Designing railway transit network with first-mile access via ride-sharing autonomous vehicle service

The rail transit system is an essential component of a sustainable transport system, but its accessibility remains a barrier owing to inadequate transit network coverage. Ride-sharing autonomous vehicle (RAV) service is anticipated to be a potentially viable option to improve the economics of rail-based transportation systems (RTS) by substantially lowering the upfront and ongoing costs. In this study, we model and evaluate the impact of RAVs on integrated RTS + first-mile autonomous vehicle design where the commuters' waiting time at railway stations is capped to ensure the level of quality service for rail. The proposed framework optimizes the RAV fleet size while allowing their relocations to tackle the vehicle imbalance problem in the network to assist operators in optimally and effectively running their services. To this end, this study intends to help the relevant Government agency in their planning, where the RTS extension design and RAV operations are planned concurrently to maximize the coverage area with improved accessibility. The modeling framework is solved using mixed-integer linear programming. We applied our proposed framework to the Doncaster railway extension in Melbourne, Australia, to demonstrate the applicability of our modeling framework. Our findings show that increasing rideshare capacity enhanced the operational performance of the integrated services, increased the percentage of AVs performing ride-sharing, and resulted in a shorter RTS alignment and that the RAV-only service is always a costlier option. Finally, the RTS alignment is found to be robust against the variations in capped commuters' waiting time, the disutility of commuter delays by the RAV service, and changes in headway between train trips.