Abstract: Chaotic mutation and local searching combine with the particle swarm optimization algorithm for multi-objective optimization. In the process of optimization, different phases of evolution adopt different optimization strategy using crowding distance for the standard. When the population has trapped into local optimum, chaotic mutation helps to break away from the local optimum. In the later phases of evolution, local searching helps to enhance the diversity and convergence of this algorithm. The experimental results show that the distribution of Pareto front is more uniform, closer to the theoretical Pareto front.

Key words: chaotic mutation, local searching, convergence, diversity, Pareto front, particle swarm optimization

摘要： 将混沌变异和局部搜索与粒子群算法相结合用于多目标寻优。寻优的过程中以拥挤距离为标准，在进化的不同阶段采用相应的优化策略。当种群陷入局部最优时用混沌变异跳出该局部最优；用局部搜索法在进化后期增强算法的多样性和收敛性。实验结果表明，该方法求得的Pareto前沿分布更加均匀，更加接近理论的Pareto前沿。

关键词: 混沌变异, 局部搜索, 收敛性, 多样性, Pareto前沿, 粒子群