A multi-strategy enhanced sine cosine algorithm for global optimization and constrained practical engineering problems

Huiling Chen,Mingjing Wang,Xuehua Zhao

DOI: https://doi.org/10.1016/j.amc.2019.124872

IF: 4.397

2020-03-01

Applied Mathematics and Computation

Abstract:<p>The Sine Cosine Algorithm (<em>SCA</em>) has received much attention from engineering and scientific fields since it was proposed. Nevertheless, when solving multimodal or complex high dimensional optimization tasks, the conventional <em>SCA</em> still has a high probability of falling into the local optimal stagnation or failing to obtain the global optimum solution. Additionally, it performspoorly in convergence. Therefore, in this study, a multi-strategy enhanced <em>SCA</em>, a memetic algorithm termed <em>MSCA</em>, is proposed, which combines multiple control mechanisms including Cauchy mutation operator, chaotic local search mechanism, opposition-based learning strategy and two operators based on differential evolution to achieve a better balance between exploration and exploitation. To verify its performance, <em>MSCA</em> was compared with 11 state-of-the-art original optimizers and variant algorithms on 23 continuous benchmark tasks including 7 unimodal tasks, 6 multimodal tasks, 10 various fixed-dimension multimodal functions, and several typical CEC2014 benchmark problems. Furthermore, <em>MSCA</em> was utilized to solve three constrained practical engineering problems including tension/compression spring design, welded beam design, and pressure vessel design. The experimental results demonstrate that the proposed algorithm <em>MSCA</em> is superior to other competitors in terms of quality of solutions and convergence speed and can serve as an effective andefficient computer-aided tool for practical tasks with complex search space.</p>

mathematics, applied