Let $(X, \Omega, \mu)$ be a $\sigma-$ finite measure space and for $\phi \in L^\infty(\mu)$ let $M_\phi:L^p(\mu) \to L^p(\mu)$ defined by $M_\phi f = \phi f $ be the multiplication operator. Give necessary and sufficient conditions on $(X, \Omega, \mu)$ and $\phi$ for $M_\phi$ to be compact.

I can show if X is infinite,$\mu$ is the lebesgue measure, $\phi=0$ then $M_\phi$ is compact or if X is finite, for every measure and $\phi$, $M_\phi$ is compact. But for other measures I can not show. Please help me. Thanks in advance.


A common way to prove that some operator $T$ is not compact is to exhibit an infinite-dimensional subspace $M$ on which $T$ has a lower bound: that is, there exists $c>0$ such that $$\|Tx\|\ge c\|x\|,\quad \forall\ x\in M \tag{1}$$ If (1) holds, then the image of unit ball under $T$ contains a ball of radius $c$ in the infinite-dimensional subspace $TM$, and therefore is not compact.

To have (1), it suffices to have $|\phi|\ge c$ on the support of $f$. This leads to a necessary condition: for every $c>0$, the subspace $$\{f\in L^p: f=0 \text{ on the set $|\phi|< c$}\}\tag2 $$ is finite-dimensional.

The finite dimensionality amounts to the requirement that the set $\{|\phi|\ge c\}$ consists of finitely many atoms.

This necessary condition is also sufficient: if it holds, the multiplication operator is a norm-limit of finite rank operators, namely multiplication by $\phi$ restricted to $\{|\phi|\ge c\}$. Indeed, let $\phi_n:= \phi \chi_{\{|\phi|\ge 1/n\}}$. The norm of multiplication by $\phi-\phi_n$ is at most $1/n$. And the multiplication by $\phi_n$ produces functions in the space (2) with $c=1/n$, which is finite-dimensional.

  • $\begingroup$ Thanks. Please introduce to me a reference about it. $\endgroup$ – niki Sep 2 '14 at 12:29
  • $\begingroup$ I think we should suppose $\{f\in L^p$: f=0 on the set $|\phi|<c\}$ is finite - dimensional, because we want to get a contradiction by $|\phi|\geq c$ on the support of f. Am I right? $\endgroup$ – niki Sep 3 '14 at 16:19
  • $\begingroup$ @niki Yes, you are right. I corrected the statement. Sorry, I don't have a reference. It's just something that most functional analysis books explain at some point. $\endgroup$ – user147263 Sep 3 '14 at 16:50
  • $\begingroup$ I think for sufficient part we can not show the multiplication operator is a norm-limit of such finite rank operators. In the other word I can not show $M_{\phi_n}$ is finite rank where $\phi_n:= \phi_{|\{|\phi|<\frac{1}{n}\}}$. Please explain more about this part. $\endgroup$ – niki Sep 4 '14 at 7:20
  • $\begingroup$ @niki You need $>1/n$ in the definition of $\phi_n$. See the revised answer, last paragraph. $\endgroup$ – user147263 Sep 4 '14 at 15:15

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.