Strategy complexity of finite-horizon Markov decision processes and simple stochastic games
Krishnendu ChatterjeeRasmus Ibsen-JensenPublished in: CoRR (2012)
Keyphrases
- markov decision processes
- stochastic games
- finite horizon
- optimal policy
- infinite horizon
- average reward
- finite state
- state space
- reinforcement learning algorithms
- policy iteration
- dynamic programming
- reinforcement learning
- multiagent reinforcement learning
- markov decision process
- average cost
- partially observable
- decision problems
- computational complexity
- expected reward
- reward function
- heuristic search
- multistage
- search space
- machine learning