Capacitated Vehicle Routing Problem (Constructive) on CVRP-Constructive N=200 (val)
32.301Mean Objective ValueOptimal
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Optimal2026.05 | 32.301 | 0 | |
| AHD Agent w/SRModel=Qwen3-4B-Instruct-2507, Eval=100, Cost ($)=0.0332026.05 | 38.155 | 18.325 | |
| AHD AgentModel=Qwen3-4B-Instruct-2507, Eval=14.6, Cost ($)=0.0042026.05 | 38.166 | 18.369 | |
| AHD AgentModel=DeepSeek-V4-Flash, Eval=20.4, Cost ($)=0.0672026.05 | 39.12 | 21.317 | |
| ReEvoModel=DeepSeek-V4-Flash, Eval=100, Cost ($)=0.0372026.05 | 39.658 | 23.054 | |
| MCTS-AHDModel=GPT-4o, Eval=100, Cost ($)=0.7022026.05 | 39.809 | 23.577 | |
| MCTS-AHDModel=DeepSeek-V4-Flash, Eval=100, Cost ($)=0.0742026.05 | 40.138 | 24.609 | |
| EOHModel=DeepSeek-V4-Flash, Eval=100, Cost ($)=0.0162026.05 | 40.216 | 24.747 | |
| ReEvoModel=GPT-4o, Eval=100, Cost ($)=0.6422026.05 | 40.521 | 25.878 | |
| EOHModel=GPT-4o, Eval=100, Cost ($)=0.3322026.05 | 40.988 | 27.34 | |
| AHD AgentModel=GPT-4o, Eval=12, Cost ($)=0.1142026.05 | 41.255 | 28.094 | |
| Baseline heuristic2026.05 | 42.466 | 32.037 | |
| CALMModel=Qwen3-4B-Instruct-2507, Eval=150, Cost ($)=0.0942026.05 | 43.236 | 34.588 |