Scribd
Upload
72 views

Homework 3

This document contains instructions for 5 homework problems on linear system theory. Problem 1 involves finding the LU decomposition of a matrix. Problem 2 asks to show that left and right eigenvectors of a matrix corresponding to different eigenvalues are orthogonal. Problem 3 covers the discrete Fourier transform (DFT) as an orthogonal basis change, including properties of the DFT matrix and using the DFT to represent circular convolution as a matrix multiplication. Problems 4 and 5 involve properties of orthogonal matrices, including reflecting a vector across a hyperplane and determining if a 2D orthogonal matrix represents a rotation or reflection.

Uploaded by

jerry
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
72 views

Homework 3

This document contains instructions for 5 homework problems on linear system theory. Problem 1 involves finding the LU decomposition of a matrix. Problem 2 asks to show that left and right eigenvectors of a matrix corresponding to different eigenvalues are orthogonal. Problem 3 covers the discrete Fourier transform (DFT) as an orthogonal basis change, including properties of the DFT matrix and using the DFT to represent circular convolution as a matrix multiplication. Problems 4 and 5 involve properties of orthogonal matrices, including reflecting a vector across a hyperplane and determining if a 2D orthogonal matrix represents a rotation or reflection.

Uploaded by

jerry
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

ELEC505/ELEC405 Handout #4

Linear System Theory Wednesday, November 25, 2020


Homework Set #3
Due: Monday, December 7, 2020.
1. LU Decomposition Exercise...
Find the L, U decomposition for
 
6 30 36
 2 10 4 
5 4 2

2. Orthogonality of Left-Right Eigenvectors


Let A ∈ Rn×n and if λ, µ are to distinct eigenvalues of A, i.e. µ 6= λ, then show that
any left eigenvector of A corresponding to µ is orthogonal to any right eigenvector of
A corresponding to λ.
3. DFT as an Orthogonal Basis Change
Consider C N the vector space of N dimensional complex vectors. We can define a basis
F={f1 , . . . , fN } where
 
fk,1
 fk,2  1 j2π(k−1)(l−1)
fk =  , fk,l = e . (1)
 
.. N
 .  N
fk,N

(a) Is F an orthogonal basis? Is it an orthonormal basis?


(b) Define the matrix
 
F= f1 f2 . . . fN . (2)
Is F unitary? What is the inverse of F?
 T
(c) Write down F for N = 4. What are the coordinates of x = 1 1 0 0
corresponding to the basis F? What is the FFT of x?
(d) Suppose {xn : n = 0, . . . 3}, {hn : n = 0, . . . 3} and {yn : n = 0, . . . 3} are discrete
time sequences of length 4. We are also given that yn is equal to the circular
convolution of xn and hn , i.e.,
yn = hn ~ xn
Defining
 

x0 y0
 x1   y1 
x= 
 x2  y=
 y2  ,

x3 y3

1
we would like to write the relation

y = Hx (3)

Find H in terms of hn , n = 0, . . . 3.
(e) Check out the following multiplication:

Hfk (4)

What is your comment about the result?


(f) Pick arbitrary values for {h0 , h1 , h2 , h3 }. Find the eigenvalues and eigenvectors
of H using MATLAB/Python/Julia, find eigenvectors and eigenvalues (e.g., in
Matlab )by typing
[V D]=eig(H)
where V is the matrix containing eigenvectors, and D is the matrix containing
eigenvalues.) Compare V with F, what is your comment.
(g) If we change the basis from the standard basis to F, what would be the relation
between y̆ and x̆, where y̆ and x̆ are the new coordinate vectors for y and x in Eq.
(3) respectively, corresponding to basis F. In other words find H̆ where y̆ = H̆x̆.
What is the effect of basis change as far as the mapping in (3) is concerned? Note
that this comment is independent of the values of {h0 , h1 , . . . , hN −1 }....

4. Reflection through a hyperplane.


Find the matrix Q ∈ Rn×n such that the reflection of x through the hyperplane
{z | aT z = 0} is given by Qx. Verify that matrix Q is orthogonal. (To reflect
x through the hyperplane means the following: find the point z on the hyperplane
closest to x. Starting from x go in the direction z − x through the hyperplane to a
point on the opposite side, which has the same distance to z as x does.)

5. Orthogonal matrices

(a) Show that if U and V are orthogonal then so is UV.


(b) Suppose that U ∈ R2×2 is orthogonal. Show that U is either a rotation or a
reflection. Make clear how you decide whether a given orthogonal U is a rotation
or reflection.

You might also like