Closed Convex Cones in LP

Tue, 02 Jan 2024 00:13:03 +0000

I was formalizing some linear programming in Lean and came across one of the most surprisingly difficult to prove theorems.

::: {#thm-closed-convex-cones}

Let $A : \mathbb{R}^m \to \mathbb{R}^n$ be a linear transformation. Then the set $$ { A x : : : x \ge 0 } $$ is closed in $\mathbb{R}^n$. :::

The naive topological approach doesn’t work as this theorem is no longer true if we replace $A$ by an arbitrary continuous map or if we replace $x \ge 0$ by an arbitrary closed set.

Convex Analysis on PostDoc Problems

Closed Convex Cones in LP