Laplacian Salp Swarm Algorithm for continuous optimization

The most essential task of any algorithm is to maintain the balance between its exploration and exploitation capabilities while solving an optimization problem. Salp Swarm Algorithm (SSA) is a population-based metaheuristic algorithm inspired by the swarming behavior of salps when foraging and navigating in oceans. SSA suffers from stagnation in local optima and poor convergence speed with low solution accuracy. This study proposes a modified SSA based on the Laplace crossover operator called Laplacian Salp Swarm Algorithm (LX-SSA). In LX-SSA, a new position updating mechanism is used by the follower salps to enhance the exploration and solution accuracy of the classical SSA. Thus, better exploration enhances the convergence speed of LX-SSA while avoiding the local optima. The performance of the proposed LX-SSA is tested on a problem set of 23 standard benchmark functions and is compared with the classical SSA along with Ant Lion Optimizer, Moth-flame Optimizer, Sine-Cosine Algorithm, and Whale Optimization Algorithm. The experimental results demonstrate that the proposed LX-SSA outperforms the classical SSA and the other considered algorithms for most of the unimodal and multimodal functions regarding convergence speed and solution precision. A non-parametric Wilcoxon rank sum statistical test has been performed to analyse the results. The experimental and statistical results conclude that the proposed LX-SSA is a promising modified version of classical SSA in terms of convergence and solution accuracy. This paper also examines some computationally expensive and challenging classical engineering design problems.