Traveling Salesman Problem on Uniform-TSP100
0Optimality GapLKH3
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LKH32024.05 | 0 | 56 | 7.7632 | |
| Concorde2024.05 | 0 | 34 | 7.7632 | |
| Att-GCN+MCTS*Search Strategy=MCTS2024.05 | 0.037 | 15 | 7.7638 | |
| GLOP (more revisions)Local Search / Refinement=more revisions2024.05 | 0.046 | 1.9 | 7.7668 | |
| POMO aug×8Data Augmentation=8x2024.05 | 0.134 | 1 | 7.7736 | |
| ICAM aug×8Data Augmentation=8x2024.05 | 0.148 | 37 | 7.7747 | |
| Pointerformer aug×8Data Augmentation=8x2024.05 | 0.163 | 49 | 7.7759 | |
| ELG aug×8Data Augmentation=8x2024.05 | 0.225 | 3 | 7.7807 | |
| DIFUSCO greedy+2-opt*Search Strategy=greedy, Local Search / Refinement=2-opt2024.05 | 0.24 | — | 7.78 | |
| BQ greedySearch Strategy=greedy2024.05 | 0.349 | 1.8 | 7.7903 | |
| ICAM2024.05 | 0.462 | 5 | 7.7991 | |
| LEHD greedySearch Strategy=greedy2024.05 | 0.577 | 27 | 7.808 | |
| ICAM single trajec.Search Strategy=single trajectory2024.05 | 0.897 | 2 | 7.8328 | |
| Efficient Decoder Scaling StrategyDepth=42, Width=512, Decoding Strategy=beam search, Beam Size=22026.02 | 1.3 | 21 | — | |
| BQDecoding Strategy=beam search, Beam Size=162026.02 | 1.5 | 17.6 | — | |
| ReEvo BUType=White-Box, Variant=Node, Setting=Single-task2026.05 | 1.58 | — | — | |
| ReEvo TDType=White-Box, Variant=Node, Setting=Single-task2026.05 | 1.71 | — | — | |
| ReEvo TD & BUType=White-Box, Variant=Node, Setting=Multi-task2026.05 | 1.78 | — | — | |
| POMOType=White-Box, Variant=Node, Setting=Single-task2026.05 | 1.85 | — | — | |
| EoH-SType=White-Box, Variant=Node, Setting=Multi-task2026.05 | 1.85 | — | — | |
| EoH-SType=Black-Box, Variant=Node, Setting=Multi-task2026.05 | 1.86 | — | — | |
| ReEvo TD & BUType=Black-Box, Variant=Node, Setting=Multi-task2026.05 | 1.94 | — | — | |
| Efficient Decoder Scaling StrategyDepth=42, Width=128, Decoding Strategy=beam search, Beam Size=22026.02 | 3.1 | 4.3 | — | |
| Efficient Decoder Scaling StrategyDepth=42, Width=512, Decoding Strategy=greedy2026.02 | 3.5 | 11 | — | |
| LEHDDecoding Strategy=beam search, Beam Size=162026.02 | 3.8 | 4.4 | — | |
| ReEvo TD & BUType=White-Box, Variant=Edge, Setting=Multi-task2026.05 | 5.2 | — | — | |
| EoH-SType=Black-Box, Variant=Edge, Setting=Multi-task2026.05 | 5.21 | — | — | |
| EoH-SType=White-Box, Variant=Edge, Setting=Multi-task2026.05 | 5.57 | — | — | |
| ReEvo TD & BUType=Black-Box, Variant=Edge, Setting=Multi-task2026.05 | 6.13 | — | — | |
| ReEvo BUType=White-Box, Variant=Edge, Setting=Single-task2026.05 | 6.18 | — | — | |
| ReEvo TDType=White-Box, Variant=Edge, Setting=Single-task2026.05 | 6.63 | — | — | |
| Efficient Decoder Scaling StrategyDepth=42, Width=128, Decoding Strategy=greedy2026.02 | 7 | 2 | — | |
| DIFUSCO + LSType=White-Box, Variant=Edge, Setting=Single-task2026.05 | 7.33 | — | — | |
| Efficient Decoder Scaling StrategyDepth=6, Width=512, Decoding Strategy=beam search, Beam Size=22026.02 | 9 | 3.2 | — | |
| SIGDDecoding Strategy=beam search, Beam Size=162026.02 | 9.2 | 28 | — | |
| Efficient Decoder Scaling StrategyDepth=6, Width=512, Decoding Strategy=greedy2026.02 | 19 | 1.7 | — | |
| Efficient Decoder Scaling StrategyDepth=6, Width=128, Decoding Strategy=beam search, Beam Size=22026.02 | 24.6 | 1.1 | — | |
| BQDecoding Strategy=greedy2026.02 | 34.8 | 1.18 | — | |
| Efficient Decoder Scaling StrategyDepth=6, Width=128, Decoding Strategy=greedy2026.02 | 46.4 | 0.34 | — | |
| LEHDDecoding Strategy=greedy2026.02 | 57.7 | 0.29 | — | |
| SIGDDecoding Strategy=greedy2026.02 | 85.1 | 2.4 | — |