Anytime Planning for Decentralized POMDPs using Expectation Maximization
Akshat KumarShlomo ZilbersteinPublished in: CoRR (2012)
Keyphrases
- expectation maximization
- partially observable markov decision processes
- dec pomdps
- em algorithm
- distributed constraint optimization
- multi agent
- belief state
- stochastic domains
- partially observable
- reinforcement learning
- planning under uncertainty
- predictive state representations
- belief space
- planning problems
- decision theoretic planning
- single agent
- optimal plans
- sequential decision making problems
- decision theoretic
- heuristic search
- mixture model
- decentralized decision making
- maximum likelihood
- generative model
- markov decision processes
- cooperative
- dynamic programming
- ai planning
- state space
- peer to peer
- blocks world
- point based value iteration
- probabilistic model
- continuous state
- markov decision problems
- domain independent
- planning domains
- decision problems
- planning process
- parameter estimation
- gaussian mixture model
- goal oriented
- infinite horizon
- gaussian mixture
- computer vision