Advancement of the search process of salp swarm algorithm for global optimization problems. (15th November 2021)

• Record Type: Journal Article
• Authors: Çelik, Emre; Öztürk, Nihat; Arya, Yogendra
• Is Part Of: Expert systems with applications. Volume 182(2021)
• Publication Date: 2021-11-15
• Subjects: Modified algorithm -- Chaos theory -- Sinusoidal map -- Mutualism -- Global optimization; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2021.115292
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A novel variant of salp swarm algorithm is proposed. The proposal fruitfully employs three simple but effective methodologies. It is applied to 35 benchmark test functions and four real-life application problems. Results are widely compared to the relevant results in literature. The findings are highly impressive, ratifying the overt potential of this work. Abstract: This paper propounds a modified version of the salp swarm algorithm (mSSA) for solving optimization problems more prolifically. This technique is refined from the base version with three simple but effective modifications. In the first one, the most important parameter in SSA responsible for balancing exploration and exploitation is chaotically changed by embedding a sinusoidal map in it to catch a better balance between exploration and exploitation from the first iteration until the last. As a short falling, SSA can't exchange information amongst leaders of the chain. Therefore, a mutualistic relationship between two leader salps is included in mSSA to raise its search performance. Additionally, a random technique is systematically applied to the follower salps to introduce diversity in the chain. This can be since there may be some salps in the chain that do not necessarily follow the leader for exploring unvisited areas of the search space. Several test problems are solved by the advocated approach and results are presented in comparison with the relevant results in the available literature. It is ascertained that mSSA, despite its simplicity, significantly outperforms not only the basic SSA but also numerous recent algorithms in terms of fruitful solution precision and convergent trend line.