Results 1 to 6 of 6

Math Help - Concave / Convex function

  1. #1
    Member
    Joined
    Nov 2008
    Posts
    171

    Concave / Convex function

    hi

    is the following function concave or convex:
    f{x,y,z} = x*y + z

    I tried to find the hessian of this but it is coming out to be 0 which means that the Hessian test doesnt say anything about convexity or concavity. So, is there any other method that can be used which can give conclusive results?

    Thanks


    (p.s. MF rocks as always thanks guys for ur replies to all my posts! )
    Follow Math Help Forum on Facebook and Google+

  2. #2
    AMI
    AMI is offline
    Junior Member AMI's Avatar
    Joined
    Jun 2009
    Posts
    40
    The Hessian is \begin{pmatrix}0&1&0\\1&0&0\\0&0&0\end{pmatrix} in any point (x,y,z)\in\mathbb{R}^3, and it is neither positive semidefinite, nor negative semidefinite,
    which implies that the function is neither convex nor concave.
    Follow Math Help Forum on Facebook and Google+

  3. #3
    Member
    Joined
    Nov 2008
    Posts
    171
    Hessian is a sufficient condition. So when |H| = 0, we cannot say anything if the function is concave or convex. Hessian can be 0 even for strictly convex functions. For example : Hessian of f=x^2 + y^2 is 0 at (0,0) but then also the function is STrictly convex
    Follow Math Help Forum on Facebook and Google+

  4. #4
    AMI
    AMI is offline
    Junior Member AMI's Avatar
    Joined
    Jun 2009
    Posts
    40
    If you have X\subset\mathbb{R}^n convex with X^\circ\neq\emptyset, f:X\to\mathbb{R}, f\in\mathrm{C}^2 and denote by Hf(x) the Hessian of f at x\in X i.e. Hf(x)=\bigg(\frac{\partial^2 f}{\partial x_i\partial x_j}(x)\bigg)_{i,j=1,\dots,n}, then:
    f is convex if and only if Hf(x) is a positive semidefinite matrix, \forall x\in X.
    Follow Math Help Forum on Facebook and Google+

  5. #5
    Member
    Joined
    Nov 2008
    Posts
    171
    in ur definition it says "If and only if the hessian is positive semidefinite" .

    so that means that when Hessian is 0 there still exists a possibility of the function being strictly convex. We cannot conclude anything by using Hessian in this case.
    Follow Math Help Forum on Facebook and Google+

  6. #6
    AMI
    AMI is offline
    Junior Member AMI's Avatar
    Joined
    Jun 2009
    Posts
    40
    I didn't say anything about strict convexity. And my last post is a proposition (a characterisation of convex functions that belong to \mathrm{C}^2), not a definition.

    Check and see that the Hessian is not positive semidefinite (in every/any point). That implies that the function is not convex. Also, the Hessian is not negative semidefinite - and that implies that the function is not concave.
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Verify These Games are Convex-Concave
    Posted in the Advanced Applied Math Forum
    Replies: 0
    Last Post: April 18th 2011, 08:44 AM
  2. Replies: 1
    Last Post: August 24th 2010, 07:06 AM
  3. Where is f(x) concave and convex?
    Posted in the Calculus Forum
    Replies: 2
    Last Post: August 26th 2009, 01:57 PM
  4. Convex and Concave Question
    Posted in the Advanced Applied Math Forum
    Replies: 0
    Last Post: March 10th 2008, 06:32 PM

Search Tags


/mathhelpforum @mathhelpforum