An uncertainty and interaction-aware framework for sustainable additive manufacturing decisions in electric vehicle applications

Expert-based sustainability assessment of additive manufacturing (AM) for electric-vehicle (EV) components frequently relies on crisp ratings due to elicitation simplicity and auditability. However, crisp evaluations conceal three critical uncertainty sources: inter-expert disagreement, heterogeneous assessor reliability, and criterion-dependent ambiguity. This paper proposes an aggregation-centered decision framework that (i) adaptively transfers crisp evaluations into trapezoidal fuzzy numbers using a Takagi–Sugeno (TS) uncertainty inference mechanism driven by dispersion, reliability, and criterion volatility; (ii) performs interaction-aware robust aggregation through a consensus-based capacity, Shapley-effective weights, and power-support adjustment to mitigate unsupported or weakly aligned opinions; and (iii) completes the decision using value–ambiguity ranking with an explicit ambiguity-aversion parameter. A full EV bracket manufacturing-route case study demonstrates deterministic matrix construction, uncertainty propagation, interaction-aware aggregation, comparative prioritized (TrIFPWA) and Einstein-induced baselines, module-ablation analysis, controlled expert-conflict stress testing, and extensive sensitivity and uncertainty analyses. Results show that the ranking favors the polymer-composite route, and the preferred alternative remains stable across the tested ranges of ambiguity aversion, interaction strength, robustness parameters, uncertainty scaling, reliability perturbation, module-ablation variants, and progressively stronger expert conflict. Under maximum controlled conflict, the full model attenuates the conflicting expert more strongly and preserves a larger top-second score gap than additive aggregation. The proposed workflow enhances interpretability, robustness, and traceability in sustainability-driven manufacturing decision making.