spectral decomposition

**bernoli** · Jun 15th 2011, 08:41 AM

Consider the spectral decomposition , let

Gj=xjyjT for j=1,2,3 ,n so that A=segma mjGj .

a-Show that each Gj has rank one and that G1,…..Gn are mutually orthogonal in the sense that GkGj=mkjGk for 1<j , k<n

b-if p(l) is any scalar polynomial show that p(A) is simple and p(A)= segma p(mj)Gj.

The matrix Gj, 1<j<n, is referred to as the constituent matrix of A associated with mj .

**topsquark** · Jun 15th 2011, 08:53 AM

Originally Posted by bernoli

Consider the spectral decomposition , let

Gj=xjyjT for j=1,2,3 ,n so that A=segma mjGj .

a-Show that each Gj has rank one and that G1,…..Gn are mutually orthogonal in the sense that GkGj=mkjGk for 1<j , k<n

b-if p(l) is any scalar polynomial show that p(A) is simple and p(A)= segma p(mj)Gj.

The matrix Gj, 1<j<n, is referred to as the constituent matrix of A associated with mj .

What have you been able to do on this problem so far?

-Dan

Thread: spectral decomposition