A Bayesian network approach to control of networked Markov decision processes.
Sachin AdlakhaSanjay LallAndrea GoldsmithPublished in: Allerton (2008)
Keyphrases
- markov decision processes
- bayesian networks
- state space
- optimal policy
- reinforcement learning
- transition matrices
- finite state
- policy iteration
- dynamic programming
- decentralized control
- planning under uncertainty
- reachability analysis
- control system
- decision theoretic planning
- average reward
- partially observable
- reinforcement learning algorithms
- state and action spaces
- factored mdps
- average cost
- infinite horizon
- finite horizon
- control strategy
- probabilistic inference
- semi markov decision processes
- markov decision process
- decision processes
- risk sensitive
- graphical models
- interval estimation
- action sets
- machine learning
- control policy
- action space
- partially observable markov decision processes
- control problems
- reward function
- optimal control