Tutorial 9 Iterative Methods and Matrix Norms
Tutorial 9 Iterative Methods and Matrix Norms
Iterative Methods
and Matrix Norms
Iterative processes
1 i n, i 1 lim u k 0
k
2
Spectral radius of a matrix
(T) max k
1 k n
The largest eigenvalue (the first after ordering) governs the dynamic
system, i.e.:
u k c1 k1 v1
following axioms:
a. Positivity : v 0; v 0v 0
b. Linearity : cv c v ;
c. Triangle inequality : vw v w ;
v
p
vp p
k 1 p
k 1
and its three special cases, p=1: sum of abs. values, p=2: Euclidian
norm, and p=∞: the maximal absolute value element of the vector.
4
Norms Equivalence
v, c v B v A C v B (2)
Proof (outline):
5
Induced Matrix Norm
The induced norm of the matrix T is the maximum factor by which it can
multiply the norm ( = the ‘length’ ) of the input vector.
Tv
T max A
max T v A
A v v A
v A
1
The last equality is due to the fact that matrices are linear operators and
vector length can be reduced.
6
The Norm Axioms are Satisfied
T A
max T v A
v A
1
n
w T v w j t ji vi t jk 1 t jk w A 0
i 1
T A
max Tu A
Tv A
0
u A
1
(b) Linearity:
cT A
max cT v A
c max T v A
c T A
v A
1 v A
1
cT v ct ij v j c t ij v j c T v
i 1..n; j1..m i 1..n ; j1..m
8
Natural Matrix Norm
The right hand side is the norm of the sum of two vectors and therefore
must satisfy the triangle inequality for the vector norm:
max A v B v
v 1 A
max A v
v 1
A
Bv A
A A
9
Matrix norm for Convergence Estimation
The norms of matrices are a useful alternative to the spectral radius in
evaluation of matrix convergence due to the following
Av A v
k k
Continuing the iterative process we obtain: A v A v
10
Spectral radius is not a norm.
The matrix norm is a sufficient but not necessary condition for testing for
convergence: on the next slide we will see an example of the converging
matrix with Euclidian norm >1.
0 1
Proof: Consider the matrix A
0 0
Its both eigenvalues are 0, and therefore the spectral radius=0. However,
A≠0 and therefore its norm must be positive.
11
Example – See the Gap
Statement: The Euclidian (p=2) norm of a matrix equals its maximum
singular value.
13
Proof of Gerschgorin Circle Theorem
v
Let v be an eigenvector of A with an eigenvalue λ. Let u be a
v
corresponding “normalized” eigenvector.