Monday 17 February 2014

real analysis - Improper integral $int_0^{1/2}frac{mathrm{d}t}{t^a lvertln(t)rvert^b}$




I'm working in this problem and I'm having some problems.




Study the convergence of this improper integral:



$$\displaystyle\int_0^{\frac12}\dfrac{\mathrm{d}t}{t^a \lvert\ln(t)\rvert^b},\quad a,b>0$$




For $\boxed{a<1}$ I've compared it with the integral
$$\int_0^{\frac12}\dfrac{\mathrm{d}t}{t^a}$$

and found that is convergent. When $\boxed{a=0}$, taking $u=\ln(t)$ and $du=\frac{\mathrm{d}t}{t}$ we have:
$$\int_{0}^{\frac12}\dfrac{\mathrm{d}t}{t\lvert\ln(t)\rvert^b}=\int_{-\infty}^{\ln(\frac12)}\dfrac{\mathrm{d}u}{u^b}$$
which is convergent for $b>1$ and divergent for $b\leq 1$ (is this correct?).
When $\boxed{a>1}$ I think that diverges, but cannot prove it. Any hint?


Answer



Set $u=-\log{t}$. Then $t = e^{-u} $, so $dt = -e^{-u} \, du$. The limits become $\infty$ and $\log{2}$, and we have
$$ \int_{\log{2}}^{\infty} u^{-b} e^{-(1-a)u} \, du $$
Now, the integrand is bounded, so the problem is only for large $u$. In particular, there are now several cases to examine:





  1. $a>1$. The integral diverges, because $e^{ku}$ grows faster than $u^{-b}$ shrinks for any $k>0$, so the integrand does not tend to zero as $u \to \infty$.

  2. $a=1$. The integral is $\int_{\log{2}}^{\infty} u^{-b} \, du$, which we know converges if and only if $b>1$.

  3. $00$, because
    $$ \int_{\log{2}}^{\infty} u^{-b} e^{-(1-a)u} \, du < \int_{\log{2}}^{\infty} (\log{2})^{-b} e^{-(1-a)u} \, du = (\log{2})^{-b} \frac{1}{1-a}. $$


No comments:

Post a Comment

real analysis - How to find $lim_{hrightarrow 0}frac{sin(ha)}{h}$

How to find $\lim_{h\rightarrow 0}\frac{\sin(ha)}{h}$ without lhopital rule? I know when I use lhopital I easy get $$ \lim_{h\rightarrow 0}...