Exponentially adaptive sine-cosine algorithm for global optimization

Sine-Cosine algorithm (SCA) is an optimization algorithm formulated based on mathematical Sine and Cosine terms. It is widely used to solve various optimization problems. However the algorithm performance in terms of accuracy is not at optimum level. This paper presents an improved SCA with a new adaptive strategy based on an exponential term. The exponential term is adopted to establish a relationship between searching agents step size and fitness cost. The agents step size is exponentially changed due to the change of the fitness cost. The proposed algorithm is tested with a set of benchmark functions in comparison to the original SCA. A statistical analysis of the algorithms performance in terms of their accuracy is conducted. A Wilcoxon Sign Rank test is adopted to check significance level of the proposed algorithm as compared to the original SCA. Based on the simulation conducted, the adaptive strategy has resulted a significance improvement of the accuracy and convergence speed