A Modified Subtraction Average-Based Optimizer for Solving Optimization Problems

Abstract

Subtractive average-based optimizer (SABO) is extensively employed to resolve global optimization problems. However, similar to other metaheuristic algorithms, the SABO algorithm also suffers from the problem of insufficient global search ability. To overcome the problems, a novel modified SABO (MSABO) is proposed, which consists of a refracted opposition-based learning (ROBL) mechanism, piecewise mapping randomized operator selection, and a Golden Sine algorithm (GSA) mechanism. The MSABO obtains high-quality populations by the ROBL mechanism, increases the diversity of the populations, and enhances the algorithm’s global searching ability. Then, the piecewise mapping operator is employed for randomized operator selection, and the GSA mechanism is utilized to optimize the searching optimum position, which makes the algorithm easily escape from the local optimum and further elevates the convergence accuracy. To prove the performance of the MSABO, the proposed algorithm and other algorithms are thoroughly analyzed with benchmark functions. Furthermore, to assess the capability of the MSABO, the proposed algorithm is utilized to address two real engineering design problems. The results indicate that the MSABO obtains better results than other state-of-the-art algorithms and is an effective approach to solving design problems.