Bus Operational Performance Failure Evaluation Method Based on Enhanced Entropy Weight–TOPSIS with Grey Relational Analysis

Abstract

Current evaluations of bus systems predominantly focus on macrolevel network performance, lacking granular tools to diagnose operational failures at the route-specific level. To address this gap, this study proposes a novel evaluation framework integrating grey relational analysis (GRA) with an entropy-weighted technique for order preference by similarity to ideal solution (TOPSIS), enabling microscopic assessment of bus operational anomalies. The model leverages multisource data (global positioning system, Integrated Circuit (IC) cards, and geographic information systems) to construct a quantifiable indicator system spanning bus operations, routes, and stops. The entropy weighting mechanism ensures objective indicator allocation, and the GRA enhances robustness against real-world data noise. Validation using empirical case studies demonstrates that the hybrid GRA–TOPSIS approach outperformed stand-alone methods, achieving an 18.1% higher median score than data envelopment analysis and reducing outlier counts by 42%. The model’s stability was confirmed by a 68% smaller interquartile range compared with conventional TOPSIS, and Kolmogorov–Smirnov testing validated score distribution normality, enabling reliable percentile-based classification.