Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X=(C([0,1]), \Vert \cdot \Vert_{\infty})$. Determine the spectrum of $$ \begin{split} M \colon & X \to X\\ & u(t) \mapsto \int_0^t h(s)u(s)ds \end{split} $$ where $h \in C([0,1])$ is fixed.

First of all, I have proved the operator is compact (by Ascoli-Arzelà). Hence $0 \in \sigma(M)$.

Now how can I find the other elements of $\sigma(M)$? I know that they are eigenvalues and they are either finite, either a sequence (converging to $0$), since $M$ is compact.

Here's what I've tried: let $\lambda \ne 0$ and eigenvalue $g(\cdot)$ a corresponding eigenvector. So I have $$ \lambda g(x) = \int_0^x h(s)g(s)ds $$ hence $g \in C^1$ and $$ \begin{cases} \lambda g'(x) = h(x)g(x)\\ g(0)=0 \end{cases} $$ Solving this Cauchy problem, I get $$ g(x)=\exp{\left( \frac{1}{\lambda}\int_0^x h(s)ds\right) } -1. $$ Now I don't manage to finish: I should find some conditions on $\lambda$, but can't see how...

Thanks in advance.

share|cite|improve this question
up vote 2 down vote accepted

Solving this Cauchy problem actually gives that $$\frac d{dx}\left(g(x)\exp\left(-\frac 1\lambda\int_0^xg(t)h(t)dt\right)\right)=0$$ so $g(x)\exp\left(-\frac 1\lambda\int_0^xg(t)h(t)dt\right)$ is constant on $[0,1]$. As $g(0)=0$, this constant is $0$ so $g$ is identically $0$ and $\lambda$ is not an eigenvalue of $M$.

share|cite|improve this answer
Here is a reason why this should not be surprising: The operator in question is a bit like a triangular matrix $G$, with “entries” $G_{xt}=g(t)$ if $t<x$, $G_{xt}=0$ otherwise. Stretching the analogy, you expect to find the eigenvalues on the diagonal; but the diagonal isn't “really there”, or more precisely, the diagonal in $[0,1]\times[0,1]$ has measure zero. (None of this is a substitute for actual proof, of course.) – Harald Hanche-Olsen Dec 25 '12 at 11:41
Hi Davide, thanks for your answer. I'm sorry, but I think there's a typo: it should be $\exp\left(\frac{-1}{\lambda}\int_0^x h(t)dt \right)$ (the integrand is not $g(x)h(x)$). Do you agree? Indeed, I was almost at the end but I made a mistake (I should have written a multiplicative constant, instead of an additive one). Thanks a lot for your help. – Romeo Dec 25 '12 at 13:31
@HaraldHanche-Olsen Very interesting, thanks for your observation. – Romeo Dec 25 '12 at 13:32

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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