摘要
针对粒子群优化算法易陷入局部最优并难以跳出的缺陷,提出了一种改进的算法。算法在采用自适应惯性权重基础上,引入粒子群局部收敛判断机制,对陷入局部最优的粒子采取新的进化模型以增加群体的多样性,进而降低了群体早熟收敛。通过新进化模型的粒子群优化算法对4个基准函数反复试验,仿真结果表明改进的粒子群算法具有更好的收敛精度并且可以有效避免早熟收敛问题,寻找到全局最优解。
For the defect of the particle swarm optimization algorithm easily falls into local optimum and difficulty jumps the global optimum, this paper propose an improved algorithm. On the base of using the adaptive inertia weight, the algorithm introduces particle swarm local convergence judgment mechanism, taking another evolutionary model for the particles which fall into local optimum. Thus it can increase the diversity of the population then effec- tively reduce the premature convergence phenomenon. The repeatedly test results on four representative benchmark functions demonstrate that the new algorithm has a better convergence accuracy and can effectively avoid the premature convergence problem in finding the global optimal solution.
出处
《成都信息工程学院学报》
2012年第6期580-584,共5页
Journal of Chengdu University of Information Technology
基金
四川省科技支撑计划资助项目(2012SZ0070)
关键词
计算机应用
智能工程
粒子群算法
惯性权重
局部最优
进化模型
computer application
intelligent engineering
particle swarm algorithm
inertia weight
local convergence
evolutionary models