您好,欢迎访问三七文档
当前位置:首页 > 电子/通信 > 综合/其它 > 势博弈理论及在移动通信中的应用
2012,,,、,,,。;;;;1,2,3,1,2,1,2,3(1.网络体系构建与融合北京市重点实验室北京100876;2.北京邮电大学信息与通信工程学院北京100876;3.中国电力科学研究院北京100192)11.1MondererShapley1996。,,,[1](generalizedordinal)、(ordinal)、(exact)(weighted)。Voorneveld,[1,2],。,。?,,:(FIP)。,。FIP,。“”,“”“”,“”,“”,“”,,,,。NeelJ“AllroadsleadtoRome()”[4],,,。,,,。,,、、。1.2,,,,,,5512012,τ=N,Y,u。,N,Yiuii(payofffunction),,[5]。1.2.1普通势博弈τ,。P∶Y→R,i∈Ny-i∈Y-i(i):ui(y-i,x)-ui(y-i,z)0,ifP(y-i,x)-P(y-i,z)0x,z∈Yi(1),i(),()。τ,:P,:eui(y)eyi0,ifeP(y)eyi0,i∈N,y∈Y(2)1.2.2加权势博弈w=(wi,…,wN),i∈Ny-i∈Y-i,τ:ui(y-i,x)-(y-i,z)=wi(P(y-i,x)-P(y-i,z)),x,z∈Yi(3)P:Y→R,τ。,τ,:eui(y)eyi=wieP(y)eyi∈∈,i∈N,y∈Y(4)1.2.3完全势博弈,wi=1,i∈N,P:Y→R,:ui(y-i,x)-ui(Y-i,z)=P(y-i,x)-P(y-i,z),Ax,z∈Yi(5)τ,。,,。,。1.2.4伪势博弈P:Y→R,i∈Ny-i∈Y-i:argmaxyi∈Yiui(yi,y-i)∩argmaxyi∈YiP(yi,Y-i)(6),τ*=〈N,Y,P〉τ=〈N,Y,ui〉。ττ*,τ,P,。1。2,,。,,。,,:—、、、、,,。2.1—(7):ui(y)=C(y)+Di(y-i)(7)C:Y→R,Di:Y-i→R,-[4]。P=C(y)。C(y),y,Di(y-i),i,。(8):ui(yi,y-i)-u(xi,y-i)=C(yi,y-i)-C(xi,y-i)i,j∈N,y∈Y(8)15522012C(y)[1]。,—。(9):ui(y)=C(y)(9)C:Y→R,,—di=0i∈N。—C(y)=0,y∈Y,。2.2(10),。ui(y)=Si(yi)(10)Si:Yi→R,,,[4]。[1]:P=ΣSi(yi)(11)2.3(12):ui(y)=j∈NΣΣiΣwijj(yi,yj)-Si(yi)(12),wij:Yi×Yj→R,,Si:Yi→R,,(yi,yj)∈Yi×Yjwij(yi,yj)=wij(yj,yi),。(BSI)(13)[4]:P=Σi-1Σwij(yi,yj)-ΣSi(yi)(13)33.1,。1Pτ=〈N,Y,{ui}ieN〉,ττ*=〈N,Y,{P}ieN〉,。,,。,。2。2,τ=〈N,Y,{ui}ieN〉,1,P(y)y*∈Yτ,。。τ=〈N,Y,{ui}ieN〉,P,P,P,,2。,,。y*=maxy∈YP(y),,,y*Y'∈Y,iui(y')ui(y*),,P(y')P(y*),P,,。3.2,,,—(Gauss-Seidel)(Jacobi)。γ=(y0,y1,...)Y,yk=(y-ik-1,x)(x∈Yi),-y0。,k,,ik。k≥1,ui(yk)ui(yk-1),γτ2i∈Ni∈Ni∈Nj∈15532012。τ,τ。,。FIP。γ=(y0,y1,...),,:P(y0)P(y1)P(y2)...(14)Y,γ,。,,,,。,,,。3.3,,。。。,,,,,[6]。,,。:(bestresponse)、(betterresponse)(gradientprojectionresponse)。(15):Di(y-i)=yi*∈Yi(y-i):y*i=argmaxyiui(yi,y-i∈∈)(15)(16):Di(y)=yi*∈Yi(y-i):ui(yi*,y-i)ui(yi,y-i∈∈)(16),ui(yi,y-i)Yi,,,〈·,·〉,·。,△iui(y):=eui(yi-y-i)eyi,,Yi,:Di(x)=yi+ρi△iui(yi,y-i,,)yi(y-i)(17),,ρi,,,cYi(y-i)Yi(y-i)C:,,cYi(y-i)=argminyi∈Yic-yi22(18)4,。,,,。,,。,,。,,[7]。,,,,。,,,。,,,[8]:P,=Objective(19),,“”[8],,、,“”,。“”1,。Ye,Y*,(20):POA=Fobj(Ye)Fobj(Y*)(20)55420125、。。,3。5.1[5,9],,。,。m。piti,gijji。,ijpijr=gijpj。ηii。i(SINR):αi=giipitΣi≠jgij+pjtηi=priiΣi≠jpijr+ηi(21),,,。,i:ui(αi,pit)。Iii:Ii=j≠iΣgijpjt=j≠iΣpijr(22)·:·;·();·:ui(αi,pit)=Ri(αi)-cpit,C。RiSlog(1+αi)。:P=log1+i∈NΣhi2pi∈∈-i∈N∑cpi(23)5.2[4,10],。:·;·Yi,Yi=[0,ymax];·:ui(Y)=-γ^-giyi1k∑j≠igiyi+∑∑η(24)yii,gii,η,K,γ^。[9],τ=〈N,P,ui∑∑i∈N〉,:p=2γ^K(∑i∑k>igigkpipk)+i∑(-gi2pi2+2γ^gipi/K)(25),,。。1VoorneveldM.Best-responsepotentialgames.EconomicsLetters,2000(66):289~2952DubeyP,HaimankoO,ZapechelnyukA.Strategiccomplementsandsubstitutesandpotentialgames.GamesandEconomicsBehavior,2006(54):77~943MondererD,ShapleyL.Potentialgames.GamesandEconomicsBehavior,1996(14)4NeelJO.Analysisanddesignofcognitiveradionetworksanddistributedradioresourcemanagementalgorithms.VirginiaPolytechnicandStateUniversityDissertation,20065ScutariG,BarbarossaS,PalomarDP.Potentialgames:aframeworkforvectorpowercontrolproblemswithcoupledconstraints.ProceedingsofICASSP2006,Toulouse,France,20066JohnsonDS,PapadimitriouCH,YannakakisM.Howeasyislocalsearch?JComputerSystemSci,1988(37):79~1007PapadimitriouFC,TalwarK.Thecomplexityofpurenashequilibria.ProcACMSympofTheoryofComputing,2004:604~6128VoorneveldM.Equilibriaandapproximateequilibriaininfnitepotentialgames.EconomicsLetters,1997(56):163~1699HeikkinenT.Apotentialgameapproachtodistributedpowercontrolandscheduling.ComputerNetworks,2006(50):2295~231110NeelJ,ReedJ,GillesR.Convergenceofcognitiveradionetworks.ProcofWirelessCommunicationsandNetworkingConference,Phoenix,AZ,USA,Nov200411,,..,2011,27(8):63~6712,,.LTE-A.,2011,27(12):39~43(收稿日期:2012-11-15)555
本文标题:势博弈理论及在移动通信中的应用
链接地址:https://www.777doc.com/doc-5206780 .html