Integrated fuzzy–rough framework for decision making under multiple uncertainties: empirical evidence from two FR-AHP case studies
Аннотация
Multi-Criteria Decision Making (MCDM) under real-world conditions involves simultaneous uncertainties(linguistic vagueness, incomplete knowledge, and inter-expert disagreement) yet most existing methods address only one of these dimensions at a time. This study proposes and benchmarks a four-step Fuzzy-Rough AHP (FR-AHP) framework that jointly handles all three uncertainty types by embedding rough set neighborhoods inside the classical Fuzzy AHP (F-AHP) pipeline. The framework is evaluated against F-AHP on two structurally contrasting published case studies: a traffic accessibility analysis with a highly divergent expert panel, and a service provider selection problem with a tighter consensus group. Results show that FR-AHP meaningfully alters priority weights and, under high expert disagreement, reverses rank order (reducing the maximum weight spread by $$33\%$$ ) while preserving the original ranking when consensus is strong. Monte Carlo sensitivity tests confirm that FR-AHP consistently produces more conservative and stable priority estimates across both scenarios, without incurring additional computational cost relative to F-AHP. Overall, FR-AHP emerges as a robust and practical extension of F-AHP for decision contexts where expert panel heterogeneity is a concern, while remaining equivalent to F-AHP when judgments are already well-aligned.