Prof. Truong Wins NSF Career Award
|Professor Van Anh Truong|
The IEOR Department congratulates Professor Van Anh Truong for receiving the prestigious National Science Foundation (NSF) Career Award for her research on "Optimization Methods to Support Real-Time Personalized Consumer Transactions".
Abstract: The goal of this Faculty Early Career Development (CAREER) Program research project is to investigate and solve in real-time important emerging scheduling, product recommendation and product framing problems involving dynamic consumer choice. Such problems abound in e-service and e-commerce ecosystems that serve millions of users in highly personalized interactions. For example, in e-service, websites must match large inventories of service slots dynamically to arriving consumers in a way that accommodates consumer preferences for service type, location, schedule, payment options, and other dimensions of service. In e-commerce, retailers must curate and frame information about potentially hundreds of products for each consumer in order to derive the most benefit from their inventory. These problems are challenging because they depend upon external information on both inventory and consumer preferences that unfolds over time, the problem size is large, and the number of potential solutions can overwhelm conventional solution methods. This research will address methods to optimize real-time decisions in interactions like these, enabling the fast-growing e-service and e-commerce economy to fully leverage data and technology to produce significant gains in both the expected benefit to the consumer and the productivity of the service provider. The educational plan will (1) create content for a pilot STEM module at Frederick Douglas Academy II, a secondary school in Harlem, NY; (2) mentor and advise the Columbia Charter of "Girls Who Code"; and (3) integrate industry experience into undergraduate and graduate education.
This award will support fundamental research on a methodological framework and technical machinery for investigating and solving online adaptive stochastic optimization problems, including dynamic bipartite matching problems, over combinatorially large structured decision spaces. Current methods are only able to solve certain limited one-shot versions of these problems that do not adapt to new information over time. The research questions that will be addressed include: 1) Can problem structure be exploited to decompose the original adaptive optimization problem into sub-problems with smaller decision spaces; 2) Can these sub-problems be adaptively solved over time while maintaining bounds on performance; 3) Can one assemble the solutions to these sub-problems into a feasible solution to the original problem; and 4) What is the performance, both theoretical and practical, of such an approach? Guided by these research questions, this CAREER award will systematically advance the state-of-the-art in online stochastic optimization and discover new fast and robust algorithms for problems in operations research and management science.