On the Controller Synthesis for Finite-State Markov Decision Processes.
Antonín KuceraOldrich StrazovskýPublished in: FSTTCS (2005)
Keyphrases
- finite state
- markov decision processes
- controller synthesis
- closed loop
- optimal policy
- multi agent
- control system
- reinforcement learning
- state space
- dynamic programming
- reinforcement learning algorithms
- transition matrices
- policy iteration
- continuous state
- decision theoretic planning
- average cost
- partially observable
- infinite horizon
- control algorithm
- action space
- average reward
- action sets
- machine learning
- partially observable markov decision processes
- decision processes
- risk sensitive
- search algorithm
- policy iteration algorithm