Stability and robustness of decision rankings across objective weighting methods in multi-criteria decision-making

Itishom Al Khoiry, Dhea Rizky Amelia

Abstract


Objective weighting methods derive criterion importance directly from data, but a method producing consistent rankings across aggregation models may not be robust to weight changes. This study jointly evaluates cross-model ranking stability and perturbation robustness in multi-criteria decision-making (MCDM) across five objective weighting methods: entropy, criteria importance through intercriteria correlation (CRITIC), method based on the removal effects of criteria (MEREC), principal component analysis (PCA), and standard deviation (SD). Two datasets were used: a renewable energy case (five alternatives and ten criteria) and a supplier-selection case (ten alternatives, eleven criteria) for external validation. Each weight vector was applied to five MCDM aggregation models: visekriterijumska optimizacija i kompromisno resenje (VIKOR), complex proportional assessment (COPRAS), multi-objective optimization on the basis of ratio analysis (MOORA), evaluation based on distance from average solution (EDAS), and technique for order preference by similarity to ideal solution (TOPSIS). Stability was measured using rank correlation, rank shift, and rank reversal; robustness through criterion-weight perturbation. The two diagnostics are positively associated but not interchangeable. Entropy and MEREC achieved the highest stability and robustness in both cases, whereas PCA was stable across models yet highly sensitive to weight perturbation, indicating that reliability cannot be judged from cross-model stability alone.

Keywords


Decision support systems; Multi-criteria decision-making; Objective weighting methods; Ranking stability; Robustness analysis; Weight perturbation

Full Text:

PDF


DOI: https://doi.org/10.11591/eei.v15i6.13862

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

Bulletin of EEI Statistics

Bulletin of Electrical Engineering and Informatics (BEEI)
ISSN: 2089-3191 , e-ISSN: 2302-9285
This journal is published by the Institute of Advanced Engineering and Science (IAES) .