基于环形拓扑结构和动态邻域的多模态多目标粒子群优化算法

Multimodal multi-objective particle swarm optimization algorithm based on ring topology and dynamic neighborhood

传统多目标优化算法在求解多模态多目标优化问题时未充分考虑解集在决策空间中的分布特性,导致种群早熟收敛且所得Pareto解集不完整.针对该问题,提出一种基于环形拓扑结构和动态邻域的多模态多目标粒子群优化算法.为更好地平衡探索与开发,将进化过程分为2个阶段:1)整个种群被划分为多个小型子种群和一个劣势子种群.采用基于空间距离的无重叠环形拓扑结构提升小型子种群的多样性,使得算法能搜索到更多Pareto最优解,并利用全局最优位置更新劣势子种群,提高搜索效率;2)所有粒子都跟随全局最优位置进行搜索,提高算法的搜索精度.同时,引入周期重组和一种新的全局最优位置更新策略,避免算法早熟收敛.仿真结果表明,所提算法可以有效解决多模态多目标优化问题.

The traditional multi-objective optimization algorithm does not fully consider the distribution of solutions in the decision space when solving multimodal multi-objective optimization problems,resulting in premature convergence and incomplete Pareto solution sets.To solve the above problem,a multimodal multi-objective particle swarm optimization algorithm based on ring topology and dynamic neighborhood is proposed.In order to better balance exploration and exploitation,the entire evolutionary process is divided into two stages.In the first stage,the population is divided into several small sub-swarms and one disadvantage sub-swarm.The non-overlapping ring topology based on spatial distance is used to improve the diversity of small subpopulations,so that the algorithm can search more Pareto optimal solutions,and use the global optimal location to update the inferior subpopulations to improve the search efficiency.In the second stage,all the particles follow the global optimal position to search and improve the search accuracy.At the same time,periodic recombination and a new global optimal position update strategy are introduced to avoid premature convergence.Simulation results show that the proposed algorithm can effectively solve multimodal multi-objective optimization problems.