IIT Bombay Research Helps Indian Railways To Streamline Timetables
Indian Railways is set to optimise its scheduling without altering tracks or trains, thanks to an innovative clustering technique developed by researchers at IIT Bombay and railway experts. The method, known as ‘dailyzing,’ groups non-daily trains to improve timetable efficiency, reduce congestion, and potentially make room for additional services.
The research, led by Prof Madhu Belur and Prof Narayan Rangaraj of IIT Bombay, in collaboration with experts from Zonal Railways and the Centre for Railway Information Systems (CRIS), focuses on restructuring how non-daily trains are scheduled. Indian Railways currently operates over 13,150 passenger trains daily, yet many services run inconsistently across the week, causing inefficiencies.
The core idea behind dailyzing is to group similar non-daily trains—those using the same resources and running within a 15-minute window on different days—into a single structured schedule. This prevents scattered scheduling, which can lead to underutilised tracks on some days and congested bottlenecks on others. By mapping these clusters onto a 24-hour schedule, planners can improve train flow and reduce conflicts.
The researchers employed Hierarchical Agglomerative Clustering (HAC), a data-driven technique, to identify patterns in train movements. The model was tested on India’s Golden Quadrilateral and Diagonals (GQD) network, which connects major cities like Delhi, Mumbai, Chennai, and Kolkata. The results showed HAC was the most effective at forming conflict-free clusters, significantly reducing timetable inconsistencies.
One of the major advantages of this system is its ability to accommodate new trains. If a cluster has fewer than seven trains—one for each day of the week—additional services can be slotted into the free days. The model also improves bottleneck management, reducing delays and maximising railway capacity.
Prof Belur noted that the system is already being used in a modified form to enhance timetabling on the GQD network. Future developments could further refine train scheduling by incorporating real-time adjustments, making the network even more seamless and adaptive.
However, the researchers acknowledge that as the model scales up, new challenges will arise. Since long-distance trains often pass through multiple congested sections, a single clustering approach may not be enough. Additionally, different railway zones currently plan their timetables independently, which means better coordination will be necessary to fully realise the benefits of dailyzing.
