2

Is this proof good? Given the problem as stated. I first define,

$$ g(x,b) = \frac{f(x)}{x}e^{ibx} $$

Which has the following property,

$$ g_b(x,b) = if(x)e^{ibx} $$

And that,

$$ |g_b(x,b)| = |if(x)e^{ibx}| = |i||f(x)||e^{ibx}| \leq |f(x)| $$

where since $f \in L^1$ we have a bounding function of $g$. Thus, we can differentiate under the integral,

$$ \frac{d}{db} \int_{-\infty}^{\infty} g(x,b)dx $$

My question before I do that, is whether or notI can split the integral like this because $f(0)=0$ and $f$ is differentiable at zero,

$$ \int_{-\infty}^{\infty} g(x,b)dx = \int_{-\infty}^{0} g(x,b)dx + \int_{0}^{\infty} g(x,b)dx $$

So that,

$$ \frac{d}{db} \int_{-\infty}^{\infty} g(x,b)dx = \frac{d}{db}\int_{-\infty}^{0} g(x,b)dx + \frac{d}{db}\int_{0}^{\infty} g(x,b)dx $$

And thus,

$$ \frac{d}{db}\int_{-\infty}^{0} g(x,b)dx + \frac{d}{db}\int_{0}^{\infty} g(x,b)dx = \int_{-\infty}^{0} g_b(x,b) dx + \int_{0}^{\infty} g_b(x,b)dx $$

And since $g_b$ is bounded by $f$ which is integrable, the two integrals exists. Thus, setting $b=0$, we attain the desired result.

IQ472
  • 173
  • 1
  • 8
  • 1
    "Is this proof good?" Much more complicated than needed, mainly. – Did Nov 16 '14 at 16:59
  • You're differentiating with respect to $t$ above, but should this be with respect to $b$? If not, it does not appear $t$ has been defined. – msteve Nov 16 '14 at 16:59
  • Ah! Sorry that was a typo. Fixed it. – IQ472 Nov 16 '14 at 17:10
  • @Did. So, the proof works? Could you explain a short way? – IQ472 Nov 16 '14 at 17:10
  • "So, the proof works?" At first glance, you seem to apply incorrectly the theorem of differentiation of integrals depending on a parameter. "Could you explain a short way?" See the answer by @user2345215. – Did Nov 16 '14 at 17:13
  • Just saw @user2345215. I appreciate the comments. How did I mis-apply the differentiation of integrals? – IQ472 Nov 16 '14 at 17:16

1 Answers1

0

Let me restate your question:

If $f$ is Lebesgue integrable, $f(0)=0$, and $f'(0)$ is finite, then $\frac{f(x)}x$ is Lebesgue integrable.

I didn't read your solution because it seemed needlessly complicated, so here's mine:

Obviously $\frac{f(x)}x$ is measurable, so we only have to find some bounds.

Firstly, $\lim_{x\to0}\frac{f(x)}x=\lim_{x\to0}\frac{f(x)-f(0)}x=f'(0)\in\mathbb R$, so it must be bounded on some neighborhood of $0$, namely there are $\delta>0$ and $K>0$ such that for all $x\in(-\delta,\delta)\setminus\{0\}$ we have $\Bigl|\frac{f(x)}x\Bigr|<K$. Therefore it is Lebesgue integrable on $(-\delta,\delta)$.

Secondly, $\Bigl|\frac{f(x)}x\Bigr|\le\frac{|f(x)|}\delta$ on $(-\infty,-\delta]\cup[\delta,\infty)$, so it's also Lebesgue integrable there since $\frac{f(x)}\delta$ is.

user2345215
  • 16,422
  • Thanks. I think I forgot that the derivative existing meant it was finite. – IQ472 Nov 16 '14 at 17:20
  • @IQ472 The derivative f'(0) is sometimes (but not always) allowed to be even $+\infty$ or $-\infty$, but differentiability surely means it's finite. But I guess your convention is that the derivative is always finite. – user2345215 Nov 16 '14 at 17:23