$q(x) =x_1^2+x_2^2+3x_3^2-2x_1x_3+2x_2x_3+x_1+x_3+5 =x^TQx+b^Tx+5$

$ \begin{matrix} x=( x_1 & x_2 & x_3 )^T\\ \end{matrix} $

the first question was find a symmetric matrix Q and a vector b. so I found that matrix by using way to solve quadratic form

$$ Q= \begin{matrix} 1 & 0 & -1 \\ 0 & 1 & 1 \\ -1 & 1 & 3 \\ \end{matrix} $$

and $$ b =\begin{matrix} (1 & 0 & 1)^T\\ \end{matrix}$$

The second question was 'find, in terms of Q and b and find the minimum of q(x)'

If $ \partial q(x)/∂x= \begin{matrix} (∂q/∂x_1& ∂q/∂x_2&∂q/∂x_3)^T\\ \end{matrix}$

how can I solve second question? how can I approach in good way?


I just wanna share my try,

I guess we can find the minimum of this function by differentiation. Let's find our critical points, that is $$ \exists x_{0} \in R^3, / \bigtriangledown q(x_{0})=0?$$ note that $q(x): R^3 \rightarrow R$ $$\bigtriangledown q(x)=(2x-2z+1,2y+2z,6z-2x+2y+1)$$ Now we should solve the following system




it's east to find $x_{o}=(-3/2 , 1,-1)$ as its solution. Thus we just need to check if the Hessian is a positive matrix (cause it implies this point is a minimum). If you derivative again, you'll find:

$ Hq(x_{o})= \begin{matrix} 2 & 0 & -2 \\ 0 & 2 & 2 \\ -2 & 2 & 6 \\ \end{matrix} $

As far as I remember there is a theorem which relate Hessian with the quadratic form, check your notes cause it will help. However, in our case $Hq(x_{0})=2Q$, then we just need to check if Q is a positive matrix. $A_{1} >0, A_{2}>0, A_{3}>0$ so Q is a positive matrix thus $x_{0}$ is the minimum of q(x) $\forall b$

  • $\begingroup$ it was really good help to understand thank you. $\endgroup$ – matthew Sep 24 '16 at 14:20
  • $\begingroup$ It's better to use correct English: "wanna" $\rightarrow$ "want to" (I am not native English speaker) $\endgroup$ – Jean Marie Sep 24 '16 at 19:29

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.