0

This exercise is from my textbook:

Let $f : \mathbb{R}^2 \rightarrow \mathbb{R}$ such that $\frac{\partial f}{\partial x}(x,y) = \frac{\partial f}{\partial y}(x,y)$ for all $(x,y)\in\mathbb{R}^2$. Show that there exists $g : \mathbb{R} \rightarrow \mathbb{R}$ such that $g$ is differentiable and $f(x,y) = g(x + y)$ for all $(x,y)\in\mathbb{R}$.

I think that the key is to define $g(x) := f(x,0)$, but I'm not sure how to show that $f$ satisfies $g(x + y) = f(x + y, 0) = f(x,y)$, Can you give me some hint? I found the same question here but I'm not allowed to use the chain rule for the composition of two functions $f\circ g$ where $f:\mathbb{R}^n \rightarrow \mathbb{R}$ and $g : \mathbb{R}^m \rightarrow \mathbb{R}^n$ that is used in the answer.

Jorge S.
  • 782
  • 2
    Are you allowed to differentiate $f(x+t, x-t)$ on $t$ (for a fixed $x$)? –  Nov 24 '22 at 05:50
  • 2
    Without the chain rule (or at least some special case of it), you can't even verify that $f(x,y)=g(x+y)$ satisfies the given PDE, can you? So it seems hard to do without it... – Hans Lundmark Nov 24 '22 at 07:01

1 Answers1

-1

This is pushing the boundaries of semantic trickery, here is a solution that is identical to the answer you posted that does not explicitly use the chain rule. This is usually solved using a change of variables; if we let:

$$ \begin{cases} u = x + y\\ v = x - y \end{cases} $$ we will get:

$$ \begin{cases} x = \frac{1}{2}(u + v)\\ y = \frac{1}{2}(u - v) \end{cases} $$

Therefore $$ f(x, y) = f\left(\frac{1}{2}(u + v), \frac{1}{2}(u - v)\right) $$

So if we let $\varphi:\mathbb{R}^2\to\mathbb{R^2}, (u, v)\mapsto \left(\frac{1}{2}(u + v), \frac{1}{2}(u - v)\right)$, we get $f(x, y) = (f\circ\varphi)(u, v)$.

Let us then define the auxiliary function $F = f\circ\varphi$, the jacobian $J$ of $F$ is the product of that of $f$ and that of $\varphi$. It is easy then to verify that this implies

$$ \begin{cases} \frac{\partial F}{\partial u} = \frac{1}{2}\left(\frac{\partial f}{\partial x} + \frac{\partial f}{\partial y}\right)\\ \frac{\partial F}{\partial v} = \frac{1}{2}\left(\frac{\partial f}{\partial x} - \frac{\partial f}{\partial y}\right) \end{cases} $$

By using the PDE, we get $\frac{\partial F}{\partial v} = \frac{1}{2}\left(\frac{\partial f}{\partial x} - \frac{\partial f}{\partial y}\right) = 0$, and hence $f(x, y) = F(u, v) = g(u) = g(x+y)$ for some differentiable function g.

0sharp
  • 64
  • 5
    That the Jacobian of a composition is the product of the Jacobians of the factors is precisely the chain rule – Mariano Suárez-Álvarez Nov 24 '22 at 07:14
  • This is why I said I am pushing the boundaries of playing semantics. – 0sharp Nov 24 '22 at 07:18
  • 2
    That's a very long winded way of saying you are not telling the truth. ;) – Pedro Nov 24 '22 at 07:19
  • And that's a very condescending way of reiterating an objection that has already been raised and acknowledged. If the OP or anyone else found this answer helpful, fantastic, else, it would still be of more use than this comment. – 0sharp Nov 24 '22 at 07:31
  • You should consider that people have to spend time reading through your work, only to realize that you've used the method that was explicitly ruled out. – JonathanZ Nov 24 '22 at 14:12