Generalized Planning on Transport
100ScaleExplicitly regularized Q-value policy (Ω^Exp.)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Explicitly regularized Q-value policy (Ω^Exp.)Architecture=R-GNN2026.03 | 100 | 90.4 | |
| Heuristically regularized Q-value policy (Ω^Heu.)Architecture=R-GNN2026.03 | 100 | 92.7 | |
| Explicitly regularized Q-value policy (Ω^Exp.)Architecture=OE2026.03 | 100 | 88.6 | |
| Heuristically regularized Q-value policy (Ω^Heu.)Architecture=OE2026.03 | 100 | 89.7 | |
| Explicitly regularized Q-value policy (Ω^Exp.)Architecture=OAE2026.03 | 100 | 89.7 | |
| Heuristically regularized Q-value policy (Ω^Heu.)Architecture=OAE2026.03 | 100 | 89.6 | |
| State-value policy (V)Architecture=R-GNN2026.03 | 72 | 46.6 | |
| Vanilla Q-value policy (Q)Architecture=OAE2026.03 | 57 | 32.1 | |
| Vanilla Q-value policy (Q)Architecture=R-GNN2026.03 | 46 | 24.7 | |
| Vanilla Q-value policy (Q)Architecture=OE2026.03 | 42 | 23.1 | |
| State-value policy (V)Architecture=OE2026.03 | 38 | 20.1 | |
| State-value policy (V)Architecture=OAE2026.03 | 19 | 9.3 |