一种改进的粒子群优化算法

针对基本粒子群算法在处理复杂问题时有可能陷入局部极小的现象,引入群体适应度方差及群体位置方差,协调算法的种群多样性,使之能有效地克服基本粒子群算法容易陷入局部收敛的问题。在算法的中后期,根据粒子的表现不同,自适应调整惯性权重,保持群体惯性权重的多样性。通过选取4个基准函数进行测试,验证了改进算法可提高粒子群算法的优化性能。

A Kind of Algorithm for the Improved Particle Swarm Optimization

When the PSO algorithm optimization is used in complex problems, it is likely to be trapped at local minima phenomenon, the exploration and exploitation ability of the algo- rithm were regulated through introducing two criteria in the evolutionary process, the popu- lation-fitness-variance and the population-position-variance to preserve population diversity , which can effectively overcome the problem of premature convergence encountered by PSO. In the middle-end of the algorithm, based on the different expression of the particle, the inertia weight adapted by itself , so it can keep the inertia weight diversity. Finally, in this paper, to test four basic math function can improve the optimization capability of it.