摘要
在求解高维空间中复杂多峰函数的优化问题时,传统的粒子群算法在收敛速度和局部搜索能力等方面表现出严重不足。针对这些问题,提出了一种基于最优评价的改进自适应粒子群算法(IAPSO),引入了改进的速度迭代公式,利用对每次迭代后种群的一系列最优值的评价来控制惯性权重的增幅,并设置对速度和位置的变异机制来防止搜索陷入局部最优。相关实验表明,在对高维空间中的复杂多峰函数进行优化求解时,改进粒子群算法IAPSO的表现比常规粒子群算法更加优越。
For complex multi peaks function with high dimensions, the classical PSOA has some serious disadvantages such as slow convergence and weak ability of local search. With respect to the deficiency this paper presents an improved adaptive particle swarm optimization algorithm based on best fitness evaluation (lAPSO) and introduces an improved velocity iterative formula. The IAPSO changes inertial weight according to the best fitness of each generation and sets up a mutation mechanism of velocity and particle position to prevent a search trapping into local optimum. Experimental results demonstrate that the performance of IAPSO is superior to the canonical PSOA.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第12期2497-2501,共5页
Systems Engineering and Electronics
关键词
粒子群算法
最优值评价
变异机制
自适应
particle swarm optimization
best fitness evaluation
mutation mechanism
adaptive