演化博弈框架下带奖惩的P2P激励机制研究  

Study of Incentive Mechanism with Punish-reward Under Evolutionary Game Framework

作　　者：陆坤[1] 王世宇[1] 谢玲[1] 李明楚[1] 

机构地区：[1]大连理工大学软件学院,大连116620

出　　处：《小型微型计算机系统》2015年第3期488-492,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61100194;61272173)资助

摘　　要：P2P系统中引入激励机制,主要为了促进节点之间相互服务,从而提升系统性能.现有的基于演化博弈的P2P激励机制,采用演化博弈的方法对P2P网络中的个体及交易建模,可以更真实反映个体行为变化的规律.然而完全按照交易历史的服务方式对经常服务的个体不够公平,对经常不提供服务的个体也过于宽容.本文在当前的激励机制中引入奖惩,通过仿真实验证明,这种带有奖惩的机制可以更好促进节点合作,也可以更快的使得系统达到最优平均收益.P2P incentive mechanism is aiming at encouraging peers to serve each other, so that the performance of system can be im- proved. Evolutionary game based incentive mechanism using evolutionary game to model peers and transactions in P2P network. It can have a better description of peers＇ behaviors changing. However,precisely provide service according to transaction history is unfair to those who always provide service and far too tolerant to those who seldom provide service. In this paper, we bring in punish and re- ward and prove it through experiments that this mechanism can better help facilitate cooperation and faster achieve best average payoff of the system.

关 键 词：P2P 激励机制 演化博弈 奖惩机制 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]