动态环境下分布式自适应粒子群优化算法  

Distributed Adaptive Particle Swarm Optimizer in Dynamic Environments

作　　者：唐剑[1,2] 史浩山[1] 杨奇[1] 邢云冰[1] 

机构地区：[1]西北工业大学电子信息学院,西安710072 [2]空军工程大学电讯工程学院,西安710077

出　　处：《系统仿真学报》2009年第17期5431-5435,共5页Journal of System Simulation

基　　金：国防基础研究项目(K1804060127);教育部博士点基金资助项目(20050699037)

摘　　要：针对现有粒子群算法的不足,提出一种基于微粒自身信息的环境变化检测方法,同时采用分布式处理模式,通过激活粒子群中的停滞粒子适应环境变化,不仅降低了的算法复杂度,而且提高了算法对复杂环境的自适应能力。对于有界连续函数,证明新算法能依概率收敛于全局极小点。应用抛物线函数和Rastrigin函数构造的复杂动态环境对该算法进行验证,并同APSO、D-PSO算法进行了对比。实验结果表明,在复杂的动态环境中,DAPSO算法具有更好的适应性。To the shortage of existing PSO, a new change detection method was proposed based on the particle self-information. By means of effectively reactivating all the particles in stagnant state with a distributed processing model, it could not only reduce the algorithm complexity, but also strengthen and enhance the adaptability of the algorithms to complex dynamic environment. It is proved that the algorithm converges in probability to a neighborhood of the global minima. To validate the performance of the DAPSO compared with the APSO and D-PSO methods, a number of experiments were performed under the complex dynamic environment which was generated by the Sphere functions and the Rastrigin function. The simulation results indicate that the proposed algorithm is more effective in complex dynamic environments.

关 键 词：粒子群优化算法 动态环境 自适应粒子群优化算法 多样性粒子群优化算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]