Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Suppose I have a singular matrix given by

$$A = \begin{pmatrix} a_{11} & a_{12} & a_{13} & a_{14}\\ a_{12} & a_{11} & a_{14} & a_{13}\\ a_{31} & a_{32} & a_{33} & a_{34}\\ a_{32} & a_{31} & a_{34} & a_{33} \end{pmatrix}$$ Which is a homogeneous system in $(x_1, x_2, x_3, x_4)^T$ where there is a variable $\Lambda$ in the coefficients that makes the matrix singular if chosen such that the determinant is zero.

Now I have the matrices: $$B = \begin{pmatrix} a_{11} + a_{12} & a_{13} + a_{14}\\ a_{31} + a_{32} & a_{33} + a_{34} \end{pmatrix}$$

$$C = \begin{pmatrix} a_{11} - a_{12} & a_{13} - a_{14}\\ a_{31} - a_{32} & a_{33} - a_{34} \end{pmatrix}$$

which are homogeneous systems in $(x_1 + x_2, x_3 + x_4)^T$ and $(x_1 - x_2, x_3 - x_4)^T$ respectively.

If I make the matrices $B$ and $C$ out of $A$, and I determine the $\Lambda$ such that those determinants are zero, can I say anything about the determinant of $A$ with those values of $\Lambda$?

The idea is that I won't have a fourth degree polynomial to solve (which Maple or Mathematica can seem to do in this case) and then I can obtain the original result by summing or subtracting and such.

share|improve this question

1 Answer 1

up vote 3 down vote accepted

Yes. If the $a_{44}$ in the lower right corner of your $A$ is a typo (which I am pretty sure about) and should be an $a_{33}$ instead, then $\det A=\det B\cdot \det C$. There are some different ways to see this; here is one:

Your matrix $A$ can be written as the block matrix $\left(\begin{array}{cc} X&Y\\ U&W\end{array}\right)$, where $X$, $Y$, $U$, $W$ are the following $2\times 2$ matrices:

$X=\left(\begin{array}{cc} a_{11}&a_{12}\\ a_{12}&a_{11}\end{array}\right)$;

$Y=\left(\begin{array}{cc} a_{13}&a_{14}\\ a_{14}&a_{13}\end{array}\right)$;

$Z=\left(\begin{array}{cc} a_{31}&a_{32}\\ a_{32}&a_{33}\end{array}\right)$;

$W=\left(\begin{array}{cc} a_{33}&a_{34}\\ a_{34}&a_{33}\end{array}\right)$.

Now, these matrices $X$, $Y$, $U$, $W$ are circulant matrices, and thus can be diagonalized by the unitary discrete Fourier transform matrix

$F_2=\frac{1}{\sqrt 2}\left(\begin{array}{cc} 1&1\\ 1&-1\end{array}\right)$.

So we have

$X=F_2\mathrm{diag}\left(a_{11}+a_{12},a_{11}-a_{12}\right)F_2^{-1}$;

$Y=F_2\mathrm{diag}\left(a_{13}+a_{14},a_{13}-a_{14}\right)F_2^{-1}$;

$Z=F_2\mathrm{diag}\left(a_{31}+a_{32},a_{31}-a_{32}\right)F_2^{-1}$;

$W=F_2\mathrm{diag}\left(a_{33}+a_{34},a_{33}-a_{34}\right)F_2^{-1}$.

As a consequence, the block matrix $A=\left(\begin{array}{cc} X&Y\\ U&W\end{array}\right)$ can be written as

$A=\left(\begin{array}{cc} F_2&0\\ 0&F_2\end{array}\right)\left(\begin{array}{cc} \mathrm{diag}\left(a_{11}+a_{12},a_{11}-a_{12}\right) & \mathrm{diag}\left(a_{13}+a_{14},a_{13}-a_{14}\right) \\ \mathrm{diag}\left(a_{31}+a_{32},a_{31}-a_{32}\right) & \mathrm{diag}\left(a_{33}+a_{34},a_{33}-a_{34}\right) \end{array}\right) \left(\begin{array}{cc} F_2&0\\ 0&F_2\end{array}\right)^{-1}$

(check this!), so that

$\det A = \det \left(\begin{array}{cc} \mathrm{diag}\left(a_{11}+a_{12},a_{11}-a_{12}\right) & \mathrm{diag}\left(a_{13}+a_{14},a_{13}-a_{14}\right) \\ \mathrm{diag}\left(a_{31}+a_{32},a_{31}-a_{32}\right) & \mathrm{diag}\left(a_{33}+a_{34},a_{33}-a_{34}\right) \end{array}\right) $.

Now, the determinant on the right hand side can be even simplified by transposing the second row with the third row and transposing the second column with the third column:

$\det A = \det \left(\begin{array}{cccc} a_{11}+a_{12} & a_{13}+a_{14} & 0 & 0 \\ a_{31}+a_{32} & a_{33}+a_{34} & 0 & 0 \\ 0 & 0 & a_{11}-a_{12} & a_{13}-a_{14} \\ 0 & 0 & a_{31}-a_{32} & a_{33}-a_{34} \end{array}\right)$.

Now the matrix on the right hand side is obviously just the block matrix $\left(\begin{array}{cc} B&0\\ 0&C\end{array}\right)$, so its determinant is $\det B\cdot \det C$.

share|improve this answer
    
It is indeed a typo, I'll study your answer. –  Jonas Teuwen Feb 17 '11 at 13:42

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.