Stability Of Salp Swarm Algorithm With Random Replacement And Double Adaptive Weighting

The salp swarm algorithm is a newly population-based search algorithm. Because the orig-inal salp swarm algorithm has low search efficiency and is easy to fall into local opti-mum, in this paper, we propose an enhanced salp swarm algorithm, which combines two strategies with the original salp swarm algorithm. One is the random replacement strategy, which can replace the current position with the optimal solution position with a certain probability of speeding up the convergence rate. The other tactic is double adaptive weight, which can expand the search scope throughout the early stages and enhance exploitation capability in the later stages. With the cooperation and guidance of the two mechanisms, the algorithm's convergence speed is accelerated, and the exploitation capacity is meritori-ously increased. The proposed method's performance is compared with three mainstream meta-heuristics and four advanced algorithms on four necessary test cases. The extensive analysis and recorded results indicate that the proposed method outperforms these algo-rithms in terms of the accuracy of the solution and convergence speed. Finally, we apply the developed method to four well-known engineering design problems (welded beam de-sign problem; cantilever beam design; I-beam design; and multiple disk clutch brake) to validate the algorithm's effectiveness for some constrained challenge. The results show that our algorithm has significant advantages in solving practical problems with constraints and unknown search spaces.(c) 2021 Elsevier Inc. All rights reserved.