Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Find all points $(x_1,x_2,\ldots,x_k) = x \in \mathbb{R}^k$ such that $f(x)=|x_1 \cdot x_2 \cdots x_k|$ is a differentiable function.

I know that if $\forall i, x_i \neq 0$ then this function is differentiable. I don't know what about rest of points.

share|improve this question
add comment

2 Answers 2

Let $(x_1,\ldots,x_k)$ be a point where $f(x)=0$. If $x_i=0$ then $\frac\partial{\partial x_j}f(x)=0$ for all $j\ne i$. Thus if exactly one coordinate are zero all other partial derivatives are zero, while for the $i$th component we have that $f$ looks like $c|x_i|$, where $c$ is the (nonzero) absolute value of the rproduct of all other coordinates, hence $\frac\partial{\partial x_i}f(x)$ does not exist and $f$ is not differentiable.

But if $r\ge 2$ coordinates are zero, all partial derivatives are zero. In fact, we can see that $\frac{f(x+h)}{||h||}\to 0$ as $h\to 0$:

  • If $r=k$, then $|f(x+h)|\le |h|_\infty^k\le ||h||^k$ and from $k= r\ge 2$, the claim follows.
  • If $r<k$, let $m=\min\{|x_i|\colon x_i\ne 0\}$. Then for $||h||<m$, we have $f(x+h)\le |h|_\infty^r (2m)^{k-r}$, whence again the claim.

Thus $f$ is totally differentiable (with derivative 0) if at least two coordinates are zero.

share|improve this answer
add comment

$f$ is differentiable at $x$ if and only if all its partial derivatives exist at $x$. If you vary $x_i$ while keeping the other components fixed, is the resulting function $f_i(x_i)$ of one variable differentiable at $x_i=0$?

If you do this, you will see that the function is differentiable only if all its variables are nonzero. Special care need to be given to the points where more than one coordinate is zero, but it is not differentiable there either, by a well-definedness-argument.

share|improve this answer
    
Your second paragraph contradicts the first: At points with multiple zeroes, all partials are zero. –  Hagen von Eitzen Oct 25 '12 at 18:01
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.