$$\lim_{(x,y) \to (0, \infty)} (xy) = [x\to\frac{1}{x}\Rightarrow x\to \infty] = \lim_{(x,y)\to(\infty, \infty)} (\frac{y}{x}) = [x = r\cos\theta, y = r\sin\theta] = \lim_{r\to\infty}\frac{r\sin\theta}{r\cos\theta} =\lim_{r\to\infty}\tan\theta$$

Therefore, limit does not exist. Is substitution in the beginning viable here?

  • 2
    $\begingroup$ Not really. If $x\to 0^+$ (from one side) then $\frac{1}{x}\to +\infty$, but if $x\to 0$ oscillating around zero, then the limit $1/x$ does not exists. $\endgroup$ – A.Γ. Dec 10 '18 at 16:29
  • $\begingroup$ Change $y$ to $1/y$. $\endgroup$ – Paracosmiste Dec 10 '18 at 16:29
  • $\begingroup$ You're assuming $x\to 0^+$. $\endgroup$ – Shaun Dec 10 '18 at 16:30
  • $\begingroup$ This limit doesn't exist or $0\cdot\infty$ would not be an indeterminate form. $\endgroup$ – egreg Dec 10 '18 at 17:05

More simply we have that as $t\to \infty$

  • $x=\frac1t\to 0 \quad y=t\to \infty$

$$xy=\frac1t \cdot t \to 1$$

  • $x=\frac1t\to 0 \quad y=t^2\to \infty$

$$xy=\frac1t \cdot t^2=t \to \infty$$

and therefore the limit doesn't exist.


The best way of saying things is this : $\lim_{(x,y) \to (0,\infty)} xy$ exists and equals $L$, if and only if for every subsequence $x_n \to 0$ and $y_n \to \infty$ we have $x_ny_n \to L$, where the latter is convergence of a sequence, in which case we have the $\epsilon-N$ definition of convergence.

Of course, one sees that taking $x_n = \frac 1n$ and $y_n = kn$ for any $k \in \mathbb R_{> 0}$, we have $x_n \to 0$, $y_n \to \infty$ and $x_ny_n \to k$. Therefore, the limit in question does not exist, since different choices of subsequences give different limit values.

As to what you have done, unfortunately as $x \to 0$ it is not true that $\frac 1x$ goes to $\infty$, since $x$ can approach $0$ from below, in which case $\frac 1x$ assumes negative values and cannot converge to any positive value, let alone positive infinity.

Furthermore, the part where you set $(x,y) = (r \cos \theta,r \sin \theta)$ : The point is that change of variable in limits only works in certain situations. In particular, since the variable $\theta$ has no particular limit as $(x,y) \to (0,\infty)$, I do not think that this change of variable is correct.


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.