Re: Determinant proof



In article <y8zsl3e3nuy.fsf@xxxxxxxxxxxxxxxxxxxx>,
Bill Dubuque <wgd@xxxxxxxxxxxxxxxxxxxx> wrote:
istvankaster@xxxxxxxxx wrote:
On nov. 9, 11:19, ptresadern@xxxxxxxxxxxxxx wrote:

I have three matrices (in Matlab notation):

a = rand(4,4); b = rand(3,3); p = rand(3,4);

I'm trying to find an analytical proof of the following statement:

det(eye(size(b))+b*p*inv(a)*p') = det(eye(size(a))+inv(a)*p'*b*p)

http://en.wikipedia.org/wiki/Sylvester%27s_determinant_theorem

I.e. det(I+AB) = det(I+BA). I don't ever recall seeing this called
"Sylvester's determinant theorem". Does anyone know a reference?

The proof is easy: in the polynomial ring Z[Aij,Bij] simply

take the det of (I+AB)A = A(I+BA), and then cancel det(A)

What is det(A) when A is non-square?

See my prior posts [1] for much more on such "universal" proofs.

Another proof arises from the following Schur decomposition,
but that is far too much effort compared to the above proof.

[ I A ] [ I 0 ] [ I 0 ] [ I A ]
[ ] = [ ] [ ] [ ]
[ B I ] [ B I ] [ 0 I-BA ] [ 0 I ]

[ I A ] [ I-AB 0 ] [ I 0 ]
= [ ] [ ] [ ]
[ 0 I ] [ 0 I ] [ B I ]

It is not too much effort if it works, which it does, whether A is
square or not. It is quite similar to the proof in the paper on
JSTOR cited on Wikipedia: <http://tinyurl.com/37zxgd>.

<http://en.wikipedia.org/wiki/Sylvester's_determinant_theorem>

Rob Johnson <rob@xxxxxxxxxxxxxx>
take out the trash before replying
to view any ASCII art, display article in a monospaced font
.



Relevant Pages

  • Re: Determinant proof
    ... I have three matrices (in Matlab notation): ... I'm trying to find an analytical proof of the following statement: ... --Bill Dubuque ...
    (sci.math)
  • Re: Determinant proof
    ... I have three matrices (in Matlab notation): ... I'm trying to find an analytical proof of the following statement: ... Append 0's to A,B till square. ...
    (sci.math)
  • Determinant proof
    ... I have three matrices (in Matlab notation): ... and I am trying to find an analytical proof of the following ... By experimentation, this seems to hold. ...
    (sci.math)
  • Re: Determinant proof
    ... I have three matrices (in Matlab notation): ... I'm trying to find an analytical proof of the following statement: ...
    (sci.math)
  • Re: Linear Algebra question: newbie
    ... Unfortunately the prime notation has a specific meaning ... in Matlab (Hermitian conjugate for complex matrices, ... A is an orthogonal matrix. ...
    (comp.soft-sys.matlab)

Quantcast