Approximation of Discounted Minimax Markov Control Problems and Zero-Sum Markov Games Using Hausdorff and Wasserstein Distances.
François DufourTomás Prieto-RumeauPublished in: Dyn. Games Appl. (2019)
Keyphrases
- markov games
- control problems
- markov decision processes
- reinforcement learning
- optimal control
- infinite horizon
- adaptive control
- markov chain
- markov decision process
- optimal policy
- dynamic programming
- state space
- markov model
- policy iteration
- learning algorithm
- worst case
- queueing networks
- reinforcement learning algorithms
- average cost
- cooperative
- finite state
- control law