Impact of the improved tuna algorithm on optimization and engineering problems

Tuna Swarm Optimization (TSO) is a powerful metaheuristic algorithm inspired by the fishing strategies of tuna fish. However, it has fundamental drawbacks such as premature convergence and local minimum trap. In this study, three new algorithms are developed to overcome these shortcomings of TSO and improve its performance: TSO-CM (TSO with Chebyshev Map), CSTSO (TSO with Cuckoo Search) and CSTSO-CM (TSO with Chebyshev Map and Cuckoo Search). TSO-CM aims to avoid the local optimum trap by diversifying the initial solutions with the chaotic structure provided by the Chebyshev Map and to increase population diversity in the search process. CSTSO avoids the problem of premature convergence thanks to the adaptive search mechanism of the Cuckoo Search algorithm, allowing it to explore a wider search space. CSTSO-CM aims to reach the global optimum faster and with higher accuracy by combining the strengths of both methods. The performance of the developed algorithms is tested on various benchmark functions and the superiority of CSTSO-CM is statistically proven by Friedman and Wilcoxon tests. In addition, evaluations on engineering optimization problems were also performed: CSTSO was applied to the Welded Beam design problem and achieved efficient results, while CSTSO-CM performed well on the Gear Train design problem. The results show that the proposed algorithms outperform existing metaheuristic algorithms in terms of search accuracy, diversity and convergence to the global optimum. In this context, it is shown that the proposed methods offer innovative and effective solutions to optimization problems. Keywords: Chebyshev Map, Cuckoo Search, Hybrid, Tuna Swarm Optimization