0

enter image description here

For this question, if I divide this function into two parts, which are $x \ge 0$ and $x<0$, then the part that doesn't include "0" will have no critical point, and I also have no idea of how to find the singular point (might be the sharp point). Can anyone help me, thanks.

vonbrand
  • 27,812
Jakoer
  • 315

1 Answers1

0

Recall that a crititcal point $p$ satisfies $\nabla f(p) = 0$ and that $\nabla f$ blows up at a singular point.

Suppose x < 0 so that $|x| = -x$. We then have $$\nabla f(x,y,z) = \Bigg(-3x^2, \quad y\Big(2-\frac{1}{\sqrt{z^2+y^2}}\Big) + 1, \quad \frac{-z}{\sqrt{z^2+y^2}}\Bigg) = 0,$$ from which we see that $x = z = 0$. This forces $y(2-\frac{1}{|y|}) + 1 = 0$.

For $y < 0$ we get $y = -1$ and $y = \frac{1}{2}$. But since we assume that $y<0$, $y = \frac{1}{2}$ does not apply.

For $y > 0$ we get $2y^2 + y + 1 = 0$. Noting that the left side has a negative determinant we conclude there are no solutions in $\mathbb{R}$ in this case.

Note that our presumption $x < 0$ in order to evaluate the absolute value $|x|$ was unnecessary since $x = 0$ from our condition $\nabla f(x,y,z) = 0$.

We have the critical point $A = (0,-1,0)$.

$\nabla f$ blows up at $(x,0,0)$ for all $x$, so we have the the set of singular points $B = (x,0,0)$.

Mussé Redi
  • 1,429
  • 1
    I know when z=y=0,∇f blows up, but I don't get why x should equals 0 at singular point. – Jakoer Mar 24 '14 at 03:24
  • You are right: I corrected my answer. – Mussé Redi Mar 24 '14 at 03:30
  • 1
    I remembered sqrt(y^2) = |y|, so here if we split y as y<0 and y>0, we get y =-1 and y does not exist. So there is only one critical point (0,-1,0). – Jakoer Mar 24 '14 at 03:46
  • For $y > 0$ there are no solutions in $\mathbb{R}$ for the equation in $y$. For $y < 0$ however, we also have a solution $y = \frac{1}{2}$ aside from the $y = -1$, that you correctly pointed out. – Mussé Redi Mar 24 '14 at 04:05
  • 1
    I don't understand why y=1/2. When y<0, y should be negtive number and if we substitute y as 1/2, then it does not equal to 0. – Jakoer Mar 24 '14 at 04:13
  • It should not, indeed. We only have $y = -1$. Sorry for all my calculation mistakes. – Mussé Redi Mar 24 '14 at 04:18
  • That's fine, thanks again. – Jakoer Mar 24 '14 at 04:22