A sine cosine algorithm guided by elite pool strategy for global optimization

When handling global optimization problems by metaheuristic algorithms (MAs), an important yet difficult assignment is to keep a tradeoff between the swarm's diversity and convergence. Hence, this paper develops an enhanced sine cosine algorithm called EPSCA to achieve the above target. In EPSCA, to balance diversity and convergence, the elite pool strategy and Brownian motion are introduced to modify the position updating formula of the original SCA. Then the pattern search method serves as a local search tool to reinforce the quality of the current best solution. Additionally, an efficient mutation operator is devised to discourage premature convergence. The comparative analysis between the developed EPSCA and other state-of-the-art techniques is conducted on 30 CEC2017 benchmarks. Besides, EPSCA is employed to identify the core parameters of photovoltaic (PV) systems. The experimental evidence based on 30 independent runs demonstrated that EPSCA can perform better than four state-of-the-art SCA variants and five other well-known methodologies. The source code of EPSCA is publicly available at https://github.com/denglingyun123/EPSCA.