摘要
为了解决动态改变惯性权重的自适应粒子群算法不易跳出局部最优的问题,提出了一种自适应变异的动态粒子群优化算法。在算法中引入了自适应学习因子和自适应变异策略,从而使算法具有动态自适应性,能够较容易地跳出局部最优。对几种典型函数的测试结果表明,该算法的收敛速度明显优于文献算法,收敛精度也有所提高。
To solve the problem that adaptive particle swarm algorithm with dynamically changing inertia weigh algorithm is apt to trap in local optimum,a dynamic particle swarm optimization algorithm with adaptive mutation is proposed.The adaptive learning factor and adaptive mutation strategy are introduced in this new algorithm,so that proposed algorithm can easily jump out of local optimum with effective dynamic adaptability.The test experiments with three well-known benchmark functions show that the convergence speed of proposed algorithm is significantly superior to existing algorithms,and the convergence accuracy of algorithm is also increased.
出处
《计算机工程与应用》
CSCD
北大核心
2011年第16期32-34,共3页
Computer Engineering and Applications
基金
国家自然科学基金No.70701016~~
关键词
粒子群优化算法
惯性权重
自适应变异
学习因子
particle swarm optimization algorithm
inertia weight
adaptive mutation
learning factor
