Vous êtes sur la page 1sur 4

BRIEF NOTES ON LINEAR ALGEBRA

DAVID M. MCCLENDON
Abstract. A review of some basic ideas of linear algebra necessary for Math
310. Warning: the examples in these notes have not been checked for arith-
metic errors.
1. Matrix vocabulary
A matrix is an array of numbers arranged in a rectangle. A matrix is called
m n if it has m rows and n columns. Throughout these notes we consider only
square matrices, that is, matrices that have the same number of rows as columns;
the set of nn matrices with real entries is called M
n
(R). Matrices will be referred
to with capital letters and their entries will be referred to with lowercase letters
(i.e. a
ij
represents the entry in row i and column j of the matrix A). Entries of
the form a
ii
are the diagonal entries of A. The trace of a matrix is the sum of its
diagonal entries. In these notes, the rows and columns of a matrix are numbered
starting with the number 1, so the left most column is the rst column.
Example: Let A =
_
_
1 2 3
4 5 6
7 8 9
_
_
. Then A M
3
(R) since A has three rows
and columns; we have a
13
= 3, a
32
= 8, etc.; the diagonal entries of A are 1, 5 and
9, so the trace of A is 1 + 5 + 9 = 15. The second row of A is 4, 5, 6. The entries
above the diagonal of A are 2, 3, 6; the entries below the diagonal are 4, 7, 8.
The letter I always refers to the identity matrix which has diagonal entries equal
to 1 and all other entries 0. The identity matrix is an example of a diagonal matrix,
which is any matrix which has only non-zero entries other than its diagonal entries.
A matrix is called triangular if either all the entries above the diagonal or all the
entries below the diagonal are zero.
The transpose of a matrix A is denoted A
T
and satises A
T
ij
= A
ji
. So if A is as
above, A
T
=
_
_
1 4 7
2 5 8
3 6 9
_
_
.
2. Determinants
Every square matrix has a magic number associated to it called its determi-
nant. The determinant of a diagonal or triangular matrix is the product of its
diagonal entries. In general, determinant formulas can be found by induction:
Date: April 23, 2008.
1
2 D. MCCLENDON
Proposition 2.1. If A is 2 2, i.e. A =
_
a b
c d
_
, then det(A) = ad bc.
Proposition 2.2. If A is 3 3, i.e. if A =
_
_
a b c
d e f
g h i
_
_
, then
det(A) = aei + bfg + cdh (ceg + afh + bdi).
For any size matrix, the determinant can be found using minors.
Denition 2.1. Given an nn square matrix A and an entry a
ij
of A, the minor
A
ij
is the (n 1) (n 1) matrix obtained by deleting the i
th
row and j
th
column
of A.
The method of using minors to nd determinants relies on the following facts:
Proposition 2.3. For any j, det(A) =

n
i=1
(1)
i+j
a
ij
det(A
ij
)
Proposition 2.4. For any i, det(A) =

n
j=1
(1)
i+j
a
ij
det(A
ij
).
and is best illustrated in an example:
Example: Find the determinant of A =
_
_
_
_
1 2 1 0
2 0 1 1
1 1 0 1
0 1 0 2
_
_
_
_
.
Solution: First, pick a row or column of A (preferably one that has several zeros
in it). I will pick the third column. Now what I will do is move down that column
entry by entry and for each entry, nd the minor of that entry. In particular, the
minor of entries a
13
= 1 and a
23
= 1 are
A
13
=
_
_
2 0 1
1 1 1
0 1 2
_
_
and A
23
=
_
_
1 2 0
1 1 1
0 1 2
_
_
.
and the determinants of these matrices are 5 and 1, respectively. Now use
Proposition 2.3 with j = 3. We have
det(A) =
4

i=1
(1)
i+3
a
i3
det(A
i3
)
= (1) det(A
13
) (1) det(A
23
) + 0 det(A
33
) 0 det(A
43
)
= (1)(5) (1)(1) + 0 = 6.
3. Eigenvalues and eigenvectors
Denition 3.1. Given A M
n
(R), a nonzero vector x R
n
is called an eigenvec-
tor of A if there is some number R such that Ax = x. Given an eigenvalue
of A, the set of vectors x which satisfy Ax = x of A is called the eigenspace of
A. The multiplicity of an eigenvalue is the dimension of its eigenspace.
Eigenvalues are found using the following propositions:
Proposition 3.1. is an eigenvalue of A if and only if det(I A) = 0.
Proposition 3.2. The sum of the eigenvalues (counting multiplicities) of a matrix
is the trace of that matrix, and the product of the eigenvalues (counting multiplici-
ties) is the determinant of the matrix.
BRIEF NOTES ON LINEAR ALGEBRA 3
Example: Find the eigenvalues and eigenvectors of A =
_
_
0 1 1
1 1 1
1 2 0
_
_
.
Solution: First, solve det(I A) = 0:
det(I A) = det
_
_
1 1
1 1 1
1 2
_
_
=
2
( 1) 1 + 2 + ( 1) 2
=
3

2
2;
setting this equal to zero and solving for we see = 0, = 2 and = 1 are
the eigenvalues. This makes sense because these numbers sum to 1 which is the
trace of A. To nd an eigenvector for each eigenvalue solve Ax = x for a nonzero
x = (a, b, c). For example, for = 2 we have
Ax = x
_
_
_
b c = 2a
a + b + c = 2b
a + 2b = 2c
which has as one of its nonzero solutions x = (a, b, c) = (0, 1, 1). Similarly, an
eigenvector for = 0 is x = (2, 1, 1) and an eigenvector for = 1 is x =
(3, 2, 1).
4. Inverses
A square matrix A is called invertible if there is another matrix A
1
(called the
inverse of A) such that AA
1
= A
1
A = I. It can be shown that an n n square
matrix is invertible if and only if det(A) = 0 if and only if 0 is not an eigenvalue of
A.
Proposition 4.1. The inverse of a diagonal matrix A, if it exists, is diagonal and
has entries which are the reciprocals of the entries of A.
Proposition 4.2. The inverse of a 2 2 matrix A =
_
a b
c d
_
, if it exists, is
A
1
=
1
adbc
_
d b
c a
_
.
To nd the inverse of a larger matrix, use one of two methods. The rst method
is to take the matrix A and form an n2n matrix by writing a copy of the identity
next to it. Then perform row operations until the identity appears in the position
originally occupied by A; the matrix in the position of the identity is A
1
.
Example: Find the inverse of A =
_
_
1 2 1
2 2 4
1 3 3
_
_
.
Solution: First write the matrix (A|I) =
_
_
1 2 1 1 0 0
2 2 4 0 1 0
1 3 3 0 0 1
_
_
. Next
perform row operations on this matrix to turn the left half of the matrix into the
identity (these row operations are the same operations you would do if you were
performing Gaussian elimination). The sequence of operations is: subtract 2 times
Row 2 from Row 1, then subtract Row 1 from Row 3, then multiply Row 2 by
4 D. MCCLENDON
(1/2), then subtract Row 2 from Row 3, then add Row 3 to Row 1, then add 3
times Row 3 to Row 2, then subtract 2 times Row 2 from Row 1. You obtain
(I|A
1
) =
_
_
1 0 0 9 3/2 5
0 1 0 5 1 3
0 0 1 2 1/2 1
_
_
so A
1
=
_
_
9 3/2 5
5 1 3
2 1/2 1
_
_
.
The second way to nd the inverse of a matrix is to rst nd the cofactor matrix:
Denition 4.1. The cofactor matrix of a square matrix A M
n
(R) is the n n
matrix cof(A) whose entries are dened by (cof(A))
ij
= (1)
i+j
det(A
ij
) (recall
that A
ij
is a minor of A as dened in the section on determinants).
Proposition 4.3. If A M
n
(R) is invertible, then A
1
=
1
det(A)
(cof(A))
T
.
Example: Find the inverse of A =
_
_
5 3 2
1 0 2
2 1 3
_
_
.
Solution: We use the preceding proposition. First, we nd the cofactor matrix
of A. The 1, 1 entry of this matrix is found by taking the determinant of the
1, 1 minor (which is 2) and multiplying by (1)
1+1
to obtain a nal answer of 2.
Repeating this procedure for all the entries, we see
cof(A) =
_
_
2 1 1
7 11 1
6 8 3
_
_
.
Observe also that det(A) = 12 2 + 10 + 9 = 5 so
A
1
=
1
det(A)
(cof(A))
T
=
1
5
_
_
2 7 6
1 11 8
1 1 3
_
_
.
5. Diagonalization
Denition 5.1. A square matrix A M
n
(R) is similar to a second matrix B
M
n
(R) if there exists an invertible matrix S M
n
(R) such that A = SBS
1
. A
square matrix A M
n
(R) is called diagonalizable if it is similar to a diagonal
matrix.
Proposition 5.1. A M
n
(R) is diagonalizable if and only if it has n linearly
independent eigenvectors; in this case A = SS
1
where is a diagonal matrix
whose diagonal entries are the eigenvalues of A and S is a matrix whose columns
are the eigenvectors of A (arranged in the same order as their respective eigenvalues
are arranged in ).
So to diagonalize a matrix, you rst nd the eigenvalues and eigenvectors; this
tells you the matrices and S. Then nd S
1
using methods of the previous
section.
Department of Mathematics, Northwestern University, Evanston, IL 60208-2370
E-mail address: dmm@math.northwestern.edu

Vous aimerez peut-être aussi