一种基于混沌变异的多目标粒子群优化算法

A mult-objective particle swarm optimization algorithm based on the chaotic mutation

针对多目标优化(multi-objective optimization problem,MOP)问题,特别是解集分布非均匀问题,提出一种基于混沌变异的优化算法。通过Pareto支配思想来决定粒子的飞行方向,在进化后期加入混沌变异操作,有效地避免早熟收敛现象;根据粒子群优化算法(particle swarm optimization,PSO)特有的记忆建立外部档案,动态引导微粒在每一次迭代的飞行方向。最后通过8个标准多目标测试函数进行测试,实验结果表明该算法是有效可行的,其性能比SPEA和NSGAII更优。

To solve the problem of the multi-objective optimization problems,especially the problem of solution set distribution non-uniform,an optimization algorithm was proposed based on the chaos mutation.In order to prevent effectively premature convergence phenomenon,the Pareto dominate ideology was used to determine the direction of particles and the chaos variable operating was added in the later evolution.According to the characteristic of particle swarm algorithm,the external files were established to dynamically lead the flight direction of the particles.The experimental results of eight standards test functions showed that the algorithm was effective and feasible,and its performance was better than SPEA and NSGAII.