Finite-Time Complexity of Online Primal-Dual Natural Actor-Critic Algorithm for Constrained Markov Decision Processes.
Sihan ZengThinh T. DoanJustin RombergPublished in: CDC (2022)
Keyphrases
- markov decision processes
- primal dual
- dynamic programming
- linear programming
- worst case
- natural actor critic
- computational complexity
- convergence rate
- state space
- learning algorithm
- policy iteration
- linear program
- algorithm for linear programming
- state variables
- convex optimization
- monte carlo
- reinforcement learning
- approximation algorithms
- finite state
- average reward
- optimal solution