CAREER: Two-Stage Competition on Spatio-Temporal Location and Pricing

Dartmouth College, Hanover NH

Investigators

Abstract

This Faculty Early Career Development Program (CAREER) project exploits the opportunities brought by the emerging technologies and combines knowledge from operations research (OR), economics and computer science to advance science in resource allocation and promote national prosperity and welfare. The project will establish the mathematical foundations for resource allocation, coupled with computational algorithms and empirical validation methods, for prescriptive and predictive analysis in determining facility locations and associated pricing. Prescriptive analysis helps competing entities in optimizing their scheduling and facility location decisions. Predictive analysis helps central authorities (e.g., government and regulators) in optimizing system efficiency and equity through infrastructure investments, incentives creation, and regulations. Under this project's integrated vision for research and education, the research program directly informs the educational program. The educational program includes the development and dissemination of an active learning module for high-school mathematics focused on underrepresented minority students; an undergraduate research internship program for underprepared prospective engineering students; and undergraduate and graduate course development in OR and game theory. These educational activities contribute toward strengthening and diversifying the pipeline of researchers and practitioners skilled in OR and analytics, in turn fostering the long-term growth of the very disciplines which the research program builds upon. Educational and outreach activities also focus on highlighting the economic, social and political complexities associated with planning and operations of civil infrastructure systems. Scheduling and facility location problems (together called as spatio-temporal location problems) share similar underlying mathematical structures that allow researchers to analyze them jointly. However, under dynamic pricing, spatio-temporal location decisions become interdependent with pricing decisions which are made at a later stage through complex interactions between competing service providers and their customers. This necessitates the use of two-stage game theoretic models. This project develops a comprehensive framework that relies on combining and mutually reinforcing mathematical, computational and empirical insights. The key innovations are the tight interplay between mathematical, computational and empirical ideas, a holistic approach that enables the optimal tradeoff between modeling precision and solution tractability, and a comprehensive testbed comprising large-scale real-world networks from aviation, railways, transit, hotels and gasoline distribution. The project not only makes deep contributions to the theoretical and applied research in competitive transportation scheduling, but also triggers further exploration enabling similar future contributions to the competitive facility location theory and applications. Additionally, this general approach could be usefully employed in other, otherwise intractable, applications of multi-stage game theory. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.