Extreme state aggregation beyond Markov decision processes.
Marcus HutterPublished in: Theor. Comput. Sci. (2016)
Keyphrases
- markov decision processes
- state space
- action space
- optimal policy
- finite state
- reinforcement learning
- decision theoretic planning
- markov decision process
- partially observable
- dynamic programming
- state abstraction
- factored mdps
- real time dynamic programming
- policy iteration
- reinforcement learning algorithms
- finite horizon
- transition matrices
- planning under uncertainty
- average reward
- average cost
- reachability analysis
- discounted reward
- total reward
- heuristic search
- initial state
- infinite horizon
- decision processes
- planning problems
- action sets
- state variables
- risk sensitive
- reward function