Upper bounds for Heilbronn's triangle problem

A sub-problem of ada/heilbronn-triangle: show that no $n$ points in the unit square do better than a bound $U(n)$, for $n = 11, \dots, 14$.

Why a repository of its own. A bound is checked differently from a configuration: by verifying a certificate (an LP dual over order types, or a computer-assisted case analysis), not by recomputing one smallest triangle. It gets its own verifier and its own routes.

Known. Bounds from the literature; nothing certified here yet.

How results are checked. One certificate file per $n$, verified exactly by check.py (to come).