Service networks with open routing by self-interested customers have drawn attention in the theoretical literature (Arlotto et al. 2019, Parlakturk and Kumar 2004). However, these networks, which range from shopping centers to amusement parks, remain challenging to explore empirically. Customers’ physical-movement trajectories simultaneously reflect their reactions to congestion, demand for complementary groups of stations, and dynamic choices about the order of station visits. As such, large-scale trajectory datasets offer tremendous opportunities to understand customer motivations and behaviors but are complex to analyze. We develop structural empirical methods to recover customer demand preferences and congestion sensitivities from diverse trajectory patterns using machine learning. Specifically, we employ adversarial neural networks to handle the high-dimensional space of (combinatorially many) trajectory types. Key innovations collapse the dynamics of customer trajectory choices into static trajectory market shares and derive theoretically efficient incentive-compatibility bounds on customers’ preferences.
Recommended citation: Moon, K. (2021). "Strategic Choices and Routing Within Service Networks: Modeling and Estimation Using Machine Learning" Working paper.