AN ARTIFICIAL NEURAL NETWORK MODEL SUPPORTED WITH HYBRID MULTI-CRITERIA DECISION-MAKING APPROACHES TO RANK LEAN TOOLS FOR A FOUNDRY INDUSTRY.

The primary objective of this study is to optimise operating efficiency and minimise waste within a core foundry shop through the application of lean manufacturing techniques. The research emphasises the significance of Artificial Neural Networks (ANNs) in aligning an expert assessment matrix with lean tool rankings, particularly in addressing the challenges associated with fu//y logic-based leanness computation. The expert assessment matrix was constructed with the entropy approach for generating weights and the TOPSIS ranking algorithm for evaluating lean tools. The use of the TOPSIS technique resulted in a notable level of agreement, with a percentage of73.42%, and a corresponding level of disagreement of26.57%, when compared to the expert evaluation matrix developed for the assessment of lean tools. The expert assessment matrix that was produced was utilised in the analysis of the efficacy of several lean tools inside a foundry core manufacturing line. The research suggests the implementation of an automated conveyor system for the transportation of several cores, which would lead to the optimisation of floor space, enhanced safety measures, and more schedule flexibility. The findings of this study reveal a significant decrease of 79.6% in non-value-added activities (NVA), a notable improvement of62.66% in process efficiency, a substantial reduction of66.66% in waiting times, a considerable decrease of 35% in personnel requirements, and a significant cost reduction of 45%. A three-month accident-free workplace demonstrated the efficacy of the safety strategy. [ABSTRACT FROM AUTHOR]