元胞多目标粒子群优化算法与其应用

Algorithm and Application of Cellular Multi-objective Particle Swarm Optimization

针对现有多目标粒子群算法多样性不佳,难以平衡多目标优化的全局搜索和局部寻优的能力,提出了一种元胞多目标粒子群算法。在分析多目标粒子算法理论基础上,该算法将元胞自动机思想融入粒子群算法,研究粒子之间相互关系和信息传递机制,并提出一种粒子飞行速度控制策略。实验证明,新算法相对于4种比较算法,在求解含有无约束和有约束的多目标优化问题时有更好的收敛性和多样性,将其应用于盘式制动器优化设计,得到的解精度更高。

For improving the diversity of existing multi-objective particle swarm optimization algorithm and keeping the balance between exploration and exploitation well, a multi-objective cellular PSO was proposed. The algorithm combined the concept of cellular automata with the multi-objective PSO theory. In addition, the relationship between the particles and the information transmission mechanism was studied, and a particle flight speed control strategy was presented. The results indicate that the improved algorithm outperforms the four compared algorithms concerning the convergence and diversity in solving multi-objective optimization problems with unconstraint and constraint. And also, the new algorithm can get more accurate solutions when applied in disc brake design problem.