Fireworks Algorithm Based on Search Space Partition

Yifeng Li,Ying Tan

DOI: https://doi.org/10.1002/int.22952

IF: 8.993

2022-01-01

International Journal of Intelligent Systems

Abstract:Fireworks algorithm is a novel swarm intelligence optimization framework which focuses on the potential of collaboration among multiple subpopulations with independent search ability. Although it has been proved to perform excellently in many tasks, the collaborative mechanism of fireworks is still quite undeveloped. In this paper, a theoretical model of fireworks algorithm based on search space partition is proposed, analyzed and implemented. The local search of each firework is replaced by covariance matrix adaptation evolution strategy for efficient exploitation. A coordination strategy inspired from Voronoi diagram is proposed to approximate the theoretical model for stable global exploration. Experimental results show that the proposed algorithm not only outperforms previous variants of fireworks algorithm significantly, but also achieves competitive results compared with state-of-the-art evolutionary algorithms, which are intensively fine-tuned on the objective functions.