Capacitated Vehicle Routing Problem on CVRP 20
6.0801Objective ValueR2E-IG-SGBS
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| R2E-IG-SGBSSearch Strategy=SGBS, Beam Width=4, Expansion Factor=42026.05 | 6.0801 | -0.048 | 4.2 | |
| LKH-3Search Strategy=LKH-32026.05 | 6.0831 | — | 3 | |
| R2E-IGSearch Strategy=Greedy2026.05 | 6.0907 | 0.1263 | 1.3 | |
| POMOSearch Strategy=Greedy2026.05 | 6.0974 | 0.2355 | 0.7 | |
| Sym-NCOSearch Strategy=Greedy2026.05 | 6.0984 | 0.2517 | 2 | |
| GurobiTime=13h2026.03 | 6.1 | 0 | — | |
| HGS2026.03 | 6.1 | 0 | 0.2 | |
| Gurobi2026.06 | 6.1 | 0 | — | |
| AMDKDSearch Strategy=Greedy2026.05 | 6.1041 | 0.3457 | 0.7 | |
| HGSTime=8.2h2026.03 | 6.11 | 0.16 | — | |
| AlignOPTvariant=STFT2026.03 | 6.13 | 0.49 | 0.051 | |
| LKH3Time=2h2026.03 | 6.14 | 0.65 | — | |
| POMO×8AugAugmentation=8x, Time=5s2026.03 | 6.14 | 0.65 | — | |
| Constraints Matrix Diffusion based Generative Neural Solver (×8Aug)Augmentation=8x, Time=6s2026.03 | 6.14 | 0.65 | — | |
| two-stage DRL-based neural heuristicTime(s)=0.022026.06 | 6.14 | 0.66 | — | |
| ReLD ×8AugAugmentation=8x, Time=3s2026.03 | 6.15 | 0.89 | — | |
| OR ToolsTime=*1000s2026.03 | 6.18 | 1.31 | — | |
| OR tools2026.03 | 6.18 | 1.3 | 0.27 | |
| AlignOPTvariant=MTFT2026.03 | 6.18 | 1.31 | 0.051 | |
| R2E-IG-SGBSSearch Strategy=SGBS, Beam Width=4, Expansion Factor=42026.05 | 6.1948 | -0.0357 | 4.2 | |
| LKH-3Search Strategy=LKH-32026.05 | 6.197 | — | 3.1 | |
| Constraints Matrix Diffusion based Generative Neural SolverTime=2s2026.03 | 6.2 | 1.64 | — | |
| GOAL2026.03 | 6.2 | 1.5 | 0.013 | |
| R2E-IGSearch Strategy=Greedy2026.05 | 6.2049 | 0.1281 | 1.2 | |
| GASETime=2s2026.03 | 6.21 | 1.8 | — | |
| POMOSearch Strategy=Greedy2026.05 | 6.212 | 0.2422 | 0.7 | |
| Sym-NCOSearch Strategy=Greedy2026.05 | 6.2129 | 0.2561 | 1.9 | |
| AMDKDSearch Strategy=Greedy2026.05 | 6.2208 | 0.3841 | 0.7 | |
| DARSearch Strategy=Greedy2026.05 | 6.2243 | 2.3226 | 0.3 | |
| AMTime(s)=0.052026.06 | 6.23 | 2.11 | — | |
| AM SamplingSearch strategy=Sampling, Time=3m2026.03 | 6.24 | 2.3 | — | |
| LNCS2026.03 | 6.25 | 2.51 | 0.315 | |
| PSOTime(s)=0.172026.06 | 6.25 | 2.38 | — | |
| E-GATTime=2s2026.03 | 6.26 | 2.6 | — | |
| ALNSTime(s)=0.632026.06 | 6.26 | 2.62 | — | |
| ELGSearch Strategy=Greedy2026.05 | 6.3189 | 3.8762 | 0.6 | |
| Parallel saving2026.03 | 6.33 | 3.85 | 0.01 | |
| DARSearch Strategy=Greedy2026.05 | 6.3356 | 2.2372 | 0.3 | |
| POMOTime=1s2026.03 | 6.35 | 4.09 | — | |
| ELG×8Aug ensemble 0.4 Local SizeAugmentation=8x, Ensemble Local Size=0.4, Time=4s2026.03 | 6.35 | 4.16 | — | |
| LEHD GreedySearch strategy=Greedy, Time=2s2026.03 | 6.36 | 4.26 | — | |
| AMTime=1s2026.03 | 6.4 | 4.97 | — | |
| ELGSearch Strategy=Greedy2026.05 | 6.4286 | 3.738 | 0.6 | |
| Omni-VRPSearch Strategy=Greedy2026.05 | 6.5057 | 6.9477 | 0.3 | |
| PtrNetTime=<1s2026.03 | 6.59 | 8.03 | — | |
| Omni-VRPSearch Strategy=Greedy2026.05 | 6.6162 | 6.7641 | 0.3 | |
| Sweep heuristic2026.03 | 7.51 | 23.17 | 0.01 | |
| ACO2026.03 | 7.72 | 26.56 | 0.8 | |
| ReEvobackbone=ACO-based2026.03 | — | 13.7 | — | |
| ReVELbackbone=ACO-based2026.03 | — | 4.57 | — |