A Mutation Lion Swarm Optimization Algorithm Based on Proportional Strategy

During the iterative process, the probability of selection is directly linked to the fitness magnitude, as each iteration of swarm intelligence optimization algorithms progressively converges towards an optimal solution. Building upon this premise, we introduce the Lion Swarm Optimization Algorithm based on the Proportional Strategy (PLSO), designed to enhance convergence speed and achieve superior optima. Inspired by the roulette strategy, our approach integrates the concept into the Lion Swarm Optimization (LSO) algorithm. In essence, it mimics the behavior of lion cubs, who learn hunting by following lionesses; however, their choice of lioness to follow is governed by rationality, favoring those with superior hunting skills. This fosters knowledge transfer among lionesses, strengthening the link between global and local optima, thereby enhancing local search capabilities and significantly accelerating convergence. In this paper, we evaluate the performance of the PLSO algorithm against six single-peak test functions, four multi-peak test functions, and four additional functions, chosen randomly. Comparative analyses are conducted with classical optimization algorithms, and the PLSO algorithm is applied to address the image segmentation problem. Our findings demonstrate the superior efficacy and robustness of the PLSO algorithm.