Job Shop Scheduling on Taillard Benchmark
1,488.9Average MakespanLinear-Complexity Unified Graph
Evaluation Results
| Method | Links | |
|---|---|---|
| Linear-Complexity Unified GraphInstance Size=15 × 15, Variant=comp, Model Type=shape-specific2026.04 | 1,488.9 | |
| Linear-Complexity Unified GraphInstance Size=15 × 15, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 1,538.3 | |
| Zhang et al.Instance Size=15 × 15, Model Type=shape-specific2026.04 | 1,547.4 | |
| FDD/MWRInstance Size=15 × 152026.04 | 1,577.4 | |
| MORInstance Size=15 × 152026.04 | 1,735.7 | |
| Linear-Complexity Unified GraphInstance Size=20 × 15, Variant=comp, Model Type=shape-specific2026.04 | 1,755.4 | |
| Zhang et al.Instance Size=20 × 15, Model Type=shape-specific2026.04 | 1,774.7 | |
| FDD/MWRInstance Size=20 × 152026.04 | 1,836.7 | |
| Linear-Complexity Unified GraphInstance Size=20 × 15, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 1,845.2 | |
| MWKRInstance Size=15 × 152026.04 | 1,896.1 | |
| MORInstance Size=20 × 152026.04 | 1,998.5 | |
| Linear-Complexity Unified GraphInstance Size=20 × 20, Variant=comp, Model Type=shape-specific2026.04 | 2,053.8 | |
| Zhang et al.Instance Size=20 × 20, Model Type=shape-specific2026.04 | 2,128.1 | |
| Linear-Complexity Unified GraphInstance Size=20 × 20, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 2,156.5 | |
| FDD/MWRInstance Size=20 × 202026.04 | 2,164.7 | |
| MWKRInstance Size=20 × 152026.04 | 2,178.9 | |
| Linear-Complexity Unified GraphInstance Size=30 × 15, Variant=comp, Model Type=shape-specific2026.04 | 2,289.1 | |
| MORInstance Size=20 × 202026.04 | 2,334.8 | |
| Linear-Complexity Unified GraphInstance Size=30 × 15, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 2,346.5 | |
| Zhang et al.Instance Size=30 × 15, Model Type=shape-specific2026.04 | 2,378.8 | |
| FDD/MWRInstance Size=30 × 152026.04 | 2,511.2 | |
| MORInstance Size=30 × 152026.04 | 2,567.6 | |
| Linear-Complexity Unified GraphInstance Size=30 × 20, Variant=comp, Model Type=shape-specific2026.04 | 2,594.2 | |
| Zhang et al.Instance Size=30 × 20, Model Type=shape-specific2026.04 | 2,603.9 | |
| MWKRInstance Size=20 × 202026.04 | 2,633.9 | |
| FDD/MWRInstance Size=30 × 202026.04 | 2,726.4 | |
| Linear-Complexity Unified GraphInstance Size=30 × 20, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 2,756 | |
| MWKRInstance Size=30 × 152026.04 | 2,985.9 | |
| MORInstance Size=30 × 202026.04 | 3,010.4 | |
| MWKRInstance Size=30 × 202026.04 | 3,265.7 | |
| Linear-Complexity Unified GraphInstance Size=50 × 15, Variant=comp, Model Type=shape-specific2026.04 | 3,349.3 | |
| Zhang et al.Instance Size=50 × 15, Model Type=shape-specific2026.04 | 3,393.8 | |
| Linear-Complexity Unified GraphInstance Size=50 × 15, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 3,423.3 | |
| Linear-Complexity Unified GraphInstance Size=50 × 20, Variant=comp, Model Type=shape-specific2026.04 | 3,566.9 | |
| Zhang et al.Instance Size=50 × 20, Model Type=shape-specific2026.04 | 3,593.9 | |
| MORInstance Size=50 × 152026.04 | 3,714.1 | |
| FDD/MWRInstance Size=50 × 152026.04 | 3,723.7 | |
| Linear-Complexity Unified GraphInstance Size=50 × 20, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 3,730.2 | |
| FDD/MWRInstance Size=50 × 202026.04 | 3,845.5 | |
| MORInstance Size=50 × 202026.04 | 4,143.6 | |
| MWKRInstance Size=50 × 152026.04 | 4,444.9 | |
| MWKRInstance Size=50 × 202026.04 | 4,677.4 | |
| Zhang et al.Instance Size=100 × 20, Model Type=shape-specific2026.04 | 6,097.6 | |
| Linear-Complexity Unified GraphInstance Size=100 × 20, Variant=comp, Model Type=shape-specific2026.04 | 6,263.4 | |
| Linear-Complexity Unified GraphInstance Size=100 × 20, Variant=single, Model Type=generalist (trained on 20x20)2026.04 | 6,423 | |
| MORInstance Size=100 × 202026.04 | 6,728.8 | |
| FDD/MWRInstance Size=100 × 202026.04 | 6,837.4 | |
| MWKRInstance Size=100 × 202026.04 | 8,040.3 |