The Student Room Group

Limit / Integral

Screen Shot 2020-07-15 at 13.44.01.png

Hi, I was just wondering how would I go about calculating this limit. At first, I was thinking I can use L'Hopital's rule, but then I would be assuming that the integral tends to infinity as x tends to infinity and I'm not sure if I can just do that in my course, without calculating it. It is a methods course. We usually don't calculate integrals which look tough, but we prove whether they are convergent or divergent. Would it be enough to say that because this integral is divergent, it must have a value of +infinity as it has a non-negative integrand. I would show it is divergent by the Direct Comparison test with the function t4\sqrt t^4. Then, after when I use the L'Hopital's rule, it is clear I would use FTC on the top and the bottom is fine, etc..
Your intuition that L'Hôpital's Rule is the way to go is correct. Because as you said, whilst the numerator is not something that is easy to work with directly, it has a very easily computable derivative via the Fundamental Theorem of Calculus. It's also good that you question whether you can justify the application of L'Hôpital here because a lot of students will simply apply this rule without bothering to check the conditions. Since this is a methods course I would expect that you'd get away with just saying that the integral 'clearly' diverges as x tends to infinity (and obviously so does x^3), so you can just apply L'Hôpital and then use a comparison of some sort to evaluate the limit. Regardless, it's worth having a think about how you would rigorously prove that the integral diverges, even though it is 'obvious'.
(edited 3 years ago)
Original post by Takeover Season
Screen Shot 2020-07-15 at 13.44.01.png

Hi, I was just wondering how would I go about calculating this limit. At first, I was thinking I can use L'Hopital's rule, but then I would be assuming that the integral tends to infinity as x tends to infinity and I'm not sure if I can just do that in my course, without calculating it. It is a methods course. We usually don't calculate integrals which look tough, but we prove whether they are convergent or divergent. Would it be enough to say that because this integral is divergent, it must have a value of +infinity as it has a non-negative integrand. I would show it is divergent by the Direct Comparison test with the function t4\sqrt t^4. Then, after when I use the L'Hopital's rule, it is clear I would use FTC on the top and the bottom is fine, etc..

Yeah all you really need to notice is that as xx \to \infty, we have

0x1+t4 dt0xt4 dtx33\displaystyle \int_0^x \sqrt{1+t^4} \ dt \sim \int_0^x \sqrt{t^4} \ dt \sim \dfrac{x^3}{3}

No need for L'Hopitals.
Original post by RDKGames
Yeah all you really need to notice is that as xx \to \infty, we have

0x1+t4 dt0xt4 dtx33\displaystyle \int_0^x \sqrt{1+t^4} \ dt \sim \int_0^x \sqrt{t^4} \ dt \sim \dfrac{x^3}{3}

No need for L'Hopitals.

True analyst's answer :biggrin:
Original post by RDKGames
Yeah all you really need to notice is that as xx \to \infty, we have

0x1+t4 dt0xt4 dtx33\displaystyle \int_0^x \sqrt{1+t^4} \ dt \sim \int_0^x \sqrt{t^4} \ dt \sim \dfrac{x^3}{3}

No need for L'Hopitals.

Yeah I suppose this is also possible (though for rigour I would use greater than or equal to signs). It's still worth mentioning that there are limits of this type where it's necessary to use FTC.
Original post by Legomenon
Yeah I suppose this is also possible (though for rigour I would use greater than or equal to signs). It's still worth mentioning that there are limits of this type where it's necessary to use FTC.

Inequalities are not too important here. All we care about is the asymptotic behaviour as xx \to \infty, so we just need to obtain a useful asymptotic expression for the integrand; doesn't really matter whether it's less than or greater than the expression we start with.
Original post by RDKGames
Inequalities are not too important here. All we care about is the asymptotic behaviour as xx \to \infty, so we just need to obtain a useful asymptotic expression for the integrand; doesn't really matter whether it's less than or greater than the expression we start with.

Right, but simply stating that 0x1+t4dt0xt4dt\int_0^x \sqrt{1+t^4}dt \sim \int_0^x \sqrt{t^4}dt is not entirely rigorous. Sure, it's 'obvious' because the '11' under the square root is insignificant, but a full proof would make the explicit comparisons 2t41+t4t4\sqrt{2t^4}\geq\sqrt{1+t^4}\geq \sqrt{t^4} to establish that the integral is asymptotic to x3x^3. Even here, one would have to note that the inequality chain only holds for t1t\geq 1 but that the remaining integral is of course o(x3)o(x^3). And then one could just apply the Sandwich Theorem instead of the somewhat heavier Limit Comparison Test.

This is why I avoided your approach, but it's equally valid.

Quick Reply

Latest