Seminar by Dr. Xiaoqi Tan (University of Toronto)

Title: Sequential Decision Analytics, Networks, and Energy

Abstract: Many real-world problems in networks and energy (e.g., network routing and renewable energy integration) involve making sequential decisions, over a certain horizon, in the presence of uncertainty. It is also not uncommon that in some situations such sequential decisions must be made with zero knowledge of the future. This leads to various types of sequential decision-making (SDM) problems with either rigid uncertainty models (e.g., stochastic optimization), zero uncertainty models (e.g., online optimization), or the mixed in between (e.g., robust optimization).

In this talk, I will first introduce a mathematical framework for modeling general SDM problems with different uncertainty models and performance metrics. Based on this framework, I will present my research regarding two SDM problems, namely, sequential multi-agent resource allocation and stochastic shortest path. For both problems, I will first motivate the research by real-world applications in networks and energy  (e.g., computer networks, data center networks, transportation networks, smart grid, and smart cities), and then introduce novel algorithm design with rigorous performance guarantee. I will conclude the talk with an outlook of future directions towards combining model-driven and data-driven approaches to build a unified framework of sequential decision analytics.

Bio: Xiaoqi Tan is currently a Post-Doctoral Fellow with The Edward S. Rogers Sr. Department of Electrical and Computer Engineering, University of Toronto. Prior to the current position, he received his Ph.D. degree in electronic and computer engineering from the Hong Kong University of Science and Technology in 2018. During his Ph.D., he was a Visiting Research Fellow of Applied Mathematics at Harvard University. His research interests include sequential decision-making and algorithmic mechanism design, with a focus on applications to computer systems and networks, energy systems, and smart cities.