A new multi-function global particle swarm optimization

作者：Ruan, Zhao-Hui; Yuan, Yuan^*; Chen, Qi-Xiang; Zhang, Chuan-Xin; Shuai, Yong; Tan, He-Ping^*

来源：Applied Soft Computing, 2016, 49: 279-291.

DOI：10.1016/j.asoc.2016.07.034

• Particle swarm optimization (PSO)
• Population density
• Global convergence capability
• Search behavior
• Search capability

In this paper, we introduce the concept of population density in PSO, and accordingly, we discuss the relationship between the search capability of PSO and the population density. From related numerical experiments, we find that the search capability of PSO becomes saturated when the population density exceeds a certain value. Accordingly, we propose a strategy that divides the particles into two parts for different functions. Thus, we propose an approach called multi-function global particle swarm optimization (MFPSO) on the basis of this strategy. Further, we carry out a series of numerical experiments to verify that MFPSO has high global convergence capability, high convergence speed, and highly reliable performance when it is used to solve complex problems.

• 出版日期2016-12
• 单位哈尔滨工业大学