Recursive stochastic games with positive rewards.
Kousha EtessamiDominik WojtczakMihalis YannakakisPublished in: Theor. Comput. Sci. (2019)
Keyphrases
- stochastic games
- markov decision processes
- nash equilibria
- reinforcement learning
- average reward
- reinforcement learning algorithms
- optimal policy
- multi agent
- state space
- learning automata
- reward function
- repeated games
- policy iteration
- nash equilibrium
- average cost
- robust optimization
- finite horizon
- long run
- partially observable
- single agent
- least squares
- upper bound
- dynamic programming
- cooperative
- knowledge base
- learning algorithm