Optimizing Sustainable Transit Bus Networks in Smart Cities
This is an online defence. Click HERE to watch the livestream.
Urban mobility has been facing several challenges in the recent years due to the increasing populations and private vehicles ownership, which led to several negative environmental and social impacts in big cities. Thus, shifting towards the more sustainable and cleaner modes of urban mobility has become an urgent need. In this dissertation, we investigate how public transit bus networks can contribute to solving this problem and enhance urban mobility systems. Namely, we study how to make transit bus networks more sustainable to the environment while offering a more convenient and better service level to passengers.
First, we study the problem of making the transition from conventional to electric transit bus networks. We formulate and implement an efficient mixed-integer linear programming optimization algorithm to optimize the charging schedules in the electrified networks. We show the significant benefits of optimizing the charging schedules from the perspective of the public transit network and electricity grid operators. Then, we develop a discrete-event simulation to assess various scenarios and evaluate the electric transit network's performance under uncertainty. To mitigate the impact of operational uncertainty, we propose a real-time decision support system that is based on online optimization. Our results show the essential need and importance of such a system, especially when the electric transit network is susceptible to high levels of operational uncertainty. Finally, we study and quantify the positive impact of coordinating electric transit bus networks' charging schedules with renewable energy generation profiles.
Second, we investigate how public transit bus networks can offer more convenient service level to passengers. We propose dynamic transit bus networks, which plan routes and schedules according to the actual observed demand and passengers' preferences. We develop a heuristic algorithm to solve the complex underlying mathematical optimization problem efficiently. Our results show an essential improvement in the service level provided to passengers and ridership of transit bus networks in the dynamic system compared to the conventional static one. We also show how passengers' flexibility can lead to significant improvement in the dynamic system's performance.