Degree of liberty of a matrix 2x2

  • Level: Graduate 
  • Thread starter Thread starter brunotolentin.4
  • Start date Start date
  • Tags Tags
    Degree Matrix
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
2 replies · 2K views
brunotolentin.4
Messages
6
Reaction score
0
How many degree of liberty exist, actually, in a matrix 2x2 ?

I think that is three! Because the conic equation can be wrote like this:

[tex] \begin{bmatrix}<br /> A & B\\ <br /> C & D<br /> \end{bmatrix}<br /> :\begin{bmatrix}<br /> x^2 & xy\\ <br /> yx & y^2<br /> \end{bmatrix}<br /> +<br /> \begin{bmatrix}<br /> E\\ <br /> F<br /> \end{bmatrix}<br /> \cdot<br /> \begin{bmatrix}<br /> x\\ <br /> y<br /> \end{bmatrix}<br /> +G=0[/tex]

But, xy = yx, thus ... + Bxy + Cyx +... = ... + (B+C)xy + ...

So: [tex] \begin{bmatrix}<br /> A & (B+C)\\ <br /> 0 & D<br /> \end{bmatrix}<br /> :\begin{bmatrix}<br /> x^2 & xy\\ <br /> yx & y^2<br /> \end{bmatrix}<br /> +<br /> \begin{bmatrix}<br /> E\\ <br /> F<br /> \end{bmatrix}<br /> \cdot<br /> \begin{bmatrix}<br /> x\\ <br /> y<br /> \end{bmatrix}<br /> +G=0[/tex]

Another example: the coefficients of the equation Ay'' + By' + Cy = 0 has three degree of liberty (A, B and C) and it can be converted in a matrix:

y' = a y + b y'
y'' = c y + d y'

So, exist more and more examples that I could give here. But, the felling that I have is the a matrix 2x2 has 3 degree of liberty, although of has four coefficients... My feeling is correct?
 
Physics news on Phys.org
You are looking for degrees of freedom.
Are the matrices you are looking at the most general?

Look for example at the general linear group ##Gl(n, \mathbb{R})##
This a group containing the ##n\times n## invertible matrices.
You can prove that this group has dimension ##n^2##.

If we take n = 2 you can show this in several ways.
The main thing is that the condition that the matrix is invertible reduces to ##\text{det}A\neq 0##.
So let's say we have such a matrix ##A = \left[a_{ij}\right]##.

The determinant condition is ##\text{det}A = a_{11}a_{22} - a_{12}a_{21} \neq 0##.
It is clear that when we know three components (##a_{11},\,a_{22}\text{ and }a_{12}##), the fourth still has a lot of freedom.
[tex]a_{21} \neq \frac{a_{11}a_{22}}{a_{12}}[/tex]

Clearly there is some symmetry in your examples.
An example is found by considering the Special linear group.
This is the subgroup ##Sl(n, \mathbb{R}) \subset Gl(n, \mathbb{R})## with ##\text{det}A = 1##.
You can see how knowledge of 3 elements gives you the fourth in the case of n = 2.

I'm not entirely familiar with your first notation (what is the colon?).
I also don't understand your point.

The second example is trivial, you start with three coefficients so that will be reflected in your matrix.
 
I agree with Joris L . The number degrees of freedom depend on the context. Would you elaborate on what you are after?