# Homework Help: Eigenvalue Factorization and Matrix Substitution

1. Mar 21, 2010

### the_dialogue

In my literature reviews I found a few things that I can't quite understand.
1. The problem statement, all variables and given/known data

I have the following equation:
http://img717.yfrog.com/img717/6416/31771570.jpg [Broken]

I'm told that by using the eigenvalue factorization:
http://img89.yfrog.com/img89/760/83769756.jpg [Broken]

, I can change the first equation to:
http://img28.imageshack.us/img28/5023/84802099.jpg [Broken]

2. The attempt at a solution

I tried changing Equation 2 to just be (A^T)A and then subbing into the first equation, but I can't quite do anything with those inverses.

Also, what does the exponent of '-2' mean in the context of a 4x4 matrix? Lastly, what is matrix U?

Thank you!

Last edited by a moderator: May 4, 2017
2. Mar 21, 2010

### gabbagabbahey

I think it's probably easiest to start from $\textbf{p}^{T}(\mathbf{\Lambda}+\lambda\textbf{I})^{-2}\textbf{q}=0$ and work your way backwards instead.

$$\textbf{C}^{-2}\equiv\textbf{C}^{-1}\textbf{C}^{-1}$$

You simply square the inverse of the matrix.

3. Mar 21, 2010

### the_dialogue

I'll give it a try gabbagabbahey. Thanks.

Any idea what the matrix "U" is?

4. Mar 21, 2010

### gabbagabbahey

It's the invertible matrix which relates the matrix $\textbf{A}^{T}\textbf{A}\mathbf{\Sigma}$ to the diagonal matrix $\mathbf{\Lambda}$ via a similarity transform. Its columns will be the eigenvectors of $\textbf{A}^{T}\textbf{A}\mathbf{\Sigma}$.

See http://en.wikipedia.org/wiki/Diagonalizable_matrix for a refresher on matrix diagonalization.

5. Mar 21, 2010

### the_dialogue

Yes I recall now. Thanks!