Game Theory and Reinforcement Learning
Interdisciplinary Areas: | Data and Engineering Applications, Autonomous and Connected Systems, Smart City, Infrastructure, Transportation, Security and Privacy |
---|
Project Description
Consider the pricing of ride-sharing in autonomous vehicles. There are a large number of vehicles and large number of passengers. How to come up with efficient pricing algorithms so far has been understood with efficient utility modeling. However, recent advances in multi-agent reinforcement learning allow for solving such problems without the knowledge of the model. Similar aspects happen for game playing between large number of players, including the case where players may be colluding. We aim to understand different game-theoretic problems through the lens of model-free reinforcement learning.
Start Date
Apr 2021
Postdoc Qualifications
The postdoc researcher should preferably have PhD in Computer Science, Electrical Engineering, Mathematics, Statistics, or related field.
Co-Advisors
Vaneet Aggarwal, vaneet@purdue.edu, School of IE and ECE, https://web.ics.purdue.edu/~vaneet
Bharat Bhargava, bbshail@purdue.edu, Dept of CS, https://www.cs.purdue.edu/people/faculty/bb
References
Model-free mean-field reinforcement learning: mean-field MDP and mean-field Q-learning
Unified Reinforcement Q-Learning for Mean Field Game and Control Problems
$Q$-Learning in a Stochastic Stackelberg Game between an Uninformed Leader and a Naive Follower
Q-learning solution for optimal consensus control of discrete-time multiagent systems using reinforcement learning
Reinforcement Learning for Mean Field Game