1
$\begingroup$

$M(n,\mathbb{F})$ denotes the set of all $n\times n$ matrices with entries in $\mathbb{F}$.

Assume that $M(n,\mathbb{F})$ is a vector space over $\mathbb{F}$. Suppose $(\mu_{ij}) \in M(n,\mathbb{F})$ where $\mu_{ij} \mu_{ji}=1$ for all $i,j$. Let $U = \{(m_{ij})\in M(n,\mathbb{F}): m_{ij} = \mu_{ij}m_{ij} \text{ for all } i,j\}$.

Suppose $\mu_{ii}=1$ for all $i$; find a basis for $U$ and find the dimension of $U$.

I wrote out few cases (when $n = \{2, 3, 4\})$ and these help me see that the dimension of $U$ should be given by the expression $\frac12 n(n+1)$.

Case $n=2$:

$\begin{pmatrix} m_{11} & m_{12} \\ m_{21} & m_{22} \end{pmatrix} = \begin{pmatrix} m_{11} & m_{12} \\ \mu_{21}m_{12} & m_{22} \end{pmatrix}$

Case $n=3$:

$\begin{pmatrix} m_{11} & m_{12} & m_{13} \\ m_{21} & m_{22} & m_{23} \\ m_{31} & m_{32} & m_{33} \end{pmatrix} = \begin{pmatrix} m_{11} & m_{12} & m_{13} \\ \mu_{21}m_{12} & m_{22} & m_{23} \\ \mu_{31}m_{13} & \mu_{32}m_{23} & m_{33} \end{pmatrix}$

Case $n=4$:

$\begin{pmatrix} m_{11} & m_{12} & m_{13} & m_{14} \\ m_{21} & m_{22} & m_{23} & m_{24} \\ m_{31} & m_{32} & m_{33} & m_{34} \\ m_{41} & m_{42} & m_{43} & m_{44} \end{pmatrix} = \begin{pmatrix} m_{11} & m_{12} & m_{13} & m_{14} \\ \mu_{21}m_{12} & m_{22} & m_{23} & m_{24} \\ \mu_{31}m_{13} & \mu_{32}m_{23} & m_{33} & m_{34} \\ \mu_{41}m_{14} & \mu_{42}m_{24} & \mu_{43}m_{34} & m_{44} \end{pmatrix}$

Doing the following

$$\begin{pmatrix} m_{11} & m_{12} \\ m_{21} & m_{22} \end{pmatrix} = \begin{pmatrix} m_{11} & m_{12} \\ \mu_{21}m_{12} & m_{22} \end{pmatrix}= \begin{pmatrix} m_{11} & 0 \\ 0 & 0 \end{pmatrix} + \begin{pmatrix} 0 & m_{12} \\ \mu_{21}m_{12} & 0 \end{pmatrix}+ \begin{pmatrix} 0 & 0 \\ 0 & m_{22} \end{pmatrix}$$

we can see that the $2\times 2$ matrices in $U$ are spanned by 3 linearly independent matrices, which agrees with the dimension listed above. Repeat this process for greater dimensions. Essentially my solution depends greatly upon the reader picking up on the pattern.

QUESTION: Is there a better way to present the solution to this problem?

1 Answers 1

1

Hint:

If $D$ is the diagonal and $U$ is the upper triangular part you can write $$M = D+U+ K_1 U^T K_2$$

Now can you find $K_1,K_2$ matrices in terms of the $\mu$s which accomplish what you want?