Abstract — Large-scale engineering optimization problems, characterized by hundreds or thousands of decision variables, complex constraints, and highly nonlinear objective functions, present significant challenges for conventional optimization methods. These problems require robust algorithms that can efficiently explore vast search spaces while preserving solution diversity. The Grey Wolf Optimizer (GWO), inspired by the hunting behavior of wolf packs, has gained popularity as an effective metaheuristic due to its simplicity, flexibility, and population-based search mechanism. However, GWO’s reliance on a single, greedy search strategy focused on the top three wolves limits its exploration and diversity, often leading to premature convergence. To address these limitations, this study proposes an Evolutionary Grey Wolf Optimizer (EGWO), which enhances GWO’s optimization capabilities for solving large-scale, constrained mechanical and structural engineering design problems. EGWO integrates a cross-search strategy along with elite and worst-of-elite memory mechanisms to guide the population toward promising regions over iterative cycles. The cross-search strategy combines average and perturbed search methods to create new candidate solutions, boosting both exploration potential and overall performance. Moreover, EGWO employs self-adaptive control parameters that dynamically balance exploration and exploitation throughout the optimization process. The proposed EGWO was evaluated using benchmark functions and constrained engineering design problems. Comparative experiments against eight state-of-the-art metaheuristic algorithms demonstrate that EGWO consistently outperforms its competitors, confirming its superior capability in solving complex constrained optimization problems.
Zamani et al. (Mon,) studied this question.