基于最优生产技术的业务过程资源动态随机优化配置

157计算机集成制造系统Vol.15No.720097ComputerIntegratedManufacturingSystemsJuly2009:1006-5911(2009)07-1414-07:20080610;:20081123Received10June2008;accepted23Nov.2008.:(70671094);(Q0921);(2009C21G2050028);;(20061084)Foundationitems:ProjectsupportedbytheNationalNaturalScienceFoundation,China(No.70671094),theScienceFundforYoungScholarsofZhejiangGongshanUniversity,China(No.Q0921),theScientificResearchFundofZhejiangProvinicalScienceandTechnologyDepartment,China(No.2009C21G2050028),theProgramforNewCenturyExcellentTalentsinUniversity,China,andtheScientificResearchFundofZhejiangProvincialEducationDepartment,China(No.20061084):(1975-),,,,,CIMSEmail:[e-mail],(,310018):,,,,,,,,,:;;;;:TP311:ADynamicalstochasticresourceallocationoptimizationforbusinessprocessbasedonoptimizedproductiontechnologyXIEYi,JUChunhua(InstituteofComputer&InformationEngineering,ZhejiangGongshangUniversity,Hangzhou310018,China)Abstract:Theexsistingresourcestaticallocationoptimizationmethodswithschemawerelackofflexibilityandadaptabilitytothedynamicsanduncertaintiesoftherealbusinessprocessexecution,anddynamicalresourceallocationpolicieswithsharedqueueorloadbalancehadlittleroomforglobaloptimization,sotraditionalresourceallocationmethods/policiescouldntguaranteeglobalperformanceoptimizationofbusinessprocessexecutionunderuncertainenvironments.Therefore,anewmethodwasproposed.Inthismethod,abusinessprocessanalysismodelfordynamicalstochasticresourceallocationoptimizationwasbuiltwithindividualresourcequeueandstochasticprobabilitytheory.Then,thedynamicalstochasticresourceallocationoptimizationmethodbasedonOptimizedProductionTechnology(OPT)waspresented.Finally,theeffectivenessandadvantageofthismethodwereverifiedbyanexampleandsimulationexperiment.Keywords:businessprocess;modelbuilding;optimization;resourceallocation;optimizedproductiontechnology0,,,,7:[1](ProgramEvaluationandReviewTechnigue,PERT)/(CriticalPathMethod,CPM),,,[2]~[5]M/M/1M/M/C,,PERT/CPM,[6][7],,,//[8][9],,,[10][11],[12][13]~[15],/,[16],/(buildtime),(schema),,,(runtime),(sharedqueue)(loadbalance)[1719]/,,(FirstinFirstService,FIFS)(ShortestProcessTime,SPT)(EarliestDueDate,EDD)(MinimamSlackTime,MST)//,//,FIFO,SPT,EDD,MST/,/,,,(individualqueue),(OptimizedProductionTechnology,OPT),,,111,,1(1)(1):(1),2,n(n3),141515pi,:,01,,n+2,()2n,n3n+22n,,,pi,,2,,(2),2a(p1+p21),2b,,,,11)顺序基本模型结构,,,,2)并行基本模型结构,,,3)选择基本模型结构,,,,4)循环基本模型结构,,,12PM=(A,C,L,R,U,,,),:(1)A;C;N=AC,n!N,|∀n|n,|n∀|n;ns!N,|∀ns|=0,,ne!N,|n∀e|=0,;a!A,|∀a|#1∃|a∀|#1;c!C,(|∀c|1∃|c∀|#1)%(|∀c|#1∃|c∀|1)(2)L!A&AA&CC&AC&Cl!L,l=n1,n2n1n2(3)R(4)U!A&R,u=a,rr/a(5)∋C({And,Or},(6)∋L([0,1];l,l∀{c,n|(c)=OR∃|c∀|1},(l)=1,l!{c,n|(c)=OR∃|c∀|1},)|c∀|i=1(li)=1(7)∋U(R+,,u!U,u=a,r,ra(a,r),a,r131fa,afa:1a,fa,a1f,a1,∗,anf14167:1b,f,a1x,x=f+x(1-p),x=f/p1c,f,a1,∗,anf1d,f,a1,∗,anp1f,p2f,∗,pnf,)pi=1,pi/,,,,,,(),2!a,rar:)r!Ra!a,r=1,Ra={r|a,r!U}3∀a,rra∀a,r=!a,r∀fa/a,r)a+!Ar!a+,r∀fa+/a+,r,Ar={a|a,r!U}4#a,/a:#a=)r!Raa,r∀∀a,r,Ra={r|a,r!U}5∃%=#a/fa221OPTE.Goldratt70,,,,OPT,22,∃a,OPT,,:maxmin{∃a}s.t.)r!Ra!a,r=1,a!A;0#!a,r#1,a,r!U,Matlab23,/,,,,,OPT,,OPT:r,,!a,ra33,(),A={a1,a2,a3,a4,a5,a6,a7,a8},(),R={r1,r2,r3,r4},1,,E,U:(c1)=(c6)=And,(c2)=(c3)=(c4)=(c5)=Or:(c2,a3)=04,(c2,a4)=06,(c5,c4)=02,(c5,c6)=08,11417151a1a2a3a4a5a6a7a8r1E(14)U(4,12)E(10)E(12)r2E(8)E(12)U(6,10)E(4)E(2)r3E(6)U(2,8)U(6,18)E(8)E(4)r4E(8)E(14)E(6)E(18)E(3)22Matlabfminmax(),OPT(!a,r),2maxmin{∃a}00852,/008522OPT!a,ra1a2a3a4a5a6a7a8r1011750030410r20006831106391r31096960316900r40882501003609∃a0085200852008520085200852008520085200852/,,,!a,r01,0~1,1ar,0ar,220#!a,r#1:!a,r=1!a,r=0OPT,33OPTa1a2a3a4a5a6a7a8r10010r200110r311001r410100OPT(POPT)OPT(DPOPT)OPT(COPT)(SQ)(LB)(FCFS),GPSS3(),()(),445:4COPTSQLBPOPTDPOPT00067380820078482840104565940105400660078375150076001003952201164848301556071501564158501173857801140014342056016649861022067254022443965016740900016300200456550233520360310792850315484540234446820229002635089403065636004059239004165318903075053103020033358023038861977051011457705226268203915682803830041765695048575595063617358706587396504886552804840050081841058199022075928076007918812005867607705800058810690306841681080888803530092411348706899775206850066714018207759359301000166943078212769807790072523425708422153550843165080084814187:40076979133908883061320892199610090200800381128092825388709360083363321109653685910977008516100230993(1)DPOPTPOPTCOPTSQLB(2)(0069,080),LBSQPOPTCOPTDPOPT;(0068,080),LBSQCOPTPOPTDPOPT(3)LB,SQCOPT,POPT,DPOPT,OPT(DPOPT),,,,/(r1r2),a1r4,,a1r2,(),SQLB/,,COPT10,0-1,/;POPT/DPOPT01,,/,,POPT/DPOPTCOPT4,OPT,/,:(1),,,,/,141915(2),,,,/(3),/,,(4),DPOPT,,/,:[1]ZHANGXiaoguang,CAOJian,ZHANGShensheng,etal.Criticalpathanalysisforstructureoptimizationofworkflow[J].JournalofShanghaiJiaotongUniversity,2004,38(1):2933(inChinese).[,,,.[J].,2004,38(1):2933.][2]CHANGDH,SONJH,KIMMH.Criticalpathidentificationinthecontextofaworkflow[J].InformationandSoftwareTechnology,2002,44(7):405417.[3]SONJH,KIMMH.Improvingtheperformanceoftimeconstrainedworkflowprocessing[J].JournalofSystemsandSoftware,2001,58