Integration of metaheuristic operators through unstructured evolutive game theory approach: a novel hybrid methodology

This paper introduces a novel approach to addressing the complexity of current optimization challenges by developing hybrid metaheuristic algorithms. These algorithms combine the strengths of different strategies to enhance solution identification and refinement. In this work, the integration of metaheuristic operators using the Unstructured Evolutive Game Theory to regulate and merge the advantageous features of the Cheetah Optimizer and Particle Swarm Optimization to solve continuous optimization problems is proposed. The Cheetah Optimizer and Particle Swarm Optimization are chosen for their superior exploitation and exploration capabilities, respectively. Our methodology unfolds in two primary phases: combination and modulation. In the combination phase, Cheetah Optimizer and Particle Swarm Optimization search strategies are merged, creating a unified population that generates new candidate solution positions based on the best solution and a modulation factor. During the modulation phase, pairwise competitions based on Unstructured Evolutive Game Theory assess candidate solutions against the objective function, adjusting their modulation factor and position accordingly. We conducted various experiments to evaluate our approach against the original Cheetah Optimizer and Particle Swarm Optimization, as well as seven other well-known metaheuristic algorithms and hybrid schemes, across 30 benchmark functions in dimensions of 30, 50, and 100. The results reveal that our hybrid scheme outperforms the comparative algorithms, demonstrating enhanced performance, effectiveness, and robustness through a detailed convergence and significance analysis.