A Stochastic Approach to Criterion Weighting in Multi-Criteria Decision Making: Entropy-Based Monte Carlo Modeling

Route selection for Unmanned Combat Aerial Vehicles (UCAVs) is of strategic importance due to its dynamism and uncertainty. This study addresses the weighting of criteria in the UCAV route planning process, methodologically advancing the findings of a previous. In the previous study, 54 criteria were analyzed using Einstein Fuzzy sets based on the evaluations of 41 experts from 14 different countries, and the 20 most critical criteria were determined. The current article accepts these "top 20 criteria" as the basic input; however, it presents a new approach to minimize the influence of subjective judgments on expert opinions and uncertainties in the decision-making process. In this context, the importance weights of the criteria were recalculated using Monte Carlo simulation and the integration of the Entropy weight method. Beta and triangular probability distributions were defined based on the entropy values of the criteria, and criterion weights were determined through simulations of 10,000 iterations. The weightings obtained from the study provide more robust results reflecting the variability in the system, as opposed to static results based solely on expert opinions. The methods used in the current literature generally disregard the uncertainties of expert opinions and express criterion weights as fixed numbers. This methodological deficiency constitutes the main motivation of our study. However, the motivation of the study is not limited to this theoretical deficiency alone. In Unmanned Combat Aerial (UAV) route planning, military operational areas are extremely dynamic, can change instantaneously, and fixed weights may be insufficient in modeling risk scenarios in the event of sudden threats. This study provides decision-makers with a realistic risk map showing lower and upper limits instead of a single fixed number. Thus, it offers a dynamic decision support infrastructure to the military command.