A causal approach to hierarchical decomposition of factored MDPs.
Anders JonssonAndrew G. BartoPublished in: ICML (2005)
Keyphrases
- hierarchical decomposition
- factored mdps
- hierarchical organization
- state space
- context specific
- markov decision processes
- approximate dynamic programming
- algebraic decision diagrams
- linear program
- finite state machines
- policy iteration
- basis functions
- stochastic processes
- optimal policy
- planning under uncertainty
- markov decision problems
- bayesian networks
- machine learning
- markov chain
- linear programming
- reinforcement learning