Sustainable Manufacturing Material Selection Using a Reliability- and Data-Maturity-Aware Multicriteria Decision Model

ABSTRACT Material selection in a digitally connected factory is not a single screening event: Evidence changes as supplier certificates, pilot runs, qualification tests, and first-article records become available. This paper develops a reliability- and data-maturity-aware multicriteria decision model for early-stage industrial material selection. The model combines fixed engineering normalization anchors, subjective priorities, entropy information, criterion-specific evidence maturity, service reliability, TOPSIS, VIKOR, Monte Carlo rank probabilities, and a confidence-based retain-or-switch rule. A literature-calibrated synthetic case evaluates six materials for a removable guarding panel used on an automated manufacturing cell. The decision is updated through five evidence gates, from literature prior to first-article production. Early evidence can favor glass-fiber polypropylene, but pilot and qualification data shift the recommendation toward recycled AA5052-H32. At the final gate, recycled AA5052-H32 has a 0.828 probability of ranking first and a 0.995 probability of ranking in the top two; the model then recommends switching from the galvanized-steel incumbent. Scenario tests show that galvanized high-strength low-alloy (HSLA) steel remains preferred for high-throughput, corrosion-critical, and cost-constrained settings. The contribution is a traceable smart-manufacturing decision architecture that links material ranking to evidence quality and production learning rather than treating literature values as equally credible and permanently fixed.