无线传感器网络CSMA博弈优化算法研究  

The WSNs CSMA Game Optimization Algorithm Research

作　　者：杨俊刚[1] 史浩山[1] 杨武[1] 

机构地区：[1]西北工业大学电子信息学院,西安710129

出　　处：《传感技术学报》2009年第12期1774-1778,共5页Chinese Journal of Sensors and Actuators

基　　金：教育部博士点基金项目资助(2005069903)

摘　　要：提出了一种通用的数学模型GO/CSMA(Game Optimization CSMA),用于优化CSMA退避机制。GO/CSMA以节点发包概率τ为优化变量,建立n人非合作混合策略博弈模型。以吞吐量和成功发包概率作为博弈效用函数为例,独立进行两个回合博弈,博弈的结果经过多约束单目标规划优化后得到折衷的τ*,并据此设计出最小竞争窗口CWmin。仿真结果表明GO/CSMA显著提高了MAC层的传输效率。GO/CSMA可进一步扩展为对网络的其它多个性能(时延、丢包率、能耗、公平性)进行折衷的博弈优化。A general mathematical model, named by GO/CSMA （Game Optimization CSMA）, was proposed to optimize CSMA backoff mechanism. GO/CSMA selects τ（the probability for sending data packet） as optimization variable,and builds a model named by n-persons non-cooperative game with mixed strategy. In this paper, as special example, the throughput and successful probability of sending packet are selected as game utility function and theyplay a game among n nodes respectively and arrive at the Nash equilibrium point respectively. Further,it come into being a compromise τ＊ between the throughput and the probability through a multi-constrained single-objective planning optimization algorithm. At last,we design the minimum contention window CWmin on basis of τ＊. The simulation experiment indicates that GO/CSMA improves the transmission efficiency of MAC protocol greatly. By extension, GO/CSMA might be used to play games and gain compromise among networks performances, such as delay, packet drop, power exhaust and fairness.

关 键 词：无线传感器网络 CSMA 博弈 优化 

分 类 号：TP393.02[自动化与计算机技术—计算机应用技术]