Find $\frac{\partial u}{\partial x}, \frac{\partial u}{\partial y}, \frac{\partial v}{\partial x}$, if $xu-yv=0$ and $yu+xv=1$
Actually, I have an answer but I don't understand anything. Not because it is Chinese but transformations. Could you explain?

- 115
-
1What have you tried so far? – Brian Borchers Oct 21 '21 at 17:33
-
1$v=\frac{x}{y}u$, substitute into the second equation and differentiate – Vasili Oct 21 '21 at 17:35
-
2"not because it is Chinese": I appreciate your humor... – Jean Marie Oct 21 '21 at 17:36
-
1Using the product rule, $(xu-yv)'=xdu+udx-ydv-vdy=0$. Do you get it now? – Vasili Oct 21 '21 at 17:41
-
@Vasya this thing I understood – sln Oct 21 '21 at 17:42
-
@Vasya but why in the second system the equation is like this – sln Oct 21 '21 at 17:44
-
1OK, $(1)'=0=(yu+xv)'$ – Vasili Oct 21 '21 at 17:45
-
@Vasya oh, I understand – sln Oct 21 '21 at 17:46
-
@Vasya but by which variable do we differentiate? Or is it a formula without differentiating variable? – sln Oct 21 '21 at 17:48
-
1@kfcgenius: This is called implicit differentiation because we do not have a function that depends on some variables but an equation. Thus, we do not differentiate using a specific variable. – Vasili Oct 21 '21 at 17:54
-
Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. – Community Oct 21 '21 at 17:55
1 Answers
The first set of equations:
$$\begin{cases}xu-yv&=&0\\yu+xv&=&1\end{cases} \tag{0}$$
can be differentiated under the form:
$$\begin{cases}(udx+xdu)-(vdy+ydv)&=&0\\(u dy +y du)+(vdx+xdv)&=&0\end{cases},$$
giving the equivalent system:
$$\begin{cases}xdu-ydv&=&-udx+vdy\\ ydu+xdv&=&-vdx-udy\end{cases} $$
as in your book. Now, write it with matrices and vectors:
$$\underbrace{\begin{pmatrix}x&-y\\y&x \end{pmatrix}}_A\begin{pmatrix}du\\dv \end{pmatrix}=\underbrace{\begin{pmatrix}-u&v\\-v&-u\end{pmatrix}}_B \begin{pmatrix}dx\\dy \end{pmatrix}$$
from which you deduce that :
$$\begin{pmatrix}du\\dv \end{pmatrix}=\underbrace{A^{-1}B}_J \begin{pmatrix}dx\\dy \end{pmatrix}$$
It remains to compute matrix $J$:
$$J=\dfrac{1}{x^2+y^2}\begin{pmatrix}x&y\\-y&x \end{pmatrix}\begin{pmatrix}-u&v\\-v&-u\end{pmatrix}$$
$$J=\dfrac{1}{x^2+y^2}\begin{pmatrix}(-ux-vy)&(vx-uy)\\(uy-vx)&(-ux-vy)\end{pmatrix}\tag{1}$$
Remark: We have used the classical formula for the inverse of a $2 \times 2$ matrix.
$J$ is the jacobian matrix of the transformation $(x,y) \to (u,v)$, whose entries are precisely:
$$J=\begin{pmatrix}\frac{\partial u}{\partial x}&\frac{\partial u}{\partial y}\\\frac{\partial v}{\partial x}&\frac{\partial v}{\partial y} \end{pmatrix}\tag{2}$$
It remains to identify the corresponding entries in (1) and (2).
Edit : important remark: My attention has been triggered by the fact that:
$$\frac{\partial u}{\partial x}=\frac{\partial v}{\partial y}, \ \ \frac{\partial u}{\partial y}=-\frac{\partial v}{\partial x}$$
which are called "Cauchy-Riemann equations", classical for analytical ("smooth") complex functions.
Nothing surprizing because the initial relationships can be written plainly :
$$(x+iy)(u+iv)=i \ \iff \ u+iv=\dfrac{i}{x+iy}$$
finally giving
$$u+iv=\dfrac{i(x-iy)}{(x+iy)(x-iy)}=\dfrac{y+ix)}{x^2+y^2}$$
Otherwise said, identifying real and imaginary parts, we have explicit expressions:
$$u=\dfrac{y}{x^2+y^2}, \ \ \ v=\dfrac{x}{x^2+y^2},\tag{3}$$
from which the partial derivatives are easily obtained...
(transformation (3) is a parent of the classical "inversion" transform)

- 81,803
-
-
And if both equations in system are both equals to 0, can I just unite them in one equation? – sln Oct 21 '21 at 18:06
-
1You could treat a single system by using complex numbers, but, as you can see, with the method I propose you, you reach the result. – Jean Marie Oct 21 '21 at 18:08
-
-
Because the formula for the inverse of a matrix $A$ begins by $\frac{1}{\det(A)}$ – Jean Marie Oct 21 '21 at 18:16
-
-
1I have taken the liberty to change your initial title "solve this task". – Jean Marie Oct 22 '21 at 08:14