MIT Operations Research Center
77 Massachusetts Ave.
Cambridge, MA 02139
zikai@mit.edu [✉️]
About Me
I am a fifth-year Ph.D. student at the MITOperations Research Center, advised by Prof. Robert Freund.
My current research interests are mainly in optimization, with twin interests in theoretical foundations and computational practice.
My research up to now has involved enhancing the scalability of optimization algorithms for solving large-scale linear and convex optimization -- both generically as well as on specific applications. I am delighted that some of my methods have already been implemented in the state-of-art commercial solvers (COPT in particular). My overarching aspiration is to extend the capability of modern optimization methods to effectively solve problems arising in operations research, data science, and machine learning/AI.
Research Assistant, MITSloan School of Management, 2020 - present Supervisor: Robert M. Freund Modern optimization methods for huge-scale operations research and machine learning problems.
Winter 2024: 15.S60 Computing in Optimization and Statistics - Instructor for the advanced optimization sessions of a computing course during MIT January Independent Activities Period. The duties are designing courses, creating teaching materials, and giving lectures.
-
Fall 2022 and Fall 2023: 15.081 Introduction to Mathematical Programming (Teaching score: 6.8/7 ) - Head TA for the MIT’s doctoral introductory optimization course (for ORC and other MIT PhD programs)
- ⭐ I was voted as the "Best TA" in the ORC Superlatives voting (photo)
Spring 2022: 15.071 The Analytics Edge (Teaching score: 6.4/7 ) - TA for MBA course in business analytics at Sloan School of Management.
Online Course in Data Science and Business Analytics - Designed course materials of a new online course in data science and business analytics.
- Supervisor: Robert Freund
Mentoring
Qiushi Han, University of Illinois Urbana-Champaign (UIUC) (Former SUFE undergraduate who transferred to UIUC and who asked me to continue to supervise his work)
First-order methods for identifying optimal basic feasible solutions of linear programs, with applications to first-order integer programming
Chenghan Xie, Stanford University (Incoming PhD student at Stanford who has asked me to collaborate with him on research)
Computational guarantees and lower bound examples of PDHG for discounted Markov decision process (MDP)
Talks
Level-Set Geometry and Improving the Performance of PDHG for Linear Optimization - OM Seminar, MIT, Cambridge, MA, September 2024
- Cornell ORIE Young Researchers Workshop, Ithaca, NY, October 2024
- INFORMS Annual Meeting, Seattle, WA, October 2024
- INFORMS Annual Meeting (Nicholson Student Paper Competition), Seattle, WA, October 2024
Theoretical Complexity of PDHG for Linear Programs with Unique Optima - Conference in honor of Yinyu Ye’s retirement, Shanghai, China, July 2024
- International Symposium on Mathematical Programming (ISMP), Montréal, Canada, July 2024
- Modeling and Optimization: Theory and Applications (MOPTA), Lehigh University, PA, August 2024
Improving the Geometry of (Conic) Linear Optimization Problems for the Primal-Dual Hybrid Gradient Method (PDHG) - Workshop on Modern Continuous Optimization, Cambridge, MA, August 2023
- ORC Student Seminar, MIT, Cambridge, MA, October 2023
- INFORMS Annual Meeting, Phoenix, AZ, October 2023
- Invited formal talk at Shanghai University of Finance and Economics, Shanghai, China, September 2023
- LIDS Conference, MIT, Cambridge, MA, February, 2024
- INFORMS Optimization Society Conference, Houston, TX, March 2024
- Modeling and Optimization: Theory and Applications (MOPTA), Lehigh University, PA, August 2024
Geometric Condition Measures for the Primal-Dual Hybrid Gradient Method for Linear Programming - SIAM Conference on Optimization (SIOPT), Seattle, WA, June 2023
- ORC Student Seminar, MIT, Cambridge, MA, March 2023
Using Taylor-Approximated Gradients to Improve the Frank-Wolfe Method for Empirical Risk Minimization
- International Conference on Continuous Optimization (ICCOPT), Lehigh University, Bethlehem, PA, July 2022
- ORC Student Seminar, MIT, Cambridge, MA, October 2022
- INFORMS Annual Meeting, Indianapolis, IN, October 2022
From an Interior Point to A Corner Point: Smart Crossover
- INFORMS Annual Meeting, Indianapolis, IN, October 2022
Interior-Point Methods Strike Back: Solving the Wasserstein Barycenter Problem
- INFORMS Annual Meeting, Seattle, WA, October 2019
- "Heart of the Machine" Seminar, Shanghai, China, October 2019
Computing the Wasserstein Barycenter Efficiently: A Structured Linear System and Customized Algorithms - Seminar of New Advances in Theory and Application of ADMM, Shanghai, China, March 2019
Service
Reviewer
Journal: SIAM Journal on Optimization (SIOPT), Mathematical Programming
Conference: ICML, NeurIPS
Session (Co-)Chair
INFORMS Annual Meeting, Seattle, WA, October 2024
International Symposium on Mathematical Programming (ISMP), Montréal, Canada, July 2024
INFORMS Optimization Society Conference, Houston, TX, March 2024
INFORMS Annual Meeting, Indianapolis, IN, October 2022