Feasibility Study: AI-Based Prediction of Train Delays

Jnction Ltd, a supplier of IT products and services to the rail industry in the UK, is working with AI and rail consultancy Interimconsult Ltd, led by AI expert Dr Paul Corcoran, who has worked in AI for over 20 years, to deliver a feasibility study into better predictions of train delays on the UK rail network.

The joint feasibility study will investigate an innovative approach to improve the efficiency and reliability of train operations through the application of artificial intelligence (AI) to forecast train delays more accurately.

Train delays often result from unforeseen events such as infrastructure issues, weather conditions, and operational disruptions, and are complex to forecast. By developing an AI-based system that can accurately predict potential delays in real-time, operators can proactively respond, minimise the impact on service quality, and improve accuracy of passenger information and hence customer experience.

Reduced delays will also reduce rail costs at a time when train operators and Network Rail are under pressure to save money. Better utilisation of crew and rolling stock, reduced passenger compensation payments, and lower overtime payments could all add up to significant savings.

The project aims to develop a system capable of accurately predicting train delays using advanced AI Deep Learning algorithms, known as Graph Neural Networks. By leveraging real-time data and historical patterns, this AI-powered study, if successfully implemented, could enhance the overall performance of train services, minimise disruptions, and improve passenger satisfaction. The study will assess the technical, economic, and commercial viability of the proposed solution, laying the groundwork for a future transformative application within the rail industry.

This particular approach has not been tried previously in the UK. The UK Rail system is hampered by old legacy IT systems and long established manual processes. The successful application of AI could have a significant modernising impact on all aspect of operations in rail, and this study could encourage greater take up of AI in an industry that has traditionally been slow to adopt new technology.

Following completion of the study the report findings will be published and disseminated to the rail industry through articles in the rail press and social media and at specialist rail conferences.