|
Jingxing Wang
About me
My research interests span machine learning, optimization, information theory and data-driven decision making. I've worked on compressive sensing algorithms, model explainability and scheduling problems.
I'm a first year Ph.D. student at UW ECE, co-advised by Vasileios Charisopoulos and Maryam Fazel. Previously, I completed my undergraduate at UC Berkeley, where I worked with Kannan Ramchandran and his group at BLISS. I also collaborated with Chiwei Yan from the department of IEOR.
Awards and Honors
Rushmer Electrical Engineering Endowment Fellowship
Publications
 |
A Dynamic Model for Airline Fleeting and Scheduling
SSRN, 2026
Chiwei Yan, Jingxing Wang, Archis Ghate
Paper
I joined Chiwei on the revision of this paper in 2024 and helped with addressing the followings: 1, Constructed two representative toy networks to establish performance guarantees and optimality gaps. Both networks reduce to linear programs involving constants that depend on actual flight demand. I extracted properties from realistic flight-demand scenarios with varying levels of volatility and proved that, in each case, our network is either optimal or has small performance gaps. 2, Built a theoretical benchmark upper bound using Information Relaxation, a framework that allows the use of future demand information with penalties. I implemented this framework by incorporating our existing Lagrangian value into the penalty calculation, resulting in provable and empirically tighter upper bounds than the Lagrangian value alone. 3, Extended prior synthetic experiments to a large-scale setting with 1,000 flights and real data, and built a new testing pipeline that learns the underlying data distribution (through clustering and a moment-matching parameter estimation) for out-of-sample validation. I presented this work at INFORMS 2025. We completed the revision and resubmitted the paper to Operations Research in January 2026.
|
Education
|
