Strategy Complexity of Finite-Horizon Markov Decision Processes and Simple Stochastic Games.
Krishnendu ChatterjeeRasmus Ibsen-JensenPublished in: MEMICS (2012)
Keyphrases
- markov decision processes
- finite horizon
- stochastic games
- optimal policy
- infinite horizon
- average reward
- state space
- reinforcement learning
- finite state
- reinforcement learning algorithms
- decision problems
- markov decision process
- policy iteration
- dynamic programming
- average cost
- multiagent reinforcement learning
- partially observable
- long run
- optimal strategy
- action space
- reward function
- partially observable markov decision processes
- function approximation
- multistage
- computational complexity
- objective function