Standardized feature extraction from pairwise conflicts applied to the train rescheduling problem

TL;DR

This paper introduces a standardized feature extraction method based on pairwise conflicts for train rescheduling, integrated into a reinforcement learning framework, and demonstrates its effectiveness through empirical testing.

Contribution

It presents a novel feature extraction approach for train rescheduling that enhances reinforcement learning performance by focusing on conflict-based observations.

Findings

The feature space yields meaningful observations for scheduling.

The reinforcement learning model learns effective policies.

Empirical results validate the approach's potential.

Abstract

We propose a train rescheduling algorithm which applies a standardized feature selection based on pairwise conflicts in order to serve as input for the reinforcement learning framework. We implement an analytical method which identifies and optimally solves every conflict arising between two trains, then we design a corresponding observation space which features the most relevant information considering these conflicts. The data obtained this way then translates to actions in the context of the reinforcement learning framework. We test our preliminary model using the evaluation metrics of the Flatland Challenge. The empirical results indicate that the suggested feature space provides meaningful observations, from which a sensible scheduling policy can be learned.