群决策专家权重自适应算法研究王俊英

2010-07-302010-09-2469775022863-306-ZT04-06-31971-、[e-mail]1946-、．*1211．4300742．443000TOPSIS。TP391A1001-3695201102-0532-04doi10．3969/j．issn．1001-3695．2011．02．035Researchonself-adjustingalgorithmofgroupdecision-makingexperts’weightWANGJun-ying12LIDe-hua11．InstituteforPatternRecognition＆ArtificialIntelligenceHuazhongUniversityofScience＆TechnologyWuhan430074China2．CollegeofComputer＆InformationTechnologyChinaThreeGorgesUniversityYichangHubei443000ChinaAbstractInordertoobtainscientificrationalgroupdecision—makingresultsandmaintainadynamicexperts’weightinfor-mationbaseinHWMEhallforworkshopofmetasynthesisengineeringthispaperproposedanexperts’weightself-adaptiveadjustmentalgorithmbasedonTOPSISmethodandgreycorrelationdegree．Thisalgorithmtookintoaccountthesimilaritybe-tweentheindividualsandgroupsofexpertsinthejudgeontheimportantattributesandorderofthealternatives．Meanwhileaccordingtodifferentdecision-makingproblemsthisalgorithmcouldselectcorrespondingparameter．Numericalexampleshowsthatthealgorithmiseffectiveanditsconvergencerateisrelativelyfast．Keywordsmultipleattributegroupdecision-makingexpertweightTOPSISdecisionrelationaldegreeadaptivelyadjust-ment、。1992“”。、。———。1abcd1。AHP2、、、、、、、。。3AHP456、group’sidealsolution47～9。28220112ApplicationResearchofComputersVol.28No.2Feb.20112。。1011。。12、、。TOPSIS13。2.1E=e1…esekωk。0≤ωk≤1∑sk=1ωk=1k=1…sωk0ek。X=x1x2…xnF=f1…fmfjλj。0≤λj≤1∑mj=1λj=1j=1…m。ekAK=aki()jn×m。akijekij。2.2TOPSIS。Rae1969。。。TOPSIS14。1AK=aki()jn×mekakijekijB=bijn×mbijbij=∑sk=1aijωk1ωkek。2AK=aki()jn×mekB=bijn×mC=ckis×n。ckiekxicki=∑mj=1bij－akij2λj槡22λjfjCek。3C=ckis×nc*=c1*…cm*c0=c10…cm0xi。ckikxici*ci0ci*=min0≤k≤scki3ci0=max0≤k≤scki44C=ckis×nc*=c1*…cm*c0=c10…cm0dk*dk0tkek、ekdk*=∑ni=1cki－ci*槡25dk0=∑ni=1cki－ci0槡26tk=dk0dk0+dk*7k=1…si=1…n。。2ekλjek。ek0≤tk≤1tk=1ektk=0ek。TOPSIS。2.3。15。。。·335·214。7。。。5AK=akijn×mekλjfjωkekakijekxifjSk=sk1sk2…ski…sknek。sk1=∑mj=1akijλj8i=1…nj=1…mk=1…s6AK=akijn×mekλjfjωkekakijekxifjG=g1g2…gi…gn。g1=∑sk=1∑mj=1akijλjωk9i=1…nj=1…mk=1…s@GRF。s'ki=skisk1i=1…n10g'i=gig1i=1…n117S'k=s'k1s'k2…s'ki…s'knekG'=g'1g'2…g'i…g'nγg'is'ki=minkmini|g'i－s'ki|+ζmaxkmaxi|g'i－s'ki||g'i－s'ki|+ζmaxkmaxi|g'i－s'ki|i=1…nk=1…s12ekxi。γk=G'S'k=1n∑ni=1γg'is'kiζ∈01i=1…nk=1…s13ek。ζζ=0.5。。。2.42.4.1ωk、、、、———。TOPSIS。TOPSIS。γk。ω'tk=tk/∑sk=1tk14ω'gk=γk/∑sk=1γk15k=1…stkekγkek。ωk=μωk0+1－μμ1ω'tk+1－μ1ω'gk16k=1…sωk0ω'k。μ0≤μ≤1μ。μ=1μ=0μ10≤μ1≤1μ1μ1μ10。2.4.2G=g1g2…gi…gnG*=g*1g*2…g*i…g*nd。d=∑ni=1gi－g*i槡217dσ。38·435·281/8A、B、C、D、E0.34、0.17、0.16、0.17、0.16A1=28877477775889976396687777879979999A2=89767．5897．56．57．597．58979．578．59789878887．586．58．58．5988A3=8789679991010899710798678899710991088879A4=5105510510551085855105101010885877858879677A5=483365646593777647810455671335535455A6=51021088105810753521025618629101531893892A7=633164111310889686884335666666677667A8=898781111154535565537566655533666553.1a1B。b2C。c34TOP-SISc*c0。d5～7TOPSIS14ω'tk。e8Sk=sk1sk2…ski…skn。nk=1…sek10SkS'k。f9G=g1g2…gi…gn。n。11G'。g12。h13。i15ω'gkk=12…s。j16。k17dσd＞σad≤σ。3.2μ=μ1=0.5ζ=0.51、2。1TOPSIS3TOPSIS5510．1250．10860．13760．121820．1250．14170．13530．139630．1250．14030．12770．133540．1250．13850．14790．143950．1250．1230．11090．116660．1250．13480．0860．112570．1250．10290．13240．116380．1250．11010．12220．11572TOPSISTOPSIS16．3856．47266．38826．4357525．69565．95025．6485．8124736．97816．99867．08977．0385247．08637．18957．20477．1987156．6256．74196．64496．6995465．97256．02966．20656．1129676．93066．98467．02927．0081312ω4＞ω2＞ω3＞ω1＞ω5＞ω7＞ω8＞ω6ω2＞ω3＞ω4＞ω6＞ω5＞ω8＞ω1＞ω7TOPSISω4＞ω1＞ω2＞ω7＞ω3＞ω8＞ω5＞ω65、6、862115．25、。4＞3＞7＞5＞1＞6＞2。。4。TOPSIS。。。540·535·25α＊＊DP/NF//sDP/NF//sDP/NF//stestn11/10/00．601/10/00．721/10/016．01testn21/10/00．671/10/00．761/10/016．67testn31/10/00．681/10/00．791/10/017．31testn42/20/00．752/20/00．832/10/018．54testn51/20/00．771/20/00．921/00/017．08testn61/10/00．891/10/01．031/10/021．72testn70/01/00．500/01/10．670/01/119．33testn80/31/00．850/01/11．320/01/123．684α＊＊、。α＊＊。α。。1AALSTWVDHEEKV．Workflowmanagementmodelsmeth-odsandsystemsM．CambridgeEnglandMITPress20021-280-85．2AALSTvanderWMPREIJERSHAWERJTERSAJMMetal．BusinessprocessmininganindustrialapplicationJ．Informa-tionSystem2007325713-732．3COOKJEWOLFAL．Discoveringmodelsofsoftwareprocessesfromevent-baseddataJ．ACMTransonSoftwareEngineeringandMethodology199873215-249．4AALSTvanderWMPWEIJTERSAJMMMARUSTERL．WorkflowminingdiscoveringprocessmodelsfromeventlogsJ．IEEETransonKnowledgeandDataEngineering20041691128-1142．5HAMMORIMHERBSTJKLEINERN．InteractiveworkflowminingC//Procofthe2ndInternationalConferenceonBusinessProcessManagement．2004211-226．6HAMMORIMHERBSTJKLEINERN．InteractiveworkflowminingrequirementsconceptsandimplementationsJ．DataandKnowledgeEngineering200656141-63．7AALSTvanderWMPWEIJTERSAJMMMARUSTERL．WorkflowminingdiscoveringprocessmodelsfromeventlogsJ．IEEETransonKnowledgeandDataEngineering20041691128-1142．8WENLi-jieAALSTvanderWMPWANGJian-min．Miningprocessmodelswithnon-free-choiceconstructsJ．DataMiningandKnowledgeDiscovery20071510145-180．9．αJ．20073081436-1445．