Multi-objective optimization and decision-making in ultrasonic vibration-assisted EDM of HARDOX 500 using NSGA-II coupled with SAW, MARCOS, and TOPSIS = Tối ưu hóa đa mục tiêu và ra quyết định trong quá trình gia công xung điện thép HARDOX 500 có trợ giúp rung siêu âm sử dụng thuật toán NSGA-II kết hợp với các phương pháp SAW, MARCOS và TOPSIS

Abstract

This paper presents a hybrid multi-objective optimization and decision-making framework for improving the performance of ultrasonic vibration-assisted electrical discharge machining (UV-EDM) of HARDOX 500 steel. The Non-dominated Sorting Genetic Algorithm II (NSGA-II), assisted by Gaussian Process Regression (GPR) surrogate models, was employed to simultaneously maximize the material removal rate (MRR) and minimize the surface roughness (Ra). The resulting Pareto front demonstrated a clear trade-off between machining productivity and surface quality. To identify the most balanced solution among the Pareto-optimal points, three multi-criteria decision-making (MCDM) techniques – Simple Additive Weighting (SAW), Measurement of Alternatives and Ranking according to Compromise Solution (MARCOS), and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) – were applied. The Spearman correlation analysis showed a perfect agreement between SAW and MARCOS (ρ = 1.000) and a high correlation (ρ ≈ 0.903) between these two methods and TOPSIS. The integration of NSGA-II with comparative MCDM analysis provides a reliable and systematic approach for selecting optimal machining parameters, enabling enhanced efficiency and surface finish in UV-EDM of hardened steels.