A robust DIBR II–MULTIMOORA–Monte Carlo framework for anti-tank system selection under battlefield uncertainty
Abstract
In this study, a robust multi-criteria model was developed for the selection of a modern anti-tank system under conditions of pronounced battlefield uncertainty. The proposed framework integrates the Interval-Based Defining Interrelationships Between Ranked Criteria II (DIBR II) method for determining criteria weights, the MULTIMOORA (Multi‐Objective Optimization by Ratio analysis plus Full Multiplicative Form) method for ranking alternatives, and Monte Carlo simulation for testing the robustness of the results. The case study includes seven coded anti-tank system alternatives (A1–A7) evaluated according to eight criteria covering tactical, technical, operational, and logistical dimensions. An expert panel of eleven specialists ranked the criteria from the most to the least important, after which the weights were calculated using the Interval-Based DIBR II method. Inter-rater agreement was very high (Kendall’s W = 0.9803, χ²(7) = 75.4848, p < 0.001), supporting the consistency of the aggregated preference structure. The Monte Carlo simulation with 1,000 iterations further confirms the robustness of the obtained ranking: A3 retains first place in approximately 89.9% of scenarios, while A7 and A2 are the most frequent competitors for second and third place. The results show that scenario-oriented validation changes the interpretation of the decision: the nominally best alternative is not necessarily absolutely dominant in every possible combat environment, but a system that maintains the leading position across a large number of simulated scenarios can be considered the most robust choice. For A3, the estimated first-place probability of 0.899 had a 95% percentile-bootstrap confidence interval of 0.880–0.917. Additional statistical comparisons and a targeted stress test were used to evaluate whether the leading position was merely an artefact of the baseline data. The contribution of the study lies in linking rank-based weight determination, a triple aggregation model, and stochastic validation within a single framework adapted to military decision-making.