Let $( \mathbb{E}\, , \leq )$ be a Riesz (vector lattice) space and let $\| \cdot \|$ be a map from $\mathbb{E}$ to $[ 0, \infty]$ that, when restricted to $$ \mathbb{B} := \{ f \in \mathbb{E} \, \colon \, \| f \| < + \infty \},$$ $ \| \cdot \|$ becomes a complete Riesz (lattice) norm on $\mathbb{B}$.

In particular, $( \mathbb{B}\, , \| \cdot \| )$ is a Banach lattice contained in $\mathbb{E}$. We denote by $\mathbb{E}_+$ and $\mathbb{B}_+$ be the positive cones of $( \mathbb{E}\, , \leq )$ and $( \mathbb{B}\, , \| \cdot \| )$, respectively.

Let $\Phi$ be an injective map from $\mathbb{E}_+$ to itself, i.e., $$ \Phi \, \colon \mathbb{E}_+ \to \mathbb{E}_+ .$$

The preimage of the positive cone $\mathbb{B}_+ \subset \mathbb{E}_+$ under $\Phi$ is the subset of $\mathbb{E}_+$ for which $$\Phi^{-1} (\mathbb{B}_+) := \{ f \in \mathbb{E}_+ \, \colon \, \Phi (\, f \,) \in \mathbb{B}_+ \}.$$

I am quite confused that whether $\Phi \left( \Phi^{-1} (\mathbb{B}_+) \right) = \mathbb{B}_+$ holds? Could anyone help me out please?

Thank you very much in advance!


This is only true if $\Phi$ is surjective.

Let $E_+= \mathbb R^n_+$. Then $\Phi$ defined by $\Phi(x) = x + e$ is injective and maps into $E_+$ for $e\ne0$ and $e\ge0$. However, it is not surjective, as $\Phi(x)\ge e$ for all $x\in E_+$.

  • $\begingroup$ Thanks so much @daw. As you mentioned, if $\Phi (\Phi^{-1} ( \mathbb{B}_+))=\mathbb{B}_+$ holds, then the mapping $\Phi \colon \mathbb{E}_+ \to \mathbb{E}_+$ must be needed to be surjective. May I confirm that is my understanding correct please? Many thanks again:) $\endgroup$ – Paradiesvogel Apr 28 '18 at 9:29
  • $\begingroup$ yes, it is correct $\endgroup$ – daw Apr 28 '18 at 12:52

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.