Temporal Planning Optimizes Railway Routes Amid Disruptions

Ensuring safety and punctuality in railway operations, especially within heterogeneous multi-gauge networks, presents significant challenges. These systems involve varying train speeds, stopping patterns, and infrastructure compatibility constraints, all of which increase coordination complexity. Single-track systems further intensify these issues as all trains share the same track, necessitating frequent and precise track switching. The unpredictability introduced by stochastic disruptions like blocked tracks, engine failures, or speed slowdowns further complicates operations, often deviating from planned timetables. Existing research primarily focuses on high-level timetabling, frequently overlooking critical operational details such as track switching coordination, which are often left to human operators and can introduce safety risks. To address this gap, this study introduces a novel framework based on temporal planning for dynamic route optimization and comprehensive disruption management in these complex railway environments. The framework models railway operations as a temporal planning problem using PDDL 2.1, explicitly incorporating gauge compatibility constraints and diverse disruption scenarios. It generates conflict-free, timestamped operational plans that detail both optimized schedules and executable action sequences. Evaluated on a benchmark of 200 instances with up to 1,000 track points and 120 trains, the framework, assessed with state-of-the-art temporal planners, effectively produces operational plans for heterogeneous systems, manages multi-gauge constraints, handles disruptions, and significantly reduces reliance on manual decision-making.