摘要: 针对粒子群优化(PSO)易陷入局部最优、收敛速度慢的现象,提出一种新的惯性权重取值方法——分段取值惯性权重(SW)方法。该方法在算法前期增加粒子多样性,后期加速算法收敛。针对PSO仅使用2个最优值寻优的问题,引入第3个最优值GB,将SW与GB结合,改进PSO的进化方程。实验结果表明,该算法解决多序列比对问题时,可以有效地避免算法早熟,并提高解的精度。
关键词:
粒子群优化算法,
分段取值惯性权重,
SW与GB的结合
Abstract: In this paper, a new method of getting inertia weight, Subsection Weight(SW) is proposed in order to solve the Particle Swarm Optimization(PSO) disadvantages which are likely to fall into local optimum and slow converge. The diversity of swarm increases at the prophase and the convergence is accelerated in the later period. Meanwhile, the combination of SW and GB can improve the evolutionary equation of PSO and makes it perform better. Experimental result shows that the algorithm can effectively avoid converging too early and increase the precision in solving multiple sequence alignment.
Key words:
Particle Swarm Optimization(PSO) algorithm,
Subsection Weight(SW),
combination of SW and GB
中图分类号:
徐小俊, 雷秀娟, 郭玲. 基于SWGPSO算法的多序列比对[J]. 计算机工程, 2011, 37(6): 184-186.
XU Xiao-Dun, LEI Xiu-Juan, GUO Ling. Multiple Sequence Alignment Based on SWGPSO Algorithm[J]. Computer Engineering, 2011, 37(6): 184-186.