Monte Carlo Analysis of Gathering and Scattering Behavior Generated by Gravitational Particle Swarm Algorithm
Monte Carlo Analysis of Gathering and Scattering Behavior Generated by Gravitational Particle Swarm Algorithm
复制标题
DOI：
10.5687/sss.2019.167
复制
发表时间：
2019-07
期刊：
影响因子：
--
通讯作者：
Yoshikazu Yamanaka;Katsutoshi Yoshida
中科院分区：
文献类型：
--
作者：
Yoshikazu Yamanaka;Katsutoshi Yoshida
在本研究中，我们提出了一种新的多模式优化(MMO)算法--重力粒子群算法(GPSA)。这用引入粒子间反平方引力的项取代了经典粒子群优化中的全局反馈项。通过蒙特卡罗模拟分析了该算法的搜索行为，并对其在一维MMO问题上的搜索性能进行了评估。实验结果表明，该算法的粒子具有聚集和散射行为，在不使用任何算法聚类的情况下，我们的算法能够fi和83%的最优解。
In this study, we propose a novel multimodal optimization (MMO) algorithm, gravitational particle swarm algorithm (GPSA). This replaces the global feedback term of a classical particle swarm optimization with a term that introduces inverse-square gravitational force between particles. We analyze GPSA’s search behavior by Monte-Carlo simulation and evaluate its search performance for a one-dimensional MMO problem. It is shown that our GPSA’s particles exhibit gathering and scattering behavior, based on which our GPSA was able to find 83% of the optimal solutions without using any algorithmic clustering procedures.