Real-Time Train Rescheduling Using ML for Delay Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In high-density urban railway lines, delays in boarding and alighting at dwelling stations cause chain delays due to inadequate recognition of preceding train delays, leading to inefficient train operation and schedule deterioration.
Innovation Solution
A real-time rescheduling method using a machine learning model to predict passenger demand and required dwell duration by collecting passenger and operation information, estimating boarding and alighting quantities, and updating the train schedule to secure additional dwell time at departure stations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If the following train operates according to the existing schedule without recognizing the delay of the host train in advance, then the train operation follows the original timetable, but the following train must wait while stopping in the operation section between stations, causing repeated acceleration and deceleration which deteriorates train service quality and schedule speed
Solution Approach 1:
The system performs preliminary actions by predicting the host train's delay duration using machine learning models before the following train reaches the dwelling station. The prediction result is used to proactively adjust the following train's schedule, including modifying dwell duration and departure time, so that the following train can smoothly enter and depart the dwelling station without unnecessary waiting and repeated acceleration-deceleration cycles.
2Reliability
If additional dwell duration is secured at the departure station to solve chain delays, then the host train can accommodate more passengers boarding and alighting, but the overall train operation time increases
Solution Approach 1:
The system dynamically adjusts the dwell duration of the following train based on real-time predictions of the host train's delay. Instead of adding fixed dwell time to all trains, the machine learning model predicts the actual delay duration and modifies the following train's schedule accordingly. This dynamic adjustment ensures that additional dwell time is only added when necessary to resolve chain delays, minimizing the impact on overall train operation time while maintaining reliability.
3Productivity
If machine learning models are used to predict passenger demand and dwell duration in real time, then the train schedule can be optimized to reduce chain delays, but the system complexity and computational requirements increase
Solution Approach 1:
The system introduces machine learning models as intermediary components between the host train's actual operation and the following train's schedule adjustment. These models act as mediators that process passenger information, train operation data, and historical patterns to predict delay duration and required dwell time. By using these intelligent intermediaries, the system achieves automated, data-driven schedule optimization without requiring complex manual coordination or centralized control systems.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
The present disclosure in some embodiments relates to a method and apparatus for real-time rescheduling for resolving a delay of a train operation. A real-time rescheduling method for resolving an operation delay of a train, the method comprising: collecting passenger information, and operation information of the train; generating a first data set including a preceding boarding and alighting quantity; estimating a host boarding and alighting quantity; estimating a required dwell duration of the train at the scheduled dwelling station; and generating a schedule updating an operation schedule of the train.