Re: eigenvalue magnitude question
- From: spellucci@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx (Peter Spellucci)
- Date: Thu, 9 Mar 2006 16:08:49 +0000 (UTC)
In article <uCRPf.35282$Ru5.34795@xxxxxxxxxxxxxxxxxxxx>,
"NoIdea" <foru@xxxxxxxxx> writes:
would the fact that the highest singular value of a real/complex matrix is
less than 1 , necessarily imply that all of its eigenvalues have magnitudes
less than 1 ?
thanks
take
A*x=lambda*x and w.l.o.g. ||x||_2=1
insert
A=U*S*V' U,V unitary
multiply by U'
S*V'*x= lambda*U'*x
take the 2-norm on both sides and use submultiplicativity
max|S(i,i)| ||V'*x||_2 >= abs(lambda)||U'*x||_2
but ||x||_2=1 and U', V' unitary
done
hth
peter
.
- References:
- eigenvalue magnitude question
- From: NoIdea
- eigenvalue magnitude question
- Prev by Date: Re: A Long Thin No Stretch No Weight Sagging Film Is Supported By Horizontal Parallel Beams and Filled With a Liquid
- Next by Date: Re: Determinant of an integer matrix
- Previous by thread: eigenvalue magnitude question
- Next by thread: Determinant of an integer matrix
- Index(es):
Relevant Pages
|