# Multi-Agent Reinforcement Learning on REF-q rac-dist (test) Source: [Multi-Agent Reinforcement Learning on REF-q rac-dist (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:52:01.727Z Evaluation of Multi-Agent Reinforcement Learning performance on the test split of the REF-q rac-dist environment, measured by episodic rewards and normalized performance. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Multi-Agent Reinforcement Learning | | Dataset (including any stated split) | REF-q rac-dist (test) | | Metric | Mean Episodic Reward (q=2) | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 12 of 12 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Mean Episodic Reward (q=2)](): Higher is better; unit: SCALAR. - [Mean Episodic Reward (q=7)](): Higher is better; unit: SCALAR. - [Mean Episodic Reward (q=10)](): Higher is better; unit: SCALAR. - [Normalized Performance](): Higher is better; unit: SCALAR. ## Recorded results | Method | Mean Episodic Reward (q=2) | Result date | Report paper | | --- | --- | --- | --- | | QMIX | 125 | Not recorded | 68fa6daa1e209131cea3b4de | | GRE-MARL | 116 | Not recorded | 68fa6daa1e209131cea3b4de | | DRE-MARL | 113 | Not recorded | 68fa6daa1e209131cea3b4de | | DRE-MARL | 82 | Not recorded | 68fa6daa1e209131cea3b4de | | DRE-MARL | 44 | Not recorded | 68fa6daa1e209131cea3b4de | | p2p-MARL | 35 | Not recorded | 68fa6daa1e209131cea3b4de | | DRE-MARL | 23 | Not recorded | 68fa6daa1e209131cea3b4de | | DRE-MARL | 11 | Not recorded | 68fa6daa1e209131cea3b4de | | MAAC | 11 | Not recorded | 68fa6daa1e209131cea3b4de | | MAPPO | -50 | Not recorded | 68fa6daa1e209131cea3b4de | | IQL | -62 | Not recorded | 68fa6daa1e209131cea3b4de | | MADDPG | -108 | Not recorded | 68fa6daa1e209131cea3b4de | ### QMIX Mean Episodic Reward (q=2): 125. Result ID: 698fd0973a6c314849fc0e93. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### GRE-MARL Mean Episodic Reward (q=2): 116. Result ID: 698fd0973a6c314849fc0e8f. [Reporting paper]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### DRE-MARL Mean Episodic Reward (q=2): 113. Result ID: 698fd0973a6c314849fc0e8b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation Scheme":"l\_MO + g\_MO","Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### DRE-MARL Mean Episodic Reward (q=2): 82. Result ID: 698fd0973a6c314849fc0e8a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation Scheme":"l\_SMO + g\_MO","Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### DRE-MARL Mean Episodic Reward (q=2): 44. Result ID: 698fd0973a6c314849fc0e8d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation Scheme":"l\_SMO","Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### p2p-MARL Mean Episodic Reward (q=2): 35. Result ID: 698fd0973a6c314849fc0e8e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### DRE-MARL Mean Episodic Reward (q=2): 23. Result ID: 698fd0973a6c314849fc0e89. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation Scheme":"l\_ss + g\_ss","Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### DRE-MARL Mean Episodic Reward (q=2): 11. Result ID: 698fd0973a6c314849fc0e8c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation Scheme":"l\_SMO + g\_ss","Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### MAAC Mean Episodic Reward (q=2): 11. Result ID: 698fd0973a6c314849fc0e92. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### MAPPO Mean Episodic Reward (q=2): -50. Result ID: 698fd0973a6c314849fc0e91. [Reporting paper]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### IQL Mean Episodic Reward (q=2): -62. Result ID: 698fd0973a6c314849fc0e94. [Reporting paper]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. ### MADDPG Mean Episodic Reward (q=2): -108. Result ID: 698fd0973a6c314849fc0e90. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Training Reward":"Individual","Evaluation Setting":"rac-dist"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.