Development of a subway operation incident delay model using accelerated failure time approaches

Area of Science:

• Transportation Engineering
• Reliability Engineering
• Operations Research

Background:

• Subway operational incidents significantly impact urban transit reliability and efficiency.
• Accurate modeling of incident-related delays is crucial for service improvement and passenger satisfaction.

Purpose of the Study:

• To develop and validate a robust model for predicting subway operational incident delays.
• To identify key factors contributing to subway incident delay duration.
• To propose mitigation strategies for reducing subway delays.

Main Methods:

• Utilized the parametric accelerated time failure (AFT) approach.
• Developed and compared six parametric AFT models (log-logistic, lognormal, Weibull) with fixed and random parameters.
• Incorporated a Weibull model with gamma heterogeneity for performance comparison.
• Employed goodness-of-fit tests to determine the most suitable model.

Main Results:

• The log-logistic AFT model with random parameters demonstrated the best fit for subway incident delay data.
• Identified power cable failure, signal cable failure, turnout communication disruption, and casualty crashes as significant contributors to longer delays.
• Vehicle failure was found to have the least impact on delay duration.

Conclusions:

• The log-logistic AFT model with random parameters provides a stable and reliable method for estimating subway incident delays.
• Findings suggest implementing wireless communication technologies to mitigate delays.
• Understanding causal factors enables targeted interventions to enhance subway operational efficiency.