Combinatorial Optimization on OVRP-Constructive
6.574Mean Objective (N=50)Optimal
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Optimal2026.05 | 6.574 | 0 | 10.636 | 0 | 18.491 | 0 | — | — | |
| AHD Agent w/SRBackbone=Qwen3-4B-Instruct-25072026.05 | 12.083 | 83.485 | 20.523 | 92.892 | 37.236 | 101.272 | 100 | 0.035 | |
| AHD AgentBackbone=Qwen3-4B-Instruct-25072026.05 | 12.103 | 83.756 | 20.551 | 93.153 | 37.247 | 101.337 | 10.7 | 0.004 | |
| CALMBackbone=Qwen3-4B-Instruct-25072026.05 | 12.83 | 94.908 | 21.966 | 106.506 | 39.403 | 113.158 | 150 | 0.08 | |
| ReEvoBackbone=DeepSeek-V4-Flash2026.05 | 12.88 | 95.659 | 22.115 | 107.953 | 39.69 | 114.745 | 100 | 0.04 | |
| AHD AgentBackbone=DeepSeek-V4-Flash2026.05 | 12.885 | 95.744 | 21.977 | 106.58 | 39.232 | 112.351 | 16.6 | 0.077 | |
| MCTS-AHDBackbone=DeepSeek-V4-Flash2026.05 | 13.039 | 98.139 | 22.11 | 107.846 | 39.682 | 114.685 | 100 | 0.054 | |
| EOHBackbone=DeepSeek-V4-Flash2026.05 | 13.05 | 98.372 | 22.35 | 110.188 | 39.724 | 114.98 | 100 | 0.016 | |
| MCTS-AHDBackbone=GPT-4o2026.05 | 13.086 | 98.942 | 22.193 | 108.69 | 39.42 | 113.459 | 100 | 0.728 | |
| EOHBackbone=GPT-4o2026.05 | 13.089 | 98.952 | 22.432 | 111.016 | 39.96 | 116.182 | 100 | 0.346 | |
| ReEvoBackbone=GPT-4o2026.05 | 13.341 | 102.901 | 22.645 | 113.117 | 40.221 | 117.773 | 100 | 0.678 | |
| AHD AgentBackbone=GPT-4o2026.05 | 14.145 | 115.091 | 23.503 | 121.105 | 41.45 | 124.441 | 7.2 | 0.087 | |
| Baseline heuristic2026.05 | 14.413 | 119.248 | 23.711 | 123.114 | 41.9 | 126.979 | — | — |