Approximate Dynamic Programming for Two-Player Zero-Sum Markov Games.
Julien PérolatBruno ScherrerBilal PiotOlivier PietquinPublished in: ICML (2015)
Keyphrases
- approximate dynamic programming
- markov games
- reinforcement learning
- markov decision processes
- policy iteration
- reinforcement learning algorithms
- markov decision process
- average cost
- dynamic programming
- control problems
- multiagent reinforcement learning
- state space
- optimal policy
- linear program
- control policy
- learning algorithm
- infinite horizon
- function approximation
- model free
- finite state
- partially observable
- optimal control
- multi agent
- nash equilibrium
- action space
- multiagent systems
- stochastic games
- machine learning
- state variables
- markov decision problems
- evolutionary algorithm