Archive-based Differential Learning Incorporated Sparrow Search Algorithm
Qianrui Yu, Ziqian Wang, Haotian Li, Yifei Yang, Zhenyu Lei, Shangce Gao · 2022
Sparrow search algorithm (SSA) is a new evolutionary algorithm that has the advantage of good exploration ability. Benefiting from its search strategy, SSA can effectively alleviate the local optima. However, it still suffers from the issue of low solution quality because of its weak exploitation ability. Therefore, we propose an archive-based differential learning incorporated sparrow search algorithm (ASSA), which introduces a local search strategy for enhancing the exploitation ability of SSA. The local search strategy searches for a superior solution in the neighborhood region of the optimal solution in each iteration and uses it to replace the original optimal solution so that SSA can find a better solution. ASSA is compared with four state-of-the-art algorithms on 29 benchmark functions of IEEE CEC2017. The experiment results demonstrate that ASSA has superior performance to its competitors.