Dual Ascent and Primal-Dual Algorithms for Infinite-Horizon Nonstationary Markov Decision Processes.
Archis GhatePublished in: SIAM J. Optim. (2023)
Keyphrases
- markov decision processes
- infinite horizon
- finite horizon
- primal dual
- policy iteration
- non stationary
- optimal policy
- dynamic programming
- linear program
- average cost
- partially observable
- linear programming
- convergence rate
- state space
- finite state
- reinforcement learning
- markov decision process
- long run
- partially observable markov decision processes
- approximation algorithms
- average reward
- inventory control
- algorithm for linear programming
- optimal control
- learning algorithm
- convex optimization
- production planning
- reinforcement learning algorithms
- markov decision problems
- least squares
- policy iteration algorithm
- dec pomdps
- stationary policies
- computational complexity
- multistage
- fixed point
- model free
- random fields