DQN | PAU | EWC0.1 | 0.5 | 1 | -0.5 | CDE0.1 | 0.5 | 1 | -0.5 | |||
---|---|---|---|---|---|---|---|---|---|---|---|---|
PMD | 0.110 | 0.131 | 0.240 | 0.183 | 0.147 | 0.104 | 0.18 | 0.253 | 0.144 max(0.94) | |||
PDM | 0.065 max(0.21) | 0.088 | 0.095 max(0.21) | 0.104 | ||||||||
DPM | 0.057 max(0.11) | 0.045 | 0.069 max(0.18) | 0.113 | ||||||||
DMP | 0.086 | 0.062 | 0.114 | 0.101 | ||||||||
MPD | 0.037 | 0.061 | 0.217 | 0.189 | ||||||||
MDP | 0.084 | 0.080 max(0.13) | 0.098 max(0.18) | 0.143 |
DQN | PAU | EWC0.1 | 0.5 | 1 | -0.5 | CDE0.1 | 0.5 | 1 | -0.5 | |||
---|---|---|---|---|---|---|---|---|---|---|---|---|
PMD | x | x | x | x | x | x | x | x | ||||
PDM | x | x | ||||||||||
DPM | x | x | ||||||||||
DMP | x | |||||||||||
MPD | x | |||||||||||
MDP | x | x |
Score and Completions Evaluation of the Pathfinding, Malfunction, Deadlock and Evaluation Environment during the training. All Evaluations were made with a 2 Layers à 1024 Neurons Sub-Environments have their first letter as Abreviations D=Deadlock, M=Malfunction, P=Pathfinding.
Evaluation for CDE just for the best Sub-Network at the moment
Evaluation for CDE just for the best Sub-Network at the moment
DQN | PAU | EWC0.1 | 0.5 | 1 | -0.5 | CDE0.1 | 0.5 | 1 | -0.5 | |||
---|---|---|---|---|---|---|---|---|---|---|---|---|
PMD | 0.10998578535891973 | 0.1311846689895474 | 0.24004000666777675 | 0.18346080796673922 | 0.14735930735930755 | 0.10354029532111728 | 0.17513736263736218 | 0.25259515570934093 | 0.14408113306522116 | |||
PDM | 0.06465595087592581 | 0.0881724260776632 | 0.09500177242112744 | 0.10410714285714319 | ||||||||
DPM | 0.05660036166365291 | 0.0448357233617716 | 0.06928034371643421 | 0.11273559978862083 | ||||||||
DMP | 0.08592777085927789 | 0.0618444846292949 | 0.11357586512866051 | 0.10062111801242253 | ||||||||
MPD | 0.03725917848055245 | 0.0609911054637867 | 0.2170003393281295 | 0.18932874354561102 | ||||||||
MDP | 0.08427672955974871 | 0.0804020100502516 | 0.09813664596273322 | 0.14285714285714285 |