Combinatorial Optimization on OP-ACO
13.463Mean (N=50)Baseline heuristic
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Baseline heuristic2026.05 | 13.463 | 16.019 | 24.38 | 26.635 | 36.675 | 41.385 | — | — | |
| AHD AgentBackbone=GPT-4o2026.05 | 13.732 | 14.344 | 24.995 | 24.787 | 38.599 | 38.297 | 7 | 0.089 | |
| AHD AgentBackbone=DeepSeek-V4-Flash2026.05 | 14.948 | 6.736 | 29.208 | 12.102 | 50.783 | 18.841 | 19 | 0.065 | |
| ReEvoBackbone=DeepSeek-V4-Flash2026.05 | 14.952 | 6.687 | 29.118 | 12.368 | 51.161 | 18.257 | 100 | 0.047 | |
| MCTS-AHDBackbone=GPT-4o2026.05 | 14.977 | 6.358 | 29.863 | 9.9 | 52.892 | 15.23 | 100 | 0.93 | |
| CALMBackbone=Qwen3-4B-Instruct-25072026.05 | 14.989 | 6.471 | 29.809 | 10.289 | 53.176 | 15.036 | 150 | 0.074 | |
| ReEvoBackbone=GPT-4o2026.05 | 15.043 | 6.136 | 29.307 | 11.822 | 51.817 | 17.183 | 100 | 0.903 | |
| EOHBackbone=DeepSeek-V4-Flash2026.05 | 15.06 | 6.071 | 30.096 | 14.022 | 53.815 | — | 100 | 0.018 | |
| EOHBackbone=GPT-4o2026.05 | 15.084 | 5.89 | 29.651 | 10.786 | 52.299 | 16.442 | 100 | 0.343 | |
| MCTS-AHDBackbone=DeepSeek-V4-Flash2026.05 | 15.141 | 5.53 | 30.27 | 8.933 | 54.298 | 13.252 | 100 | 0.074 | |
| AHD AgentBackbone=Qwen3-4B-Instruct-25072026.05 | 15.173 | 5.34 | 30.364 | 8.624 | 54.555 | 12.837 | 13.6 | 0.005 | |
| AHD Agent w/SRBackbone=Qwen3-4B-Instruct-25072026.05 | 15.361 | 4.184 | 30.97 | 6.601 | 55.836 | 10.467 | 100 | 0.034 | |
| Optimal2026.05 | 16.019 | 0 | 33.234 | 0 | 62.595 | 0 | — | — |