ResearchBenchmarksMulti-Agent Reinforcement Learning on sce3m (test)Multi-Agent Reinforcement Learning on sce3m (test)Follow100Win RateMAPPO92.294.22596.2598.275Jun 29, 2026Evaluation resultsMethodMethodLinksWin RateFinal ScoreMAPPO2026.0610074.406IQL2026.0610048.082VDN2026.0610047.62QMIX2026.0699.756.587QTRAN2026.0699.757.92COMA2026.0698.955.375HRL-IM/CBS2026.0692.547.158