I am not a student majoring in math, so I apologize if my question is stupid to you.
Assume we have ${\bf X_1}$, ${\bf X_2}$, and ${\bf A}$ that are all Hermitian and positive semidefinite matrices with rank $1$. Or, more precisely, $\bf A$, ${\bf X_1}$ and ${\bf X_2}$ are formed by
$${\bf A}={\bf aa}^H \qquad \qquad {\bf X_1}={\bf x}_1 {\bf x}_1^H \qquad \qquad {\bf X_2}={\bf x}_2 {\bf x}_2^H$$
Is it true that
if $\mbox{tr}({\bf A} {\bf X_1}) \geq \mbox{tr}({\bf A} {\bf X_2})$, then $\mbox{tr}({\bf X_1}) \geq \mbox{tr}({\bf X_2})$?
if $\mbox{tr}({\bf X_1}) \geq \mbox{tr}({\bf X_2})$, then $\mbox{tr}({\bf A} {\bf X_1}) \geq \mbox{tr}({\bf A} {\bf X_2})$?
I assume this is an extension of the linear case where we have y=a*x with a>0, and it is clearly true if we have x1>x2 then y1>y2, but I am not sure if this still holds for matrix case. I have tried to search for 'matrix function' and 'trace inequalities' but I still couldn`t get an answer.
Thanks in advance!
The first claim is false, consider for example $$A = \begin{pmatrix} 1 & 0 \\ 0 & 0\end{pmatrix} \qquad X_1 = \begin{pmatrix}0 & 0 \\ 0 & 0\end{pmatrix} \qquad X_2 = \begin{pmatrix}0 & 0 \\ 0 & 2\end{pmatrix}.$$