Re: [R] eigenvalues of a circulant matrix

From: Globe Trotter <itsme_410_at_yahoo.com>
Date: Tue 03 May 2005 - 00:22:35 EST

Agreed. As may a circulant matrix if a_i = a_{p-i+2}

>
> A circulant matrix is an n x n matrix whose rows are composed of
> cyclically shifted versions of a length-n vector. For example, the
> circulant matrix on the vector (1, 2, 3, 4) is
>
> 4 1 2 3
> 3 4 1 2
> 2 3 4 1
> 1 2 3 4
>
> So circulant matrices are a special case of Toeplitz matrices.
> However a circulant matrix cannot be symmetric.
>
> The eigenvalues of the forgoing circulant matrix are 10, 2 + 2i,
> 2 - 2i, and 2 --- certainly not roots of unity.

The eigenvalues are 4+1*omega+2*omega^2+3*omega^3. omega=cos(2*pi*k/4)+isin(2*pi*k/4) as k ranges over 1, 2, 3, 4, so the above holds.

 Bellman may have
> been talking about the particular (important) case of a circulant
> matrix where the vector from which it is constructed is a canonical
> basis vector e_i with a 1 in the i-th slot and zeroes elsewhere.

No, that is not true: his result can be verified for any circulant matrix, directly.

> Such a matrix is in fact a unitary matrix (operator), whence its
> spectrum is contained in the unit circle; its eigenvalues are indeed
> n-th roots of unity.
>
> Such matrices are related to the unilateral shift operator on
> Hilbert space (which is the ``primordial'' Toeplitz operator).
> It arises as multiplication by z on H^2 --- the ``analytic''
> elements of L^2 of the unit circle.
>
> On (infinite dimensional) Hilbert space the unilateral shift
> looks like
>
> 0 0 0 0 0 ...
> 1 0 0 0 0 ...
> 0 1 0 0 0 ...
> 0 0 1 0 0 ...
> . . . . . ...
> . . . . . ...
>
> which maps e_0 to e_1, e_1 to e_2, e_2 to e_3, ... on and on
> forever. On (say) 4 dimensional space we can have a unilateral
> shift operator/matrix
>
> 0 0 0 0
> 1 0 0 0
> 0 1 0 0
> 0 0 1 0
>
> but its range is a 3 dimensional subspace (e_4 gets ``killed'').
>
> The ``corresponding'' circulant matrix is
>
> 0 0 0 1
> 1 0 0 0
> 0 1 0 0
> 0 0 1 0
>
> which is an onto mapping --- e_4 gets sent back to e_1.
>
> I hope this clears up some of the confusion.
>
> cheers,
>
> Rolf Turner
> rolf@math.unb.ca

Many thanks and best wishes!



R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Received on Tue May 03 00:27:12 2005

This archive was generated by hypermail 2.1.8 : Fri 03 Mar 2006 - 03:31:31 EST