Optimizing Local Satisfaction of Long-Run Average Objectives in Markov Decision Processes.
David KlaskaAntonín KuceraVojtech KurVít MusilVojtech RehákPublished in: AAAI (2024)
Keyphrases
- average cost
- long run
- markov decision processes
- optimal policy
- short run
- average reward
- infinite horizon
- expected cost
- finite horizon
- finite state
- discounted reward
- initial state
- reinforcement learning
- state space
- control policy
- stationary policies
- dynamic programming
- transition matrices
- policy iteration
- multistage
- inventory level
- sample path
- reachability analysis
- queueing networks
- exchange rate
- decision problems
- planning under uncertainty
- linear program
- state and action spaces
- least squares