Optimization of Modern Manufacturing Processes Using Three Multi-Objective Evolutionary Algorithms

Abstract

The optimization in manufacturing processes refers to the investigation of multiple responses simultaneously. Therefore, it becomes very necessary to introduce a technique that can solve the multiple response optimization problem efficiently. In this study, an attempt has been taken to find the application of three newly introduced multi-objective evolutionary algorithms, namely multi-objective dragonfly algorithm (MODA), multi-objective particle swarm optimization algorithm (MOPSO), and multi-objective teaching-learning-based optimization (MOTLBO), in the modern manufacturing processes. For this purpose, these algorithms are used to solve five instances of modern manufacturing process—CNC process, continuous drive friction welding process, EDM process, injection molding process, and friction stir welding process—during this study. The performance of these algorithms is measured using three parameters, namely coverage to two sets, spacing, and CPU time. The obtained experimental results initially reveal that MODA, MOPSO, and MOTLBO provide better solutions as compared to widely used nondominated sorting genetic algorithm II (NSGA-II). Moreover, this study also shows the superiority of MODA over MOPSO and MOTLBO while considering coverage to two sets and CPU time. Further, in terms of spacing a marginally inferior performance is observed in MODA as compared to MOPSO and MOTLBO.