Optimal Markov Policies for Finite-Horizon Constrained MDPs With Combined Additive and Multiplicative Utilities.
Uday Kumar MVeeraruna KavithaSanjay P. BhatNandyala HemachandraPublished in: IEEE Control. Syst. Lett. (2023)
Keyphrases
- finite horizon
- optimal policy
- optimal stopping
- markov decision processes
- control policies
- average cost
- markov decision process
- infinite horizon
- stochastic inventory control
- periodic review
- single product
- inventory models
- dynamic programming
- inventory control
- decision problems
- finite state
- expected reward
- multistage
- state space
- average reward
- long run
- reinforcement learning
- markov chain
- inventory policy
- policy iteration
- single item
- lot size
- lost sales
- optimal control
- learning algorithm
- holding cost
- state dependent
- markov model
- partially observable
- reward function
- finite number