1
$\begingroup$

I'm stuck on part b) of this problem from the 1989 edition of Elements of Calculus and Analytic Geometry by Thomas and Finney, Chapter 12 (Power Series), Section 6 (Indeterminate Forms). Part a) asks to prove $$ \lim_{x\to\infty}\int_{0}^{x} e^{t^2}\, dt$$ which is straightforward. I substituted a power series, integrated a few terms, and saw that it $\rightarrow \infty$.
Part b) asks to find $$ \lim_{x\to\infty} x \int_{0}^{x} e^{t^2-x^2}\, dt$$ I've tried it a couple ways, first rewriting the integrand as $\frac{e^{t^2}}{e^{x^2}}$. I then treated the denominator as a constant and brought it in front of the integral with the $x$. I then integrated a power series for the remaining integral; I keep getting an answer of $\infty$, but the given answer is $\frac{1}{2}$.

Alternatively, I tried substituting $(t^2-x^2)$ into a generic $e^x$ power series before integrating, but that also gives me $\infty$. (I'm too slow with LaTeX to show all my work here.) I can't find my mistake, and I can't think of what else to try.

Is my technique faulty or is there some methodological hint?

$\endgroup$
  • 1
    $\begingroup$ Try L'Hospital's rule. $\endgroup$ – James Oct 13 '15 at 19:36
8
$\begingroup$

You can use l'Hospital, $$ \lim_{x\to+\infty}\frac{\int_0^x e^{t^2}\,dt}{e^{x^2}/x}=\lim_{x\to+\infty}\frac{e^{x^2}}{2e^{x^2}-e^{x^2}/x^2}=\lim_{x\to+\infty}\frac{1}{2-1/x^2}=\frac{1}{2}. $$

$\endgroup$
  • 1
    $\begingroup$ Magnificent! Thank you. $\endgroup$ – DBS Oct 13 '15 at 19:53
3
$\begingroup$

If you wish to avoid the usage of L'hospital rule using simple estimates you may proceed as follows:

For a lower bound \begin{aligned} x e^{-x^2}\int^x_0 e^{t^2}\,dt \geq e^{-x^2}\int^x_0 t e^{t^2}\,dt=\frac12(1-e^{-x^2})\xrightarrow{x\rightarrow\infty}\frac12 \end{aligned}

For an upper bound

\begin{aligned} x e^{-x^2}\int^x_0 e^{t^2}\,dt &=x\int^x_0 e^{(t-x)(t+x)}\,dt\leq x\int^x_0 e^{(t-x)2x}\,dt\\ &=x e^{-2x^2}\int^x_0e^{2xt}\,dt= \frac{1-e^{-2x^2}}{2}\xrightarrow{x\rightarrow\infty}\frac12 \end{aligned}

The advantage of not using L'Hospital rule here is that you also obtain rate of convergence. The function $F(x)=x e^{-x^2}\int^x_0 e^{t^2}\,dt$ converges to$\frac12$ much faster than $G(x)=\frac{1}{2-x^{-2}}$ does.

$\endgroup$

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.