全局混沌杂交的粒子群优化算法及应用

An Entirely Chaotic and Cross Particle Swarm Optimization Algorithm and Application

提出一种新的混沌粒子群优化算法(EC-CPSO),该算法在基本混沌粒子群优化算法(CPSO)基础之上,将粒子速度计算公式中的随机数用混沌随机序列来替代,同时应用早熟判断机制,在对最优粒子进行混沌化处理之外,对其余粒子进行杂交处理,提高了算法的寻优能力,有效避免算法陷入局部最优并防止过早收敛.将之用于(N+M)容错系统优化模型证明该算法与CPSO相比具有一定的优势.

An entirely chaotic and cross particle swarm optimization（EC-CPSO） algorithm is proposed based on the chaos particle swarm optimization（CPSO） algorithm.In order to improve the searching efficiency and deal with the problems of trapped in local convergence and premature,the random numbers of the classical PSO algorithm are substituted by chaotic sequences.By means of the premature judging method,when the algorithm gets into the local convergence,EC-CPSO can start the chaos researching for the best particle and cross the other particles.The experimental results demonstrate that the new algorithm has better performance for solving the optimal model-（N＋M） fault-tolerant system than the CPSO algorithm.