Control of Markov decision processes from PCTL specifications.
Morteza LahijanianSean B. AnderssonCalin BeltaPublished in: ACC (2011)
Keyphrases
- markov decision processes
- optimal policy
- transition matrices
- decision theoretic planning
- finite state
- policy iteration
- reinforcement learning algorithms
- decentralized control
- state space
- dynamic programming
- reinforcement learning
- finite horizon
- average cost
- decision processes
- factored mdps
- planning under uncertainty
- average reward
- partially observable
- action space
- infinite horizon
- optimal control
- risk sensitive
- action sets
- reachability analysis
- state and action spaces
- reward function
- model based reinforcement learning
- control problems
- control strategy
- monte carlo
- decision makers
- control system