Research map / Bandits and online learning / RL theory and MDPs
Learning Infinite-horizon Average-reward Markov Decision Process with Constraints
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use policy optimization with a bounded-scale action-value estimator and an exploration bonus | Ergodic CMDPs · Ergodic CMDPs | regret: \tilo{T^{2/3}} -> \tilo{\sqrt{T}} | Helped |
| Use policy optimization with a bounded-scale action-value estimator and an exploration bonus | Ergodic CMDPs · Ergodic CMDPs | constraint violation: \tilo{T^{2/3}} -> constant | Helped |
| Add span constraints on reward and cost value functions (ablation) | Finite Horizon Approximation for CMDP · Weakly communicating CMDPs | regret: \tilO{ (1+\text{\rm sp}^\star_r)(S^2A)^{1/3}T^{2/3} } -> \tilo{ \text{\rm sp}^\star_r S\sqrt{AT} } | Helped |
| Add span constraints on reward and cost value functions (ablation) | Finite Horizon Approximation for CMDP · Weakly communicating CMDPs | constraint violation: \tilO{ (1+\text{\rm sp}^\star_c)(S^2A)^{1/3}T^{2/3} } -> \tilo{ \text{\rm sp}^\star_c S\sqrt{AT} } | Helped |
Closest papers by meaning
- Learning General Parameterized Policies for Infinite Horizon Average Reward Constrained MDPs via Primal-Dual Policy Gradient Algorithm (NeurIPS 2024 · similarity 0.93)
- Learning Weakly Communicating Average-Reward CMDPs: Strong Duality and Improved Regret (NeurIPS 2026 · similarity 0.93)
- Regret Analysis of Policy Gradient Algorithm for Infinite Horizon Average Reward Markov Decision Processes (AAAI 2024 · similarity 0.92)
- Achieving Tractable Minimax Optimal Regret in Average Reward MDPs (NeurIPS 2024 · similarity 0.92)
- Efficient Exploration in Average-Reward Constrained Reinforcement Learning: Achieving Near-Optimal Regret With Posterior Sampling (ICML 2024 · similarity 0.91)
- Global Convergence for Average Reward Constrained MDPs with Primal-Dual Actor Critic Algorithm (NeurIPS 2025 · similarity 0.91)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.