QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning

Benchmark Model Rank Results
smac-on-smac-def-armored-parallelQMIX#2Median Win Rate: 75.0
smac-on-smac-def-armored-sequentialQMIX#9Median Win Rate: 0.0
smac-on-smac-def-infantry-parallelQMIX#3Median Win Rate: 95.0
smac-on-smac-def-infantry-sequentialQMIX#5Median Win Rate: 96.9
smac-on-smac-def-outnumbered-parallelQMIX#1Median Win Rate: 30.0
smac-on-smac-def-outnumbered-sequentialQMIX#5Median Win Rate: 0.0
smac-on-smac-off-complicated-parallelQMIX#3Median Win Rate: 0.0
smac-on-smac-off-complicated-sequentialQMIX#1Median Win Rate: 87.5
smac-on-smac-off-distant-parallelQMIX#3Median Win Rate: 0.0
smac-on-smac-off-distant-sequentialQMIX#1Median Win Rate: 93.8
smac-on-smac-off-hard-parallelQMIX#3Median Win Rate: 0.0
smac-on-smac-off-hard-sequentialQMIX#1Median Win Rate: 96.9
smac-on-smac-off-near-parallelQMIX#1Median Win Rate: 95.0
smac-on-smac-off-near-sequentialQMIX#1Median Win Rate: 90.6
smac-on-smac-off-superhard-parallelQMIX#3Median Win Rate: 0.0
smac-on-smac-off-superhard-sequentialQMIX#3Median Win Rate: 0.0