For a bounded function $\operatorname{F}: \mathbb{R}_{\,\ge\ 0} \to \mathbb{R}$ ( not necessarily non-negative ), is it true that $$ \int_{0}^{\infty}\frac{x^{k}\,s}{(s^{2} + x^{2})^{\left(k + 3\right)/2}\,\,}\, \operatorname{F}\left(x\right)\,{\rm d}x = 0\quad \forall s > 0 \iff \operatorname{F} \equiv 0 $$ where $k \in \mathbb{N}$ is a positive constant $?$ Of course, one implication ($\leftarrow$) is true. What about the other one $?$.

  • $\begingroup$ Do you mean $F:\mathbb{R}^{\geq 0} \to \mathbb{R}^{\geq 0}$ above? $\endgroup$ – anomaly Jan 23 at 15:19
  • 1
    $\begingroup$ @anomaly No, F does not need to be positive $F: \mathbb{R}_{\ge 0} \to \mathbb R$ and is just bounded. $\endgroup$ – Jun Jan 23 at 15:21
  • $\begingroup$ Does $F$ need to be continuous? I'm pretty sure I could construct some sort of pathological function that would break this idea, but it wouldn't be continuous. EDIT: In response to your question edit, if we require $F\geq 0$ the statement is very easy to prove. $\endgroup$ – K.defaoite Jan 23 at 15:36
  • $\begingroup$ @K.defaoite No $F$ does not need to be continuous (although I would also be interested in seeing a positive result assuming continuity) $\endgroup$ – Jun Jan 23 at 15:47
  • $\begingroup$ The converse is always false. Think about it like this: if I tell you that some weighted average of $F$ is zero, can you say that $F$ is always equal to zero? $\endgroup$ – Thomas Bakx Jan 23 at 15:56


$$f_s(x) = \frac{x^ks}{(s^2+x^2)^{(k+3)/2}}$$

Notice that any such function vanishes at infinity. Let $A$ be the subalgebra of $\textbf{C}_0((0, \infty), \mathbb{R})$ generated by the family $f_s$. All these functions are positive on $(0, \infty)$ so the family doesn't vanish at any point, and it clearly separates points.

Thus by the locally compact version of the Stone-Weierstrass Theorem, $A$ is dense in $\textbf{C}_0((0, \infty), \mathbb{R})$, and $F$ satisfies your integral equation with $f_s$ replaced by any element of $A$. Thus as a functional, using the Riesz Representation Theorem (using the uniqueness part) you get that $F \equiv 0$ a.e.

  • $\begingroup$ If you don't have the "locally compact version of Stone-Weierstrass" then just do your approximation on successively larger intervals $[\frac{1}{n}, n]$. $\endgroup$ – John Samples Jan 23 at 18:59
  • $\begingroup$ Thanks! Could you add some details about why the assumptions of the Stone-Weierstrass theorem are satisfied. In particular, why does the family $f_s$ generate a subalgebra of $C_0((0,\infty),\mathbb R)$? Why does it separate the points? $\endgroup$ – Jun Jan 27 at 10:26
  • $\begingroup$ You take the sub-algebra generated by the $f_n$, i.e. all finite linear combinations over $\mathbb{R}$. It's an algebra by definition. It separates any positive real (the case at $r = 0$ doesn't matter because a point is a null-set) because for any $r > 0$ you can find different $f_s$ that take different values on $r$. I mean, just write out the equation. If you're feeling very lazy, you can just add in functions $g_s = (1/2)f_s$ and take the algebra generated by both the $f$'s and $g$'s. Can you check the answer as correct? That's what you do on MSE when you get a correct answer. $\endgroup$ – John Samples Jan 29 at 15:33

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.