Data driven models for congested urban mobility networks

Abstract

Building efficient and sustainable urban mobility systems have both societal and environmental implications. The urban population is growing at a fast rate for the past couple of decades (55% of the world population lives in urban areas as of 2018). Fast urbanization is causing additional challenges to urban mobility systems since increases in demand are out-pacing capacity improvements. This causes congestion in urban mobility systems which is reflected as traffic in road networks and crowding in public transit systems. Further-more, the widespread use of internet technologies and smartphone apps led to the advent of new mobility modes that had disruptive impacts on the urban mobility system. Although emerging technologies brought additional challenges, they also brought new opportunities as unprecedented amounts of data is available for research and development. The primary motivation of this dissertation is making use of these new data sources by developing data-driven models to help the new challenges faced by urban mobility systems. We propose methodologies for both traditional mobility modes (public transit) and emerging modes(shared mobility systems) that aim at increasing sustainability. We consider public transit crowding with a specific focus on urban heavy rail systems. Crowding at stations and on trains is a concern due to its impact on safety, service quality, and operating efficiency. The safety aspect of crowding is further emphasized during the COVID-19 pandemic. Denied boarding is considered a key measure of the impact of crowding on customers. Denied boarding probabilities represent the fraction of passengers experiencing denied boarding on platforms different times. We deal with the problem of estimating denied boarding probabilities for a given platform using automated data sources (smart card and train movement data). We also develop methods towards real-time information provision which would enable informed travel decisions and encourage behavior change for passengers. The thesis also contributes to the overall urban mobility crowding and sustainability theme from the point of view of the emerging mobility services. Shared mobility on demand (MoD) operations are receiving increased attention as many high-volume ride-hailing companies are offering shared services (e.g. UberPool, LyftLine) at an increasing rate. Also, the advent of autonomous vehicles (AVs) promises further operational opportunities to benefit from these developments as AVs enable a centrally operated and fully connected fleet. We propose a novel operational formulation to perform shared ride-matching which aims to reduce system-wide vehicle miles traveled (VMT).--Author's abstract