A POMDP Extension with Belief-dependent Rewards.
Mauricio Araya-LópezOlivier BuffetVincent ThomasFrançois CharpilletPublished in: NIPS (2010)
Keyphrases
- reinforcement learning
- belief state
- belief space
- markov decision processes
- point based value iteration
- partially observable markov decision processes
- reward function
- fully observable
- sequential decision making under uncertainty
- multi agent
- state space
- dynamical systems
- planning under uncertainty
- partially observable
- hidden state
- markov decision process
- policy evaluation
- learning algorithm
- belief revision
- knowledge base