Design and implementation of a network-aware automated bus scheduling system for optimizing operational efficiency and financial performance

Abstract

This work develops decision support models for bus schedule optimization to improve the existing public transport (PT) systems schedule. It particularly addresses the issue of overlapping trips, where multiple buses are scheduled to travel the same route simultaneously. This occurs due to the absence of network-aware scheduling, a significant problem of operational inefficiency in PT systems. To remove such overlapping trips, our approach presents mixed-integer linear programming (MILP) models and innovative solution algorithms for efficient rescheduling and redeployment. To prove the efficacy of the developed model, we use data from Bengaluru Metropolitan Transport Corporation (BMTC), India’s largest urban bus fleet operator, and optimize the schedule with no additional capital expenditure and minimal changes to the existing schedule. The work is carried out in close consultation with BMTC and led to the development of a computer-based decision support bus-schedule optimization toolkit (B-SOT). The B-SOT automates the schedule development process using simple CSV files as input and output, making it easy to use for officials at all levels. The toolkit is sufficiently comprehensive to be applied to other cities with contextual adjustments. Overall, the decision support models and the computer-based toolkit developed are novel, modular, and quite useful for PT decision-making to optimize bus schedules while freeing up some buses for deployment elsewhere.