Measuring high-speed train delay severity: Static and dynamic analysis.

Area of Science:

• Operations Research
• Transportation Science
• Railway Engineering

Background:

• Effective management of train delays is crucial for operational efficiency in high-speed railway systems.
• Existing methods may not adequately capture the complexity and cascading effects of train delays.
• Accurate prediction and classification of delay severity are needed to optimize mitigation strategies.

Approach:

• Developed static and dynamic models using real-world high-speed railway delay data.
• Utilized K-means clustering for classifying delays into four risk-based levels based on key indicators.
• Employed Markov chains for sequential dynamic analysis, considering China's railway context and peak travel periods.

Key Points:

• Key indicators include initial delay duration, station impacts, and cascading effects.
• K-means clustering provided an optimized, four-level categorization of train delays.
• Dynamic analysis using Markov chains captured temporal variations, especially during the Spring Festival travel rush.

Conclusions:

• The combined static and dynamic approach offers a robust method for categorizing and managing train delays.
• Enhanced classification improves the ability to identify and predict delay extents, leading to more efficient mitigation.
• This research contributes to bolstering railway operational efficiency and resilience in complex delay scenarios.