Jefferson Huang
Postdoctoral Associate
School of Operations Research and Information Engineering
Cornell University
Office: 237 Rhodes Hall
Email: firstname dot lastname at cornell dot edu
In the Fall of 2018, I will be joining the Operations Research Department at the Naval Postgraduate School as a tenure-track Assistant Professor.
Research Interests:
- Applied Probability
- Optimization Under Uncertainty
- Operations Management
- Reinforcement Learning, Approximate Dynamic Programming
- Computational Complexity of Algorithms
Working Papers:
- J. Huang, D. G. Down, M. E. Lewis and C. Wu. Dynamically scheduling and maintaining a flexible server. Submitted, 2018.
Publications:
- J. Huang, D. G. Down, M. E. Lewis and C. Wu. Dynamic scheduling and maintenance for a two-class queue with a deteriorating server. Proceedings of the 2018 American Control Conference, Milwaukee, WI, USA, June 27-29, 2018. proceedings
- E. A. Feinberg and J. Huang. Reduction of total-cost and average-cost MDPs with weakly continuous transition probabilities to discounted MDPs. Operations Research Letters, 46(2):179-184, 2018. arXiv journal
- E. A. Feinberg and J. Huang. Strongly polynomial algorithms for transient and average-cost MDPs. ACM SIGMETRICS Performance Evaluation Review, 45(2):6-8, 2017. pdf
- E. A. Feinberg and J. Huang. On the reduction of total-cost
and average-cost MDPs to discounted MDPs. To appear in Naval
Research Logistics (Special Issue Dedicated to Uriel G. Rothblum),
2017. arXiv journal
- R. Zhang, J. Huang and T. Kumar. Preventive leak detection
for high pressure gas transmission networks. AAAI-17 Workshop on AI and OR for Social Good,
2017. pdf
- E. A. Feinberg, J. Huang and B. Scherrer. Modified policy iteration algorithms are not strongly polynomial for discounted dynamic programming. Operations Research Letters, 42(6-7):429-431, 2014. pdf
- E. A. Feinberg and J. Huang. The value iteration algorithm is not strongly polynomial for discounted dynamic programming. Operations Research Letters, 42(2):130-131, 2014. pdf
- E. A. Feinberg and J. Huang. Strong polynomiality of policy iterations for average-cost MDPs modeling replacement and maintenance problems. Operations Research Letters, 41(3):249-251, 2013. pdf
Talks:
- Dynamic scheduling and maintenance for a two-class queue with a deteriorating server. American Control Conference, Milwaukee, WI, June 28, 2018. slides
- Dynamic scheduling and maintenance of a deteriorating server.
- INFORMS International Conference, Taipei, Taiwan, June 19, 2018. slides
- CORS Annual Conference, Halifax, NS, Canada, June 4, 2018. slides
- AMS Spring Eastern Sectional Meeting, Northeastern University, Boston, MA, April 21, 2018. slides
- Seminar on Combinatorics, Games and Optimisation, London School of Economics and Political Science, London, UK, March 28, 2018. slides
- Near-optimal control of queueing systems via approximate one-step policy improvement. Reinforcement Learning for Processing Networks Seminar, Cornell University, Ithaca, NY, March 21, 2018. slides
- Solving Markov decision processes. Operations Research Seminar, Naval Postgraduate School, Monterey, CA, February 8, 2018. slides
- Optimality of a priority policy for a server scheduling problem with a deteriorating server. INFORMS Annual Meeting, Houston, TX, October 22, 2017. slides
- On the reduction of total cost and average cost MDPs to discounted MDPs. INFORMS Applied Probability Society Conference, Northwestern University, Evanston, IL, July 12, 2017. slides
- Strongly polynomial algorithms for transient and average-cost MDPs. Workshop on MAthematical performance Modeling and Analysis (MAMA), University of Illinois at Urbana-Champaign, Urbana, IL, June 5, 2017. slides
- Reducing undiscounted Markov decision processes and stochastic games with unbounded costs to discounted ones. Northeast Regional Conference on Optimization and Optimal Control under Uncertainty, IBM T. J. Watson Research Center, Yorktown Heights, NY, December 8, 2016. slides
- Reductions of undiscounted Markov decision processes and
stochastic games to discounted ones. INFORMS Annual Meeting,
Nashville, TN, November 16, 2016. slides
- Computational complexity estimates for value and policy
iteration algorithms for total-cost and average-cost Markov decision
processes. AI Seminar, University of Alberta, Edmonton, AB, Canada, May 5, 2016. slides
- Recovering bandits. SBU Algorithms Reading Group, Stony Brook, NY, Feb. 26, 2016. slides
- Markov decision processes and complexity theory: some research directions. SBU AMS Graduate Reading Group, Stony Brook, NY, Feb. 10, 2016. slides
- Computational complexity estimates for policy and value iteration algorithms for total-cost and average-cost Markov decision processes. INFORMS Annual Meeting, Philadelphia, PA, Nov. 2, 2015. slides
- Nonlinear methods for temporal causal discovery. Industries & Solutions Summer Intern Seminar Series, IBM T. J. Watson Research Center, Yorktown Heights, NY, August 13, 2015.
- Computational complexity estimates for value and policy iteration algorithms for total-cost and average-cost Markov decision processes. AP for Lunch Seminar, IBM T. J. Watson Research Center, Yorktown Heights, NY, July 29, 2015. slides
- Computational complexity estimates for policy and value iteration algorithms for total-cost and average-cost Markov decision processes. The Fifth International Workshop in Sequential Methodologies, New York, NY, June 23, 2015. slides
- Reduction of average-cost Markov Decision Processes to discounting under an accessibility condition. INFORMS Annual Meeting, San Francisco, CA, Nov. 10, 2014. slides
- Recent progress on the complexity of solving Markov Decision Processes. Prelim talk, Jan. 24, 2014. slides writeup
Teaching:
Patents:
- F. Heng, J. Huang, T. Kumar, and R. Zhang. System and Method for Forecasting Leaks in a Fluid-Delivery Pipeline Network. US Patent Application No. 15/134868. 2016. pdf
PhD Dissertation: Complexity estimates and reductions to discounting for total and average-reward Markov decision processes and stochastic games. Dept. of Applied Mathematics and Statistics, Stony Brook University, 2016. pdf
Master's Thesis: Piecewise truckload network
procurement. Dept. of Civil & Environmental Engineering, Massachusetts Institute of Technology,
2011. link
During the summer of 2015, I interned with
the Smarter Energy group
at IBM
Research in Yorktown Heights, NY.
From 2008 to 2009, I worked
at Cambridge Systematics in
Oakland, CA.
I earned a B.S. in Civil & Environmental Engineering from UC
Berkeley in 2008.