Chapter 7 Rank 1 perturbation Brauers Flashcards
Give two ways to use real numbers to manipulate matrix eigenvalues
If I am manipulating the eigenvalues of A I can use:
A + bI
or
bA where b is a real number
Why would we want to manipulate eigenvalues
As we want the ideal situation to find the limit of a matrix where we are dealing with a positive stochastic matrix ideally with fast convergence
How do the eigenvalue of AB and BA relate?
They have the same non zero eigenvalues
What do we know about a rank one stochatic matrix?
K has all entries 1/m so eigenvalues are 1 with multiplicity 1 and 0 with multiplicty m-1
Any rank one matrix of the for 1/n (e)(e^transpose) or ew^tranpose where (w^tranpose)(e)=1 has limit of what?
ew^tranpose … aka itself
State brauers theorem
Let A be any nxn matrix with eigenvalues k,,…,kn and v being the eigenvector Av =kv
Let B be the nxn rank 1 matrix vy^tranpose.
Then the eigenvalues of A + B are the same as the
eigenvalues of A except k1 is replace by k1+ y^tranpose(v). Eigenvalues: k, + y^tranpose(v) ,…, kn (only 1 transformed)
If i multiply matrix A by b what are the eigenvalues of bA
B multiplied by eigenvalues of A