I have function $f(x,y)$ with $x\geq0$ and $y\geq0$. The Hessian matrix of this function has following properties.

1- $f_{xx}>0$

2- Determinant of the Hessian matrix is zero.

Can we say that the function $f(x,y)$ is jointly convex over $(x,y)$ for all of the desired domain?

  • $\begingroup$ you want determinan be zero at all points in domain ? $\endgroup$ – Red shoes Feb 21 '18 at 0:27
  • $\begingroup$ @Redshoes I do not want. But actually for $f(x,y)$ the determinant of the Hessian matrix is zero over all the domain. $\endgroup$ – Frank Moses Feb 21 '18 at 0:30
  • $\begingroup$ @Redshoes I have also clarified in my edited post. $\endgroup$ – Frank Moses Feb 21 '18 at 0:33

Yes. The Hessian of $f$ is semi positive definite every on the domain , because it passes the Sylvester's criterion .

So $f$ is jointly convex on its domain.

  • $\begingroup$ shouldn't we check the (2,2) element according to Sylvester's criterion? $\endgroup$ – LinAlg Feb 21 '18 at 1:50
  • $\begingroup$ @LinAlg you don't have to. even you check it is positive because $$f_{yy} = \frac{{f_{xy}}^2 }{f_{xx}}$$ $\endgroup$ – Red shoes Feb 21 '18 at 2:45

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.