蚁群算法参数组合的博弈优化  

Combination of ant colony algorithm parameters optimization based on game theory

作　　者：郜庆市[1] 孙树栋[1] 韩青[1] 钟尧[1] 

机构地区：[1]西北工业大学机电学院,西安710072

出　　处：《计算机工程与应用》2013年第21期51-55,共5页Computer Engineering and Applications

基　　金：西北工业大学创业种子基金(No.Z2013035)

摘　　要：针对蚁群优化算法参数组合选取的问题,提出了一种基于博弈论的蚁群算法参数优化模型。由于算法各个参数之间相互依赖、相互影响的关系,将各参数作为博弈论中的局中人,利用算法收敛时间与各个参数之间的数学关系,将其转化为博弈模型中参数的收益函数,求解出算法的最优参数组合。仿真结果表明,该模型能够方便有效求解出蚁群算法的最优参数组合。For the problem of ant colony algorithm combined parameters selection, a model of optimum combined parameters selection based on game theory is proposed. Due to the parameters influence and dependence on each other, they are as players. In this model, using the payoff function that transforms from the math relation between convergence time and parameters can obtain the best combination of parameters. Simulation results show that this model can get the combined parameters effectively.

关 键 词：蚁群算法 博弈论 参数组合优化 收益函数 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]