基于自适应小生境粒子群算法的多重Nash均衡求解  

SOLVING MULTIPLE NASH EQUILIBRIA BASED ON ADAPTIVE NICHED PARTICLE SWARM ALGORITHM

作　　者：贾文生[1,2] 向淑文[1,2] 杨剑锋[2] 

机构地区：[1]贵州大学理学院,贵州贵阳550025 [2]贵州大学计算机科学学院,贵州贵阳550025

出　　处：《计算机应用与软件》2015年第1期247-250,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(11161008);教育部博士点基金项目(20115201110002);贵州省自然科学基金项目(黔科合J字[2012]2139号);贵州大学青年基金项目(2010021)

摘　　要：针对n人非合作博弈多重Nash均衡求解问题,提出一种自适应小生境粒子群算法。该算法融合了序列小生境技术、粒子群优化算法的思想,并加入了变异算子和自动生成小生境半径机制,使得所有粒子尽可能分布到整个搜索空间的不同局部峰值区域,从而有效地求得博弈问题的多重Nash均衡。最后给出几个数值算例,计算结果表明所提出的算法具有较好的性能。This paper presents an adaptive niched particle swarm algorithm for solving the multiple Nash equilibria of n-persons＇ noncooperative game. The algorithm combines the ideas of the sequential niched technique and the particle swarm algorithm, and the mutation operator and the mechanism of automatic niche radius generation are added as well, so that all the particles are distributed to different local peak regions of entire search space as possible as can, thereby the multiple Nash equilibria of the game problem are effectively achieved. Finally, some numerical examples are given and the computation results demonstrate that the proposed algorithm has quite good performance.

关 键 词：小生境技术 粒子群算法 自适应 非合作博弈 NASH均衡 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]