Re: Stability of a Companion Matrix
- From: Roger Bagula <rlbagulatftn@xxxxxxxxx>
- Date: Mon, 13 Mar 2006 15:45:09 GMT
Erasmus wrote:
Linearizing a nonlinear differenceThe characteristic polynomial :
equation within neighborhood of its fixed
point yields an n x n companion matrix A =
[ 0 1 0 ... 0 ]
[ 0 0 1 ... 0 ]
[ 0 0 0 ... 0 ]
[ ... ... ... ... ... ]
[ 0 0 0 0 1 ]
[ x_1 x_2 x_3 ... x_n ]
where each of the elements of the last row is
x_i = [1 - 1 / (1 + r)] a_i
a_i = any real number
for i = 1 ... n.
If r = 0, then x_i = 0 for all i and
the matrix has all eigenvalues at the
origin.
My question is, what is the
range of r such that every eigenvalue
of the matrix is within the unit circle?
Thanks.
z^n-Sum[x[m+1]*z^m,{m,0,n}]=0
You have stated more or less the Salem Polynomial problem
for the x[n] being Integers.
You want all the roots inside the unit disk right?
It might be easier for real number x[n],
than integers... but I would look at Salem numbers if I were you.
.
- Follow-Ups:
- Re: Stability of a Companion Matrix
- From: Roger Bagula
- Re: Stability of a Companion Matrix
- References:
- Stability of a Companion Matrix
- From: Erasmus
- Stability of a Companion Matrix
- Prev by Date: Re: Stability of a Companion Matrix
- Next by Date: nonlinear multiobjective test problems
- Previous by thread: Re: Stability of a Companion Matrix
- Next by thread: Re: Stability of a Companion Matrix
- Index(es):
Relevant Pages
|