I'm struggling with the following problem:

Let $f\colon\mathbb{R}^2\to\mathbb{R}$ be a twice continuously differentiable function satisfying $$f(0,y)=0\mbox{ for all }y\in\mathbb{R}$$

(a) Show that $f(x,y) = xg(x,y)$ for all pairs $(x,y)\in\mathbb{R}^2$, where $g$ is the function given by $$g(x,y) = \int\limits_0^1\frac{\partial f}{\partial x}(tx,y)dt$$

(b) Show that $g$ is continuously differentiable and that for all $x\in\mathbb{R}$ $$g(0,y) = \frac{\partial f}{\partial x}(0,y),~\frac{\partial g}{\partial y}(0,y) = \frac{\partial^2 f}{\partial x\partial y}(0,y)$$

(c) If $\frac{\partial f}{\partial x}(0,0)\neq0$ there is a neighborhood $V$ of $(0,0)$ in $\mathbb{R}^2$ such that $f^{-1}(0)\cap V = V\cap \{x=0\}$

(d) If $\frac{\partial f}{\partial x}(0,0) = 0$ and $\frac{\partial^2 f}{\partial x\partial y}(0,0)\neq0$ there is a neighborhood $V$ of $(0,0)$ in $\mathbb{R}^2$ such that $f^{-1}(0)\cap V$ consists of the union of the set $V\cap\{x=0\}$ with a curve through $(0,0)$, whose tangent at $(0,0)$ is not vertical (not parallel to the $y$-axis)

Here are my attempts:

(a) Note that $\frac{\partial f}{\partial x}(tx,y) = \frac{t}{x}\cdot\frac{\partial f}{\partial t}(tx,y)$ and so using the integration by parts $$g(x,y) = \int\limits_0^1\frac{\partial f}{\partial x}(tx,y)dt = \int\limits_0^1 \frac{t}{x}\cdot\frac{\partial f}{\partial t}(tx,y)dt = \frac{1}{x}\left(f(x,y) - \int\limits_0^1 f(tx,y)dt\right)$$ it suffices to show that $\int\limits_0^1 f(tx,y)dt = 0$, however, I can't see how to proceed with it.

(b) Expressions for $g(0,y)$ and $\frac{\partial g}{\partial y}(0,y)$ is the result of a straightforward calculation using the definiton of $g(x,y)$, but I have troubles showing that $g$ is continuously differentiable. What am I supposed to do -- prove that the partial derivatives exist and they are continuous? If so, can someone write down the details?

(c) One can fix $y=0$ and consider a function $f(x,0)\colon\mathbb{R}\to\mathbb{R}$. Then, since $\frac{\partial f}{\partial x}(0,0)\neq0$, the Inverse Function Theorem says that there exists a neighborhood $U$ of $0\in\mathbb{R}$ such that $f\colon U\to f(U)$ is the bijection. Denote $V = U\times\{0\}$, so $V\subset\mathbb{R}^2$ and we are done. Am I right here?

(d) It seems like using the result of part (b) we can do the same trick with the function $g(0,y)\colon\mathbb{R}\to\mathbb{R}$, but I can't finish the proof.

Any help will be really appreciated. Thanks a lot in advance.

  • 4
    $\begingroup$ Can you make the title somehow informative and descriptive, rather than subjective? (Do describe the equation; Don't say "tricky problem".) $\endgroup$ – Asaf Karagila May 17 at 11:10
  • 1
    $\begingroup$ @AsafKaragila, I did it. $\endgroup$ – Hasek May 17 at 17:52

I think you are misunderstanding something with the integral. $\frac{\partial f}{\partial x}(tx,y)$ doesn't mean the derivative with respect to the $x$ inside, it means the derivative of the function with respect to the first argument. In other words the integral really evaluates to

$$g(x,y) = \int_0^1 \frac{\partial f}{\partial x}(tx,y)\:dt = \begin{cases} \frac{1}{x}f(tx,y)\Bigr|_0^1 = \frac{f(x,y)-f(0,y)}{x} = \frac{1}{x}f(x,y) & x\neq0 \\ \int_0^1 \frac{\partial f}{\partial x}(0,y) \:dt = \frac{\partial f}{\partial x}(0,y) & x= 0 \\ \end{cases}$$

therefore we have in both cases that $f(x,y) = xg(x,y)$.

In order to show $g$ is continuous differentiable it is enough (sufficient but not necessary) to show that the partial derivatives of $g$ exist and are continuous.

The other questions are just application of the inverse function theorem on $g$ and its first partial, which is why the question has you prove differentiability first. I believe your proof is valid for part (c), so just continue on with (d).

| cite | improve this answer | |
  • $\begingroup$ Part (d) is still unclear for me. If I will consider $g(0,y)\colon\mathbb{R}\to\mathbb{R}$, then, since $\frac{\partial g}{\partial y}(0,0)\neq0$, from the Inverse Function Theorem for $g(0,y)$ there exist a neighborhood $W\subset\mathbb{R}$ such that $g\colon W\to g(W)$ is a bijection, i.e. $g(0,0)=0$ will have exactly one $y=0$ as a preimage. This is where I'm lost. I can't link it with the preimage of $f$ (not $g$!) and I can`t understand where should this curve with a non-vertical tangent come from. Can you please elaborate on it? $\endgroup$ – Hasek May 17 at 21:24

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.