无限重复流速与拥塞控制博弈实例的应用.pdfVIP

第 35 卷第 6 期 Vol. 35 No.6 在 Ih 大学学提〈自然科学版) 2∞5 年 11 月 10日四AL OF SOUτHEAST UNIVERSITY (Na能ral Science Edi岳阳) Nov. 2∞5 无限重复流速与掘塞控制博弈实侧的应用 南军吴家呆刘业 〈东南大学汁算机网络和信息集成教育部重点实验室，南京 210仿岳〉 (东南大学汁算机科学与工程系，南京 210096) 擂妻:在讨论了一次博弈模型和流速均衡解析解将基础上，对基于无破重复博弈模型的流速与 拥塞控制行为进行了研究.提出了重复和无限重复流速与拥塞控制溥弈模型，论证了重复溥弈 NEP 的存在性和最优性，讨论了重复博弈中贴现国子的意义;透过2 个无限重复流速控制博弈 的实例(2 个端系统为N 个端系统) ，介绍了在非合作的流速与掘塞控制中规莲(约束)端系统的 行为方法，以维护和实茸整个流速与揭塞控制系统效恶的最佳. 关键词:无限重复博弈; Nash 均衡;效用;流速与拥塞控制 中噩分类号: TP393 文献标iR码:A 文章编号: 10m -0505(2005)06-0853-05 Application of examples of infinite repeated t10w and conges柱。n control game Tao Jun Wu Jiagao Liu Ye (Key Laboratory of Compnter Net自Orks and Information Intεgration of Miní甜y of Education , Southeast University , Nanjing 21009岳， Chìna) ([比阿反ment of Computer Science 血d Engineering , South回到 University ， Nanjing 21部)96 ， China) Abstract: After the one-shot game model and flow equilibrium solution are discussed ，位le behaviors of flow and congestion control based on infmite repeated g忽ne model further is studied. Subsequently , repeated and infmite repeated flow and congestion control game model is presented. τhe existence and optimization of the Nash equilibrium point is proved. The discount factor in repeated g缸ne is discussed. Finally , through two ex挝nples of repea就d flow control g缸ne with two users and N users , this 红ticle explains how to regularize or res时ct 也e users behaviors in non- cooperative flow 扭d congestion control r毛spec也vely to optimize the utility of whole system. Key words: infinite repeated game; Nash eq也librium; utility; flow and congestion control 网络视颜会议、远程教学等不响应搪塞