1
$\begingroup$

This is a book example (not a homework question) about implicit differentiation on a composite of functions in $n$-dimensional space.

But my book explains this example in a very unclear manner. So I really appreciate your help in clarifications, since I'm very lost now >_<

The task is as follow:

Given:

$F(x,y,u,v) = x^2 + ux + y^2 + v$

$G(x,y,u,v) = x + yu + v^2 + x^2 v$

Goal: Find $\frac{\partial u}{\partial x}$ by 2 methods:

(1) by implicit differentiation

(2) explicitly solve for $u$

So far, I used the Jacobian to get the partial derivative, where:

$$\frac{\partial u}{\partial x} = \frac{-\left|\frac{\partial (F,G)}{\partial (x,v)}\right|}{\left|\frac{\partial (F,G)}{\partial (u,v)}\right|}$$

after some prep-calculations for partials of $F$ with respect to $x, u, v$; and partials of $G$ with respect to $x, u, v$

So my question is, is this method considered (1) or (2)? What does it really mean to "explicitly solve" for $u$? Just treat $x,y,v$ as constants and solve for $u$ in $F$ and in $G$?

My friend said if I do method (2), I shall get a different answer from my answer using Jacobian, but I don't know what to do to get the 2nd answer.

Would someone please help me on this question?
Thank you in advance

$\endgroup$

1 Answer 1

0
$\begingroup$

I rewrite the equations for readability:
$F(x,y,u,v)=x^2+ux+y^2+v$,
$G(x,y,u,v)=x+yu+v^2+x^2v$.

Maybe for method (2), the idea is to eliminate $v$, then solve for $u$. So we can take $v=F-x^2-ux-y^2$, then substitute that into the second equations, expand/simplify, and try to solve for $x$. I suppose it is possible you will have a quartic equation in $x$, so if you don't see a trick for solving it, you may try to have some software solve that algebraic equation.

$\endgroup$

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.