Primal-Dual Regression Approach for Markov Decision Processes with General State and Action Spaces.
Denis BelomestnyJohn SchoenmakersPublished in: SIAM J. Control. Optim. (2024)
Keyphrases
- markov decision processes
- state and action spaces
- primal dual
- action space
- state space
- linear programming
- optimal policy
- finite state
- reinforcement learning
- linear program
- average reward
- markov decision problems
- special case
- policy iteration
- dynamic programming
- convex optimization
- planning under uncertainty
- partially observable markov decision process
- decision processes
- decision theoretic
- approximation algorithms
- partially observable
- average cost
- reinforcement learning algorithms
- markov decision process
- planning problems
- partially observable markov decision processes
- convergence rate
- multistage