摘要
提出一种基于博弈论的多目标量子粒子群算法。算法中将每个目标函数看成是一个智能体,智能体控制种群往自己最有利的方向进行搜索,然后将它看成是参与博弈的一个参与人。采用存在一个博弈序列的重复博弈模型,在重复博弈中,并不是每次博弈都产生最大效益,而是要总的效益最大化。将算法用于求解多目标0/1背包问题。仿真实验结果表明,该算法能够找到接近Pareto最优前端的更好的解,同时维持解分布的均匀性。
This paper presents a multi-objective quantum Particle Swarm Optimization(PSO) based on game theory.The algorithm for each objective function will be seen as an agent,agent to control the populations of the direction of their most advantageous to search,and then participate in it as a game participant.With the existence of a sequence games of repeated game model,in repeated game,not every game has produced the maximum benefit,but to the overall maximum benefit.And the algorithm is solving multi-objective 0/1 knapsack problem.The simulation results show that this algorithm can be found near the Pareto optimal front of a better solution,while maintaining the uniformity of the distribution solution.
出处
《计算机工程与应用》
CSCD
北大核心
2011年第26期43-45,65,共4页
Computer Engineering and Applications
基金
河南省科技厅科技攻关项目(No.092102110274)
关键词
量子粒子群
多目标优化
背包问题
博弈论
Quantum Particle Swarm Optimization(QPSO)
multi-objective optimization
knapsack problem
game theory