Traveling Salesman Problem on TSP n=100 10K instances (test)
7.76Objective ValueConcorde
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Concorde2021.10 | 7.76 | — | — | — | — | |
| LKH2021.10 | 7.76 | 0 | 6 | — | — | |
| ConcordeMethod type=Heuristic2022.05 | 7.76 | 0 | — | — | 3 | |
| LKH3Method type=Heuristic2022.05 | 7.76 | 0 | — | — | 21 | |
| ConcordeTime=(82m)2022.07 | 7.765 | — | — | — | — | |
| LKH3Time=(8h)2022.07 | 7.765 | 0 | — | — | — | |
| DPDPTime=(2h)2022.07 | 7.765 | 0.001 | — | — | — | |
| ConcordeEvaluation Protocol=Inference (10k instances)2022.10 | 7.765 | 0 | — | — | — | |
| LKH3Evaluation Protocol=Inference (10k instances)2022.10 | 7.765 | 0 | — | — | — | |
| Concorde2024.05 | 7.765 | — | — | — | 82 | |
| LKH32024.05 | 7.765 | 0 | — | — | 8 | |
| DPDP2024.05 | 7.765 | 0.4 | — | — | 2 | |
| COMPASS2024.05 | 7.765 | 0.2 | — | — | 2 | |
| LR+SGBS+EAS(short)Leader Reward=True, duration=short2024.05 | 7.765 | 0.07 | — | — | 25 | |
| LR+SGBS+EASLeader Reward=True, duration=long2024.05 | 7.765 | 0.02 | — | — | 2 | |
| SGBS+EASStage=Converged, Time=(15h)2022.07 | 7.766 | 0.024 | — | — | — | |
| LR+POMOLeader Reward=True2024.05 | 7.766 | 1.4 | — | — | 1 | |
| SGBS+EASStage=Intermediate, Time=(3h)2022.07 | 7.767 | 0.035 | — | — | — | |
| EASStage=Converged, Time=(15h)2022.07 | 7.768 | 0.044 | — | — | — | |
| SGBS+EASduration=long2024.05 | 7.768 | 4.5 | — | — | 2 | |
| EASStage=Intermediate, Time=(3h)2022.07 | 7.769 | 0.053 | — | — | — | |
| SGBSHyperparameters=(10,10), Time=(9m)2022.07 | 7.769 | 0.058 | — | — | — | |
| DACTData Augmentation=x42021.10 | 7.77 | 0.09 | 14 | 3.5 | — | |
| DPDPInference Steps=100k2021.10 | 7.77 | 0 | 8 | — | — | |
| PoppyEvaluation Protocol=Inference (10k instances), Population size=162022.10 | 7.77 | 0.07 | — | — | — | |
| Poppy 16k=162024.05 | 7.77 | 7 | — | — | 1 | |
| SGBS+EAS(short)duration=short2024.05 | 7.77 | 6.3 | — | — | 37 | |
| DACTTime=(8h)2022.07 | 7.771 | 0.089 | — | — | — | |
| POMOSearch Strategy=sampling, Time=(3h)2022.07 | 7.771 | 0.078 | — | — | — | |
| POMO(sampling)mode=sampling2024.05 | 7.771 | 7.8 | — | — | 3 | |
| PoppyEvaluation Protocol=Inference (10k instances), Population size=82022.10 | 7.772 | 0.09 | — | — | — | |
| POMOSearch Strategy=greedy, Time=(1m)2022.07 | 7.776 | 0.144 | — | — | — | |
| POMOmode=greedy2024.05 | 7.776 | 14.4 | — | — | 1 | |
| PoppyEvaluation Protocol=Inference (10k instances), Population size=42022.10 | 7.777 | 0.15 | — | — | — | |
| POMOData Augmentation=x82021.10 | 7.78 | 0.18 | 2 | — | — | |
| POMOEvaluation Protocol=Inference (10k instances), Number of samples=162022.10 | 7.785 | 0.27 | — | — | — | |
| Neural-2-OptInference Steps (T)=6k2021.10 | 7.79 | 0.39 | 3.5 | — | — | |
| DACTInference Steps (T)=10k2021.10 | 7.79 | 0.37 | 3.5 | 50 | — | |
| MDAM-BS2021.10 | 7.79 | 0.38 | 44 | — | — | |
| Sym-NCORollout strategy=multi-start, Sampling size=1002022.05 | 7.79 | 0.39 | — | — | 13 | |
| MDAMEvaluation Protocol=Inference (10k instances), Search strategy=beam search2022.10 | 7.79 | 0.38 | — | — | — | |
| POMOEvaluation Protocol=Inference (10k instances), Ensemble=true, Number of samples=162022.10 | 7.79 | 0.33 | — | — | — | |
| POMOEvaluation Protocol=Inference (10k instances)2022.10 | 7.796 | 0.41 | — | — | — | |
| POMO2021.10 | 7.8 | 0.46 | 12 | — | — | |
| POMORollout strategy=multi-start, Sampling size=1002022.05 | 7.8 | 0.44 | — | — | 13 | |
| MDAMSearch strategy=beam search, Beam width=30, Ensemble size=52022.05 | 7.8 | 0.48 | — | — | 20 | |
| Neural-2-OptInference Steps (T)=3k2021.10 | 7.81 | 0.61 | 1.5 | — | — | |
| DACTInference Steps (T)=5k2021.10 | 7.81 | 0.61 | 1.5 | 25 | — | |
| 2-Opt-DLEvaluation Protocol=Inference (10k instances)2022.10 | 7.83 | 0.87 | — | — | — | |
| Sym-NCORollout strategy=greedy2022.05 | 7.84 | 0.94 | — | — | 2 | |
| POMORollout strategy=greedy2022.05 | 7.85 | 1.04 | — | — | 2 | |
| GCN-BS2021.10 | 7.87 | 1.39 | 40 | — | — | |
| LIHEvaluation Protocol=Inference (10k instances)2022.10 | 7.87 | 1.42 | — | — | — | |
| Wu et al.Inference Steps (T)=5k2021.10 | 7.88 | 1.54 | 2 | — | — | |
| DACTInference Steps (T)=1k2021.10 | 7.89 | 1.62 | 20 | 5 | — | |
| MDAMRollout strategy=greedy, Ensemble size=52022.05 | 7.93 | 2.19 | — | — | 36 | |
| MDAMEvaluation Protocol=Inference (10k instances), Search strategy=greedy2022.10 | 7.93 | 2.19 | — | — | — | |
| AM-sampling2021.10 | 7.94 | 2.26 | 1 | — | — | |
| AMRollout strategy=multi-start, Sampling size=12802022.05 | 7.94 | 2.26 | — | — | 41 | |
| Wu et al.Inference Steps (T)=1k2021.10 | 8.02 | 3.31 | 24 | — | — | |
| OR-Tools2021.10 | 8.06 | 3.86 | 23 | — | — | |
| AMRollout strategy=greedy2022.05 | 8.12 | 4.53 | — | — | 2 | |
| S2V-DQNRollout strategy=greedy2022.05 | 8.31 | 7.03 | — | — | — | |
| PointerNetRollout strategy=greedy2022.05 | 8.6 | 6.9 | — | — | — | |
| Att-GCRN+MCTSEvaluation Protocol=Inference (10k instances)2022.10 | — | 0.037 | — | — | — | |
| CVAE-Opt-DE2021.10 | — | 0.34 | — | — | — |