Find $L=\lim\limits_{n \to \infty} \int_0^{n a} \exp\left(-\dfrac{t}{1+\frac{b t}{n}}\right) dt$, where $a>0$, $b>0$.

I can't see what is dominating function, but I feel that I have to use Dominated convergence theorem. Any kind of help is welcome.


Rewrite the integral as

$$\int_0^{n a} dt \: \exp\left(-\dfrac{t}{1+\frac{b t}{n}}\right) = n a \int_0^1 du \: \exp\left(-\dfrac{n a u}{1+a b u}\right)$$

As $n \rightarrow \infty$, the contributions from the integral come mainly from the neighborhood near $u=0$. In this limit, then

$$L = \lim_{n \rightarrow \infty} n a \int_0^{\infty} du \: e^{- n a u} = 1$$

  • $\begingroup$ i have a small q about your ans: i to got to the point where the integral of u is between 0 and 1/ in the next step you take L to be the same integral but u is now from o to infinity. why is that? is that a mistake? because my ans is 0... thanks! $\endgroup$ – user61202 Feb 6 '13 at 11:03
  • $\begingroup$ We can set the upper limit to $\infty$ because the error in doing so is exponentially small as $n \rightarrow \infty$. In fcat, you can see that in evaluating the integral with the finite limit. $\endgroup$ – Ron Gordon Feb 6 '13 at 11:12

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.