Strong polynomiality of policy iterations for average-cost MDPs modeling replacement and maintenance problems.
Eugene A. FeinbergJefferson HuangPublished in: Oper. Res. Lett. (2013)
Keyphrases
- average cost
- markov decision processes
- optimal policy
- markov decision process
- infinite horizon
- finite horizon
- finite state
- long run
- decision problems
- finite number
- approximate dynamic programming
- policy iteration
- linear programming
- reinforcement learning
- linear program
- optimal control
- markov decision problems
- markov decision chains
- inventory models
- state space
- initial state
- probabilistic planning
- reinforcement learning problems
- real time dynamic programming
- stochastic shortest path
- partially observable markov decision processes
- multistage
- dynamic programming
- action space
- planning under uncertainty
- special case
- lower bound
- continuous state spaces
- decision making
- learning algorithm