基于粒子群与聚类的多目标优化算法

针对粒子群优化算法容易陷入局部最优的问题，提出了一种基于粒子群优化与分解聚类方法相结合的多目标优化算法。算法基于参考向量分解的方法，通过聚类优选粒子策略来更新全局最优解。首先，通过每条均匀分布的参考向量对粒子进行聚类操作，来促进粒子的多样性。从每个聚类中选择一个具有最小聚合函数适应度值的粒子，以平衡收敛性和多样性。动态更新全局最优解和个体最优解，引导种群均匀分布在帕累托前沿附近。通过仿真实验，与4种粒子群多目标优化算法进行对比。实验结果表明，提出的算法在27个选定的基准测试问题中获得了20个反世代距离(IGD)最优值。

Multi-objective Optimization Algorithm Based on Particle Swarm and Clustering

To solve the problem that the particle swarm optimization algorithm is easy to fall into the local optimum.A multi-objective optimization algorithm based on the combination of particle swarm optimization and clustering method is proposed.The algorithm is based on the method of reference vector decomposition, and the global optimal solution is updated through the clustering optimization particle strategy.First, the particles are clustered by each uniformly distributed reference vector to promote the diversity of particles. A particle with the smallest aggregation function fitness value is selected from each cluster in order to balance convergence and diversity.The global optimal solution and the individual optimal solution are dynamically updated, and the population is guided to be evenly distributed near the Pareto front.It is compared with the four particle swarm multi-objective optimization algorithms through simulation experiments.Experimental results show that the proposed algorithm obtains 20 IGD optimal values on 27 selected benchmark problems.