1.) The Reverse Triangle Inequality
[edit]
Consider two vectors x and y. Without loss of generality, say
Since
is a vector norm, the triangle inequality holds. Then if you think of x as x=x-y+y
Since
,
and thus
. So, finally:
2.) An Induced Matrix Norm
[edit]
If [s1,s2,...,sn] are the sums of the values of A (that is, sk=sum(i=1 to n)(a_{i,k))
One definition of an induced matrix norm is that
It follows immediately that if A=I, then
However,
, so the Frobenius norm is not an induced norm.
4.) Spectral Radius
[edit]
For nxn matrix A and
we can find a natural norm satisfying
The left hand equality we verified in class. The difficult bit is the right-hand equality.
We can find a non-singular matrix P such that
Where
... Argh. Incomplete, this proof is. The completion can be found here
5.) Inequities are Sharp
[edit]
This inequality came about due to the study of error in linear systems. To be specific, given a system
we want to know what will happen if the value of x changes by some error. That is, given an error e, we want to know the relative error in respect to the perturbed system
. By subtraction, it follows that
.
To the end of studying this, we know four things:
By composing the 1st and 4th inequalities, we find:
Shuffle around the terms a bit:
By composing the 2nd and 3rd inequalities, we find:
Shuffle around the terms a bit:
And thus we arrive at the desired inequality:
A simple system that satisfies equalities on all counts is this:
Let
It follows that
for both the 2-norm and the
-norm, and all inequalities are satisfied.
6.) A Symmetric Matrix Has Real Eigenvalues
[edit]
First consider a general Hermitian matrix A. If
is an eigenvalue of A with associated eigenvector x, then
So, where H indicates the transpose conjugate, and * the conjugate:
Also, since
,
So we then have
, so
, and thus
is real.
Now consider a symmetric matrix B. Since the conjugate of a real matrix is the matrix itself,
, so B is Hermitian, and thus has real eigenvalues.
7.) The Gershgorin Theorem
[edit]
Consider a n x n matrix A and
. Then let
be the union of p Gershgorin circles with
disjoint from
, the remaining n-p Gershgorin circles. For
, let us define
with
Then B(1)=A, and B(0) is the diagonal matrix with diagonal entries corresponding to the diagonal entries of A. Therefore, each of the eigenvalues of B(0) is the center of a Gershgorin circle. So, exactly p of the eigenvalues of B(0) lie in
. The eigenvalues of B are the zeros of its characteristic polynomial, which is a polynomial with coefficients that are continuous variables of
. Thus, the zeros of the characteristic polynomial are also continuous functions of
. As
goes from 0 to 1 the eigenvalues of
move along continuous paths in the complex plane. At the same time, the radii of the Gershgorin circles increase from 0 to their full radius. Since p of the eigenvalues lie in
when
and these disks are disjoint from
, the p eigenvalues must still lie in the union of the disks when
.
So
Then note:
Finally, we arrive at:
9.) More on Block Matrices
[edit]
a.) False! Consider the following submatrices:
Obviously,
, so
. However, if you consider the whole matrix
Then you see that
by expansion along the 3rd row and then the 2nd row.
b.) Also false! Consider the following matrices:
Then
Now,
,
, but
10.) Operation Count
[edit]
The calculation for the diagonal entries of the Cholesky decomposition is:
The only place in here that multiplications occur is in the sum - one for each term of the summation, for a total number of
multiplications
The calculation for the entries below the diagonal is:
To determine the number of multiplications here, first consider how many multiplications/divisions occur assuming you are in the kth column. You will always have one division, plus one multiplication from each term of the sum, of which there are k-1. So each entry below the diagonal in the first column will take one multiplication/division, and there are n-1 of these. Each entry in the second column will take two multiplications/divisions, and there are n-2 of them. Continuing onward with the yields this for the total number of multiplications/divisions in the non-diagonal entries:
So the total number of multiplications and divisions involved is: