Optimal Decision Tree Policies for Markov Decision Processes.
Daniël VosSicco VerwerPublished in: CoRR (2023)
Keyphrases
- markov decision processes
- optimal policy
- average cost
- stationary policies
- total reward
- discounted reward
- dynamic programming
- decision trees
- average reward
- finite horizon
- markov decision process
- policy iteration algorithm
- reinforcement learning
- control policies
- discount factor
- decision processes
- finite state
- action sets
- state space
- policy iteration
- expected reward
- reward function
- action space
- transition matrices
- risk sensitive
- decision problems
- reachability analysis
- decision theoretic planning
- partially observable markov decision processes
- optimal control
- markov decision problems
- macro actions
- optimality criterion
- infinite horizon
- reinforcement learning algorithms
- model based reinforcement learning
- control policy
- machine learning
- factored mdps
- planning under uncertainty
- initial state
- partially observable
- long run
- finite number
- decentralized control
- state and action spaces
- optimal solution
- semi markov decision processes
- state abstraction
- supply chain
- multistage
- total cost
- convergence rate