Robust multi-criteria decision-making of additive manufacturing processes under preference variability: a Monte-Carlo and graph-based framework

Metal Additive Manufacturing (AM) process selection is affected by uncertainty in techno-economic data and decision makers’ preferences. When competing technologies achieve similar overall performances, deterministic Multi-Criteria Decision-Making methods may produce strict rankings that do not adequately reflect the robustness of the underlying evidence. This work proposes a methodological framework for interpreting probabilistic Analytic Hierarchy Process results obtained through Monte Carlo simulation. Pairwise comparisons are modelled as random variables, generating distributions of global scores for the evaluated alternatives. Rather than relying only on mean scores or rank acceptability measures, the proposed framework analyses pairwise outscoring probabilities derived from the simulated scores. These probabilities are jointly represented through a dendrogram and a graph structure, allowing the identification of robust dominance relationships and equivalence classes among alternatives. The methodology is illustrated through a literature-based case study comparing Bound Metal Deposition and single- and multi-laser Laser Powder Bed Fusion systems under different industrial production scenarios. The results show that the proposed post-analysis framework reveals which ranking distinctions remain supported under uncertainty and which alternatives cannot be reliably distinguished. In particular, high-production scenarios reveal weak preferential separation between BMD and multi-laser L-PBF, suggesting that both technologies should be regarded as equivalent candidates rather than forcing a strict ranking. The proposed framework is particularly suited to support decisions involving emerging additive manufacturing technologies, where limited industrial maturity and uncertain techno-economic information often make strict deterministic rankings difficult to justify.