I've trying to find this:

$$\lim_{(x,y)\rightarrow(0,0)} (x+y^2)\sin(\frac{1}{xy})$$ And I've just converted to polar coordinates: $$\lim_{\rho\rightarrow 0} (\rho \cos \theta + \rho^2 \sin^2 \theta)\sin(\frac{1}{\rho^2 \cos \theta \sin \theta}) =$$ $$\lim_{\rho\rightarrow 0} \rho \cos \theta \sin(\frac{1}{\rho^2 \cos \theta \sin \theta}) + \lim_{\rho\rightarrow 0} \rho^2 \sin^2 \theta \sin(\frac{1}{\rho^2 \cos \theta \sin \theta}) $$

Then, it follows that $|\rho\cos\theta\sin(\frac{1}{\rho^2 \cos \theta \sin \theta})| \leq |\rho \cos \theta| \leq |\rho|$, so the first limit must converge to zero. Something similar can be said about the second limit: $|\rho ^2 \sin ^2 \theta \sin(\frac{1}{\rho \cos \theta \sin \theta}) \leq |\rho ^2 \sin ^2 \theta| \leq |\rho^2|$, so the second limit must also be equal to zero. Nevertheless, WolframAlpha says the limit does not converge because it's path-dependent. Have I gone wrong somewhere? I don't see a $\theta$ in that limit which may cause any problems.

PS: I'm aware that the $\sin(\frac{1}{\rho^2 \cos \theta \sin \theta})$ will oscillate very wildly as $\rho$ approaches zero, but in the end, the sine function is bounded, and I still don't undertand where's my mistake.

  • $\begingroup$ $sin(\frac{1}{xy})$ is limited by $-1$ and $1$. And you have $\lim_{(x,y)\rightarrow (0,0)}(x+y^{2})=0$. $\endgroup$ – Mateus Rocha Jul 3 '18 at 15:19

Hint: we have $$|x+y^2||\sin(\frac{1}{xy})|\le |x|+y^2$$ and this tends to Zero for $x,y$ tends to Zero.

  • $\begingroup$ Then, shouldn't that be enough to say that the limit tends to zero? Why does WolframAlpha say that the limit does not exist? $\endgroup$ – mar Jul 3 '18 at 15:24
  • $\begingroup$ I don't know, the Limit is Zero,this says my proof, the reason is that $$|\sin(\frac{1}{xy})|\le 1$$ $\endgroup$ – Dr. Sonnhard Graubner Jul 3 '18 at 15:27
  • $\begingroup$ I agree with you. My doubt was coming from the fact that WolframAlpha says otherwise, which I highlighted with bold letters. I also used the fact that the sine function is bounded, no matter how wildly it may oscillate. I just read in other threads that WolframAlpha isn't very good at calculating multivariable limits, so it just may be that WA isn't giving me a good result (as it usually does). $\endgroup$ – mar Jul 3 '18 at 15:44
  • $\begingroup$ Maple does it correct! $\endgroup$ – Dr. Sonnhard Graubner Jul 3 '18 at 17:21

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.