摘要
针对粒子群优化算法(PSO)缺少跳出局部最优的机制而易出现早熟问题,提出一种新的混沌粒子群优化算法(NCPSO)。该算法引入混沌扰动更新粒子的位置,避免搜索陷入局部最优,再嵌入判断早熟停滞的方法,一旦检测到早熟现象,使用逃逸策略来增大粒子群的多样性。最后用3个常用的测试函数进行仿真,实验结果表明:NCPSO算法比PSO算法、CPSO算法有更高的寻优精度和更快的收敛速度。
The paper proposes a new Chaotic Particle Swarm Optimization algorithm in allusion to the defect that the PS0 algorithm lacked the mechanism of leaving aside the local optimization to appear premature . The algorithm introduces chaotic perturbation into renewing particle location to avoid search in the local, and a method that identified premature stagnation is embedded, so once premature stagnation happened, escape strategy for guaranteeing the particles diversity could be used. Finally, three familiar test functions are simulated to show that NCPSO achieves better and faster convergence than PS and CPSO.
出处
《智能计算机与应用》
2017年第2期63-65,69,共4页
Intelligent Computer and Applications
关键词
粒子群算法
混沌扰动
逃逸策略
早熟
Particle Swarm Algorithm
chaotic perturbation
escape strategy
premature
