A memory guided sine cosine algorithm for global optimization

Real-world optimization problems demand an algorithm which properly explores the search space to find a good solution to the problem. The sine cosine algorithm (SCA) is a recently developed and efficient optimization algorithm, which performs searches using the trigonometric functions sine and cosine. These trigonometric functions help in exploring the search space to find an optimum. However, in some cases, SCA becomes trapped in a sub-optimal solution due to an inefficient balance between exploration and exploitation. Therefore, in the present work, a balanced and explorative search guidance is introduced in SCA for candidate solutions by proposing a novel algorithm called the memory guided sine cosine algorithm (MG-SCA). In MG-SCA, the number of guides is decreased with increase in the number of iterations to provide a sufficient balance between exploration and exploitation. The performance of the proposed MG-SCA is analysed on benchmark sets of classical test problems, IEEE CEC 2014 problems, and four well known engineering benchmark problems. The results on these applications demonstrate the competitive ability of the proposed algorithm as compared to other algorithms.