Lévy-inspired Salp Swarm Optimization Algorithm

Salps are barrel-shaped sea tunicates that typically form chains and float freely in the water. In the Salp Swarm Algorithm (SSA), the positions of all follower salps are adjusted relative to each other so that they move toward the main salp. However, this tendency often causes the leading salp to trap in local optima due to increased exploitation. To address this issue, this paper presents a modified variant of SSA called the Lévy-inspired Salp Swarm Optimization Algorithm (LSSA). The proposed method employs the Lévy distribution to enhance convergence precision. The efficacy of LSSA is validated on 18 standard benchmark problems, with results compared to five other recent algorithms regarding mean fitness. Additionally, the proposed method is statistically validated using Wilcoxon test and box plots. The experimental outcomes demonstrate that LSSA outperforms the compared methods on more than 85% of test functions.