Model-based reinforcement learning for infinite-horizon approximate optimal tracking.
Rushikesh KamalapurkarLindsey AndrewsPatrick WaltersWarren E. DixonPublished in: CoRR (2015)
Keyphrases
- infinite horizon
- model based reinforcement learning
- markov decision processes
- finite horizon
- optimal control
- dynamic programming
- single item
- stochastic demand
- average cost
- optimal policy
- long run
- fixed cost
- inventory policy
- holding cost
- state space
- single product
- inventory control
- reinforcement learning
- finite state
- markov decision problems
- markov decision process
- partially observable
- expected cost
- lot size
- initial state
- control strategy
- multistage
- linear programming
- search algorithm
- optimal solution