Enhanced sine cosine algorithm using opposition learning, adaptive evolution and neighborhood search strategies for multivariable parameter optimization problems

Abstract

Sine cosine algorithm (SCA), an emerging metaheuristic method, is usually limited by the local convergence and search stagnation defects in multivariable optimization problems. To improve the SCA performance, this study proposes an enhanced sine cosine algorithm (ESCA) using several modified strategies, including the opposition learning strategy for enlarging search range, the adaptive evolution strategy for improving global exploration, the neighborhood search strategy for increasing population diversity, and the greedy selection strategy for guaranteeing solution quality. ESCA and several metaheuristics methods are used to solve a group of numerical optimization problems. The experimental results indicate that in terms of solution efficiency and convergence rate, ESCA outperforms several traditional methods for multivariable parameter optimization problems. Then, several engineering optimization problems are employed to further test the feasibility of the ESCA method in practical applications. The simulations show that for various performance evaluation indexes, ESCA can produce high-quality solutions with better objective values compared to the control methods. Thus, a simple but powerful tool is developed to address the complex multivariable parameter optimization problems.