Abstract
To address the problem that the sand cat swarm optimization (SCSO) algorithm experiences a decline in convergence speed and a tendency to fall into local optima during the iteration process, this paper proposes a multi-strategy enhanced sand cat swarm optimization (MESCSO) algorithm to improve its ability to escape local optima and enhance convergence efficiency. Firstly, an improved sine mapping combined with random opposition-based learning (ISMROBL) is employed during population initialization to enhance the uniformity and diversity of initial solutions. Secondly, a nonlinear decreasing parameter is introduced to dynamically balance global exploration and local exploitation. Thirdly, generalized quadratic interpolation (GQI) is incorporated to strengthen global search capability, while the improved mean differential mutation (IMDM) strategy enhances local exploitation. Finally, accelerated opposition-based learning (AOBL) is applied to refine individual positions and improve the algorithm's ability to escape local optima. Experimental results on 23 standard benchmark functions and the CEC2014 benchmark functions show that MESCSO achieves superior performance compared to nine algorithms. In addition, MESCSO is tested on five constrained engineering design problems. The results demonstrate that, compared with SCSO, MESCSO yields improvements of 0.53%, 1.47%, 0.03%, 0.41%, and 0.31%, respectively, thereby confirming its effectiveness and applicability.