Fuzzy reinforcement learning control for decentralized partially observable Markov decision processes.
Rajneesh SharmaMatthijs T. J. SpaanPublished in: FUZZ-IEEE (2011)
Keyphrases
- reinforcement learning
- partially observable markov decision processes
- continuous state
- dec pomdps
- optimal policy
- partially observable domains
- multi agent
- markov decision processes
- finite state
- partially observable
- optimal control
- decentralized control
- hidden state
- policy search
- partially observable environments
- planning under uncertainty
- function approximation
- state space
- control policy
- belief state
- partial observability
- partially observable stochastic games
- fully observable
- action selection
- dynamical systems
- reinforcement learning algorithms
- policy gradient
- control system
- belief space
- average reward
- temporal difference
- stochastic domains
- actor critic
- decision problems
- dynamic programming
- machine learning
- state action
- control policies
- action space