Research map / Multi-agent systems and games / Cooperative multi-agent RL
Grid-Wise Control for Multi-Agent Reinforcement Learning in Video Game AI
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 |
|---|---|---|---|
| replace independent learning (IQL/IAC) with grid-wise centralized-conv policy GridNet | GridNet vs IQL/IAC · 5I/3I2Z cross combat | winning rate: 1.00 | Helped |
| compare GridNet with Central-V in head-to-head play | GridNet · 5I | winning rate against Central-V: 0.73 | Helped |
| compare GridNet with CommNet in head-to-head play | GridNet · 3I2Z | winning rate against CommNet: 0.68 | Helped |
| compare GridNet with Central-V on 3I2Z | GridNet · 3I2Z | winning rate against Central-V: 0.88 | Helped |
| transfer 5I/3I2Z-trained policies to 20I (more agents) | Central-V · 20I vs AN | winning rate: 0.55 | Hurt |
| transfer 5I/3I2Z-trained policies to mixed 5I5Z | Central-V · 5I5Z vs AN | winning rate: 0.02 | Hurt |
| transfer to larger scenarios with GridNet | GridNet · 10I/20I vs AN/HR | winning rate: 1.00 | Helped |
| transfer to mixed-unit large scenarios with GridNet | GridNet · 10I10Z vs HR | winning rate: 0.71 | Mixed |
| transfer IQL/IAC policies to new scenarios | IQL · 10I vs AN | winning rate: 0.03 | Hurt |
| train on self-play versus AN/HR on MAB | GridNet · MAB cross combat | winning rate: 0.82 | Helped |
| compare AN-trained and HR-trained policies on MAB | GridNet · MAB cross combat | winning rate: 0.48 | No clear effect |
| play GridNet against expert human players | GridNet · MAB | winning rate: 100% | Helped |
Closest papers by meaning
- Graph Convolutional Reinforcement Learning (ICLR 2020 · similarity 0.88)
- Foundations of Multi-Agent Learning in Dynamic Environments: Where Reinforcement Learning Meets Strategic Decision-Making (AAAI 2025 · similarity 0.88)
- Towards Efficient Collaboration via Graph Modeling in Reinforcement Learning (AAAI 2025 · similarity 0.87)
- Enhancing Cooperative Multi-Agent Reinforcement Learning with State Modelling and Adversarial Exploration (ICML 2025 · similarity 0.87)
- Investigating Relational State Abstraction in Collaborative MARL (AAAI 2025 · similarity 0.87)
- Grounded Answers for Multi-agent Decision-making Problem through Generative World Model (NeurIPS 2024 · similarity 0.87)
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.