Anda di halaman 1dari 17

3.

2 Matrix Algebra
Properties of Addition, Scalar Multiplication

Theorem 1 Let A, B and C denote arbitrary


m × n matrices; let c, d denote scalars. Then

• A + B = B + A Commutativity

• A + (B + C) = (A + B) + C Associativity

• There is an m × n matrix O such that O +


A = A for each A.

• For each A there is an m × n matrix −A


such that A + (−A) = O.

• c(A + B) = cA + cB Distributivity

1
• (c + d)A = cA + dA Distributivity

• (cd)A = c(dA)

• 1A = A

Notation:

Recall A = [aij ]m×n, (A)ij = aij .

The ithe row of A: denoted by Ai or rowi(A);

The jth column of A: denoted by aj or colj (A)

2
A+B =B+A:

Proof:

A + B = [aij + bij ] = [bij + aij ] = B + A

OR

(A + B)ij = (A)ij + (B)ij = (B)ij + (A)ij =


(B + A)ij

A + (−A) = O:

Proof:

Define −A = [−aij ].
Then A + (−A) = [aij + (−aij )] = [0ij ] = 0.

3
If A1, A2, . . . , Ak are matrices of the same size
and c1, c2, . . . , ck are scalars, we may form the
linear combination

c1A1 + c2A2 + · · · + ck Ak
The scalars c1, c2, . . . , ck are the coefficients
of the linear combination.

Example(Example
" #
3.16) " #
1 0 0 −1
Let A1 = , A2 = and A3 =
0 1 1 0
" #
1 1
. Write B, C as linear combinations of
0 1
A1, A2 and A3, if possible, where
" # " #
2 3 2 5
B= C=
−4 2 0 3

4
The span of a set of matrices is the set of all
linear combinations of the matrices.

Example(Example
" #
3.17) " #
1 0 0 −1
Let A1 = , A2 = and A3 =
0 1 1 0
" #
1 1
. Describe the span of the matrices A1, A2
0 1
and A3.

5
Matrices A1, A2, . . . , Ak of the same size are lin-
early independent if the only solution of the
equation

c1 A 1 + c2 A 2 + · · · + ck A k = O
is the trivial solution: c1 = c2 = · · · = ck = 0.

If there are nontrivial coefficients that satisfy


the equation then A1, A2, . . . , Ak are called lin-
early dependent.

Example " # " #


1 0 0 −1
Let A1 = , A2 = and A3 =
0 1 1 0
" #
1 1
. Determine whether or not A1, A2 and
0 1
A3 are linearly independent.

6
Properties of Matrix Multiplication

Theorem 2 Let A, B and C be matrices (whose


sizes are such that the indicated operations can
be performed) and let k be a scalar. Then

1. A(BC) = (AB)C Associativity

2. A(B + C) = AB + AC Left distributivity

3. (A + B)C = AC + BC Right distributivity

4. k(AB) = (kA)B = A(kB)

5. ImA = A = AIn if A is m × n

7
(A + B)C = AC + BC :
Proof

((A + B)C)ij

= rowi(A + B) • colj (C)


= [rowi(A) + rowi(B)] • colj (C)
= [rowi(A) • colj (C)] + [rowi(B) • colj (C)]
= (AC)ij + (BC)ij
Thus the result follows.

AIn = A :
Proof

AIn = A e1 ... e2 ... · · · ... en


h i

Ae1 ... Ae2 ... · · · ... Aen


h i
=

a1 ... a2 ... · · · ... an


h i
= =A

8
Example(Example
" #
3.20) " #
1 2 4 3
Let A = and B = . Confirm prop-
3 4 2 1
erty 4 of matrix multiplication listed above for
these two matrices if k = 2.

9
Some basic rules for matrix multiplication:

1. Matrix multiplication is not commutative,


i.e., in general AB 6= BA.
" #" # " # " # " #" #
1 2 1 0 1 0 1 2 1 0 1 2
= 6= =
3 4 0 0 3 0 0 0 0 0 3 4

2. If AB = O it does not imply that A = O


or B = O (that is, neither needs to equal
the zero matrix).
" #" # " #
1 −1 1 1 0 0
=
−1 1 1 1 0 0

3. AB = AC does not imply that B = C.


" #" # " #" #
1 2 1 0 1 2 0 0
= 1 0
2 4 0 0 2 4 2

10
Properties of the Transpose

Theorem 3 Let A and B be matrices (whose


sizes are such that the indicated operations can
be performed) and let k be a scalar. Then

1. (AT )T = A

2. (A + B)T = AT + B T

3. (kA)T = k(AT )

4. (AB)T = B T AT

5. (Ar )T = (AT )r for all nonnegative integers


r.

11
(AT )T = A :

Proof:

((AT )T )ij = (AT )ji = (A)ij

Let A be n × n. (Ar )T = (AT )r for r = 0, 1, . . .:

Proof (Mathematical Induction on r):

Prove r = 0 is true:
(A0)T = (In)T = In = (InT )0

Assume r = k is true
i.e. (Ak )T = (AT )k

Prove that r = k + 1 is true:

(Ak+1)T = (Ak A)T = AT (Ak )T Rule 4


= AT (AT )k Inductive Hypothesis
= (AT )k+1

12
Example " # " #
1 3 2 2 0 1
Let A = and B = Confirm
0 −1 1 1 4 3
the properties of the transpose listed above for
these two matrices.

13
Theorem 4 1. If A is a square matrix, then
A + AT is a symmetric matrix.

2. For any matrix A, AAT and AT A are sym-


metric matrices.

Example(Example
" #
3.21)
1 2
Let A = . Show that A + AT , AAT and
3 4
AT A are symmetric.

14
A square matrix is called upper triangular if all
the entries below the main diagonal are zero.
 
∗ ∗ ···
∗ ∗
0 ∗ · · ·∗ ∗
 
0 0 . . .
 ... ... 

 ... ...
 
 ∗ ∗ 
0 0 ··· 0 ∗
where the entries marked ∗ are arbitrary.
   
1 7 2 1 0 2
0 2 3  , 0 0 3
   
0 0 −1 0 0 2
Similarly the following are lower triangular
   
1 0 0 5 0 0
5 2 0 , 3 0 0 
   
3 4 −1 2 3 −1

15
A square matrix is called skew-symmetric if
AT = −A.

Exercise
Is the following matrix skew-symmetric?
 
0 1 2
−1 0 5
 
2 5 0

Exercise
Show that if A is square then A − AT is skew-
symmetric.

16
The trace of an n × n matrix A = [aij ] is the
sum of the entries on its main diagonal and is
denoted by tr(A). That is,

tr(A) = a11 + a22 + · · · + ann

Exercise
If A and B are n × n matrices, prove the fol-
lowing properties of the trace:

1. tr(A + B) = tr(A) + tr(B)

2. tr(kA) = ktr(A) where k is a scalar.

17

Anda mungkin juga menyukai