马氏模型PSO及其随机过程分析  

Particle swarm optimization algorithm based on markov model and its stochastic process analysis

作　　者：袁代林[1,2] 陈虬[1] 

机构地区：[1]西南交通大学力学与工程学院,成都610031 [2]西南交通大学数学学院,成都610031

出　　处：《计算机工程与应用》2009年第31期49-52,共4页Computer Engineering and Applications

基　　金：西南交通大学校基金项目(No.2008B07)

摘　　要：受遗传算法马氏模型理论分析的启发,提出了一种便于用马氏过程理论分析的微粒群算法。该算法中的个体仅记忆群体在进化过程中有限步内的信息,忘掉以前的信息,以建立算法的马氏过程数学模型。通过函数优化的数值模拟验证了新算法具备优良的寻优能力,同时论证了新算法是齐次马氏过程。Inspired by the theoretic analysis of genetic algorithm based on markov process,a new form of particle swarm optimization algorithm is advanced,which is convenient for analysis by the theory of markov process.The particle of new algorithm only memorizes the information of swarm in finite steps,and forgets the old information.Then the markov process model is established.The simulations of functions optimization show that the new algorithm has good ability to find the global solution,and the homogeneous markov process is got from the new algorithm.

关 键 词：微粒群算法 马氏过程 函数优化 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]