3
$\begingroup$

What are the critical points of the function: $f(x,y,z)=\frac{1}{z^2+x^2+1+y^2}$? Identify as a local minima, maxima, or saddle points.

So I know you have to take the gradient and set it equal to $(0,0,0)$. That will get you all your critical points. How do I identify it as a local minima, maxima, or a saddle point?

$\endgroup$
2
  • $\begingroup$ That Wikipedia article claims that the determinant test only works with functions of two variables. $\endgroup$
    – NicNic8
    Commented May 20, 2018 at 16:27
  • $\begingroup$ @rbird The determinant alone can't tell you for three variables because a positive determinant could mean one positive and two negative eigenvalues (saddle point) or all positive (local minimum) $\endgroup$ Commented May 20, 2018 at 19:54

2 Answers 2

3
$\begingroup$

One way is to find the Hessian and determine its curvature by looking at its eigenvalues.

If the eigenvalues are all positive, then the function as positive curvature at that point, and you've found a minima.

If the eigenvalues are all negative, then negative curvature. And if they're mixed, then it's a saddle point.

$\endgroup$
2
  • $\begingroup$ can you do this for me? I think I have an answer, however I am unsure if it is correct. I want to double check. $\endgroup$ Commented May 20, 2018 at 16:24
  • $\begingroup$ @MathWannaBe graph your answer using Wolfram Alpha and see if you're correct! :) $\endgroup$
    – NicNic8
    Commented May 20, 2018 at 16:28
3
$\begingroup$

There are no minima and no saddle points, the only maxima is at $x=y=z=0$, with the value of $1$.

The reason:

The function is spherically symmetric to the origin, so we can substitute $r=\sqrt{x^2+y^2+z^2}$. Doing that, we can see, that we have essentially a function of $\frac{1}{r^2+1}$, where r is the distance of the point from the origin. This function is monotonically decreasing from $1$ to $0$, while $r$ is going from zero to infinite. Its derivative is never zero (except the origin).

$\endgroup$
1
  • $\begingroup$ @Dirk Oops, thanks! I fixed it. $\endgroup$
    – peterh
    Commented May 20, 2018 at 16:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .