Part A: Working With Matrices

NATIONAL UNIVERSITY OF SINGAPORE
Department of Mathematics
MA1506 Laboratory 3 (MATLAB)
Part A: Working With Matrices
MATLAB actually stands for matrix laboratory, and as the name suggests, it was designed
for working with matrices.
We can input an m n matrix A by
A = [ row 1; row 2; ... ; row m ]
where the n entries of each row are separated by one or more blank spaces. For example:
>> A=[3 2 -1 ; 0 1 0 ; 1 2 2]
The following commands perform basic operations on matrices A and B:
A+B matrix addition
A-B matrix subtraction
tA scalar multiplication, with t scalar
AB matrix multiplication
A
n raising a square matrix A to a positive integral power n

A transpose of A
inv(A) inverse of an invertible square matrix A
det(A) compute the determinant of a square matrix A
trace(A) compute the trace of a square matrix A
Practice
1. We shall construct a matrix M to forecast weather. To predict the weather 4 days
from now and 30 days from today, we can do the following:
>> M=[ 0.6 0.3 ; 0.4 0.7 ]
>> M^4
>> M^30
2. We know that rotation about z-axis and x-axis in 3 dimensions do not commute.
Use MATLAB to verify that the two matrix products are really dierent.
1 0 0
0 0 1
0 1 0
0 1 0
1 0 0
0 0 1
0 1 0
1 0 0
0 0 1
1 0 0
0 0 1
0 1 0
.
1
>> A=[ 1 0 0 ; 0 0 -1 ; 0 1 0]
>> B=[ 0 -1 0 ; 1 0 0 ; 0 0 1]
>> A*B
>> B*A
3. Compute the transpose of the matrix M =
1 2 4
6 8 9
and N =
1 2 4
6 8 9
2 1 0
.
Verify that N
T
+ N is symmetric and N
T
N is anti-symmetric.
>> M=[ 1 2 4 ; 6 8 9]
>> M
>> N=[ 1 2 4 ; 6 8 9 ; 2 1 0]
>> N + N
>> N - N
4. Using M and N dened previously, predict what happens if we try to perform the
matrix addition M + N and the matrix multiplications MN and NM. Verify your
prediction.
>> M + N
>> M*N
>> N*M
5. Determine if the following matrices, C1 =
2 7 5
1 3 1
4 13 3
and C2 =
2 7 5
1 3 1
4 13 4
are invertible.
>> C1=[ 2 7 5 ; 1 3 -1 ; 4 13 3]
>> C2=[ 2 7 5 ; 1 3 -1 ; 4 13 4]
>> det(C1)
>> det(C2)
>> inv(C2)
>> inv(C2)*C2
6. Let D =
5 7 9
8 8 1
20 4 6
and E =
1 2 4
1 1 1
1 1 0
. Verify that (DE)

1
= E
1
D
1
and
(DE)
T
= E
T
D
T
.
2
>> D=[ 5 7 9 ; 8 8 1 ; 20 4 6]
>> E=[ 1 2 4 ; 1 1 1 ; 1 1 0]
>> inv(D*E)
>> inv(E)*inv(D)
>> (D*E)
>> E*D
7. Recall that in previous labs, we dened an array of values with
>> x = 0: 0.2 : 1
>> x*x
We should really view this as a 1 6 row vector x. (Double click on x.) Hence we
get an error when we multiply x to itself. To get the dot product x x, when x is a
row vector, we use xx
T
, i.e.
>> x*x
What will happen if we use x
T
x?
>> x*x
8. What happens now if y is a row vector?
>> y = [2 ; 1 ; 5]
>> y*y
>> y*y
9. Consider the following linear system of equations.
x
1
x
2
+ x
3
= 4
x
1
+ x
2
= 1
x
1
+ 2x
2
x
3
= 0.
We rewrite this system as a matrix equation Ax =
b and calculate the determinant

of A. Note that
b is a column vector.
>> A= [1 -1 1; 1 1 0; 1 2 -1]
>> b= [ 4; 1; 0]
>> det(A)
3
Since det(A) = 0, the matrix is non-singular. We can then solve the system by
nding the inverse of A. The required solution is x = A
1
b.
>> x= inv(A)*b
10. In 1966, Leontief used his input-output model to analyze the Israeli economy by
dividing it into three segments: Agriculture (A), Manufacturing (M), and Energy
(E), as shown in the following technology matrix.
Output \ Input A M E
A $0.30 $0.00 $0.00
M $0.10 $0.20 $0.20
E $0.05 $0.01 $0.02
The export demands on the Israeli economy are listed as follows: Agriculture: $140
million, Manufacturing: $20 million and Energy: $2 million.
To nd the total output for each sector required to meet both internal and external
demand, we must solve the following system
A = 0.30A + 0.00M + 0.00E + 140
M = 0.10A + 0.20M + 0.20E + 20
E = 0.05A + 0.01M + 0.02E + 2.
Using the technology matrix T, we have x = (I
3
T)
1
b.
>> T=[ .3 0 0 ; .1 .2 .2 ; .05 .01 .02]
>> b=[140; 20; 2]
>> x= inv(eye(3)- T)*b
The required output is approximately, A = $200 m , M = $53 m and E = $13 m.
Note that eye(3) is the MATLAB command for the 3 3 identity matrix.
4
Part B: Eigenvectors And Eigenvalues
We know that the matrix
1 2
2 2
has an eigenvector
2
1
with corresponding eigen-

value 2. This can be easily computed using the following commands:
>> A=[ 1 2 ; 2 -2]
>> [P D]=eig(A)
The second command computes the eigenvectors of the matrix A and stores them as column
vectors in the matrix P. At the same time, the corresponding eigenvalues are stored as the
diagonal entries of the matrix D. To work with the eigenvector corresponding to eigenvalue
2, we extract the second column of P and call it v.
>> v = P(:,2)
>> A*v
From our understanding of eigenvectors, Av should give us 2v. Note that v is not
2
1
but a multiple of it. Remember that eigenvectors are never unique, and the eig function
will compute eigenvectors with lengths 1. To get our familiar
2
1
, we multiply the vector

v by the scalar 1/v(2), where v(2) is the second coordinate of the vector v
>> y = v/v(2)
>> A*y
We should recognize that that the eig function is actually trying to diagonalize the matrix
A. Recall that A = PDP
1
, where D is the diagonal matrix with eigenvalues of A as its
entries, and P is a square matrix where the columns are the corresponding eigenvectors.
Verify this by
>> P*D*inv(P)
5
Exercise 3
1. Compute the determinant of
1 1 13 6.5 1.5
1 1 0 0 0
1 2 1 3 1
1 7 1 4 9
1.5 2 21 3 1
.
(i) 615.75
(ii) -516.5
(iii) 765.0
(iv) 716.25
2. A car rental agency has three branches A, B and C. The company policy allows cars
to be rented from and returned to any one of the three branches. A statistical study
revealed that the chances of cars being returned to the same branch where they were
rented are 70%, 50% and 40% respectively. There is a 20% likelihood of cars rented
from branch A being returned to branch B. A 30% chance of cars rented from branch
B being returned to branch C and also a 30% chance of cars rented from branch C
being returned to branch A. Assuming the company started with 100 cars at each
branch, in the long run, approximately how many cars will remain at branch C?
(i) 23
(ii) 70
(iii) 134
(iv) 32
3. A small countrys economy is divided into three segments: Electronics (E), Manufac-
turing (M), and Pharmaceutical (P), as shown in the following technology matrix.
Output \ Input E M P
E $0.15 $0.20 $0.00
M $0.10 $0.30 $0.00
P $0.12 $0.18 $0.40
6
The export demands are as follows: Electronics: $80 million, Manufacturing: $20
million and Pharmaceutical: $10 million.
The electronics output is approximately (nearest million)
(i) $ 104 million
(ii) $ 100 million
(iii) $ 89 million
(iv) $ 80 million
4. Find the eigenvalues and a matrix P that diagonalizes
5 6 2
0 1 8
1 0 2
.
2 1
5 2
.
5 0 0
1 5 0
0 1 5
.
7. In order to nd the eigenvalues of a matrix A, we solve the characteristic equation
0 = det(A I) = c
n
n
+ c
n1
n1
+ + c
1
+ c
0
,
which is a polynomial equation in . The coecients of the characteristic polynomial
can be found with the command
>> poly(A)
There is a remarkable theorem called the Cayley-Hamilton Theorem which states
that a square matrix A satises its characteristic equation. Hence
c
n
A
n
+ c
n1
A
n1
+ + c
1
A + c
0
I
n
= 0.
Verify this for the three matrices given above.
The End
7

Part A: Working With Matrices

Uploaded by

Copyright:

Available Formats

Part A: Working With Matrices

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Part A: Working With Matrices

Uploaded by

Copyright:

Available Formats

NATIONAL UNIVERSITY OF SINGAPORE

n raising a square matrix A to a positive integral power n

. Verify that (DE)

b and calculate the determinant

with corresponding eigen-

, we multiply the vector

You might also like