeobservedisthesimplicityofrepresentationofinteractionsusingthe

DSM. Figure 8 represents the precedencediagramof the interactions illustrated in

figure6.Wenoticethatiscomplicatedtoanalyzetheinteractionsofprojectbymeans

of precedence diagram indicated by PMBOK (2000). Observing figure 7 we easily

distinguish which are the activities that can be executed in a parallel way. Also,

accordingly to the principles ofConcurrent Engineering, sequential activities can be

donewithsomedegreeofparallelismoroverlapping,butitmustbeanalyzed.Inorder

tousetheoverlapping,Yassineetal(1999)affirmthatitisimportanttoestimatetwocharacteristics:upstreaminformationvariabilityanddownstreamtasksensitivity.Todo

itisnecessarymaptheinteractions.Therefore,tomaptheinteractionsisvitaltothe

establishmentofsomeprinciplesofConcurrentEngineering.

Finally,theactivitieslistedinDSMunmaskedthesimplicityoftheShipDesignSpiral.

ToaddressthecharacteristicslistedintheSpiralwenecessitiesmanyactivitiesand

therearealotofinteractionsbetweenthem.Therefore,itisnecessarytoexplodethe

Spiralintoactivitiestomanagethedesign.Furthermore,theactivitieslistedarethatof

8/6/2019 Paper 127[1]

12/13




thefirstcycleofthespiral.Forothercycles,thenumberofactivitieswillbegreaterand

themostofthemwillchange,requiringanalyzingtheplanningoftheproject.

5555 ConclusionsConclusionsConclusionsConclusions

Thispaper integrates theconceptsandactivitiesof threemethodologiesandtheories

about PDP: Project Management, EngineeringDesign and Concurrent Engineering.

Also,weshowthatinteractionanalysisbasedonnecessaryactivitiestoaccomplishthe

design plays a vital role on the project. It is the first step to improve ship design

methodology.

However,thispaperonlypresentssomeexamplesofanalysisandconclusionsaboutinteraction and the model presented. Many other conclusions can be taken from

interactionanalysis.Alongerworkcouldexploretheestablishmentofrequirementsand

analyze the necessary activities to accomplish those requirements. Also,

communications and organizational planning can be deeper analyzed by means of

interaction analysis. Finally, the concepts presented can be extrapolated to other

productdevelopmentprocess.

AAAAcknowledgmentscknowledgmentscknowledgmentscknowledgmentsWe are grateful to the Brazilian Navy, to the Production Engineering Program of

Federal University of Rio de Janeiro and to the GEPRO (Gesto Integrada do

DesenvolvimentodeProdutosIndustriais)formakepossiblethisresearch.Thiswork

doesnotnecessarilyreflecttheopinionof theBrazilianNavyandtheauthorsassume

theresponsibilityfortheinformationcontainedinthispaper.

ReferencesReferencesReferencesReferences

BROWNING, T. R. (2001) - Applying the Design Structure Matrix to System

Decomposition and Integration Problems: A Review and New Directions. IEEE

TransactionsonEngineeringManagement.Vol.48,n.3,p.292-306.

BOGOSIANNETO,S.(2005).ProjetodeConcepodeumNavioPatrulhadeFormas

No-ConvencionaiscomCaractersticasSubmersveis:ViabilidadeeManobrabilidade.

Portuguese. 410p.Thesis (DoctorofScience inOceanEngineering).UFRJ,Riode

Janeiro.

8/6/2019 Paper 127[1]

13/13




CHO,S.H.(2001).AnIntegratedMethodforManagingComplexEngineeringProjects

Using theDesignStructureMatrix andAdvancedSimulation.Dissertation (Masterof

Science).MassachusettsInstituteofTechnology.

EPPINGER, S.D., WHITNEY, D.E., YASSINE, A.A. (2004) - The Design Structure

MatrixDSMhomepage..

EVANS,J.H.(1959)-BasicDesignConcepts.NavalEngineersJournal.Nov.,p.671-

678.

KRISHNANV.,ULRICHK.T.(2001)-ProductDevelopmentDecisions:AReviewoftheLiterature.ManagementScience.Vol.47,n.1,p.121.

LAVERGHETTA,T.A.(1998)-DynamicsofNavalShipDesign:ASystemsApproach.

187 p. Dissertation (Naval Engineer e M.Sc. in Ocean Systems Management).

MassachussetsInstituteofTechnology.

LEWISE.V.(Editor)(1988).PrinciplesofNavalArchitecture.Vol.I,IIandIII.Second

Edition.TheSocietyofNavalArchitectsandMarineEngineers,NewJersey.

PMBOK(2000)-ProjectManagementBodyofKnowledge.PortugueseVersion.V.1.0.

159p.ProjectManagementInsitute.MinasGerais.Brazil.

ZANGWILL,W.I.(1992).ConcurrentEngineering:ConceptsandImplementation.IEEE

EngineeringManagementReview.Vol.20,n.4,p.40-52.

YASSINE,A.,BRAHA,D.(2003)-ConcurrentEngineeringandtheDesignStructureMatrix.ConcurrentEngineeringResearchandApplications.Vol.11,n.3,p.165-176.

paper 127[1]

Documents