1
$\begingroup$


I am currently learning about different kinds of functional calculus and I was wondering if I could get something cleared up.
The first type of functional calculus we learned about was holomorphic functional calculus, i.e., given a self adjoint operator $H$ and a function $f$ that is holomorphic in a neighbourhood of the spectrum, we can define $$f_\Gamma(H)=\int_{\Gamma}f(\lambda)(\lambda-H)^{-1}d\lambda $$ where $\Gamma$ is a contour contained in region of holomorphy of $f$ that encloses the spectrum.
We also talked about Borel functional calculus, i.e., for a self adjoint operator $H$ we can define $$f(H)=\int_{\sigma(H)} f(\lambda) dP(\lambda)$$ where $f$ is a borel function and we integrate over a projection valued measure with respect to Borel sets.
I've been told that the two functional calculi actually coincide, i.e., $$f_\Gamma(H)=f(H)$$for functions and operators that satisfy the requirements. I was wondering if there was an easy way to show that this was the case. Thank you for any help!

$\endgroup$
1
$\begingroup$

If $H$ is bounded and selfadjoint with spectrum $\sigma$, then $$ (\lambda I-H)^{-1}=\int_{\sigma}\frac{1}{\lambda-\mu}dE(\mu),\;\;\; \lambda\notin\sigma. $$ Suppose $\Gamma$ is a positively-oriented simple closed rectifiable curve in $\mathbb{C}$ which contains $\sigma$ in its interior, and suppose that $f$ is holomorphic on an open neighborhood of $\Gamma$ and its interior. Then $$ \begin{align} f(H)x & =\frac{1}{2\pi i}\oint_{\Gamma}f(\lambda)(\lambda I-H)^{-1}x\,d\lambda \\ & = \frac{1}{2\pi i}\oint_{\Gamma}f(\lambda)\left[\int_{\sigma}\frac{1}{\lambda-\mu}dE(\mu)x\right]d\lambda \\ & = \int_{\sigma}\left[\frac{1}{2\pi i}\oint_{\Gamma}\frac{f(\lambda)}{\lambda-\mu}\,d\lambda\right]dE(\mu)x \\ & = \int_{\sigma}f(\mu)dE(\mu)x. \end{align} $$

$\endgroup$

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.