一种改进的多目标粒子群优化算法

为了增强多目标粒子群优化算法的收敛性与多样性,提出一种改进的多目标粒子群算法.采用Kent映射对种群进行初始化,并将目标空间均匀划分为若干扇形区域;基于一种新的多样性和收敛性判定标准,选取合适的收敛性最优解和多样性最优解,并提出一种改进的粒子群更新公式进行全局搜索;采用聚类算法对外部种群与坐标轴夹角进行分析,维护外部种群.通过标准测试函数的仿真实验,与多目标优化算法基本MOPSO(Multi-objective Particle Swarm Optimization Algorithm)和NSGA-II(Nondominated Sorting Genetic Algorithm II)进行对比,结果表明了该改进算法的有效性.

Improved multi-objective particle swarm optimization algorithm

In order to enhance the convergence and diversity of multi-objective particle swarm optimization algorithm, an improved multi-objective particle swarm optimization algorithm was proposed. The Kent mapping was used to initialize the population, and the target space was divided into several fan-shaped regions evenly. A new diversity and convergence criteria was proposed to select the optimal solutions. An improved particle swarm update formula was used for global search. The clustering algorithm was used to analyze the angles between external population and the axis, and ensure the diversity of external population. Compared with the multi-objective particle swarm optimization algorithm and the nondominated sorting genetic algorithm II, the experiment of benchmark functions simulation verifies the effectiveness of the improved algorithm.