Knowledge-Based Systems, vol.271, 2023 (SCI-Expanded)
The present study aims to optimize the engineering design and manufacturing problems with a novel hybrid optimizer named: AOA-NM (Arithmetic optimization-Nelder mead). To overcome the local optima trap shortcoming and improve the solution quality of a recently introduced arithmetic optimization algorithm (AOA), the Nelder–Mead local search methodology has been incorporated into the basic AOA framework. The objective of the proposed hybridization approach was to facilitate the refinement of the exploration–exploitation behaviour of the AOA search. In the numerical validation stage, numerous multidimensional benchmarks from the CEC2020 were used as challenging testing functions to investigate the suggested AOA-NM optimizer. To investigate the viability of the proposed hybridized algorithm in real-world applications, it is investigated for ten constrained engineering design problems, and the performance was contrasted with other distinguished metaheuristics extracted from the literature. Additionally, a hands-on manufacturing problem of milling process parameter optimization and vehicle structure shape optimization is posed and solved at the forefront to evaluate both AOA and AOA-NM efficacy. The proficiency of the AOA-NM algorithm, in terms of both solution quality and stability, is confirmed by performed comparative analysis and found to be robust in handling challenging practical issues.