Seokwoo Kim
Seokwoo Kim
Hi, my name is Seokwoo Kim and I am a Ph.D. candidate in the department of industrial and management engineering at Pohang University of Science and Technology (POSTECH), under the supervision of Prof. Dong Gu Choi. I am pursuing an academic career upon completion of my doctoral studies.
My research interests lie in the field of Operations Research (OR), with a focus on equilibrium problems under uncertainty, a synthesis of two research areas---optimization under uncertainty and game theory. During my PhD journey, I have been working on robust/bilevel optimization and cooperative game theory in energy markets and service systems.
Currently, I am particularly interested in integrating these methodologies to address the following research directions:
Distributionally robust bilevel optimization
Design of stochastic systems and equilibria through robust/bilevel optimization lens, e.g., markov decision process design
Computational methods for cooperative games, e.g., optimal cost allocation problem, stabilizing grand coalition
Energy markets applications
Seokwoo Kim, Dong Gu Choi, “A Sample Robust Optimal Bidding Model for a Virtual Power Plant”, European Journal of Operational Research, 2024, 316(3), 1101-1113 [link]
Seokwoo Kim, Seungmin Baik, Dong Gu Choi, “A Bi-level Service Capacity Planning Model with Interdependent MDP Followers” [preprint]
Seokwoo Kim, Yuin Moon, Mohammad Reza Hesamzadeh, Jan Kronqvist, Dong Gu Choi, “Profit Allocation in Energy Community Games under Non-convex Operations: A Duality-based Perspective” [preprint]
Seokwoo Kim, Dong Gu Choi, “A Distributionally Robust Model for Stochastic Bilevel Programs under Ambiguous Beliefs”
We study a novel class of distributionally robust bilevel programs in which the leader and the follower hold distinct probability distributions (beliefs) about common exogenous uncertainty parameters. The follower solves a two-stage stochastic linear program under his/her private belief, which is unobservable to the leader. The resulting belief ambiguity makes the uncertain rational reaction set in the upper level problem, producing \textit{decision uncertainty} of the follower. We consider an ambiguity-averse leader who hedges against both data and decision uncertainty. Under mild assumptions, we show that the robust counterpart is well-defined and that the problem can be $\Sigma_2^p$-hard. We then provide a minimax reformulation amenable to Benders decomposition, and validate the framework on network interdiction applications.
Yuin Moon, Seokwoo Kim, Dong Gu Choi, "Courier Scheduling in Last-mile Delivery Considering Show-up Dynamics", under review
Youngdoo Choi, Sanghyeon Bae, Seokwoo Kim, Dong Gu Choi, "Nested Benders' Decomposition Approach for Bidding Problem of VPPs in Day-Ahead and Real-Time Markets with Negative Prices", under review [preprint]
Invited reviewer for International Journal of Production Research
Here's my google scholar link and LinkedIn
You can reach me at: seokwookim@postech.ac.kr
*header image: courtesy of Whanki Kim