Glowworm Swarm Optimization Algorithm Based on Hierarchical Multi-subgroups

摘要

Glowworm swarm optimization algorithm has been widely used in the multimodal function optimization. However, this method becomes very time-consuming and has low convergence speed and convergence accuracy due to the fact that the number of glowworms needs to increase accordingly when the number of peaks is increased. To solve the problem, a new algorithm called glowworm swarm optimization algorithm, which is based on hierarchical multi-subgroup (HGSO), is presented in this paper. The algorithm initializes the position of glowworms by chaotic mapping for underlying subgroups and improves the convergence performance. For top swarm, the operation of selection and crossover is incorporated to enhance the diversity of glowworms and adaptive step size update strategy is proposed. Finally, several typical test functions are simulated and the simulation results show that computing time is greatly reduced and the convergence speed and accuracy is increased when improved algorithm is used in solving multimodal functions.