On Linear Programming for Constrained and Unconstrained Average-Cost Markov Decision Processes with Countable Action Spaces and Strictly Unbounded Costs.
Huizhen YuPublished in: Math. Oper. Res. (2022)
Keyphrases
- average cost
- markov decision processes
- action space
- linear programming
- dynamic programming
- state and action spaces
- finite state
- state space
- optimal policy
- policy iteration
- continuous state
- reinforcement learning
- finite horizon
- infinite horizon
- control policies
- linear program
- risk sensitive
- continuous state spaces
- markov decision problems
- planning under uncertainty
- initial state
- decision processes
- markov decision process
- reinforcement learning algorithms
- control policy
- partially observable
- action sets
- np hard
- function approximators
- learning algorithm
- supply chain
- multi agent