Development of RSM-based multi-attribute genetic algorithm (RAGA) for optimisation of bead profile and hardness of Al4043 weld deposits

Multi-attribute decision-making in manufacturing is often hindered by nonlinear process-response relationships and the inadequate integration of weighted performance criteria. To overcome these limitations, this study proposes the Response Surface Methodology-based Multi-Attribute Genetic Algorithm (RAGA), a novel optimisation methodology developed for single-weld-bead deposition of Al4043 alloy using gas metal arc-based processes. Unlike conventional techniques such as Artificial neural network (ANN)-Genetic algorithm (GA), n-ary Technique for order of preference by similarity to ideal solution (TOPSIS), and Multi objective GA (MOGA), RAGA offers three key advancements: (i) it uses analytically derived weights from the Analytic Hierarchy Process into the genetic algorithm, ensuring criteria relevance is explicitly considered; (ii) it integrates response surface models to efficiently handle non-linearities in process-response mapping; and (iii) it introduces user-defined optimisation intentions (maximisation/minimisation) to tailor objectives to optimisation problems. Comparative assessments reveal that RAGA achieves up to 20% higher optimisation accuracy than other methods, showcasing both precision and flexibility. Case studies in welding, cladding, and additive manufacturing confirm RAGA’s adaptability across materials and processes. Sensitivity analysis identifies current and travel speed as dominant input variables influencing bead geometry and hardness. These findings position RAGA as a robust, scalable tool for solving complex multi-attribute optimisation challenges in advanced manufacturing.