Heilbronn's triangle problem
Heilbronn's triangle problem in the unit square: configurations of 11 to 16 points, every one checked in exact arithmetic

| 1 | --- |
| 2 | number: 7 |
| 3 | title: "Get within 5% of a best known value with a search of your own" |
| 4 | state: open |
| 5 | author: bo |
| 6 | created_at: 2026-11-11T03:35:20Z |
| 7 | labels: "open,size-m" |
| 8 | --- |
| 9 | |
| 10 | **Open.** `examples/search.py` reaches about 87% of the best known value for n = 11 and about half of it for n = 16. How close does a better search get, and at what cost? |
| 11 | |
| 12 | **Do.** Write a search (restarts, basin hopping, a local refinement of the smallest triangles that tie: your choice), run it on one instance, and push the configuration with the search and its seed in the same folder. |
| 13 | |
| 14 | **Closes when** one route holds a configuration found by its own search whose `verify` status is at least 95% of the best known value: 0.0352 for n = 11, 0.0310 for n = 12, 0.0257 for n = 13, 0.0231 for n = 14, 0.0202 for n = 15, 0.0196 for n = 16. A route that tried and fell short is a result too: say how far it got. |