Research map / Multi-agent systems and games / Games and equilibria
Faster Last-iterate Convergence of Policy Optimization in Zero-Sum Markov Games
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 symmetric entropy-regularized OMWU policy updates with slower-timescale value updates | Policy optimization in discounted zero-sum Markov games · Discounted zero-sum Markov games | iterations to ε-NE: \widetilde{\mathcal{O}}\Big(\frac{|\mathcal{S}|}{(1-\gamma)^5\epsilon}\Big) versus prior bounds including \widetilde{\ma | Helped |
| Use symmetric entropy-regularized OMWU policy updates with slower-timescale value updates | Policy optimization in episodic zero-sum Markov games · Episodic zero-sum Markov games | iterations to ε-NE: \widetilde{\mathcal{O}}\big(\frac{H^3}{\epsilon}\big) versus \widetilde{\mathcal{O}}\big(\frac{H^4}{\epsilon}\big) and \ | Helped |
| Replace constant value-update learning rate α_t = ητ with a vanishing learning-rate sequence | Entropy-regularized OMWU for discounted Markov games · Discounted zero-sum Markov games | convergence rate | Hurt |
| Set regularization sufficiently small to translate ε-QRE convergence into ε-NE convergence | Entropy-regularized OMWU · Discounted and episodic zero-sum Markov games | iterations to ε-NE: \widetilde{\mathcal{O}}\left(\frac{|\mathcal{S}|}{(1-\gamma)^5\epsilon}\right); \widetilde{\mathcal{O}}\big(\frac{H^3}{\ | Helped |
Closest papers by meaning
- Policy Optimization for Markov Games: Unified Framework and Faster Convergence (NeurIPS 2022 · similarity 0.94)
- Optimistic Policy Gradient in Multi-Player Markov Games with a Single Controller: Convergence beyond the Minty Property (AAAI 2024 · similarity 0.92)
- Incentivize without Bonus: Provably Efficient Model-based Online Multi-agent RL for Markov Games (ICML 2025 · similarity 0.92)
- Sample-Efficient Multi-Agent RL: An Optimization Perspective (ICLR 2024 · similarity 0.91)
- Improving Sample Efficiency of Model-Free Algorithms for Zero-Sum Markov Games (ICML 2024 · similarity 0.91)
- Solving Zero-Sum Convex Markov Games (ICML 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.