E-NSGA-II for machining process parameters optimization

Abstract

Optimization of machining process parameters is important to improve the machining performances. There are two consecutive ways to improve the machining performances namely modeling followed by optimization. In this study, modeling technique, namely regression is used to develop the machining model and optimization technique, multi objective genetic algorithm (MoGA) to optimize the machining process. Known as a popular MoGA, non dominated sorting genetic algorithm II (NSGA-II) is able to produce many sets of solutions with good spread of solutions from the Pareto optimal front in one time run. However, the confusion of selecting the best solution has led to the idea of using genetic algorithm (GA) and weight sum average (WSA) based as the preference points for NSGA-II. In this study, GA, WSA and combination of GA-WSA are selected as point to direct the best solutions among Pareto optimal front of NSGA-II. The machining processes for this study are cobalt bonded tungsten carbide electrical discharge machining and powder mixed electrical discharge machining. Surface roughness and material removal rate are the machining performances considered. GA-NSGA-II, WSA-NSGA-II and GA-WSA-NSGA-II known as enhanced NSGA-II (E-NSGA-II) are proposed. Two datasets from previous studies are used in this study. The results are compared with the previous studies and statistical analyses are performed to describe the significant of techniques proposed. In conclusion, E-NSGA-II is an improved technique that can increase the ability to provide best sets of optimal solutions and better stable process parameters values based on selected performance measurements compared to the previous techniques proposed.