MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

I wants to understand the integrals of the form

$$\lambda (x) = \int_{0}^{\infty} \frac{\sin^{2} \alpha x}{\alpha^{2}} d\mu(\alpha), (\mu(0)=0)$$ where $\mu(\alpha)$ is a non decreasing function such that the integral converges for all real $x.$

Trivial Example: If we take, $\mu(\alpha)= \alpha,$ then $\lambda (x)= |x|\frac{\pi}{2}.$

My Questions are:

(1) Can you give few more examples of non decreasing function $\mu(\alpha)$, so that we know the exact value of $\lambda (x).$ ?

(2) Suppose the above integral ($\lambda (x)$) converges for the given non decreasing function $\mu(\alpha).$ Can we expect(to evolute) to get the exact(precise) values of $\lambda(x), (x\in \mathbb R)$ ?

(3) Why $\lambda (x)$ in the literature it is known as "negative definite function" ; can you motivates me bit ? (For instance, Arne Beurling, in his paper, "on the spectral synthesis of bounded functions" called this $\lambda (x)$ as negative definite functions)

(4) Does there exists any well-known analogue of ``negative definite functions" defined in $\mathbb R^{2} (\mathbb R^{n})$ ? If not, what one can expect ?


share|cite|improve this question
You should read some references on Stieltjes integral, for example this. For 1) you can take $\mu = \frac23 x^{3/2}$ and get $\sqrt{x\pi}$. For 2), if $\mu$ is $C^1$, $d\mu = \mu^\prime(\alpha) d\alpha$, so you can decide from this subcase. – username May 10 '14 at 9:20

(1) Denote by $\mu(\alpha)$ the Radon-Nikodym derivative $\frac{d\mu}{d\alpha}$. Notice that if we extend $\mu(\alpha)$ to the negative reals by $\mu(-\alpha)=\mu(\alpha)$, the integral becomes \begin{equation}-2\int_{-\infty}^{\infty}e^{-i\alpha x}\frac{\mu(\frac{\alpha}{2})}{\alpha^2}d\alpha+\frac{1}{2}\int_{0}^{\infty}\frac{\mu(\alpha)}{\alpha^2}d\alpha.\end{equation} Therefore, what needs to be computed is the Fourier transform of $\frac{\mu(\frac{\alpha}{2})}{\alpha^2}$. For example, if $\mu(\frac{\alpha}{2})=\alpha^{2-p}$ where $0<p<2$, the Fourier transform is \begin{equation} \frac{2(\sin \frac{\pi p}{2})\Gamma(1-p)}{|\alpha|^{1-p}}, \end{equation} (after that the original integral is easy to compute), and you can find more examples here. For example, the case $\mu(\alpha)=\alpha^2\log(1+|\alpha|)$ can be computed explicitly (although the result will be a distribution).

(2) Obviously, there is no general method for these integrals, but if you know the function $\mu$ well enough, and if it is analytic, you can compute the integral by the residue theorem. So for exaple if $\mu(\alpha)$ is an even rational function such that the integral converges, it can be computed.

(3) By Bochner's theorem, the Fourier transform of a nonnegative function is positive definite, so its negative is negative definite.

(4) Negative definite functions can even be defined on any group, see this.

share|cite|improve this answer

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.