infinite-horizon
In reinforcement learning, a setting where the decision-making process is considered over an indefinite future, requiring strategies that account for long-term rewards rather than short-term gains.
- Exploration from a Primal-Dual Lens: Value-Incentivized Actor-Critic Methods for Sample-Efficient Online RL
- Global Convergence for Average Reward Constrained MDPs with Primal-Dual Actor Critic Algorithm
- Offline Actor-Critic for Average Reward MDPs
- On the Convergence of Single-Timescale Actor-Critic
- Regret Analysis of Average-Reward Unichain MDPs via an Actor-Critic Approach