Combinatorial Optimization on MKP-ACO
16.497Mean Value (N=100)Baseline heuristic
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Baseline heuristic2026.05 | 16.497 | 3.686 | 41.069 | 7.689 | 52.783 | 8.048 | — | — | |
| EOHBackbone=GPT-4o2026.05 | 16.874 | 1.464 | 42.667 | 3.922 | 55.13 | 2.882 | 100 | 0.227 | |
| CALMBackbone=Qwen3-4B-Instruct-25072026.05 | 16.906 | 1.517 | 42.913 | 3.521 | 55.408 | 2.427 | 150 | 0.053 | |
| MCTS-AHDBackbone=GPT-4o2026.05 | 16.923 | 1.439 | 42.81 | 3.375 | 52.662 | 2.861 | 100 | 0.844 | |
| MCTS-AHDBackbone=DeepSeek-V4-Flash2026.05 | 16.941 | 1.311 | 42.651 | 4.13 | 55.569 | 2.451 | 100 | 0.05 | |
| AHD AgentBackbone=GPT-4o2026.05 | 16.943 | 1.168 | 43.201 | 2.91 | 55.674 | 2.178 | 5.8 | 0.058 | |
| ReEvoBackbone=GPT-4o2026.05 | 16.952 | 1.225 | 42.591 | 4.303 | 52.476 | 7.074 | 100 | 0.637 | |
| ReEvoBackbone=DeepSeek-V4-Flash2026.05 | 16.957 | 1.249 | 43.141 | 3.011 | 55.783 | 1.85 | 100 | 0.033 | |
| EOHBackbone=DeepSeek-V4-Flash2026.05 | 17.01 | 0.921 | 43.324 | 2.607 | 55.813 | 1.818 | 100 | 0.011 | |
| AHD AgentBackbone=DeepSeek-V4-Flash2026.05 | 17.021 | 0.86 | 43.301 | 2.659 | 55.789 | 1.906 | 15.2 | 0.04 | |
| AHD AgentBackbone=Qwen3-4B-Instruct-25072026.05 | 17.029 | 0.843 | 43.522 | 2.166 | 56.186 | 1.188 | 12.4 | 0.004 | |
| AHD Agent w/SRBackbone=Qwen3-4B-Instruct-25072026.05 | 17.058 | 0.669 | 43.84 | 1.442 | 56.565 | 0.592 | 100 | 0.033 | |
| Optimal2026.05 | 17.162 | 0 | 44.507 | 0 | 57.007 | 0 | — | — |