Single, Multi-, and Many-Objective Optimization of Manufacturing Processes Using Two Novel and Efficient Algorithms with Integrated Decision-Making

This is the authors' abstract. We don't add key points for this paper.

Manufacturing processes are inherently complex, multi-objective in nature, and highly sensitive to process parameter settings. This paper presents two simple and efficient optimization algorithms—Best–Worst–Random (BWR) and Best–Mean–Random (BMR)—developed to solve both constrained and unconstrained optimization problems of manufacturing processes involving single, multi-, and many-objectives. These algorithms are free from metaphorical inspirations and require no algorithm-specific control parameters, which often complicate other metaheuristics. Extensive testing reveals that BWR and BMR consistently deliver competitive, and often superior, performance compared to established methods. Their multi- and many-objective extensions, named MO-BWR and MO-BMR, respectively, have been successfully applied to tackle 2-, 3-, and 9-objective optimization problems in advanced manufacturing processes such as friction stir processing (FSP), ultra-precision turning (UPT), laser powder bed fusion (LPBF), and wire arc additive manufacturing (WAAM). To aid in decision-making, the proposed BHARAT can be integrated with MO-BWR and MO-BMR to identify the most suitable compromise solution from among a set of Pareto-optimal alternatives. The results demonstrate the strong potential of the proposed algorithms as practical tools for intelligent decision-making in real-world manufacturing applications.