Thesis (Ph. D.)--University of Pennsylvania, 1966. / Includes index. eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references (leaves xx-xlii).

Thesis (Ph. D.)--University of Hong Kong, 2010. / Includes bibliographical references (p. 274-305). Also available in print.

Motivated by the imbalance between demand (i.e., passenger requests) and supply (i.e., available vehicles) in the ride-hailing market and severe traffic congestion faced by modern cities, this dissertation aims to improve the efficiency of the sharing economy by building an agent-based methodological framework for optimal decision-making of distributed agents (e.g., autonomous shared vehicles), including passenger-seeking and route choice. Furthermore, noticing that city planners can impact the behavior of agents via some operational measures such as congestion pricing and signal control, this dissertation investigates the overall bilevel problem that involves the decision-making process of both distributed agents (i.e., the lower level) and central city planners (i.e., the upper level). First of all, for the task of passenger-seeking, this dissertation proposes a model-based Markov decision process (MDP) approach to incorporate distinct features of e-hailing drivers. The modified MDP approach is found to outperform the baseline (i.e., the local hotspot strategy) in terms of both the rate of return and the utilization rate. Although the modified MDP approach is set up in the single-agent setting, we extend its applicability to multi-agent scenarios by a dynamic adjustment strategy of the order matching probability which is able to partially capture the competition among agents. Furthermore, noticing that the reward function is commonly assumed as some prior knowledge, this dissertation unveils the underlying reward function of the overall e-hailing driver population (i.e., 44,000 Didi drivers in Beijing) through an inverse reinforcement learning method, which paves the way for future research on discovering the underlying reward mechanism in a complex and dynamic ride-hailing market. To better incorporate the competition among agents, this dissertation develops a model-free mean-field multi-agent actor-critic algorithm for multi-driver passenger-seeking. A bilevel optimization model is then formulated with the upper level as a reward design mechanism and the lower level as a multi-agent system. We use the developed mean field multi-agent actor-critic algorithm to solve for the optimal passenger-seeking policies of distributed agents in the lower level and Bayesian optimization to solve for the optimal control of upper-level city planners. The bilevel optimization model is applied to a real-world large-scale multi-class taxi driver repositioning task with congestion pricing as the upper-level control. It is disclosed that the derived optimal toll charge can efficiently improve the objective of city planners. With agents knowingwhere to go (i.e., passenger-seeking), this dissertation then applies the bilevel optimization model to the research question of how to get there (i.e., route choice). Different from the task of passenger-seeking where the action space is always fixed-dimensional, the problem of variable action set emerges in the task of route choice. Therefore, a flow-dependent deep Q-learning algorithm is proposed to efficiently derive the optimal policies for multi-commodity multi-class agents. We demonstrate the effect of two countermeasures, namely tolling and signal control, on the behavior of travelers and show that the systematic objective of city planners can be optimized by a proper control.

Hundreds of billions of dollars are spent every year on developing new urban transport infrastructure (Hannon et al., 2020). At the same time, we know that the transportation network is a crucial determinant of the spatial organization of economic activity within a city. For this reason, it is important to understand the effects of investing in the transport infrastructure of a city on the spatial distribution of the residents of the city, as well as the welfare implications of these investments. In this dissertation, I will explore these questions in the context of a large transport infrastructure investment in the city of Buenos Aires, Argentina. I will then use this setting to study the political implications of these investments in transport infrastructure on the incumbent party that is in office when the investments are made. In the first chapter I study how improvements in the urban transport infrastructure affect the spatial sorting of residents with different levels of income and education within a city. In particular, I study the effects of the construction of a bus rapid transit system (BRT) on the spatial reorganization of residents within the city of Buenos Aires, Argentina. To do so, I leverage an individual level panel data set of more than two million residents with which I can describe intra-city migration patterns at a very fine spatial scale. With these data, I employ an instrumental variables identification strategy to study how the increases in commuter market access produced by the new transportation network led to changes in the spatial sorting of residents in the city. In this chapter, I find evidence that the construction of the BRT increased the spatial segregation between high and low-skilled residents within the city. In the second chapter of the dissertation, I quantify the welfare effects of improving the urban transit infrastructure of a city once we take into account the patterns of spatial sorting found in Chapter 1. To do so, I develop a dynamic quantitative spatial equilibrium model of a city with heterogeneous workers. I use this quantitative framework to quantify the welfare effects of the BRT system built in Buenos Aires. This framework, allows me to measure the average welfare gains for residents that were living near the BRT lines before these were built. I find that welfare gains were very similar between high- and low-skilled workers living in the same locations before the BRT system was built, but that gains were very different within skill levels across locations. Residents living near a BRT line in neighborhoods with the lowest share of high-skilled residents saw welfare gains close to 1% on average, while residents living near a BRT line in neighborhoods with the highest high-skilled share saw welfare gains around 0.5% on average. Finally, in the third chapter I study the political consequences of public investment in transport infrastructure for the incumbent party that is in office when the investment is made. In particular, I use the same BRT construction in the city of Buenos Aires to study its effects on the incumbent party’s vote share in subsequent elections after the BRT lines were built. I use a staggered difference-in-difference estimation approach in order to capture the treatment effects relative to the time that each BRT line was either built or announced. I find that the incumbent party decreased its vote share in the districts closer to an opening BRT line in the election prior to the opening, but increased its vote share in the first election following the opening of a BRT line. However, when defining the treatment timing based on the year in which each line was announced, the effects of the BRT line on the incumbent party’s vote share appear to be negative and decreasing with the number of year from the announcement.