2

I was given the function:

$f(x,y)=5+2x+4y+x^2+y^2+(x^2y^4)^\frac15$

I need to show it is differentiable at $(0,0)$.

I started using the method of differentials and infinitesimal functions:

$\Delta f=f(0+\Delta x,0+\Delta y)-f(0,0)$

$\Delta f=2\Delta x + 4 \Delta y + (\Delta x)^2 + (\Delta y)^2 + ((\Delta x)^2(\Delta y)^4)^\frac15$

Now I cannot see a way how I can obtain infinitesimal functions* $\alpha$ and $\beta$ such that:

$\Delta f=2\Delta x + 4 \Delta y + \alpha\Delta x + \beta \Delta y$, where $\alpha,\beta=o(\rho)$

(Where $\rho^2 = x^2+y^2$, and $o$ is the Landau little-$o$ notation defined as $\alpha=o(\rho)\iff\lim \alpha/\rho=0$).

I managed to get 3 "infinitesimal functions": $\alpha=\Delta x$, $\beta=\Delta y$, and $\gamma=(\Delta x^2\Delta y^{-1})^\frac15$ but I am having trouble showing $\gamma=o(\rho)$ because frankly it doesn't seem to be.

Any hints? If this is not possible, are there any other methods that one would recommend?

*My prof defines infinitesimal functions as: $\alpha(x)$ is infinitesimal at $a$ if $\lim_{x\rightarrow a}\alpha(x)=0$

amWhy
  • 209,954
Jepsilon
  • 718
  • When you say infinitesimals, what kind of entry you mean? – Sudix Jun 08 '18 at 15:47
  • @Sudix Our prof defined infinitesimal functions as: a function $\alpha(x)$ is called infinitesimal at a point $a$ if $\lim_{x\rightarrow a}\alpha(x)=0$ – Jepsilon Jun 08 '18 at 16:01
  • Can't you just set up the finite difference limits for $\partial_x F(x,y)$ and $\partial_y F(x,y)$ and show that both limits exist? – D.B. Jun 08 '18 at 16:11
  • @D.B. So rather than try both variables at once try them separately? Would that imply the same thing? – Jepsilon Jun 08 '18 at 16:13
  • No. Sorry, I'm wrong. I just read somewhere about how existence of partial derivatives does not imply differentiability. Here is an answer that is related to your question and might help: https://math.stackexchange.com/questions/427552/how-to-examine-if-multivariable-functions-are-differentiable – D.B. Jun 08 '18 at 16:24
  • Hmm. What's wrong with using three infinitesimal functions $\alpha(x) = x^2 , \beta(y) = y^2 , \gamma(x,y) = x^2y^4$? – Sudix Jun 08 '18 at 16:24
  • @D.B. no: the function that is zero everywhere on the plane besides on the parabolic arc $(x,x^2)$, with $x > 0$, where it has value $1$, has both partial derivatives at $(0,0)$, and they agree, but it is not differentiable there because it is not continuous. – giobrach Jun 08 '18 at 16:25
  • I somehow already have 4 close votes on this for being "off topic" idk why. My question is certainly on topic and of interest to people and it far from being a "just do my work for me" post because I have provided what I have done so far. Anyone care to provide me with an explanation of how am I "off topic"? Rather than just voting to close to get badges, it would be better for everyone to tell me how to improve the question because we're all here to learn. – Jepsilon Jun 09 '18 at 10:26
  • @Jepsilon I agree with you. I will move to reopen. – zhw. Jun 09 '18 at 16:10
  • Is $\rho=(x^2+y^2)^{1/2}$? – zhw. Jun 09 '18 at 16:13
  • @zhw. yes forgot to mention that – Jepsilon Jun 09 '18 at 16:22
  • Thanks for finally editing your post, @Jepsilon. You'll see you get further on this site when you give more details and context in your post. – amWhy Jun 09 '18 at 17:01
  • @amWhy It would have been much more helpful to post a comment with what exactly was wrong with the post so I could fix it immediately, rather than having to wait a day and then reading about the improvements in the closing message, no? – Jepsilon Jun 09 '18 at 17:21
  • No one here owes you anything; if you don't read the site's standards prior to posting, you remain responsible for poor posts. Others may, or may not, want to waste their breath, for the 200 thousandth time, or not telling you why. In any case, when a question is closed, as was your's, you are given a highlighted message explaining the reason for the closure. – amWhy Jun 09 '18 at 17:29
  • In case you didn't see it: "This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level." – amWhy Jun 09 '18 at 17:30
  • 1
    Jepsilon: You've addressed those issues, through two edits to your post, which immediately sent the post to the reopening review queue, in which users can vote to reopen, or leave closed. You made good edits, and as a result, your post was (is) reopened. – amWhy Jun 09 '18 at 17:52

3 Answers3

1

You seem to have the wrong definition: You need only $\Delta f=2\Delta x + 4 \Delta y + \alpha\Delta x + \beta \Delta y$, where $\alpha,\beta=o(1).$ An equivalent form is $\Delta f=2\Delta x + 4 \Delta y + o(\rho),$ which I think is more standard. If you use the latter form, you need only verify that

$$ (\Delta x)^2 + (\Delta y)^2 + ((\Delta x)^2(\Delta y)^4)^\frac15=o(\rho).$$

zhw.
  • 105,693
  • 1
    OK down voter, please explain. – zhw. Jun 09 '18 at 02:39
  • 1
    I guess some people on this site are more concerned about earning badges rather than actually discussing ways to improve posts/answers. The question already has 4 close votes and a down vote too – Jepsilon Jun 09 '18 at 10:32
  • I changed my answer to address your approach. – zhw. Jun 09 '18 at 16:48
  • @Jepsilon And many people have no more badges to earn through reviewing, but still serve this site by reviewing anyway. Please do not overgeneralize on the motives of votes that you can't know, or to cover your butt by finger pointing and blaming the messengers. – amWhy Jun 09 '18 at 16:58
  • And no, @zhw, I did not downvote your post. – amWhy Jun 09 '18 at 16:58
0

We know that a function f(x,y) is differentiable at a point (x,y) if it's left hand derivative is equal to it's right hand derivative (please read the differentiation chapter taught in your respective school/college).

That is, $$lim_{h\to 0}\frac{f(x+h,y+h)-f(x,y)}{h}=lim_{h\to 0}\frac{f(x-h,y-h)-f(x,y)}{-h}$$

Where h is a positive number that lies between 0 and 1 and can be approximately equated to 0. I leave the calculation part up to you(Just put 0 in place of x and y and apply the limits) You will get a value 6 in the left hand derivative and the same value for the right hand derivative which proves that the function is indeed differentiable at (0,0). I will be very happy to post the calculation if you are not getting it. Hope this solves your query.

Govind
  • 1
  • This is not correct. You've only shown that a certain directional derivative exists at $(x,y).$ – zhw. Jun 09 '18 at 19:24
0

Another approach using the definition of differentiability:

To use the horrible limit

$$\lim_{(x,y)\to(0,0)}{\dfrac{f(x,y)-[f(0,0)+f'_x(0,0)(x-0)+f'_y(0,0)(y-0)]}{\sqrt{{(x-0)}^2+{(y-0)}^2}}}\tag 1$$

we need first to calculate the partial derivatives at $(0,0)$ using the fact that $f$ is $C^1(\text{Dom}(f))$. Let's begin.

Remember that $f(x,y)=5+2x+4y+x^2+y^2+\sqrt[5]{x^2y^4}$.

Finding the domain of the function

$\text{Dom}(f)=\mathbb R^2$ because is a sum of continuous.

Finding partial derivatives

$$\begin{array}{llc} f'((0,0);(a,b))&&= \\ &\displaystyle\lim_{h\to 0}{\dfrac{f(ah,bh)-f(0,0)}{h}}&= \\ &\displaystyle\lim_{h\to 0}{\dfrac{\left(5+2ah+4bh+a^2h^2+b^2h^2+\sqrt[5]{a^2h^2b^4h^4}\right)-5}{h}}&= \\ &\displaystyle\lim_{h\to 0}{\dfrac{h\left(2a+4b+a^2h+b^2h+\frac{\sqrt[5]{a^2b^4h^6}}{h}\right)}{h}}&= \\ &\displaystyle\lim_{h\to 0}{\left(2a+4b+a^2h+b^2h+\sqrt[5]{a^2b^4h}\right)}&= \\ &2a+4b. \end{array}$$

Thus the directional derivatives are continuous for all $(x,y)\in\mathbb R^2$, in particular for $(x,y)=(0,0)$, thus $f(x,y)\in C^1\left(\mathbb R^2\right)$ so

$$\begin{matrix} f'((0,0);(1,0))&=&f'_x(0,0)&=&2\cdot 1+4\cdot 0&=&2, \\ f'((0,0);(0,1))&=&f'_y(0,0)&=&2\cdot 0+4\cdot 1&=&4. \end{matrix}$$

Proving the differentiability

From $(1)$ $$\begin{array}{lcc} \displaystyle\lim_{(x,y)\to(0,0)}{\dfrac{5+2x+4y+x^2+y^2+\sqrt[5]{x^2y^4}-[5+2(x-0)+4(y-0)]}{\sqrt{x^2+y^2}}} &=& \\ \displaystyle\lim_{(x,y)\to(0,0)}{\dfrac{x^2+y^2+\sqrt[5]{x^2y^4}}{\sqrt{x^2+y^2}}}&\underbrace =_{(x,\,y)\;=\;(\rho\cos\theta,\,\rho\sin\theta)}&\\ \displaystyle{\lim_{\rho\to 0^+}}_{\theta\in[0,2\pi)}{\dfrac{\rho^2\cos^2\theta+\rho^2\sin^2\theta+\sqrt[5]{\rho^2\cos^2\theta\cdot\rho^4\sin^4\theta}}{\sqrt{\rho^2\cos^2\theta+\rho^2\sin^2\theta}}}&=&\\ \displaystyle{\lim_{\rho\to 0^+}}_{\theta\in[0,2\pi)}{\dfrac{\rho^2+\rho\sqrt[5]{p\cos^2\theta\cdot\sin^4\theta}}{\rho}}&=&\\ \displaystyle{\lim_{\rho\to 0^+}}_{\theta\in[0,2\pi)}{\left(\rho+\sqrt[5]{p\cos^2\theta\cdot\sin^4\theta}\right)}&=&\\ 0,&& \end{array}$$ hence $$\boxed{f(x,y)\quad\text{is differentiable at}\quad(0,0)}.$$

manooooh
  • 2,269