Hybrid evolutionary grey wolf optimizer for constrained engineering problems and multi-unit production planning

A novel approach called hybrid evolutionary GWO (HE-GWO) is proposed to enhance the effectiveness of the grey wolf optimizer (GWO) for handling dynamic landscapes. This approach combines the strengths of the canonical GWO with Differential Evolution (DE) to achieve a high-level hybridization. HE-GWO enhances the leadership-based guidance system and incorporates an evolutionary dimensional hunting system with population subgrouping to increase the diversity of the wolfpack. Dynamic tuning settings are specifically devised to regulate the balance between exploration and exploitation, and to facilitate the evolution of the wolfpack by increasing its diversity. The proposed method is verified through rigorous benchmarking tests using three recent benchmarking suites, including CEC2020, CEC2019, and the 100-digit competition, as well as the CEC2021 benchmarking suites. Subsequently, a comparative analysis is performed on five standard engineering problems. The proposed HE-GWO is evaluated against a total of 23 meta-heuristics, which include the standard GWO, eight advanced and recent variants GWO, ten recent meta-heuristics, and four state-of-the-art advanced meta-heuristics. Additionally, the winners of relevant contests are also included in the comparison. Furthermore, the effectiveness of the suggested approach is verified through eight instances of a multi-unit production planning problem. This problem consists of 270 decision variables, including 162 integer variables and 108 continuous variables. The experiments involve applying a combination of penalty and constraint correction techniques to handle the constraints. HE-GWO has consistently outperformed state-of-the-art algorithms and competition winners in benchmarking, achieving the highest profitability in seven out of eight situations for the multi-unit production planning problem.