A Multi-Stage Adaptive Sequential Parameter Exploration Hunger Games Search Algorithm for Solving Complex Optimization Problems

Hunger Games Search (HGS) is a newly developed metaheuristic algorithm that models the hunger-driven activities and behaviors of animals. It incorporates the concept of hunger to devise an adaptive weight that mimics the impact of hunger on each search step. In this paper, a Multi-Stage Adaptive Sequential Parameter Exploration Hunger Games Search Algorithm (MASPE-HGSA) is proposed to alleviate the shortcomings of the original HGS in terms of insufficient optimization and convergence accuracy. In MASPE-HGSA, a Multi-Stage adaptive sequential parameter exploration is proposed to improve the search performance of the algorithm as well as to increase the search accuracy and global search capability, which can commendably achieve the balance of exploration and exploitation. The effectiveness of MASPE-HGSA is verified by comparing with original HGS algorithm and several classical algorithms using 23 benchmark functions, CEC2014 test set and three advanced algorithmic problems from classical engineering. Experimental results show that the performance of MASPE-HGSA is significantly better than other similar algorithms. The proposed algorithm can effectively search for high-quality solutions and prevent premature convergence, with better convergence robustness than the original HGS algorithm. In addition, an analysis of probability-based algorithm mechanisms approaching zero is presented in this paper from a theoretical perspective, while providing a reference and verification method for the design of algorithm mechanisms.