CAJ | 학술논문

针对普通粒子群优化算法难以在动态环境下有效逼近最优位置的问题,提出一种动态粒子群优化算法.设置敏感粒子和响应阈值,当敏感粒子的适应度值变化超过响应阈值时,按一定比例重新初始化种群和粒子速度.设计双峰DFI动态模型,用于验证该算法的性能,仿真实验结果表明其动态极值跟踪能力较强.
침대보통입자군우화산법난이재동태배경하유효핍근최우위치적문제,제출일충동태입자군우화산법.설치민감입자화향응역치,당민감입자적괄응도치변화초과향응역치시,안일정비례중신초시화충군화입자속도.설계쌍봉DFI동태모형,용우험증해산법적성능,방진실험결과표명기동태겁치근종능력교강.
Aiming at the problem that normal Particle Swarm Optimization(PSO) algorithm can not approach the best position ffectively in dynamic environment,this paper proposes a dynamic PSO algorithm.It sets sensing particle and response threshold.When sensing particle's fitness change exceeds response threshold,the algorithm reinitializes the swarm and particle velocity.It designs double-hump DF1 dynamic model to validate the capability of this algorithm.Simulation experimental results show that it has high ability of dynamic extremum tracing.