Variance-Reduced Policy Gradient Approaches for Infinite Horizon Average Reward Markov Decision Processes.
Swetha GaneshWashim Uddin MondalVaneet AggarwalPublished in: CoRR (2024)
Keyphrases
- markov decision processes
- average reward
- infinite horizon
- policy gradient
- optimal policy
- long run
- policy iteration
- finite horizon
- state space
- stochastic games
- finite state
- actor critic
- reinforcement learning
- reinforcement learning algorithms
- state action
- total reward
- partially observable
- dynamic programming
- average cost
- discounted reward
- markov decision process
- discount factor
- action space
- reward function
- partially observable markov decision processes
- machine learning
- model free
- optimal control
- decision problems
- sufficient conditions
- monte carlo
- variance reduction
- rl algorithms
- markov chain
- stationary policies
- decision makers
- search space
- objective function
- bayesian networks