7
$\begingroup$

I’ve seen $e^{\frac{d}{dx}}$ yield the shift operator, but would $e^{I}$ where $If(x)\equiv \int f(x)dx$ yield something interesting as well (or even something at all)? I’d imagine that it’s defined as the integral operator is better behaved than the differentiation operator, but I it interesting and does this operator have a name?

$e^I$ is just shorthand for $\sum_{n=0}^{\infty} \frac{I^n}{n!}$ as is oftentimes done in operator theory and related fields.

$\endgroup$
1
  • $\begingroup$ Whether the formal power series yields a well-defined operator will depend on the function space you're working with; a sufficient pair of conditions for the series to be well-defined is for your space to be a Banach space on which $I$ is bounded. These properties guarantee that the partial sums of the series are a Cauchy sequence in a complete space. One concrete example of such a space are continuous functions on a compact interval under the supremum norm. I don't know whether $e^I$ is interesting so I'll leave that for someone else. $\endgroup$
    – user7530
    Jun 29 at 6:11

1 Answer 1

5
$\begingroup$

$e^I$ has a nice expression in terms of an integral kernel, but I don't know of any uses for it.

Rather than work with quotient vector spaces, let $$(If)(x)=\int_0^x{f(t)\,dt}$$ Then, for all $n\in\mathbb{Z}^+$, $$(I^nf)(x)=\int_0^x{\frac{(x-t)^{n-1}}{(n-1)!}f(t)\,dt}$$ a result due to Cauchy and easy to prove by induction.

Applying Cauchy's formula and noting that power series converge uniformly, \begin{align*} (e^If)(x)&=f(x)+\sum_{n=1}^{\infty}{\frac{1}{n!}\int_0^x{\frac{(x-t)^{n-1}}{(n-1)!}f(t)\,dt}} \\ &=f(x)+\int_0^x{f(t)\sum_{n=1}^{\infty}{\frac{(x-t)^{n-1}}{(n-1)!n!}}\,dt} \\ &=f(x)+\int_0^x{\frac{I_1(2\sqrt{x-t})}{\sqrt{x-t}}f(t)\,dt} \end{align*} where $I_1$ solves the ODE \begin{gather*} x^2I_1''(x)+xI_1'(x)-(1+x^2)I_1(x)=0 \\ I_1(0)=0 \end{gather*}

$\endgroup$
2
  • $\begingroup$ Do you know any interpretation of the final ODE? $\endgroup$
    – Bob
    Jun 29 at 7:20
  • 1
    $\begingroup$ @Bob: It's just the modified Bessel ODE. They're special functions that show up whenever you have a power series with a double-factorial in the denominator (like here); IIRC they have physical interpretation as the hyperbolic harmonics: the radial factor in solutions to Laplace's equation with hyperbolic symmetry. $\endgroup$ Jun 30 at 1:44

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.