Matrices

What is a matrix?

A Matrix is just rectangular arrays of items

A typical matrix is a rectangular array of
numbers arranged in rows and columns.

 21 62 33 93
 
A   44 95 66 13 
3x4
77 38 79 33
 Sizing a matrix

By convention matrices are “sized” using

the number of rows (m) by number of
columns (n). 7 3 2 
 21 62 33 93 B  8 4 1 
A   44 95 66 13  3x3
 6 5 9 
3x4
77 38 79 33
11 4
14 7 
C  D  17 
1 x1
4x2 16 8 
 
 22 3
“Special” Matrices
Square matrix: a square matrix is an mxn
matrix in which m = n. 7 3 2 
B  8 4 1 
3 x3
 6 5 9 
Vector: a vector is an mxn matrix where
either m OR n = 1 (but not both).
12 
9
X   Y   7 22 14
4 x1  4  1x 3
 
0
 Addition, Scalar Multiplication, and
Multiplication of Matrices
• aij: the element of matrix A in row i and column j.
• For a square nn matrix A, the main diagonal is:
 a11 a12  a1n 
a a22  a2 n 
A   21
     
 
an1 an 2  ann 

Definition
Two matrices are equal if they are of the same size and if their
corresponding elements are equal.

Thus A = B if aij = bij  i, j. ( for every, for all)

 Addition of Matrices
Definition
Let A and B be matrices of the same size.
Their sum A + B is the matrix obtained by adding together the
corresponding elements of A and B.
The matrix A + B will be of the same size as A and B.
If A and B are not of the same size, they cannot be added, and we
say that the sum does not exist.

Thus if C  A  B, then cij  aij  bij i,j.

 Example
Let A   1 4 7, B   2 5  6, and C   5 4.
0  2 3  3 1 8  2 7 
Determine A + B and A + C, if the sum exist.
Solution
(1) A  B   1 4 7    2 5  6
0  2 3  3 1 8
 1  2 4  5 7  6
0  3  2  1 3  8
  3 9 1.
 3  1 11
(2) Because A is 2  3 matrix and C is a 2  2 matrix, there are
not of the same size, A + C does not exist.
 Scalar Multiplication of matrices
Definition
Let A be a matrix and c be a scalar. The scalar multiple of A by c,
denoted cA, is the matrix obtained by multiplying every element
of A by c. The matrix cA will be the same size as A.
Thus if B  cA, then bij  caij i, j.

Example

Let A   1  2 4.
7  2 0
 3  1 3  (2) 3  4  3  6 12
3A   .
3  7 3  ( 3) 3  0 21  9 0
Observe that A and 3A are both 2  3 matrices.
 Subtraction
Definition
We now define subtraction of matrices in such a way that makes it
compatible with addition, scalar multiplication, and negation. Let
A – B = A + (–1)B
Example

Suppose A  5 0  2 and B  2 8  1.

3 6  5 0 4 6
 5  2 0  8  2  ( 1) 3  8  1
A B   .
3  0 6  4  5  6 3 2  11
 Multiplication of Matrices
Definition
Let the number of columns in a matrix A be the same as the
number of rows in a matrix B. The product AB then exists.
Let A be mn matrix, B be nk matrix,
The product matrix C=AB
 b1 j 
b 
cij  ai1  ain   2j
ai 2  ai1b1 j  ai 2b2 j    ainbnj
  
 
 bnj 
C is a mk matrix.
If the number of columns in A does not equal the number of row B,
we say that the product does not exist.
 Example
 1 3 5 0 1
Let A    ,B  , and C  6  2 5.
 2 0 3  2 6 
Determine AB, BA, and AC , if the products exist.
Sol.
AB   1 3 5 0 1
2 0 3  2 6
 5   0 1  
 1 3   1 3  1 3   
3  2  6  
 
 5   0 1  
 2 0  2 0
2 
2 0 
 3  6  
 (1 5)  (3  3) (1 0)  (3  (2)) (11)  (3  6)
 
(2  5)  (0  3) (2  0)  (0  (2)) (2  1)  (0  6)
14  6 19
 .
10 0 2 
BA and AC do not exist. Note. In general, ABBA.
 Example
 2 1
Let A   7 0 and B   1 0. Determine AB.
 3 5
 3  2
 2 1 1 2 10 
  3 5 
 2 1
 1 0   1 0 
AB   7 0     7 0   
7 05  
 3 5   3
 3  2  

   

 3  2   1  3  2  0 
  3 5 
  2  3 0  5  1 5
   7  0 0  0    7 0
 3  6 0  10   3  10
Example
Let C = AB, A   2 1 and B   7 3 2 Determine c23.
 3 4  5 0 1
  2  (3  2)  (4  1)  2
c23  3 4
 1
 Size of a Product Matrix
If A is an m  r matrix and B is an r  n matrix, then AB will be an
m  n matrix.

A B = AB
mr rn mn

Example
If A is a 5  6 matrix and B is an 6  7 matrix.
Because A has six columns and B has six rows. Thus AB exits.
And AB will be a 5  7 matrix.
 Special Matrices
Definition
A zero matrix is a matrix in which all the elements are zeros.
A diagonal matrix is a square matrix in which all the elements
not on the main diagonal are zeros.
An identity matrix is a diagonal matrix in which every diagonal
element is 1.
a11 0  0 1 0  0
0 0  0 0 a22  0  I n  0 1  0
0mn  0 0  0 A    
        
0 0  0   0 0  1
0 0  ann 

zero matrix identity matrix

diaginal matrix A
 Theorem
Let A be m  n matrix and Omn be the zero m  n matrix. Let B be
an n  n square matrix. On and In be the zero and identity n  n
matrices. Then
A + Omn = Omn + A = A
BOn = OnB = On
BIn = InB = B
Example
Let A  2 1  3 and B   2 1.
4 5 8  3 4
2 1  3 0 0 0 2 1  3
A  O23     A
4 5 8 0 0 0 4 5 8
 2 1 0 0 0 0
BO2         O2
 3 3 0 0 0 0
 2 1 1 0  2 1
BI 2     B
 3 4 0 1  3
  4
Example
Let A be a matrix whose third row is all zeros. Let B be any
matrix such that the product AB exists.
Prove that the third row of AB is all zeros.
Solution

 b1i   b1i 
b  b 
( AB )3i  [a31 a32  a3n ]  2i   [0 0  0]  2i   0, i.
 
   
b
 ni  bni 
 Algebraic Properties of Matrix
Operations
Theorem
Let A, B, and C be matrices and a, b, and c be scalars. Assume that the
size of the matrices are such that the operations can be performed.
Properties of Matrix Addition and scalar Multiplication
1. A + B = B + A Commutative property of addition
2. A + (B + C) = (A + B) + C Associative property of addition
3. A + O = O + A = A (where O is the appropriate zero matrix)
4. c(A + B) = cA + cB Distributive property of addition
5. (a + b)C = aC + bC Distributive property of addition
6. (ab)C = a(bC)
 Theorem
Let A, B, and C be matrices and a, b, and c be scalars. Assume that the
size of the matrices are such that the operations can be performed.
Properties of Matrix Multiplication
1. A(BC) = (AB)C Associative property of multiplication
2. A(B + C) = AB + AC Distributive property of multiplication
3. (A + B)C = AC + BC Distributive property of multiplication
4. AIn = InA = A (where In is the appropriate zero matrix)
5. c(AB) = (cA)B = A(cB)
Note: AB BA in general. Multiplication of matrices is not
commutative.
Proof of Thm. (A+B=B+A)
Consider the (i,j)-th elements of matrices A+B and B+A:
( A  B ) ij  aij  bij  bij  aij  ( B  A) ij .
A+B=B+A

Example
Let A   1 3, B  3  7 , and C  0  2.
 4 5 8 1 5  1
A  B  C   1 3  3  7   0  2
 4 5 8 1 5  1

  1  3  0 3  7  2  4  6.
 4  8  5 5  1  1 9 5
 Example
 4
Let A   1 2, B   0 1 3, and C   1 .
  Compute ABC.
3  1  1 0  2
 0
Sol. Which method is better?
(1) (AB)C
AB   1 2  0 1 3   2 1  1.
3  1  1 0  2  1 3 11
 2 1  1  4  9
( AB )C     1   .
 1 3 11 0  1
 
(2) A(BC)

 4
BC   0 1 3   1    1 
 1 0 2 0  4
 
A( BC )   1 2   1   9.  A(BC) is better.
3  1  4  1
 Remark
In algebra we know that the following cancellation laws apply.
If ab = ac and a  0 then b = c.
If pq = 0 then p = 0 or q = 0.
However the corresponding results are not true for matrices.
AB = AC does not imply that B = C.
PQ = O does not imply that P = O or Q = O.

Example
 1 2   1 2  3 8 
(1) Consider the matrices A   , B   2 1, and C   3  2.
 2 4     
3 4
Observe that AB  AC    , but B  C.
6 8 
 1  2  2  6
(2) Consider the matrices P    , and Q   .
 2 4 1  3
Observe that PQ  O, but P  O and Q  O.
 Powers of Matrices
Definition
If A is a square matrix, then
Ak  
AA
 
A
k times

Theorem
If A is an n  n square matrix, r and s are nonnegative integers,
then
1. ArAs = Ar+s.
2. (Ar)s = Ars.
3. A0 = In (by definition)
 Example
 1  2
If A    , compute A 4
.
 1 0
Solution  1  2  1  2  3  2
A 
2
    
  1 0   1 0    1 2 
 3  2  3  2  11  10
A 
4
    .
 1 2   1 2   5 6

Example Simplify the following matrix expression.

A( A  2 B )  3B(2 A  B )  A2  7 B 2  5 AB
Solution
A( A  2 B )  3B(2 A  B )  A2  7 B 2  5 AB
 A2  2 AB  6 BA  3B 2  A2  7 B 2  5 AB
 3 AB  6 BA  4 B 2
 Systems of Linear Equations

A system of m linear equations in n variables as follows

a11 x1    a1n xn  b1
   
am1 x1    amn xn  bm
Let
 a11  a1n   x1   b1 
A      , X    , and B    
am1  amn   xn  bm 
     
We can write the system of equations in the matrix form
AX = B
 Idempotent and Nilpotent Matrices

Definition
(1) A square matrix A is said to be idempotent if A2=A.
(2) A square matrix A is said to nilpotent if there is a
p
positive integer p such that A =0. The least integer p such that
p
A =0 is called the degree of nilpotency of the matrix.

Example
3  6 2 3  6
(1) A    ,A    A.
 1  2  1  2

3  9 2 0 0
(2) B    ,B  . The degree of nilpotency : 2
 1  3 0 0 
 Transpose of a Matrix

Definition
The transpose of a matrix A, denoted At, is the matrix whose
columns are the rows of the given matrix A.
i.e., A : m  n  At : n  m, ( At )ij  A ji i, j.

Example
A   2 7 , B   1 2  7, and C   1 3 4.
 8 0 4 5 6

 1 4   1
At  2  8 C   3.
t
B t
  2 5 
7 0  4
 7 6
 Theorem (Properties of Transpose)

Let A and B be matrices and c be a scalar. Assume that the sizes

of the matrices are such that the operations can be performed.
1. (A + B)t = At + Bt Transpose of a sum
2. (cA)t = cAt Transpose of a scalar multiple
3. (AB)t = BtAt Transpose of a product
4. (At)t = A
Proof for (AB)t = BtAt
 b1i 
b 

( AB)t ij  ( AB) ji  a j1 a j 2  a jn   2i 

 
bni 
 a j1b1i  a j 2b2i    a jn bni

( B t At )ij  [row i of B t ] [column j of At ]  [column i of B]t [row j of A]t

 a j1 
a 
 b1i b2i  bni     a j1b1i  a j 2b2i    a jnbni
j2

  
 
 a jn 
 Symmetric Matrix
Definition
A symmetric matrix is a matrix that is equal to its transpose.

A  At , i.e., aij  a ji i, j

Example match
1 0 2 4
 0 1  4  0 9
2 5
 1 7 8 7 3
 5  4 2 2  3
 4 8  3  4
3
 9 3 6 match
 Example
Let A and B are symmetric matrices of the same size. Prove that
the product AB is symmetric if and only if AB = BA.
Proof
*We have to show (a) AB is symmetric  AB = BA,
and the converse, (b) AB is symmetric  AB = BA.
() Let AB be symmetric, then
AB= (AB)t by definition of symmetric matrix
= BtAt by Thm (3)
= BA since A and B are symmetric
() Let AB = BA, then
(AB)t = (BA)t
= AtBt by Thm (3)
= AB since A and B are symmetric
 Example
Let A be a symmetric matrix. Prove that A2 is symmetric.

Proof

( A )  ( AA)  ( A A )  AA  A
2 t t t t 2
Definition
Let A be a square matrix. The trace of A, denoted tr(A) is the sum
of the diagonal elements of A. Thus if A is an n  n matrix.
tr(A) = a11 + a22 + … + ann

Example
4 1  2
Determine the trace of the matrix A  2  5 6.
7 3 0
Solution
We get
tr ( A)  4  (5)  0  1.
 Theorem (Properties of Trace)
Let A and B be matrices and c be a scalar. Assume that the sizes
of the matrices are such that the operations can be performed.
1. tr(A + B) = tr(A) + tr(B)
2. tr(AB) = tr(BA)
3. tr(cA) = c tr (A)
4. tr(At) = tr(A)
Proof of (1)
Since the diagonal element of A + B are (a11+b11), (a22+b22), …,
(ann+bnn), we get
tr(A + B) = (a11 + b11) + (a22 + b22) + …+ (ann + bnn)
= (a11 + a22 + … + ann) + (b11 + b22 + … + bnn)
= tr(A) + tr(B).
 Example of (2) tr(AB)=tr(BA)
1 3
   3 2  1
A 2 0 , B   
  1 0 1 
 1  2

0 2 2
   8 11 
AB   6 4  2, BA   
  2  5
 1  2  1

tr ( AB)  3  tr ( BA)
 The Inverse of a Matrix
Definition
Let A be an n  n matrix. If a matrix B can be found such that
AB = BA = In, then A is said to be invertible and B is called the
inverse of A. If such a matrix B does not exist, then A has no
inverse. (denote B = A1, and Ak=(A1)k )
Example  2 1 
Prove that the matrix A   1 2 has inverse B   3  1 .
3 4  2 2 
Proof
 2 1
AB   1 2  3  1    1 0  I 2
3 4  2 2 0 1
 2 1  1 2  1 0
BA   3  1    I2
 2 23 4 
   0 1

Thus AB = BA = I2, proving that the matrix A has inverse B.
 Theorem
The inverse of an invertible matrix is unique.

Proof
Let B and C be inverses of A.
Thus AB = BA = In, and AC = CA = In.
Multiply both sides of the equation AB = In by C.
C(AB) = CIn
(CA)B = C
In B = C
B=C
Thus an invertible matrix has only one inverse.
 Gauss-Jordan Elimination for finding
the Inverse of a Matrix
Let A be an n  n matrix.
1. Adjoin the identity n  n matrix In to A to form the matrix
[A : In].
2. Compute the reduced echelon form of [A : In].
If the reduced echelon form is of the type [In : B], then B is
the inverse of A.
If the reduced echelon form is not of the type [In : B], in that
the first n  n submatrix is not In, then A has no inverse.

An n  n matrix A is invertible if and only if its reduced echelon

form is In.
 Example
 1  1  2
Determine the inverse of the matrix A   2  3  5
 1 3 5
Solution
 1  1  2 1 0 0  1 1  2 1 0 0
[ A : I 3 ]   2  3  5 0 1 0 R2  (2) R10  1  1  2 1 0
 1 3 5 0 0 1 R3  R1 0 2 3 1 0 1
  1  1  2 1 0 0  1 0 1 3  1 0
( 1)R2 0 1 1 2  1 0 R1  R2 0 1 1 2  1 0
0 2 3 1 0 1 R3  (2) R2 0 0 1  3 2 1
 1 0 0 0 1 0
R1  R3 0 1 0 5  3  1
R2  (1)R3 0 0 1  3 2 1

1
 0 1 1
Thus, A   5  3  1.
 3 2 1
 Example
Determine the inverse of the following matrix, if it exist.
 1 1 5
A   1 2 7
2  1 4
Solution
 1 1 5 1 0 0  1 1 5 1 0 0
[ A : I 3 ]   1 2 7 0 1 0 R2  (1)R1 0 1 2  1 1 0
2  1 4 0 0 1 R3  (2)R1 0  3  6  2 0 1
 1 0 3 2  1 0
R1  ( 1)R2 0 1 2  1 1 0
R3  3R2 0 0 0  5 3 1
There is no need to proceed further.
The reduced echelon form cannot have a one in the (3, 3) location.
The reduced echelon form cannot be of the form [In : B].
Thus A–1 does not exist.
 Properties of Matrix Inverse
Let A and B be invertible matrices and c a nonzero scalar, Then
1. ( A1 ) 1  A 4. ( An ) 1  ( A1 ) n
1 1 1
2. (cA)  A 5. ( At ) 1  ( A1 ) t
1
c 1 1
3. ( AB)  B A
Proof
1. By definition, AA1=A1A=I.
2. (cA)( 1c A1 )  I  ( 1c A1 )(cA)
3. ( AB)( B 1 A1 )  A( BB 1 ) A1  AA1  I  ( B 1 A1 )( AB)
4. A n ( A 1 ) n  
A
 A 
A 1 1
A
  I  ( A 1 ) n A n
n times n times
5. AA1  I , ( AA1 ) t  ( A1 ) t At  I ,
A1 A  I , ( A1 A) t  At ( A1 ) t  I ,
Example
If A  4 1, then it can be shown that A1   1  1. Use this
 3 1  3 4
information to compute ( At ) 1.
Solution
t
t 1  1 1
1 t
(A )  (A )     1 3
.
3 4 1 4
 Theorem
Let AX = Y be a system of n linear equations in n variables.
If A–1 exists, the solution is unique and is given by X = A–1Y.

Proof
(X = A–1Y is a solution.)
Substitute X = A–1Y into the matrix equation.
AX = A(A–1Y) = (AA–1)Y = InY = Y.

(The solution is unique.)

Let X1 be any solution, thus AX1 = Y. Multiplying both sides of
this equation by A–1 gives
A–1AX1= A–1Y
InX1 = A–1Y
X1 = A–1Y.
 Example
x1  x2  2 x3  1
Solve the system of equations 2 x1  3x2  5 x3  3
 x1  3x2  5 x3  2
Solution
This system can be written in the following matrix form:
 1  1  2  x1   1
 2  3  5  x2    3
 1 3 5  x3   2
If the matrix of coefficients is invertible, the unique solution is
 x1   1  1  2 1  1
 x2    2  3  5  3
 x   1 3 5  2
 3
This inverse has already been found. We get
 x1   0 1 1  1  1
 x2    5  3  1  3   2
 x   3 2 1  2  1
 3
The unique solution is x1  1, x2  2, x3  1.
 Elementary Matrices
Definition
An elementary matrix is one that can be obtained from the
identity matrix In through a single elementary row operation.

Example 1 0 0 
R2  R3 E1  0 0 1
0 1 0
1 0 0 
1 0 0
I 3  0 1 0 5R2
E2  0 5 0
0 0 1
0 0 1

1 0 0 
R2+ 2R1
E3  2 1 0
0 0 1
 Elementary Matrices
elementary row operation, elementary matrix。

a b c  1 0 0
R2  R3  g h i   0 0 1  A  E1 A

d e f  0 1 0
a b c
a b c  1 0 0
A  d e f 
5R2
5d 5e 5 f   0 5 0  A  E2 A

 g h i   g h i  0 0 1

 a b c  1 0 0 
d  2a e  2b f  2c   2 1 0  A  E3 A
R2+ 2R1 
 g h i  0 0 1
 Notes for elementary matrices
Each elementary matrix is invertible.
Example
I  E1  E1  I , i.e., E2 E1  I
R1 2 R 2 R1 2 R 2

1 0 0  1 2 0  1  2 0
I  0 1 0 E1  0 1 0 E2  0 1 0
0 0 1 0 0 1 0 0 1

If A and B are row equivalent matrices and A is

invertible, then B is invertible.
Proof
If A  …  B, then
B=En … E2 E1 A for some elementary matrices En, … , E2 and E1.
So B1 = (En … E2 E1A)1 =A1E11 E21 … En1.