Reduction of total-cost and average-cost MDPs with weakly continuous transition probabilities to discounted MDPs.
Eugene A. FeinbergJefferson HuangPublished in: Oper. Res. Lett. (2018)
Keyphrases
- average cost
- total cost
- transition probabilities
- markov decision problems
- markov decision processes
- markov chain
- finite horizon
- finite state
- markov decision process
- holding cost
- long run
- optimal solution
- optimal policy
- initial state
- random walk
- service level
- production cost
- state space
- finite number
- infinite horizon
- action space
- risk sensitive
- markov models
- planning horizon
- dynamic programming
- linear programming
- optimal control
- steady state
- reward function
- inventory level
- multistage
- policy iteration
- linear program
- lead time
- stationary distribution
- sufficient conditions
- setup cost
- lot size
- link structure
- reinforcement learning
- objective function