二进制粒子群算法中V型转换函数的应用分析

二进制粒子群算法(BPSO)由于规则简单、参数设置较少等优点被广泛应用到各领域,但是其具有过强的全局搜索能力,缺乏局部的搜索能力等缺陷。针对BPSO存在的缺陷很多文献提出了改进方法,但是针对转换函数的改进较少。通过定义粒子间的距离来分析出BPSO所存在的缺陷,从而进一步分析BPSO中S型转换函数的缺点,并且有针对性地提出更符合BPSO要求的V型转换函数。实验结果表明,所提V型转换函数能克服原始BPSO的缺陷,相比S型转换函数以及现有文献所提的V型转换函数更能提升算法的性能,得到更高的分类准确率。

APPLICATION ANALYSIS OF V-SHAPED TRANSFER FUNCTION IN BINARY PARTICLE SWARM OPTIMIZATION

Binary particle swarm optimization(BPSO)is widely used in various fields due to its simple rules and few parameter settings,but it has too strong global search ability and lacks local search ability.Many literatures have proposed improvement methods for the defects of BPSO,but few for the transfer function.This paper analyzes the defects of BPSO by defining the distance between particles,so as to further analyze the shortcomings of S-shaped transfer function in BPSO,and puts forward a V-shaped transfer function more in line with the requirements of BPSO.The experiments show that the V-shaped transfer function proposed in this paper can overcome the defects of the original BPSO,and can improve the performance of the energy algorithm and obtain higher classification accuracy than the S-shaped transfer function and the V-shaped transfer function proposed in existing literatures.