Infrastructure Investments for Optimal Rail REPLACEment Bus Services

Have you ever had your train replaced by a bus because of unforeseen disruptions or scheduled maintenance? Behind every alternative route, there is a team rerouting and rescheduling with little to no automatic decision support. The REPLACE project aims to improve rail mobility through innovative AI-based methods and optimalization. When scheduled maintenance or unforeseen disruptions occur, the railway infrastructure managers and railway undertakings take steps to ensure that all passengers comfortably reach their destinations. Trains are rerouted and rescheduled, and buses is deployed to bridge passengers between disconnected locations. This requires snap judgement by experts with little to no decision support across different actors, leading to inefficient solutions. In turn, this drastically diminishes the attractiveness of the rail transport system. As an attractive public transport service, also in times of disruption, impacts the overall satisfaction of citizens on and off travel, it is essential to avoid congestion and provide quality service. Having a properly scaled and located replacement services requires continual, significant investments when determining which infrastructure investments are the most cost-effective for improving rail replacement services. Making use of advanced mathematical optimization techniques, REPLACE aims to promote a more efficient replacement service by complementing practical and professional competence in areas such as infrastructure planning, disruption management, and timetabling. REPLACE will provide a comprehensive model for infrastructure investment decisions and rail replacement bus services, along with an optimization-based solution algorithm. The model is intended to be integrated in planning tools, allowing infrastructure managers to better plan the necessary enhancements needed for providing passengers with an effective, sustainable, and future-oriented replacement service.

Unforeseen disruptions or scheduled maintenance activities often result in track blockages along railway lines. When such blockages occur, railway infrastructure managers (e.g., Bane NOR) and railway undertakings (e.g., Vy and Flytoget) implement recovery measures to ensure all passengers can comfortably reach their destinations. Initially, trains are rerouted and rescheduled, and a temporary timetable is made available to the public (Train Timetabling Problem, TTP). Then, a fleet of buses is deployed to bridge passengers between disconnected locations, requiring careful planning of dimension, routing, and scheduling (Bus Bridging Problem, BBP). This replacement service necessitates dedicated infrastructure in both stations and adjacent roads, such as additional platforms, parking facilities, etc. The REPLACE project aims to develop optimization-based AI methods to compute optimal infrastructure investments (IIP) for train replacement services. This requires individual methods for TTP, BBP and IIP, finally integrated in a unified model, resulting in large-scale, optimization problem instances. Solving such complex instances requires new mathematical approaches. Currently, the three problems are managed manually and separately, by distinct units and even in different companies. Furthermore, existing practices do not leverage optimization or other AI methods to identify high-quality solutions. REPLACE seeks to advance the mathematical knowledge necessary to tackle these challenging problems. By collaborating with key industry stakeholders, REPLACE will develop models that address all relevant aspects of underlying real-life problems. By involving experts from University of L'Aquila and prestigious advisors on the scientific board, REPLACE will develop state-of-the-art optimization methods crucial for handling hard instances. The engagement of numerous stakeholders will ensure rapid dissemination of results, within both academic and industrial communities.