Document is current

Article History

Received: 16 October 2025

Accepted: 12 March 2026

First Online: 22 April 2026

Declarations

:

: The authors declare no competing interests.

: This article does not contain any studies with human participants or animals performed by the author.

: To enhance transparency and reproducibility, we specify how all computations were performed and which tools were used.Software and environment. All calculations and simulations were carried out using custom scripts written by the authors in Python (version 3.11). Core numerical operations relied on NumPy, while data handling and aggregation used pandas. Statistical tests (Friedman and post-hoc paired Wilcoxon signed-rank tests with Holm adjustment) were computed with SciPy. All figures were generated with Matplotlib.Randomness and repeatability. For Monte Carlo perturbation experiments, we applied ± 5% multiplicative perturbations to each entry of the raw decision matrix using a uniform distribution. All random draws were generated with a fixed pseudo-random seed (seed = 42) to ensure exact reproducibility of reported frequencies and summary statistics.Computation protocol. For each Monte Carlo run, the perturbed matrix was normalized, method-specific scores were computed (GRAD, TOPSIS, VIKOR, and CoCoSo), and ranks were obtained by sorting scores in descending order (rank 1 = best). Frequencies (e.g., top-1 or top-2 occurrence) and rank variability (mean ± SD) were then aggregated across N  = 10000 runs.Availability. The scripts used to reproduce all tables and figures (including Table 9; Fig. 2) can be shared upon request. All parameter settings used in the experiments are reported in the manuscript to enable independent replication.