I want to prove the following theorem:


Let $X$ be a topological vector space. If $C \subseteq X$ is a convex subset such that $0 \in C^\circ$, then

  1. the Minkowski functional $p_C$ is sublinear
  2. $ C^\circ = \{ x \in X : p(x) < 1 \} $
  3. $ \overline{C} = \{ x \in X : p(x) \leq 1 \} $

What I know:

The assumption $0 \in C^\circ$ implies that $C$ is absorbing. Since $C$ is also convex, the function $p_C$ is sublinear. Now, let $ A = \{ x \in X : p_C(x) < 1 \} $ and $ B = \{ x \in X : p_C(x) \leq 1 \} $. Since $p_C$ is sublinear, we have $ p_A = p_B = p_C $.

The problem is that $ p_A = p_B = p_C $ does not directly imply $ C^\circ = A $ and $ \overline{C} = B $. How should I proceed?


I will give a proof of 2) and leave 3) to you.

If $x \in C^{0}$ then $(1+\frac 1 n) x \in C$ for $n$ sufficiently large and this implies $p((1+\frac 1 n) x) \leq 1$. It follows that $p(x) \leq \frac 1 {1+\frac 1 n}<1$.

Now let $p(x) <1$. Let $0<r<1-p(x)$ and consider $x+ rC^{0}$. This is an open set containing $x$. let us show that it is contained in $C$. For $c \in C^{0}$ we have $p(x+rc) \leq p(x)+rp(c)\leq p(x)+r <1$ To finish the proof we need the fact that $p(y) <1$ implies $y \in C$. By definition of $p$ there exists $t<1$ such that $y \in tC$. But then $y=t(\frac y t)+(1-t)0 \in C$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.