n人随机重复博弈下的策略选择研究

3611Vol.36,No.11200611JOURNALOFUNIVERSITYOFSCIENCEANDTECHNOLOGYOFCHINANov.2006:025322778(2006)1121171206n3,,(,230027):.,n(n2personstochasticiteratedprisonerdilemma,NSIPD).NSIPD,,.,.,TIT2FOR2TAT(TFT),.:;;:TP18;N94:AResearchonstrategyselectioninn2personstochasticiteratedgameZHANGSi2hai,XUMin,WANGXu2fa(Dept.ofComputerSci.&Tech.,USTC,Hefei230027,China)Abstract:Thestrategyselectionin22personiteratedprisonerdilemmawasextendedandtheconceptofn2personstochasticiteratedprisonerdilemma(NSIPD)wasproposedtoaccommodatetonewdevelopmentarisingfromtechnologyadvancement.InNSIPDeveryonepshistoryinformationofchoicecouldbeseenbyallplayers.Ateachgameround,twoplayerswereselectedrandomlytoplayPDgame.AcomputertournamentwasusedtosearchforthebeststrategyinNSIPD.TheexperimentalresultsshowthatTIT2FOR2TATandALWAYS_DEFECTarenotalwaysthebeststrategy,andthatthebeststrategyiscloselycorrelatedwithrepeatedtimesandstrategydistribution.Keywords:cooperation;prisonerdilemma;iteratedgames0,[1][2][3],.,,,.,TIT2FOR2TAT(TFT)[4].,.,,,.,[57].3:2005207218;:2006204207:(60428202).:,,1974..:,.E2mail:[e-mail]:,.E2mail:[e-mail],,,.,.[8]n(NIPD)[9,10].,Internet,,,.,e2bay,eachnet.,.,,.n,n,,.,,[1].,(iteratedprisonerdilemma,IPD),NSIPD,TIT2FOR2TAT,,.1111n,.Sifi.,.Axelrod[3],.,.:C(),D(),U().27(U).,CUD,15.CD.fi(1im),.1[11].1121:RAN,,;.=015.1Fig.1Payoffmatrixofprisonersdilemmagame1Tab.1Strategylist1TITFORTATTFT[2]2ALLD,3ALLC,4RAN,5N_F[2]61F_172F_2Hij,1in,1j3,i,ji,jF_1:count=0;for(k=1;k4;k++)if(Hik=D)count++;ifcount1returnC;returnD;,jF2:count=0;for(k=1;k4;k++)if(Hik=D)count++;ifcount2returnC;returnD1131121,.S={si1i7}.W={wi1i7},.,,,.271136.2.2Tab.2ListofweightdistributionTFTW10.150.350.05ALLDW20.150.050.35ALLCW30.150.10.2RANW40.150.10.1N_FW50.150.10.1F_1W60.150.10.1F_2W70.10.20.12,5000,RAN=0.5..211214,TFT,ALLDALLC.211214,1100,,,.,1000,1000,.,.215.211TIT2FOR2TAT(TFT)TFT2.2TFT(,)Fig.2WinandlosttimesofTFT2,TFT,,50%,.,TFT.,TFT,.,TFT,TFT,.,TFT,TFT,.212ALLDALLD3.3ALLDFig.3WinandlosttimesofALLD,,,ALLD,100,100%.,TFTN_F,,,,.ALLD,.,,ALLD,TFT,N_F,F_1,F_2;,TFT,N_F,F_1,F_2TFT,N_F,F_1,F_2,.,ALLD,,,,30%.,ALLD,,TFT,F_1,N_F,F_2,RAN,ALLD,ALLD.371111n,.,,ALLD.213ALLCALLC4.4ALLCFig.4WinandlosttimesofALLC,,ALLC.,ALLC.3080%;20%;40%.0,,,70%50%.,,.5RANFig.5WinandlosttimesofRAN2145RAN,6N_F.:.,,,,.N_F,ALLC.6N_FFig.6WinandlosttimesofN_F21579.,().,,,0,,.,ALLC,F_1,TFT,F_2,N_F,RAN,ALLD;(8),ALLC,F_1,TFT,F_2,RAN,N_F,ALLD;(9)N_F,F_1,F_2,TFT,ALLD,ALLC,RAN.:,RAN,.,F_1/F_2/TFT/N_FALLD(),F_1/F_2/TFT/N_FALLC().RAN,,,.Axelord,TFT4711367Fig.7Averagepayoffsunderevendistribution8Fig.8Averagepayoffsundercooperativedistribution9Fig.9Averagepayoffsunderdefectdistribution,ALLC,,N_F.TFT.,.216,.,TFT,,TFT.ALLD,,,ALLD,.,ALLD1.TFTALLD,TFTALLD.,,35%.ALLC50%.,ALLD,N_F,ALLD,ALLCRAN.21779,..,ALLD,1..ALLC,,70%,.,,,(TFT,N_F),.,,.,,ALLD,ALLC.ALLD,.,,.571111n3n,.,TFT..Axelrod.,.(TFT,N_F),.,..,,,..,.(References)[1].[M],:,1995.[2]LIANGLiang,WANGZhi2qiang,TANGWei2jun,etal.Economicmodelofsupplychainnetworkdesignbasedonassignmentgame[J].JournalofUniversityofScienceandTechnologyofChina,2005,5(6):9192926.,,,.[J].,2005,35(6):9192926.[3]HauertC,SzaboG..Gametheoryandphysics[J].J.Phys.,2005,73(5):4052414.[4]AvnetJ,FriedmanA,etal.Repeatedone2shotprisonerspdilemmawithreputations[C]SantaFeInstituteCSSS04.Qingdao,China,2004.[5]AxelrodR.TheEvolutionofCooperation[M].NewYork:BasicBooks,1984.[6]AxelrodR.TheComplexityofCooperation[M].NewJersey:PrincetonUniversityPress,1997.[7]AxelrodR.Theevolutionofstrategiesintheiteratedprisonerdilemma[C]DavisL.GeneticAlgorithmandSimulatedAnnealing.SanMateo,CA:MorganKauffmann,1987:32241.[8]MacyM.Naturalselectionandsociallearninginprisonerpsdilemma[J].SociologicalMethods&Research,1996,25(1):1032137.[9]YaoX.Evolutionarystabilityinthen2personiteratedprisonerpsdilemma[J].BioSystems,1996,37(3):1892197.[10]YaoX,DarwinP.Anexperimentalstudyofn2personprisonerpsdilemmagames[J].Informatica,1994,18:4352450.[11][],[].[M].,.:,2002.671136