573 views

1 Answer

Best answer
0 0 votes

I am considering that b and B are same.

 

$f(x) =x^{T}Ax + B^{T}x +c$

A n*n symmetric real matrix $A$ is said to be positive definite if $x^{T}Ax >0$ for all non zero $x$ in $R^{n}$

$B$ and $x$ are vectors of dimensions n *1.


Let $n =2$ ,$A_{2*2}$ = $\begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}$, $B_{2*1}$ = $\begin{bmatrix} m\\n \end{bmatrix}$

Now since $x^{T}Ax >0$ and $c$ is a constant value so we should try to make $b^{T}x$ as negative so that we could get a minimum value for f(x).

So either option b or option c should be the answer.


Option B

$x =-(A^{T}A)^{-1}B$

$\Rightarrow x = -\left ( \begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}\begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}\right )^{-1}$ $\begin{bmatrix} m\\n \end{bmatrix}$

$\Rightarrow x = -\left ( \begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}\right )^{-1}$ $\begin{bmatrix} m\\n \end{bmatrix}$

$\Rightarrow x =- \frac{\left ( \begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}\right )}{1}$ $\begin{bmatrix} m\\n \end{bmatrix}$

$\Rightarrow x = \begin{bmatrix} -m\\-n \end{bmatrix}$

Also, $x^{T} =\begin{bmatrix} -m & -n \end{bmatrix}$

$x^{T}Ax$ = $\begin{bmatrix} -m & -n \end{bmatrix}\begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}\begin{bmatrix} -m\\-n \end{bmatrix}$=$m^{2}+n^{2}$

$B^{T}x$ = $\begin{bmatrix} m & n \end{bmatrix}\begin{bmatrix} -m\\-n \end{bmatrix}$=$-m^{2}-n^{2}$

$\therefore$ $f(x)$ =$x^{T}Ax + B^{T}x +c$ = $m^{2}+n^{2}-m^{2}-n^{2}+c$= $c$


Option C

$x$ =$-\left ( \frac{A^{-1}B}{2}\right )$

$\Rightarrow x$ = $-\left ( \frac{\begin{bmatrix} 1 & 0\\ 0 & 1 \end{bmatrix}^{-1}\begin{bmatrix} m\\n \end{bmatrix} }{2}\right )$=$-\left ( \frac{\begin{bmatrix} 1 & 0\\ 0 & 1 \end{bmatrix}\begin{bmatrix} m\\n \end{bmatrix} }{2}\right )$=$ -\frac{\begin{bmatrix} m\\n \end{bmatrix} }{2}$=$\begin{bmatrix} -\frac{m}{2}\\-\frac{n}{2} \end{bmatrix}$

Also, $x^{T} =\begin{bmatrix} -\frac{m}{2} & -\frac{n}{2} \end{bmatrix}$

$x^{T}Ax$ = $\begin{bmatrix} -\frac{m}{2} & -\frac{n}{2} \end{bmatrix}\begin{bmatrix} 1 &0 \\ 0 & 1 \end{bmatrix}\begin{bmatrix} -\frac{m}{2}\\-\frac{n}{2} \end{bmatrix}$=$\frac{m^{2}}{4}+\frac{n^{2}}{4}$

$B^{T}x$ = $\begin{bmatrix} m & n \end{bmatrix}\begin{bmatrix} -\frac{m}{2}\\-\frac{n}{2} \end{bmatrix}$=$-\frac{m^{2}}{2}-\frac{n^{2}}{2}$

$\therefore$ $f(x)$ =$x^{T}Ax + B^{T}x +c$ = $\frac{m^{2}}{4}+\frac{n^{2}}{4}-\frac{m^{2}}{2}-\frac{n^{2}}{2}+c$= $-\frac{m^{2}}{2}-\frac{n^{2}}{2}+c$


$\because$ $-\frac{m^{2}}{2}-\frac{n^{2}}{2}+c < c$

$\therefore$ Option C should be the right answer.

• edited by
Position:
Show:

Related questions

0 0 votes
0 0 answers
875
875 views
0 0 votes
1 1 answer
1.2k
1.2k views
sup739 asked Feb 18, 2024
1,181 views
Is Vector Subspace, Span, Basis, Dimension part of the gate CSE engineering mathematics linear algebra syllabus ?
1 1 vote
1 1 answer
329
329 views
kabir123 asked Jan 7
329 views
<! StartFragment >A device is constantly transferring the data with 20KBps speed to memory using DMA. Assume that the transmission time to memory from IO once the data (...
1 1 vote
2 2 answers
415
415 views
Furious asked Oct 2, 2025
415 views
what does the following algorithm approximate?(Assume m>1,Epsilon>0) x=m; y=1; while (x-y>Epsilon) { x=(x+y)/2; y=m/x;}print(x);a)log...