Exercises and Problems in Linear Algebra

John M. Erdman
Portland State University

Version September 18, 2011

2010
c John M. Erdman

E-mail address: [email protected]

 Contents

PREFACE v

Chapter 1. SYSTEMS OF LINEAR EQUATIONS 1

1.1. Background 1
1.2. Exercises 2
1.3. Problems 5
1.4. Answers to Odd-Numbered Exercises 6

Chapter 2. ARITHMETIC OF MATRICES 7

2.1. Background 7
2.2. Exercises 8
2.3. Problems 10
2.4. Answers to Odd-Numbered Exercises 12

Chapter 3. ELEMENTARY MATRICES; DETERMINANTS 13

3.1. Background 13
3.2. Exercises 15
3.3. Problems 19
3.4. Answers to Odd-Numbered Exercises 20

Chapter 4. VECTOR GEOMETRY IN Rn 21

4.1. Background 21
4.2. Exercises 22
4.3. Problems 24
4.4. Answers to Odd-Numbered Exercises 25

Chapter 5. VECTOR SPACES 27

5.1. Background 27
5.2. Exercises 28
5.3. Problems 31
5.4. Answers to Odd-Numbered Exercises 32

Chapter 6. SUBSPACES 33
6.1. Background 33
6.2. Exercises 34
6.3. Problems 38
6.4. Answers to Odd-Numbered Exercises 39

Chapter 7. LINEAR INDEPENDENCE 41

7.1. Background 41
7.2. Exercises 43
7.3. Problems 45
7.4. Answers to Odd-Numbered Exercises 47

Chapter 8. BASIS OF A VECTOR SPACE 49

iii
iv CONTENTS

8.1. Background 49
8.2. Exercises 50
8.3. Problems 52
8.4. Answers to Odd-Numbered Exercises 53
Bibliography 55
Index 57
 PREFACE
preface
This collection of exercises is designed to provide a framework for discussion in a junior level
linear algebra class such as the one I have conducted fairly regularly at Portland State University.
There is no assigned text. Students are free to choose their own sources of information. Students
are encouraged to find sources whose writing style they find congenial, whose emphasis matches
their interests, and whose price fits their budgets. The short introductory background section in
these exercises which precede each assignment are intended to fix notation and provide “official”
definitions and statements of important theorems for the exercises and problems which follow.
There are a number of excellent online texts which are available free of charge. Among the best
Hefferon:2006
are Linear Algebra [8] by Jim Hefferon,
http://joshua.smcvt.edu/linearalgebra
Beezer:2004
A First Course in Linear Algebra [2] by Robert A. Beezer,
http://linear.ups.edu/download/fcla-electric-2.00.pdf
Dawkins:2007
and Linear Algebra [6] by Paul Dawkins.
http://tutorial.math.lamar.edu/pdf/LinAlg/LinAlg_Complete.pdf
Bogacki:2005
Another very useful online resource is Przemyslaw Bogacki’s Linear Algebra Toolkit [3].
http://www.math.odu.edu/~bogacki/lat
And, of course, many topics in linear algebra are discussed with varying degrees of thoroughness
wiki:xxx
in the Wikipedia [13]
http://en.wikipedia.org
Mathworld:xxx
and Eric Weisstein’s Mathworld [12].
http://http://mathworld.wolfram.com
Among the dozens and dozens of linear algebra books that have appeared, two that were written
before “dumbing down” of textbooks became fashionable are especially notable, in my opinion,
for the Halmos:1958
clarity of their authors’ mathematical vision: Paul Halmos’s Finite-Dimensional Vector
HoffmanK:1971
Spaces [7] and Hoffman and Kunze’s Linear Algebra [9]. Some students, especially mathematically
inclined ones, love these books, but others find them hard to read. Give them a look if youCurtis:1984
have
the chance. Another excellent traditional text is Linear Algebra: An Introductory Approach [5] by
Charles W. Curits. And for those of you more interested in applications both Elementary Linear
AntonR:2000
Algebra: Applications Version
Strang:1980
[1] by Howard Anton and Chris Rorres and Linear Algebra and its
Applications [11] by Gilbert Strang are loaded with applications.
If you find the level at which many of the current linear algebra texts are written depressingly
pedestrian and the endless computations irritating, you might examine some of the more advanced
Roman:2005
texts. Two excellent ones are Steven Roman’s Advanced
Brown:1988
Linear Algebra [10] and William C.
Brown’s A Second Course in Linear Algebra [4].

v
 CHAPTER 1

SYSTEMS OF LINEAR EQUATIONS

ems_lin_eqns
1.1. Background
Topics: systems of linear equations; Gaussian elimination (Gauss’ method), elementary row op-
erations, leading variables, free variables, echelon form, matrix, augmented matrix, Gauss-Jordan
reduction, reduced echelon form.

ge009def 1.1.1. Definition. We will say that an operation (sometimes called scaling) which multiplies a
row of a matrix (or an equation) by a nonzero constant is a row operation of type I. An
operation (sometimes called swapping) that interchanges two rows of a matrix (or two equations) is
a row operation of type II. And an operation (sometimes called pivoting) that adds a multiple
of one row of a matrix to another row (or adds a multiple of one equation to another) is a row
operation of type III.

1
2 1. SYSTEMS OF LINEAR EQUATIONS

1.2. Exercises
(1) Suppose that L1 and L2 are lines in the plane, that the x-intercepts of L1 and L2 are 5
and −1, respectively, and that the respective y-intercepts are 5 and 1. Then L1 and L2
intersect at the point ( , ).

(2) Consider the following system of equations.


w + x + y + z = 6

w +y+z =4 (∗)

w +y =2

(a) List the leading variables .

(b) List the free variables .
(c) The general solution of (∗) (expressed in terms of the free variables) is
( , , , ).
(d) Suppose that a fourth equation −2w + y = 5 is included in the system (∗). What is
the solution of the resulting system? Answer: ( , , , ).
(e) Suppose that instead of the equation in part (d), the equation −2w − 2y = −3 is
included in the system (∗). Then what can you say about the solution(s) of the
resulting system? Answer: .

(3) Consider the following system of equations:


x + y + z = 2

x + 3y + 3z = 0 (∗)

 x+ 3y+ 6z = 3

(a) Use Gaussian elimination to put the augmented coefficient matrix into row echelon
1 1 1 a
form. The result will be 0 1 1 b  where a =
 ,b= , and c = .
0 0 1 c
(b) Use Gauss-Jordan reduction to put the augmented
 coefficient matrix in reduced row
1 0 0 d
echelon form. The result will be 0 1 0 e  where d = , e= , and
0 0 1 f
f = .
(c) The solutions of (∗) are x = ,y= , and z = .

(4) Consider the following system of equations.

0.003000x + 59.14y = 59.17
5.291x − 6.130y = 46.78.
(a) Using only row operation III and back substitution find the exact solution of the
system. Answer: x = ,y= .
(b) Same as (a), but after performing each arithmetic operation round off your answer to
four significant figures. Answer: x = ,y= .
 1.2. EXERCISES 3

(5) Find the values of k for which the system of equations

x + ky = 1


kx + y = 1

has (a) no solution. Answer: .

(b) exactly one solution. Answer: .

(c) infinitely many solutions. Answer: .

(d) When there is exactly one solution, it is x = and y = .

(6) Consider the following two systems of equations.


 x+ y+ z =6

x + 2y + 2z = 11 (1)

 2x + 3y − 4z = 3

and

 x+ y+ z =7

x + 2y + 2z = 10 (2)

 2x + 3y − 4z = 3

Solve both systems simultaneously by applying Gauss-Jordan reduction to an appro-

priate 3 × 5 matrix.
 

(a) The resulting row echelon form of this 3 × 5 matrix is  .

 

(b) The resulting reduced row echelon form is  .

(c) The solution for (1) is ( , , ) and the solution for (2) is ( , , ).

(7) Consider the following system of equations:


 x − y − 3z = 3

2x + z=0

 2y + 7z = c

(a) For what values of c does the system have a solution? Answer: c = .
(b) For the value of c you found in (a) describe the solution set geometrically as a subset
of R3 . Answer: .
(c) What does part (a) say about the planes x − y − 3z = 3, 2x + z = 0, and 2y + 7z = 4
in R3 ? Answer: .
4 1. SYSTEMS OF LINEAR EQUATIONS

(8) Consider the following system of linear equations ( where b1 , . . . , b5 are constants).
 u + 2v − w − 2x + 3y = b1





 x − y + 2z = b2
2u + 4v − 2w − 4x + 7y − 4z = b3




 −x + y − 2z = b4

3u + 6v − 3w − 6x + 7y + 8z = b5
(a) In the process of Gaussian elimination the leading variables of this system are
and the free variables are .
(b) What condition(s) must the constants b1 , . . . , b5 satisfy so that the system is consis-
tent? Answer: .
(c) Do the numbers b1 = 1, b2 = −3, b3 = 2, b4 = b5 = 3 satisfy the condition(s) you
listed in (b)? . If so, find the general solution to the system as a function
of the free variables. Answer:
u=
v=
w=
x=
y=
z= .

(9) Consider the following homogeneous system of linear equations (where a and b are nonzero
constants). 
 x + 2y
 =0
ax + 8y + 3z = 0

 by + 5z = 0
(a) Find a value for a which will make it necessary during Gaussian elimination to inter-
change rows in the coefficient matrix. Answer: a = .
(b) Suppose that a does not have the value you found in part (a). Find a value for b so
that the system has a nontrivial solution.
Answer: b = 3c + d3 a where c = and d = .
(c) Suppose that a does not have the value you found in part (a) and that b = 100.
Suppose further that a is chosen so that the solution to the system is not unique.

The general solution to the system (in terms of the free variable) is α1 z , − β1 z , z
where α = and β = .
 1.3. PROBLEMS 5

1.3. Problems
(1) Give a geometric description of a single linear equation in three variables.
Then give a geometric description of the solution set of a system of 3 linear equations in
3 variables if the system
(a) is inconsistent.
(b) is consistent and has no free variables.
(c) is consistent and has exactly one free variable.
(d) is consistent and has two free variables.
(2) Consider the following system of equations:
−m1 x + y = b1

(∗)
−m2 x + y = b2
(a) Prove that if m1 6= m2 , then (∗) has exactly one solution. What is it?
(b) Suppose that m1 = m2 . Then under what conditions will (∗) be consistent?
(c) Restate the results of (a) and (b) in geometrical language.
6 1. SYSTEMS OF LINEAR EQUATIONS

1.4. Answers to Odd-Numbered Exercises

(1) 2, 3
(3) (a) 2, −1, 1
(b) 3, −2, 1
(c) 3, −2, 1
(5) (a) k = −1
6 −1, 1
(b) k =
(c) k = 1
1 1
(d) ,
k+1 k+1
(7) (a) −6
(b) a line
(c) They have no points in common.
(9) (a) 4
(b) 40, −10
(c) 10, 20
 CHAPTER 2

ARITHMETIC OF MATRICES
matrices
2.1. Background
Topics: addition, scalar multiplication, and multiplication of matrices, inverse of a nonsingular
matrix.

2.1.1. Definition. Two square matrices A and B of the same size are said to commute if AB =
BA.

2.1.2. Definition. If A and B are square matrices of the same size, then the commutator (or
Lie bracket) of A and B, denoted by [A, B], is defined by
[A, B] = AB − BA .

2.1.3. Definition. The trace of a square matrix A, denoted by tr A, is the sum of the diagonal
entries of the matrix. That is, if A = [aij ] is an n × n matrix, then
n
X
tr A := ajj .
j=1

2.1.4. Definition. The transpose of an n × n matrix A = aij is the matrix At = aji obtained
   

by interchanging the rows and columns of A. The matrix A is symmetric if At = A.

mo005thm 2.1.5. Proposition. If A is an m × n matrix and B is an n × p matrix, then (AB)t = B t At .

7
8 2. ARITHMETIC OF MATRICES

2.2. Exercises
   
1 0 −1 2 1 2  
 0 3 1 −1 3 −1 3 −2 0 5
(1) Let A = 
 2 4 0
, B = 
0 −2, and C = 1 0 −3 4 .

3
−3 1 −1 2 4 1
(a) Does the matrix D = ABC exist? If so, then d34 = .
(b) Does the matrix E = BAC exist? If so, then e22 = .
(c) Does the matrix F = BCA exist? If so, then f43 = .
(d) Does the matrix G = ACB exist? If so, then g31 = .
(e) Does the matrix H = CAB exist? If so, then h21 = .
(f) Does the matrix J = CBA exist? If so, then j13 = .
" #
1 1  
1 0
(2) Let A = 21 12 , B = , and C = AB. Evaluate the following.
0 −1
2
 2   
   
(a) A37 = 


 (b) B 63 = 




   
   
(c) B 138 = 


 (d) C 42 = 




Note: If M is a matrix M p is the product of p copies of M .

 
1 1/3
(3) Let A = . Find numbers c and d such that A2 = −I.
c d
Answer: c = and d = .
(4) Let A and B be symmetric n × n-matrices. Then [A, B] = [B, X], where X = .
(5) Let A, B, and C be n × n matrices. Then [A, B]C + B[A, C] = [X, Y ], where X =
and Y = .
 
1 1/3
(6) Let A = . Find numbers c and d such that A2 = 0. Answer: c = and
c d
d= .
 
1 3 2
(7) Consider the matrix a 6 2 where a is a real number.
0 9 5
(a) For what value of a will a row interchange be required during Gaussian elimination?
Answer: a = .
(b) For what value of a is the matrix singular? Answer: a = .
   
1 0 −1 2 1 2  
 0 3 1 −1 3 −1 3 −2 0 5
(8) Let A = 2 4 0
, B = 
0 −2, C = 1 0 −3 4 , and

3
−3 1 −1 2 4 1
M = 3A3 − 5(BC)2 . Then m14 = and m41 = .
(9) If A is an n × n matrix and it satisfies the equation A3 − 4A2 + 3A − 5In = 0, then A is
nonsingular
 2.2. EXERCISES 9

and its inverse is .

 B, and C be n × n matrices. Then [[A, B], C] + [[B, C], A] + [[C, A], B] = X, where
(10) Let A,

 
X=

.


(11) Let A, B, and C be n × n matrices. Then [A, C] + [B, C] = [X, Y ], where X =

and
Y = .
 
 
1 0 0 0
 1 1 0 0
 
 
 4
(12) Find the inverse of  1 1  . Answer:  .
3 1 0
  
3
1 1 1
 
2 2 2 1

(13) The matrix  1 1 1

1 2 3 4
1 1 1 1
H =  21 3 4 5

1 1 1
3 4 5 6
1 1 1 1
4 5 6 7
is the 4 × 4 Hilbert matrix. Use Gauss-Jordan elimination to compute K = H −1 . Then
K44 is (exactly) . Now, create a new matrix H 0 by replacing each entry in H
by its approximation to 3 decimal places. (For example, replace 16 by 0.167.) Use Gauss-
Jordan elimination again to find the inverse K 0 of H 0 . Then K44
0 is .
(14) Suppose that A and B are symmetric n × n matrices. In this exercise we prove that AB
is symmetric if and only if A commutes with B. Below are portions of the proof. Fill in
the missing steps and the missing reasons. Choose reasons from the following list.
(H1) Hypothesis that A and B are symmetric.
(H2) Hypothesis that AB is symmetric.
(H3) Hypothesis that A commutes with B.
(D1) Definition of commutes.
(D2) Definition of symmetric.
mo005thm
(T) Proposition 2.1.5.
Proof. Suppose that AB is symmetric. Then
AB = (reason: (H2) and )
= B t At (reason: )
= (reason: (D2) and )
So A commutes with B (reason: ).
Conversely, suppose that A commutes with B. Then
(AB)t = (reason: (T) )
= BA (reason: and )
= (reason: and )
Thus AB is symmetric (reason: ). 
10 2. ARITHMETIC OF MATRICES

2.3. Problems
(1) Let A be a square matrix. Prove that if A2 is invertible, then so is A.
Hint. Our assumption is that there exists a matrix B such that
A2 B = BA2 = I .
We want to show that there exists a matrix C such that
AC = CA = I .
Now to start with, you ought to find it fairly easy to show that there are matrices L and
R such that
LA = AR = I . (∗)
A matrix L is a left inverse of the matrix A if LA = I; and R is a right inverse
of A if AR = I. Thus the problem boils down to determining whether A can have a left
inverse and a right inverse that are different. (Clearly, if it turns out that they must be
the same, then the C we are seeking is their common value.) So try to prove that if (∗)
holds, then L = R.

(2) Anton speaks French and German; Geraldine speaks English, French and Italian; James
speaks English, Italian, and Spanish; Lauren speaks all the languages the others
 speak
except French; and no one speaks any other language. Make a matrix A = aij with
rows representing the four people mentioned and columns representing the languages they
speak. Put aij = 1 if person i speaks language j and aij = 0 otherwise. Explain the
significance of the matrices AAt and At A.

(3) Portland Fast Foods (PFF), which produces 138 food products all made from 87 basic
ingredients, wants to set up a simple data structure from which they can quickly extract
answers to the following questions:
(a) How many ingredients does a given product contain?
(b) A given pair of ingredients are used together in how many products?
(c) How many ingredients do two given products have in common?
(d) In how many products is a given ingredient used?
In particular, PFF wants to set up a single table in such a way that:
(i) the answer to any of the above questions can be extracted easily and quickly (matrix
arithmetic permitted, of course); and
(ii) if one of the 87 ingredients is added to or deleted from a product, only a single entry
in the table needs to be changed.
Is this possible? Explain.
mo005thm
(4) Prove proposition 2.1.5.

(5) Let A and B be 2 × 2 matrices.

(a) Prove that if the trace of A is 0, then A2 is a scalar multiple of the identity matrix.
(b) Prove that the square of the commutator of A and B commutes with every 2 × 2
matrix C. Hint. What can you say about the trace of [A, B]?
(c) Prove that the commutator of A and B can never be a nonzero multiple of the identity
matrix.
 2.3. PROBLEMS 11

(6) The matrices that represent rotations of the xy-plane are

 
cos θ − sin θ
A(θ) = .
sin θ cos θ

(a) Let x be the vector (−1, 1), θ = 3π/4, and y be A(θ) acting on x (that is, y = A(θ)xt ).
Make a sketch showing x, y, and θ.
(b) Verify that A(θ1 )A(θ2 ) = A(θ1 + θ2 ). Discuss what this means geometrically.
(c) What is the product of A(θ) times A(−θ)? Discuss what this means geometrically.
(d) Two sheets of graph paper are attached at the origin and rotated in such a way that
the point (1, 0) on the upper sheet lies directly over the point (−5/13, 12/13) on the
lower sheet. What point on the lower sheet lies directly below (6, 4) on the upper
one?
(7) Let
0 a a2 a3 a4
 
0 0 a a2 a3 
 
2
A=  0 0 0 a a .
0 0 0 0 a 
0 0 0 0 0
The goal of this problem is to develop a “calculus” for the matrix A. To start, recall
1
(or look up) the power series expansion for . Now see if this formula works for
1−x
the matrix A by first computing (I − A)−1 directly and then computing the power series
expansion substituting A for x. (Explain why there are no convergence difficulties for the
series when we use this particular matrix A.) Next try to define ln(I + A) and eA by
means of appropriate series. Do you get what you expect when you compute eln(I+A) ? Do
formulas like eA eA = e2A hold? What about other familiar properties of the exponential
and logarithmic functions?
Try some trigonometry with A. Use series to define sin, cos, tan, arctan, and so on. Do
things like tan(arctan(A)) produce the expected results? Check some of the more obvious
trigonometric identities. (What do you get for sin2 A + cos2 A − I? Is cos(2A) the same
as cos2 A − sin2 A?)
A relationship between the exponential and trigonometric functions is given by the
famous formula eix = cos x + i sin x. Does this hold for A?
Do you think there are other matrices for which the same results might hold? Which
ones?
(8) (a) Give an example of two symmetric matrices whose product is not symmetric.
Hint. Matrices containing only 0’s and 1’s will suffice.
(b) Now suppose that A and B are symmetric n×n matrices. Prove that AB is symmetric
if and only if A commutes with B.
Hint. To prove that a statement P holds “if and only if” a statement Q holds you must
first show that P implies Q and then show that Q implies P. In the current problem, there
are 4 conditions to be considered:
(i) At = A (A is symmetric),
(ii) B t = B (B is symmetric),
(iii) (AB)t = AB (AB is symmetric), and
(iv) AB = BA (A commutes with B).
Recall also themo005thm
fact given in
(v) theorem 2.1.5.
The first task is to derive (iv) from (i), (ii), (iii), and (v). Then try to derive (iii) from (i),
(ii), (iv), and (v).
12 2. ARITHMETIC OF MATRICES

2.4. Answers to Odd-Numbered Exercises

(1) (a) yes, 142
(b) no, –
(c) yes, −45
(d) no, –
(e) yes, −37
(f) no, –
(3) −6, −1
(5) A, BC
(7) (a) 2
(b) −4
1
(9) (A2 − 4A + 3In )
5
(11) A + B, C
(13) 2800, −1329.909
 CHAPTER 3

ELEMENTARY MATRICES; DETERMINANTS

determinants
3.1. Background
Topics: elementary (reduction) matrices, determinants.

The following definition says that we often regard the effect of multiplying a matrix M on the
left by another matrix A as the action of A on M .
3.1.1. Definition. We say that thematrix
 A acts on the matrix M to produce the matrix N if
0 1
N = AM . For example the matrix acts on any 2 × 2 matrix by interchanging (swapping)
1 0
    
0 1 a b c d
its rows because = .
1 0 c d a b

em168not 3.1.2. Notation. We adopt the following notation for elementary matrices which implement type I
row operations. Let A be a matrix having n rows. For any real number r 6= 0 denote by Mj (r) the
em168z
n × n matrix which acts on A by multiplying its j th row by r. (See exercise 1.)

em167not 3.1.3. Notation. We use the following

ge009def
notation for elementary matrices which implement type II
row operations. (See definition 1.1.1.) Let A be a matrix having n rows. Denote by Pij the n × n
em167z
matrix which acts on A by interchanging its ith and j th rows. (See exercise 2.)

em169not 3.1.4. Notation. And we use the following notation for elementary matrices which implement
ge009def
type III row operations. (See definition 1.1.1.) Let A be a matrix having n rows. For any real
number r denote by Eij (r) the n × n matrix which acts on A by adding r times the j th row of A
em169z
to the ith row. (See exercise 3.)

3.1.5. Definition. If a matrix B can be produced from a matrix A by a sequence of elementary

row operations, then A and B are row equivalent.

Some Facts about Determinants

det001thm1 3.1.6. Proposition. Let n ∈ N and Mn×n be the collection of all n × n matrices. There is exactly
one function
det : Mn×n → R : A 7→ det A
which satisfies
(a) det In = 1.
(b) If A ∈ Mn×n and A0 is the matrix obtained by interchanging two rows of A, then det A0 =
− det A.
(c) If A ∈ Mn×n , c ∈ R, and A0 is the matrix obtained by multiplying each element in one
row of A by the number c, then det A0 = c det A.
(d) If A ∈ Mn×n , c ∈ R, and A0 is the matrix obtained from A by multiplying one row of A
by c and adding it to another row of A (that is, choose i and j between 1 and n with i 6= j
and replace ajk by ajk + caik for 1 ≤ k ≤ n), then det A0 = det A.
13
 14 3. ELEMENTARY MATRICES; DETERMINANTS

3.1.7. Definition. The unique function det : Mn×n → R described above is the n × n determi-
nant function.
det001thm2 3.1.8. Proposition. If A = [a] for a ∈ R (that is, if A ∈ M1×1 ), then det A = a; if A ∈ M2×2 ,
then
det A = a11 a22 − a12 a21 .
det001thm3 3.1.9. Proposition. If A, B ∈ Mn×n , then det(AB) = (det A)(det B).
det001thm4 det At = det A. (An obvious corollary of this: in
3.1.10. Proposition. If A ∈ Mn×n , then det001thm1
conditions (b), (c), and (d) of proposition 3.1.6 the word “columns” may be substituted for the
word “rows”.)
3.1.11. Definition. Let A be an n × n matrix. The minor of the element ajk , denoted by Mjk , is
the determinant of the (n − 1) × (n − 1) matrix which results from the deletion of the jth row and
kth column of A. The cofactor of the element ajk , denoted by Cjk is defined by
Cjk := (−1)j+k Mjk .
det001thm5 3.1.12. Proposition. If A ∈ Mn×n and 1 ≤ j ≤ n, then
Xn
det A = ajk Cjk .
k=1

This is the (Laplace) expansion of the determinant along the jth row.
det001thm4
In light of 3.1.10, it is clear that expansion along columns works as well as expansion along
rows. That is,
Xn
det A = ajk Cjk
j=1
for any k between 1 and n. This is the (Laplace) expansion of the determinant along the kth
column.
det001thm6 3.1.13. Proposition. An n × n matrix A is invertible if and only if det A 6= 0. If A is invertible,
then
A−1 = (det A)−1 C t
 
where C = Cjk is the matrix of cofactors of elements of A.
 3.2. EXERCISES 15

3.2. Exercises
em168z (1) Let A be a matrix with 4 rows. The matrix M3 (4) which multiplies the 3rd row of A by 4

  em168not
is 

. (See 3.1.2.)


(2) Let A be nd th
em167z  a matrix with4 rows. The matrix P24 which interchanges the 2 and 4 rows
  em167not
of A is 

. (See 3.1.3.)


rd
em169z  4 rows. Thematrix E23 (−2) which adds −2 times the 3 row of
(3) Let A be a matrix with

  em169not
A to the 2nd row is 

. (See 3.1.4.)


 

11 = 

(4) Let A be
 the 4 × 4 elementary
 matrix E43 (−6). Then A 
 and


A−9 = 
 
.
 

 

−9
 
(5) Let B be  4 × 4 matrix P24 . Then B = 
 the elementary
  and

 
B 10 = 

.


 

4
 
(6) Let C be  4 × 4 matrix M3 (−2). Then C = 
 the elementary
  and


C −3 = 
 
.
 

 
1 2 3
 0 −1 1 
(7) Let A = −2 1
 and B = P23 E34 (−2)M3 (−2)E42 (1)P14 A. Then b23 =
0
−1 2 −3
and b32 = .
(8) We apply Gaussian elimination (using type III elementary row operations only) to put a
4 × 4 matrix A into upper triangular form. The result is
 
1 2 −2 0
0 −1 0 1
E43 ( 52 )E42 (2)E31 (1)E21 (−2)A = 
0 0 −2 2  .


0 0 0 10
Then the determinant of A is .
16 3. ELEMENTARY MATRICES; DETERMINANTS

(9) The system of equations: 


 2y+3z = 7
x+ y− z = −2

 −x+ y−5z = 0

is solved
 by applying Gauss-Jordan
 reduction to the augmented coefficient matrix
0 2 3 7
A =  1 1 −1 −2. Give the names of the elementary 3 × 3 matrices X1 , . . . , X8
−1 1 −5 0
which implement the following reduction.
     
1 1 −1 −2 1 1 −1 −2 1 1 −1 −2
X1 X2 X3
A −−−− → 0 2 3 7  −−−− → 0 2 3 7  −−−−→ 0 2 3 7
−1 1 −5 0 0 2 −6 −2 0 0 −9 −9
     
1 1 −1 −2 1 1 −1 −2 1 1 −1 −2
X4 X5 X6
−−−−→ 0 2 3 7  −−−− → 0 2 0 4  −−−−→ 0 1 0 2
0 0 1 1 0 0 1 1 0 0 1 1
   
1 1 0 −1 1 0 0 −3
X7 X8
−−−−→ 0 1 0 2  −−−− → 0 1 0 2  .
0 0 1 1 0 0 1 1
Answer: X1 = , X2 = , X3 = , X4 = ,

X5 = , X6 = , X7 = , X8 = .

(10) Solve the following equation for x:

 
3 −4 7 0 6 −2
2 0 1 8 0 0
 
3 4 −8 3 1 2 
det 27
 = 0. Answer: x = .
 6 5 0 0 3
3 x 0 2 1 −1
1 0 −1 3 4 0

 
0 0 1
(11) Let A = 0 2 4. Find A−1 using the technique of augmenting A by the identity matrix
1 2 3
I and performing Gauss-Jordan reduction on the augmented matrix. The reduction can
be accomplished by the application of five elementary 3 × 3 matrices. Find elementary
matrices X1 , X2 , and X3 such that A−1 = X3 E13 (−3)X2 M2 (1/2)X1 I.
(a) The required matrices are X1 = P1i where i = , X2 = Ejk (−2) where j =
and k = , and X3 = E12 (r) where r = .
 

(b) And then A−1 = 

 
.
 

1 t t2 t3
 
 t 1 t t2  p
t2 t 1 t  = (1 − a(t)) where a(t) =
(12) det  and p = .


t3 t2 t 1
 3.2. EXERCISES 17

(13) Evaluateeach of the following

 determinants.
6 9 39 49
5 7 32 37
(a) det 3 4 4 4  =
 .
1 1 1 1
 
1 0 1 1
1 −1 2 0 
(b) det 2 −1 3 1  =
 .
4 17 0 −5
 
13 3 −8 6
 0 0 −4 0 
(c) det 
 1 0 7 −2 =
 .
3 0 2 0
 
5 4 −2 3
5 7 −1 8 
(14) Let M be the matrix 
5
.
7 6 10
5 7 1 9
(a) The determinant of M  can be expressed
 as the constant 5 times the determinant of
3 1 5
the single 3 × 3 matrix 3 .
3
(b) The determinant of this 3 × 3 matrix can
 be expressed
 as the constant 3 times the
7 2
determinant of the single 2 × 2 matrix .
2
(c) The determinant of this 2 × 2 matrix is .
(d) Thus the determinant of M is .
 
1 2 5 7 10
1 2 3 6 7
 
1
(15) Find the determinant of the matrix  1 3 5 5. Answer: .
1 1 2 4 5
1 1 1 1 1

(16) Find the determinants of the following matrices.

   
−73 78 24 −73 78 24
A = 92 66
 25 and B =  92 66 25  .
−80 37 10 −80 37 10.01
Hint. Use a calculator (thoughtfully). Answer: det A = and det B = .

(17) Find the determinant of the following matrix.

 83 
283 5 π 347.86 × 1015
3136 56 5 cos(2.7402) 
 .
6776 121 11 5 
2464 44 4 2
Hint. Do not use a calculator. Answer: .
18 3. ELEMENTARY MATRICES; DETERMINANTS

(18) Suppose that A is a square matrix with determinant 7. Then

(a) det(P24 A) = .
(b) det(E23 (−4)A) = .
(c) det(M3 (2)A) = .
 3.3. PROBLEMS 19

3.3. Problems
(1) For this problem assume that we know the following: If X is an m × m matrix, if Y is
an m × n matrix
 and if 0 and I are zero and identity matrices of appropriate sizes, then
X Y
det = det X.
0 I
Let A be an m × n matrix and B be an n × m matrix. Prove carefully that
 
0 A
det = det AB .
−B I
  
0 A I 0
Hint. Consider the product .
−B I B I
(2) Let A and B be n × n-matrices. Your good friend Fred R. Dimm believes that
 
A B
det = det(A + B) det(A − B).
B A
He offers the following argument to support this claim:
 
A B
det = det(A2 − B 2 )
B A
= det[(A + B)(A − B)]
= det(A + B) det(A − B) .
(a) Comment (helpfully) on his “proof”. In particular, explain carefully why each of
the three steps in his “proof” is correct or incorrect. (That is, provide a proof or a
counterexample to each step.)
(b) Is the result he is trying to prove actually
  true? 
I B A+B 0
Hint: Consider the product .
0 A−B 0 I
(3) Let x be a fixed real
 number
 which is not an integer multiple of π. For each natural
number n let An = ajk be the n × n-matrix defined by

0,
 for |j − k| > 1
ajk = 1, for |j − k| = 1

2 cos x, for j = k.


sin(n + 1)x
Show that det An = . Hint. For each integer n let Dn = det An and prove that
sin x
Dn+2 − 2Dn+1 cos x + Dn = 0.
(Use mathematical induction.)
20 3. ELEMENTARY MATRICES; DETERMINANTS

3.4. Answers to Odd-Numbered Exercises

 
1 0 0 0
0 1 0 0
(1) 
0

0 4 0
0 0 0 1
 
1 0 0 0
0 1 −2 0
(3) 
0

0 1 0
0 0 0 1
   
1 0 0 0 1 0 0 0
0 0 0 1 0 1 0 0
(5) 
0
,  
0 1 0 0 0 1 0
0 1 0 0 0 0 0 1
(7) −8, −1
(9) P12 , E31 (1), E32 (−1), M3 (− 19 ), E23 (−3), M2 ( 12 ), E13 (1), E12 (−1)
3, 2, 3, −2
(11) (a)  
1 −1 1
(b) −2 21 0
1 0 0
(13) 100, 0, −72
(15) −10
(17) 6
 CHAPTER 4

VECTOR GEOMETRY IN Rn
tor_geometry
4.1. Background
Topics: inner (dot) products, cross products, lines and planes in 3-space, norm of a vector, angle
between vectors.

4.1.1. Notation. There are many more or less standard notations for the inner product of two
vectors x and y. The two that we will use interchangeably in these exercises are x · y and hx, yi.
ipr011def 4.1.2. Definition. If x is a vector in Rn , then the norm (or length) of x is defined by
p
kxk = hx, xi .
ipr007def 4.1.3. Definition. Let x and y be nonzero vectors in Rn . Then ](x, y), the angle between x
and y, is defined by
hx, yi
](x, y) = arccos
kxk kyk
ipr010thm 4.1.4. Theorem (Cauchy-Schwarz inequality). If x and y are vectors in Rn , then
|hx, yi| ≤ kxk kyk .
(We will often refer to this just as the Schwarz inequality.)

21
22 4. VECTOR GEOMETRY IN Rn

4.2. Exercises
√
(1) The angle between the vectors (1, 0, −1, 3) and (1, 3, 3, −3) in R4 is aπ where a = .
√
(2) Find the angle θ between the vectors x = (3, −1, 1, 0, 2, 1) and y = (2, −1, 0, 2, 2, 1)
in R6 . Answer: θ = .
(3) If a1 , . . . , an > 0, then
n
X n
X 
1
aj ≥ n2 .
ak
j=1 k=1
The proof of this is obvious from the Cauchy-Schwarz inequality when we choose the
vectors x and y as follows:
x= and y = .
(4) Find all real numbers α such that the angle between the vectors 2i + 2j + (α − 2)k and
2i + (α − 2)j + 2k is π3 . Answer: α = and .
(5) Which of the angles (if any) of triangle ABC, with A = (1, −2, 0), B = (2, 1, −2), and
C = (6, −1, −3), is a right angle? Answer: the angle at vertex .
(6) The hydrogen atoms of a methane molecule (CH4 ) are located at (0, 0, 0), (1, 1, 0), (0, 1, 1),
and (1, 0, 1) while the carbon atom is at ( 21 , 12 , 12 ). Find the cosine of the angle θ between
two rays starting at the carbon atom and going to different hydrogen atoms.
Answer: cos θ = .
(7) If a, b, c, d, e, f ∈ R, then
p p
|ad + be + cf | ≤ a2 + b2 + c2 d2 + e2 + f 2 .
The proof of this inequality is obvious since this is just the Cauchy-Schwarz inequality
where x = ( , , ) and y = ( , , ).
(8) The volume of the parallelepiped generated by the three vectors i + 2j − k, j + k, and
3i − j + 2k is .
(9) The equations of the line containing the points (3, −1, 4) and (7, 9, 10) are
x−3 y−j z−k
= =
2 b c
where b = ,c= ,j= , and k = .
(10) The equations of the line containing the points (5, 2, −1) and (9, −4, 1) are
x−h y−2 z−k
= =
a −3 c
where a = ,c= ,h= , and k = .
(11) Find the symmetric equations of the line containing the point (1, 0, −1) which is parallel
x−4 2y − 3 3z − 7
to the line = = .
2 5 6
x−h y−j z+1
Answer: = = where a = ,b= ,h= , and j = .
a b 4
(12) The equation of the plane containing the points (0, −1, 1), (1, 0, 2), and (3, 0, 1) is x + by +
cz = d where b = ,c= , and d = .
(13) The equation of the plane which passes through the points (0, −1, −1), (5, 0, 1), and
(4, −1, 0) is ax + by + cz = 1 where a = ,b= , and c = .
a
(14) The angle between the planes 4x + 4z − 16 = 0 and −2x + 2y − 13 = 0 is π where
b
a= and b = .
 4.2. EXERCISES 23

(15) Suppose that u ∈ R3 is a vector which lies in the first quadrant of the xy-plane and has
length 3 and that v ∈ R3 is a vector that lies along the positive z-axis and has length 5.
Then
(a) ku × vk = ;
(b) the x-coordinate of u × v is 0 (choose <, >, or =);
(c) the y-coordinate of u × v is 0 (choose <, >, or =); and
(d) the z-coordinate of u × v is 0 (choose <, >, or =).
7
√
(16) Suppose√that u and v are vectors in R both of length 2 2 and that the length of u − v
is also 2 2. Then ku + vk = and the angle between u and v is .
24 4. VECTOR GEOMETRY IN Rn

4.3. Problems

2
(1) Show that if a, b, c > 0, then 12 a + 13 b + 61 c ≤ 12 a2 + 13 b2 + 16 c2 .
(2) Show that if a1 , . . . , an , w1 , . . . , wn > 0 and nk=1 wk = 1, then
P
X n 2 X n
ak wk ≤ ak 2 wk .
k=1 k=1

∞
X
(3) Prove that if (a1 , a2 , . . . ) is a sequence of real numbers such that the series ak 2 con-
∞
X 1 k=1
verges, then the series ak converges absolutely.
k
k=1
You may find the following steps helpful in organizing your solution.
(i) First of all, make sure that you recall the difference between a sequence of numbers
∞
X
(c1 , c2 , . . . ) and an infinite series ck .
k=1
(ii) The key to this problem is an important theorem from third term Calculus:
A nondecreasing sequence of real numbers converges if and only if it is bounded. (∗)
(Make sure that you know the meanings of all the terms used here.)
X∞
(iii) The hypothesis of the result we are trying to prove is that the series ak 2 converges.
What, exactly, does this mean? k=1
Xn
(iv) For each natural number n let bn = ak 2 . Rephrase (iii) in terms of the se-
k=1
quence (bn ).
(v) Is the sequence (bn ) nondecreasing?
(vi) What, then, does (∗) say about the sequence (bn )?
n
X 1
(vii) For each natural number n let cn = . What do we know about the sequence
k2
k=1
(cn ) from third term Calculus? What does (∗) say about the sequence (cn )?
∞
X 1
(viii) The conclusion we are trying to prove is that the series ak converges absolutely.
k
What does this mean? k=1
n
X 1
(ix) For each natural number n let sn = |ak |. Rephrase (viii) in terms of the se-
k
quence (sn ). k=1
(x) Explain how for each n we may regard the number sn as the dot product of two
vectors in Rn .
(xi) Apply the Cauchy-Schwarz inequality to the dot product in (x). Use (vi) and (vii) to
establish that the sequence (sn ) is bounded above.
(xii) Use (∗) one last time—keeping in mind what you said in (ix).
 4.4. ANSWERS TO ODD-NUMBERED EXERCISES 25

4.4. Answers to Odd-Numbered Exercises

3
(1)
4
√ √ √

√1 , √1 , . . . , √1


(3) a1 , a2 , . . . , an , a1 a2 an

(5) B
(7) a, b, c, d, e, f
(9) 5, 3, −1, 4
(11) 4, 5, 1, 0
(13) 1, 3, −4
(15) (a) 15
(b) >
(c) <
(d) =
 CHAPTER 5

VECTOR SPACES
ector_spaces
5.1. Background
Topics: real and complex vector spaces, vectors, scalars.

In the following definition F may be taken to be an arbitrary field. For this collection of
exercises, however, we will be interested in only two cases, F = R (the field of real numbers)and
F = C (the field of complex numbers).
vs001def 5.1.1. Definition. A vector space is a set V together with operations of addition and scalar
multiplication which satisfy the following axioms:
(1) if x, y ∈ V , then x + y ∈ V ;
(2) (x + y) + z = x + (y + z) for every x, y, z ∈ V (associativity);
(3) there exists 0 ∈ V such that x + 0 = x for every x ∈ V (existence of additive identity);
(4) for every x ∈ V there exists −x ∈ V such that x + (−x) = 0 (existence of additive
inverses);
(5) x + y = y + x for every x, y ∈ V (commutativity);
(6) if α ∈ F and x ∈ V , then αx ∈ V ;
(7) α(x + y) = αx + αy for every α ∈ F and every x, y ∈ V ;
(8) (α + β)x = αx + βx for every α, β ∈ F and every x ∈ V ;
(9) (αβ)x = α(βx) for every α, β ∈ F and every x ∈ V ; and
(10) 1 x = x for every x ∈ V .
When F = R we speak of V as a real vector space and when F = C we call it a complex
vector space.
5.1.2. Definition. An n × n-matrix is nonsingular if its determinant is not zero.

27
28 5. VECTOR SPACES

5.2. Exercises
(1) Let V be the set of all real numbers. Define an operation of “addition” by
x
y = the maximum of x and y
for all x, y ∈ V . Define an operation of “scalar multiplication” by
α x = αx
for all α ∈ R and x ∈ V .
Under the operations
and
vs001def
the set V is not a vector space. The vector space
axioms (see 5.1.1 (1)–(10) ) which fail to hold are , , , and .
(2) Let V be the set of all real numbers x such that x ≥ 0. Define an operation of “addition”
by
x
y = xy + 1
for all x, y ∈ V . Define an operation of “scalar multiplication” by
α x = α2 x
for all α ∈ R and x ∈ V .
Under the operations
and the setvs001def
V (is/is not) a vector space. If it is
not, list all the vector space axioms (see 5.1.1 (1)–(10) ) which fail to hold. Answer: The
axioms which are not satisfied are .
(3) Let V be R2 , the set of all ordered pairs (x, y) of real numbers. Define an operation of
“addition” by
(u, v)
(x, y) = (u + x + 1, v + y + 1)
for all (u, v) and (x, y) in V . Define an operation of “scalar multiplication” by
α (x, y) = (αx, αy)
for all α ∈ R and (x, y) ∈ V .
Under the operations
and
vs001def
the set V is not a vector space. The vector space
axioms (see 5.1.1 (1)–(10) ) which fail to hold are and .
(4) Let V be R2 , the set of all ordered pairs (x, y) of real numbers. Define an operation of
“addition” by
(u, v)
(x, y) = (u + x, 0)
for all (u, v) and (x, y) in V . Define an operation of “scalar multiplication” by
α (x, y) = (αx, αy)
for all α ∈ R and (x, y) ∈ V .
Under the operations
and
vs001def
the set V is not a vector space. The vector space
axioms (see 5.1.1 (1)–(10) ) which fail to hold are , , and .
(5) Let V be the set of all n × n matrices of real numbers. Define an operation of “addition”
by
A
B = 12 (AB + BA)
for all A, B ∈ V . Define an operation of “scalar multiplication” by
α A=0
for all α ∈ R and A ∈ V .
Under the operations
and
vs001def
the set V is not a vector space. The vector space
axioms (see 5.1.1 (1)–(10) ) which fail to hold are , , and .
 5.2. EXERCISES 29

(6) Below are portions of proofs of four results about vector spaces which establish the fact
that multiplying a vector x by the scalar −1 produces −x, the additive inverse of x. Fill
in the missing steps and the missing reasons. Choose reasons from the following list.
(H) Hypothesis
vs001def
(1)–(10) Vector space axioms, see 5.1.1
(PA) Proposition A
(PB) Proposition B
(PC) Proposition C
(RN) Property of the Real Numbers

vs_prop_a 5.2.1. Proposition (A). A vector x in a vector space V has at most one additive inverse.
That is, if y and z are vectors in V such that x + y = 0 and x + z = 0, then y = z.
Proof. Suppose that x + y = 0 and x + z = 0. Then
y= (reason: )
= y + (x + z) (reason: )
= (reason: (2) )
= (x + y) + z) (reason: )
= (reason: (H) )
= (reason: (5) )
=z (reason: ).


vs_prop_b 5.2.2. Proposition (B). If x ∈ V where V is a vector space and x + x = x, then x = 0.

Proof. If x ∈ V and x + x = x, then
x=x+0 (reason: )
= (reason: (4) )
= (x + x) + (−x) (reason: )
= (reason: (H) )
=0 (reason: ).


vs_prop_c 5.2.3. Proposition (C). If x is a vector in a vector space V , then 0x = 0.

Proof. If x ∈ V , then
0x = (0 + 0) x (reason: )
= (reason: (8) )
Thus 0x = 0 (reason: ). 
 30 5. VECTOR SPACES

5.2.4. Proposition (D). If x is a vector in a vector space V , then (−1)x is −x, the
additive inverse of x.
Proof. If x ∈ V , then
x + (−1) · x = (reason: (10) )


= 1 + (−1) · x (reason: )
=0·x (reason: )
=0 (reason: ).
It then follows immediately from that (−1) · x = −x. 

(7) In this exercise we prove that multiplying the zero vector by an arbitrary scalar produces
the zero vector. For each step of the proof give the appropriate reason. Choose reasons
from the following list.
vs001def
(1)–(10) Vector space axioms 5.1.1.
vs_prop_b
(PB) Proposition 5.2.2
(RN) Property of the Real Numbers
vs_prop_e 5.2.5. Proposition (E). If 0 is the zero vector in a vector space and α is a scalar, then
α · 0 = 0.
Proof. Let 0 be the zero vector of some vector space. Then for every scalar α
α · 0 = α · (0 + 0) reason:
=α·0+α·0 reason:
It then follows immediately from that α · 0 = 0. 

(8) In this exercise we prove that the product of a scalar and a vector is zero if and only if
either the scalar or the vector is zero. After each step of the proof give the appropriate
reason. Choose reasons from the following list.
(H) Hypothesis.
vs001def
(1)–(10) Vector space axioms 5.1.1.
vs_prop_c vs_prop_e
(PC),(PE) Propositions 5.2.3 and 5.2.5, respectively.
(RN) Property of the Real Numbers.
5.2.6. Proposition. Suppose that x is a vector and α is a scalar. Then αx = 0 if and
only if α = 0 or x = 0.
Proof. We have already shown in and that if α = 0 or x = 0,
then αx = 0.
To prove the converse we suppose that αx = 0 and that α 6= 0; and we prove that
x = 0. This conclusion results from the following easy calculation:
x=1·x reason:
 
1
= ·α ·x reasons: and
α
1
= · (α · x) reason:
α
1
= ·0 reason:
α
=0 reason: . 
 5.3. PROBLEMS 31

5.3. Problems
(1) Prove that if V is a vector space, then its additive identity is unique. That is, show that
e are vectors in V such that x + 0 = x for all x ∈ V and x + 0
if 0 and 0 e = x for all x ∈ V ,
then 0 = 0.
e
(2) Let V be the set of all real numbers x such that x > 0. Define an operation of “addition”
by
x
y = xy
for all x, y ∈ V . Define an operation of “scalar multiplication” by
α x = xα
for all α ∈ R and x ∈ V .
Prove that under the operations
and the set V is a vector space.
(3) With the usual operations of addition and scalar multiplication the set of all n×nvs001def
matrices
of real numbers is a vector space: in particular, all the vector space axioms (see 5.1.1 (1)–
(10) ) are satisfied. Explain clearly why the set of all nonsingular n × n matrices of real
numbers is not a vector space under these same operations.
32 5. VECTOR SPACES

5.4. Answers to Odd-Numbered Exercises

(1) (3), (4), (7), (8)
(3) (7), (8)
(5) (2), (4), (10)
(7) (3), (7), (PB)
 CHAPTER 6

SUBSPACES
subspaces
6.1. Background
Topics: subspaces of a vector space

6.1.1. Definition. A nonempty subset of M of a vector space V is a subspace of V if it is closed

under addition and scalar multiplication. (That is: if x and y belong to M , so does x + y; and if
x belongs to M and α ∈ R, then αx belongs to M .
6.1.2. Notation. We use the notation M  V to indicate that M is a subspace of a vector space V .
sbs027not 6.1.3. Notation. Here are some frequently encountered families of functions:
F = F[a, b] = {f : f is a real valued function on the interval [a, b]} (6.1.1)
P = P[a, b] = {p : p is a polynomial function on [a, b]} (6.1.2)
P4 = P4 [a, b] = {p ∈ P : the degree of p is less than 4} (6.1.3)
Q4 = Q4 [a, b] = {p ∈ P : the degree of p is equal to 4} (6.1.4)
C = C[a, b] = {f ∈ F : f is continuous} (6.1.5)
D = D[a, b] = {f ∈ F : f is differentiable} (6.1.6)
K = K[a, b] = {f ∈ F : f is a constant function} (6.1.7)
B = B[a, b] = {f ∈ F : f is bounded} (6.1.8)
J = J [a, b] = {f ∈ F : f is integrable} (6.1.9)
(A function f ∈ F is bounded if there exists a number M ≥ 0 such that |f (x)| ≤ M for all x in
Rb
[a, b]. It is (Riemann) integrable if it is bounded and a f (x) dx exists.)
6.1.4. Definition. If A and B are subsets of a vector space then the sum of A and B, denoted by
A + B, is defined by
A + B := {a + b : a ∈ A and b ∈ B}.
6.1.5. Definition. Let M and N be subspaces of a vector space V . If M ∩N = {0} and M +N = V ,
then V is the (internal) direct sum of M and N . In this case we write
V = M ⊕ N.
In this case the subspaces M and N are complementary and each is the complement of the
other.

33
34 6. SUBSPACES

6.2. Exercises
(1) One of the following is a subspace of R3 . Which one?
The set of points (x, y, z) in R3 such that
(a) x + 2y − 3z = 4.
x−1 y+2 z
(b) = = .
2 3 4
(c) x + y + z = 0 and x − y + z = 1.
(d) x = −z and x = z.
(e) x2 + y 2 = z.
x y−3
(f) = .
2 5
Answer: ( ) is a subspace of R3 .

(2) The smallest subspace of R3 containing the vectors (2, −3, −3) and (0, 3, 2) is the plane
whose equation is ax + by + 6z = 0 where a = and b = .

(3) The smallest subspace of R3 containing the vectors (0, −3, 6) and (0, 1, −2) is the line
whose equations are x = a and z = by where a = and b = .

(4) Let R∞ denote the vector space of all sequences of real numbers. (Addition and scalar
multiplication are defined coordinatewise.) In each of the following a subset of R∞ is
described. Write yes if the set is a subspace of R∞ and no if it is not.
(a) Sequences that have infinitely many zeros (for example, (1, 1, 0, 1, 1, 0, 1, 1, 0, . . . )).
Answer: .
(b) Sequences which are eventually zero. (A sequence (xk ) is eventually zero if there is
an index n0 such that xn = 0 whenever n ≥ n0 .) Answer: .
(c) P
Sequences that are absolutely summable. (A sequence (xk ) is absolutely summable if
∞
k=1 |xk | < ∞.) Answer: .
(d) Bounded sequences. (A sequence (xk ) is bounded if there is a positive number M such
that |xk | ≤ M for every k.) Answer: .
(e) Decreasing sequences. (A sequence (xk ) is decreasing if xn+1 ≤ xn for each n.)
Answer: .
(f) Convergent sequences. Answer: .
(g) Arithmetic progressions. (A sequence (xk ) is arithmetic if it is of the form (a, a + k,
a + 2k, a + 3k, . . . ) for some constant k.) Answer: .
(h) Geometric progressions. (A sequence (xk ) is geometric if it is of the form (a, ka, k 2 a, k 3 a, . . . )
for some constant k.) Answer: .

(5) Let M and N be subspaces of a vector space V . Consider the following subsets of V .
(a) M ∩ N . (A vector v belongs to M ∩ N if it belongs to both M and N .)
(b) M ∪ N . (A vector v belongs to M ∪ N if it belongs to either M or N .)
(c) M + N . (A vector v belongs to M + N if there are vectors m ∈ M and n ∈ N such
that v = m + n.)
(d) M − N . (A vector v belongs to M − N if there are vectors m ∈ M and n ∈ N such
that v = m − n.)
Which of (a)–(d) are subspaces of V ?
Answer: .
 6.2. EXERCISES 35
sbs027not
(6) For a fixed interval [a, b], which sets of functions in the list 6.1.3 are vector subspaces of
which?
Answer:

       .

(7) Let M be the plane x + y + z = 0 and N be the line x = y = z in R3 . The purpose of

this exercise is to confirm that R3 = M ⊕ N . This requires establishing three things: (i)
M and N are subspaces of R3 (which is very easy and which we omit); (ii) R3 = M + N ;
and (iii) M ∩ N = {0}.
(a) To show that R3 = M + N we need R3 ⊆ M + N and M + N ⊆ R3 . Since M ⊆ R3
and N ⊆ R3 , it is clear that M + N ⊆ R3 . So all that is required is to show that
R3 ⊆ M + N . That is, given a vector x = (x1 , x2 , x3 ) in R3 we must find vectors
m = (m1 , m2 , m3 ) in M and n = (n1 , n2 , n3 ) in N such that x = m + n. Find two
such vectors.
1
Answer: m = ( , , )
3
1
and n = ( , , ).
3
(b) The last thing to verify is that M ∩ N = {0}; that is, that the only vector M and N
have in common is the zero vector. Suppose that a vector x = (x1 , x2 , x3 ) belongs to
both M and N . Since x ∈ M it must satisfy the equation

x1 + x2 + x3 = 0. (1)

since x ∈ N it must satisfy the equations

x1 = x2 and (2)
x2 = x3 . (3)

Solve the system of equations (1)–(3).

Answer: x = ( , , ).
(8) Let C = C[−1, 1] be the vector space of all continuous real valued functions on the interval
[−1, 1]. A function f in C is even if f (−x) = f (x) for all x ∈ [−1, 1]; it is odd if f (−x) =
−f (x) for all x ∈ [−1, 1]. Let Co = {f ∈ C : f is odd } and Ce = {f ∈ C : f is even }. To
show that C = Co ⊕ Ce we need to show 3 things.
(i) Co and Ce are subspaces of C. This is quite simple: let’s do just one part of the proof.
We will show that Co is closed under addition. After each step of the following proof
indicate the justification for that step. Make your choices from the following list.

(A) Arithmetic of real numbers.

(DA) Definition of addition of functions.
(DE) Definition of “even function.”
(DO) Definition of “odd function.”
(H) Hypothesis (that is, our assumptions or suppositions).
36 6. SUBSPACES

Proof. Let f, g ∈ Co . Then

(f + g)(−x) = f (−x) + g(−x) reason:
= −f (x) + (−g(x)) reason: and
= −(f (x) + g(x)) reason:
= −(f + g)(x). reason:
Thus f + g ∈ Co . reason . 

(ii) Co ∩ Ce = {0} (where 0 is the constant function on [−1, 1] whose value is zero). Again
choose from the reasons listed in part (i) to justify the given proof.

Proof. Suppose f ∈ Co ∩ Ce . Then for each x ∈ [−1, 1]

f (x) = f (−x) reason:
= −f (x). reason:
Thus f (x) = 0 for every x ∈ [−1, 1]; that is, f = 0. reason: . 

(iii) C = Co + Ce . To verify this we must show that every continuous function f on [−1, 1]
can be written as the sum of an odd function j and an even function k. It turns
out that the functions j and k can be written as linear combinations of the given
function f and the function g defined by g(x) = f (−x) for −1 ≤ x ≤ 1. What are
the appropriate coefficients?
Answer: j= f+ g
k= f+ g.

(9) Let M be the line x = y = z and N be the line x = 21 y = 13 z in R3 .

(a) The line M is the set of all scalar multiples of the vector ( 1 , , ).
(b) The line N is the set of all scalar multiples of the vector ( 1 , , ).
(c) The set M + N is (geometrically speaking) a 3
in R ; its equation is ax +
by + z = 0 where a = and b = .
(10) Let M be the plane x − y + z = 0 and N be the plane x + 2y − z = 0 in R3 . State in one
short sentence how you know that R3 is not the direct sum of M and N .
Answer: .
x y z
(11) Let M be the plane 2x − 3y + 4z + 1 = 0 and N be the line = = in R3 . State in
4 2 3
one short sentence how you know that R3 is not the direct sum of M and N .
Answer: .

(12) Let M be the plane x + y + z = 0 and N be the line x − 1 = 21 y = z + 2 in R3 . State in

one short sentence how you know that R3 is not the direct sum of M and N .
Answer: .
x y z
(13) Let M be the line x = y = z and N be the line = = in R3 . State in one short
4 2 3
sentence how you know that R3 is not the direct sum of M and N .
Answer: .
 6.2. EXERCISES 37

(14) Let M be the plane x + y + z = 0 and N be the line x = − 34 y = 3z. The purpose of this
exercise is to see (in two different ways) that R3 is not the direct sum of M and N .
(a) If R3 were equal to M ⊕ N , then M ∩ N would contain only the zero vector. Show
that this is not the case by finding a nonzero vector x in R3 which belongs to M ∩ N .
Answer: x = ( , ,1).
(b) If R3 were equal to M ⊕ N , then, in particular, we would have R3 = M + N . Since
both M and N are subsets of R3 , it is clear that M + N ⊆ R3 . Show that the reverse
inclusion R3 ⊆ M + N is not correct by finding a vector x ∈ R3 which cannot be
written in the form m + n where m ∈ M and n ∈ N .
Answer: x = (−6, 8, a) is such a vector provided that a 6= .
(c) We have seen in part (b) that M + N 6= R3 . Then what is M + N ?
Answer: M + N = .
38 6. SUBSPACES

6.3. Problems
(1) Let M and N be subspaces of a vector space V . Consider the following subsets of V .
(a) M ∩ N . (A vector v belongs to M ∩ N if it belongs to both M and N .)
(b) M ∪ N . (A vector v belongs to M ∪ N if it belongs to either M or N .)
(c) M + N . (A vector v belongs to M + N if there are vectors m ∈ M and n ∈ N such
that v = m + n.)
(d) M − N . (A vector v belongs to M − N if there are vectors m ∈ M and n ∈ N such
that v = m − n.)
For each of the sets (a)–(d) above, either prove that it is a subspace of V or give a
counterexample to show that it need not be a subspace of V .
(2) Let C = C[0, 1] be the family of continuous real valued functions on the interval [0, 1].
Define
f1 (t) = t and f2 (t) = t4
for 0 ≤ t ≤ 1. Let M be the set of all functions of the form αf1 + βf2 where α, β ∈ R.
And let N be the set of all functions g in C which satisfy
Z 1 Z 1
tg(t) dt = 0 and t4 g(t) dt = 0.
0 0
Is C the direct sum of M and N ? (Give a careful proof of your claim and illustrate it
with an example. What does your result say, for instance, about the function h defined
by h(t) = t2 for 0 ≤ t ≤ 1.)
(3) Let V be a vector space. T
(a) Let M be a family of subspaces of V . Prove that the intersection M of this family
is itself a subspace of V .
(b) Let A be a set of vectors in V . Explain carefully why it makes sense to say that the
intersection of the family of all subspaces containing A is “the smallest subspace of
V which contains A”.
(c) Prove that the smallest subspace of V which contains A is in fact the span of A.
(4) In R3 let M be the line x = y = z, N be the line x = 21 y = 13 z, and L = M + N . Give a
careful proof that L = M ⊕ N .
(5) Let V be a vector space and suppose that V = M ⊕ N . Show that for every v ∈ V there
exist unique vectors m ∈ M and n ∈ N such that v = m + n. Hint. It should be clear
that the only thing you have to establish is the uniqueness of the vectors m and n. To this
end, suppose that a vector v in V can be written as m1 + n1 and it can also be written
as m2 + n2 where m1 , m2 ∈ M and n1 , n2 ∈ N . Prove that m1 = m2 and n1 = n2 .
 6.4. ANSWERS TO ODD-NUMBERED EXERCISES 39

6.4. Answers to Odd-Numbered Exercises

(1) (d)
(3) 0, −2
(5) (a), (c), and (d)
(7) (a) 2x1 −x2 −x3 , −x1 +2x2 −x3 , −x1 −x2 +2x3 , x1 +x2 +x3 , x1 +x2 +x3 , x1 +x2 +x3 ,
(b) 0, 0, 0
(9) (a) 1, 1
(b) 2, 3
(c) plane, 1, −2
(11) M is not a subspace of R3 .
(13) M + N is a plane, not all of R3 .
 CHAPTER 7

LINEAR INDEPENDENCE
independence
7.1. Background
Topics: linear combinations, span, linear dependence and independence.

7.1.1. Remark. Some authors of linear algebra texts make it appear as if the terms linear de-
pendence and linear independence, span, and basis pertain only to finite sets of vectors. This is
extremely misleading. The expressions should make sense for arbitrary sets. In particular, do not
be misled into believing that a basis for a vector space must be a finite set of vectors (or a sequence
of vectors). While it is true that in most elementary linear algebra courses the emphasis is on
the study of finite dimensional vector spaces, bases for vector spaces may be very large indeed. I
recommend the following definitions.
7.1.2. Definition. Recall that a vector P y is a linear combination of vectors x1 , . . . , xn if there
αn such that y = nk=1 αk xk . Note: linear combinations are finite sums. The
exist scalars α1 , . . . P
linear combination nk=1 αk xk is trivial if all the coefficients α1 , . . . αn are zero. If at least one
αk is different from zero, the linear combination is nontrivial.
7.1.3. Example. In R2 the vector (8, 2) is a linear combination of the vectors (1, 1) and (1, −1)
because (8, 2) = 5(1, 1) + 3(1, −1).
7.1.4. Example. In R3 the vector (1, 2, 3) is not a linear combination of the vectors (1, 1, 0) and
(1, −1, 0).
7.1.5. Definition. Suppose that A is a subset (finite or not) of a vector space V . The span of
A is the set of all linear combinations of elements of A. Another way of saying the same thing:
the span of A is the smallest subspace of V which contains A. (That these characterizations are
sbs032p
equivalent is not completely obvious. Proof is required. See problem ??. We denote the span of A
by span A. If U = span A, we say that A spans U or that U is spanned by A.
7.1.6. Example. For each n = 0, 1, 2, . . . define a function pn on R by pn (x) = xn . Let P be the
set of polynomial functions on R. It is a subspace of the vector space of continuous functions on R.
Then P = span{p0 , p1 , p2 . . . }. The exponential function exp, whose value at x is ex , is not in the
span of the set {p0 , p1 , p2 . . . }.
7.1.7. Definition. A subset A (finite or not) of a vector space is linearly dependent if the zero
vector 0 can be written as a nontrivial linear combination of elements of PA; that is, if there exist
vectors x1 , . . . , xn ∈ A and scalars α1 , . . . , αn , not all zero, such that nk=1 αk xk = 0. A subset
of a vector space is linearly independent if it is not linearly dependent.
Technically, it is a set of vectors that is linearly dependent or independent. Nevertheless, these
terms are frequently used as if they were properties of the vectors themselves. For instance, if
S = {x1 , . . . , xn } is a finite set of vectors in a vector space, you may see the assertions “the set S is
linearly independent” and “the vectors x1 , . . . xn are linearly independent” used interchangeably.
7.1.8. Example. The (vectors going from the √

origin to) points
√

on the unit circle in R2 are linearly
dependent. Reason: If x = (1, 0), y = − 21 , 23 , and z = 12 , 23 , then x + y + (−1)z = 0.
41
42 7. LINEAR INDEPENDENCE

7.1.9. Example. For each n = 0, 1, 2, . . . define a function pn on R by pn (x) = xn . Then the set
{p0 , p1 , p2 , . . . } is a linearly independent subset of the vector space of continuous functions on R.
 7.2. EXERCISES 43

7.2. Exercises
(1) Show that in the space R3 the vectors x = (1, 1, 0), y = (0, 1, 2), and z = (3, 1, −4) are
linearly dependent by finding scalars α and β such that αx + βy + z = 0.
Answer: α = ,β= .

(2) Let w = (1, 1, 0, 0), x = (1, 0, 1, 0), y = (0, 0, 1, 1), and z = (0, 1, 0, 1).
(a) We can show that {w, x, y, z} is not a spanning set for R4 by finding a vector u in
R4 such that u ∈
/ span{w, x, y, z}. One such vector is u = (1, 2, 3, a) where a is any
number except .
(b) Show that {w, x, y, z} is a linearly dependent set of vectors by finding scalars α, γ,
and δ such that αw + x + γy + δz = 0.
Answer: α = ,γ= ,δ= .
(c) Show that {w, x, y, z} is a linearly dependent set by writing z as a linear combination
of w, x, and y. Answer: z = w+ x+ y.

(3) Let p(x) = x2 + 2x − 3, q(x) = 2x2 − 3x + 4, and r(x) = ax2 − 1. The set {p, q, r} is
linearly dependent if a = .

(4) Show that in the vector space R3 the vectors x = (1, 2, −1), y = (3, 1, 1), and z = (5, −5, 7)
are linearly dependent by finding scalars α and β such that αx + βy + z = 0.
Answer: α = ,β= .

(5) Let f1 (x) = sin x, f2 (x) = cos(x+π/6), and f3 (x) = sin(x−π/4) for 0 ≤ x ≤ 2π. Show that
{f1 , f2 , f3 } is linearly dependent by finding constants α and β such that αf1 −2f2 −βf3 = 0.
Answer: α = and β = .

(6) In the space C[0, π] let f , g, h, and j be the vectors defined by

f (x) = 1
g(x) = x
h(x) = cos x
x
j(x) = cos2
2
for 0 ≤ x ≤ π. Show that f , g, h, and j are linearly dependent by writing j as a linear
combination of f , g, and h.
Answer: j = f + g+ h.

(7) Let u = (λ, 1, 0), v = (1, λ, 1), and w = (0, 1, λ). Find all values of λ which make {u, v, w}
a linearly dependent subset of R3 . Answer:

(8) Let u = (1, 0, −2), v = (1, 2, λ), and w = (2, 1, −1). Find all values of λ which make
{u, v, w} a linearly dependent subset of R3 . Answer:

(9) Let p(x) = x3 − x2 + 2x + 3, q(x) = 3x3 + x2 − x − 1, r(x) = x3 + 2x + 2, and s(x) =

7x3 + ax2 + 5. The set {p, q, r, s} is linearly dependent if a = .
44 7. LINEAR INDEPENDENCE

(10) In the space C[0, π] define the vectors f , g, and h by

f (x) = x
g(x) = sin x
h(x) = cos x
for 0 ≤ x ≤ π. We show that f , g, and h are linearly independent. This is accomplished
by showing that if αf + βg + γh = 0, then α = β = γ = 0. So we start by supposing that
αf + βg + γh = 0; that is,
αx + β sin x + γ cos x = 0 (1)
for all x ∈ [0, π].
(a) We see that γ must be zero by setting x = in equation (1).
Now differentiate (1) to obtain
α + β cos x = 0 (2)
for all x ∈ [0, π].
(b) We see that α must be zero by setting x = in equation (2).
Differentiate (2) to obtain
−β sin x = 0 (3)
for all x ∈ [0, π].
(c) We conclude that β = 0 by setting x = in (3).
 7.3. PROBLEMS 45

7.3. Problems
(1) In the space C[0, 1] define the vectors f , g, and h by

f (x) = x
g(x) = ex
h(x) = e−x

for 0 ≤ x ≤ 1. Use the definition of linear independence to show that the functions f , g,
and h are linearly independent.
(2) Let a, b, and c be distinct real numbers. Use the definition of linear independence to give
a careful proof that the vectors (1, 1, 1), (a, b, c), and (a2 , b2 , c2 ) form a linearly independent
subset of R3 .
(3) Let {u, v, w} be a linearly independent set in a vector space V . Use the definition of
linear independence to give a careful proof that the set {u + v, u + w, v + w} is linearly
independent in V .
(4) You are the leader of an engineering group in the company you work for and have a
routine computation that has to be done repeatedly. At your disposal is an intern, Kim,
a beginning high school student, who is bright but has had no advanced mathematics. In
particular, Kim knows nothing about vectors or matrices.
Here is the computation that is needed. Three vectors, a, b, and c are specified
in R5 . (Denote their span by M .) Also specified is a (sometimes long) list of other vectors
S = {v1 , v2 , . . . , vn } in R5 . The problem is to
(1) determine which of the vectors in S belong to M , and
(2) for each vector vk ∈ S which does belong to M
find constants α, β, and γ such that vk = αa + βb + γc.
Kim has access to Computer Algebra System (Maple, or a similar program) with a Lin-
ear Algebra package. Write a simple and efficient algorithm (that is, a set of instructions)
which will allow Kim to carry out the desired computation repeatedly. The algorithm
should be simple in the sense that it uses only the most basic linear algebra commands
(for example, Matrix, Vector, Transpose, RowReducedEchelonForm, etc. in Maple). Re-
member, you must tell Kim everything: how to set up the appropriate matrices, what
operations to perform on them, and how to interpret the results. The algorithm should
be as efficient as you can make it. For example, it would certainly not be efficient for Kim
to retype the coordinates of a, b, and c for each new vk .
Include in your write-up an actual printout showing how your algorithm works in some
special case of your own invention. (For this example, the set S need contain only 5 or 6
vectors, some in U , some not.)

(5) The point of this problem is not just to get a correct answer to (a)–(c) below using tools
you may have learned elsewhere, but to give a careful explanation of how to apply the
linear algebra techniques you have already encountered to solve this problem in a systematic
fashion. For background you may wish to read a bit about networks
Hefferon:2006
and Kirchhoff ’s laws
AntonR:2000
(see, for example, [8] Topic: Analyzing Networks, pages 72–77 or [1] Electrical Networks,
pages 538–542).
Consider an electrical network having four nodes A, B, C, and D connected by six
branches 1, . . . , 6. Branch 1 connects A and B; branch 2 connects B and D; branch 3
connects C and B; branch 4 connects C and D; branch 5 connects A and C; and branch 6
connects A and D.
46 7. LINEAR INDEPENDENCE

The current in branch k is Ik , where k = 1, . . . , 6. There is a 17 volt battery in

branch 1 producing the current I1 which flows from A to B. In branches 2, 4, and 5 there
are 0.5 ohm resistors; and in branches 1, 3, and 6 there are 1 ohm resistors.
(a) Find the current in each branch. (Explain any minus signs which occur in your
answer.)
(b) Find the voltage drop across each branch.
(c) Let pn be the potential at node n = A, B, C, D. The voltage drop across the branch
connecting node j to node k is the difference in the potentials at nodes j and k.
Suppose the network is grounded at D (so that pD = 0). Find the potential at the
other nodes.
 7.4. ANSWERS TO ODD-NUMBERED EXERCISES 47

7.4. Answers to Odd-Numbered Exercises

(1) −3, 2
(3) 7
√ √
(5) 3 − 1, 6
√ √
(7) − 2, 0, 2
(9) −3
 CHAPTER 8

BASIS OF A VECTOR SPACE

bases
8.1. Background
Topics: basis, dimension.

8.1.1. Definition. A set B (finite or not) of vectors in a vector space V is a basis for V if it is
linearly independent and spans V .
8.1.2. Example. The vectors e1 = (1, 0, 0), e2 = (0, 1, 0), and e3 = (0, 0, 1) constitute a basis for
the vector space R3 .
8.1.3. Example. For each n = 0, 1, 2, . . . define a function pn on R by pn (x) = xn . Then the set
{p0 , p1 , p2 , . . . } is a basis for the vector space P of polynomial functions on R.
Two important facts of linear algebra are that regardless of the size of the space every vector
space has a basis and that every subspace has a complement.
8.1.4. Theorem. Let B be a linearly independent set of vectors in a vector space V . Then there
exists a set C of vectors in V such that B ∪ C is a basis for V .

8.1.5. Corollary. Every vector space has a basis.

8.1.6. Corollary. Let V be a vector space. If M  V , then there exists N  V such that
M ⊕N =V.

The next theorem says that any two bases for a vector space are the same size.
8.1.7. Theorem. If B and C are bases for the same vector space, then there is a one-to-one
correspondence from B onto C.

8.1.8. Definition. A vector space V is finite dimensional if it has a finite basis. Its dimension
(denoted by dim V ) is the number of elements in the basis. If V does not have a finite basis it is
infinite dimensional.

8.1.9. Theorem. If M and N are subspaces of a finite dimensional vector space, then
dim(M + N ) = dim M + dim N − dim(M ∩ N ) .

49
50 8. BASIS OF A VECTOR SPACE

8.2. Exercises
(1) Let u = (2, 0, −1), v = (3, 1, 0), and w = (1, −1, c) where c ∈ R. The set {u, v, w} is
linearly independent in R3 provided that c is not equal to .

(2) Let M2×2 be the vector space of all 2 × 2 matrices of real numbers.
(a) The dimension of M2×2 is .
(b) A basis for M2×2 is

(3) Let T2 be the vector space of all 2 × 2 matrices of real numbers with zero trace.
(a) The dimension of T2 is .
(b) A basis for T2 is

(4) Let S3 be the vector space of all symmetric 3 × 3 matrices of real numbers.
(a) The dimension of S3 is .
(b) A basis for S3 is

(5) In R4 let M be the subspace spanned by the vectors (1, 1, 1, 0) and (0, −4, 1, 5) and let
N be the subspace spanned by (0, −2, 1, 2) and (1, −1, 1, 3). One vector which belongs to
both M and N is (1, , , ). The dimension of M ∩ N is and the
dimension of M + N is .
 
u −u − x
(6) Let U be the set of all matrices of real numbers of the form and V be the
0 x
 
v 0
set of all real matrices of the form . Find bases for U, V, U + V, and U ∩ V.
w −v

Answer: Basis for U: Basis forV:

Basis for U + V: Basis for U ∩ V:

 8.2. EXERCISES 51

(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
52 8. BASIS OF A VECTOR SPACE

8.3. Problems

(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
 8.4. ANSWERS TO ODD-NUMBERED EXERCISES 53

8.4. Answers to Odd-Numbered Exercises

(1)
(3)
(5)
(7)
(9)
(11)
(13)
(15)
(17)
(19)
(21)
(23)
(25)
(27)
(29)
 Bibliography

1. Howard Anton and Chris Rorres, Elementary Linear Algebra: Applications Version, eighth ed., John Wiley and
Sons, New York, 2000. v, 45
2. Robert A. Beezer, A First Course in Linear Algebra, 2004,
http://linear.ups.edu/download/fcla-electric-2.00.pdf. v
3. Przemyslaw Bogacki, Linear Algebra Toolkit, 2005, http://www.math.odu.edu/~bogacki/lat. v
4. William C. Brown, A second Course in Linear Algebra, John Wiley, New York, 1988. v
5. Charles W. Curtis, Linear Algebra: An Introductory Approach, Springer, New York, 1984. v
6. Paul Dawkins, Linear Algebra, 2007, http://tutorial.math.lamar.edu/pdf/LinAlg/LinAlg_Complete.pdf. v
7. Paul R. Halmos, Finite-Dimensional Vector Spaces, D. Van Nostrand, Princeton, 1958. v
8. Jim Hefferon, Linear Algebra, 2006, http://joshua.smcvt.edu/linearalgebra. v, 45
9. Kenneth Hoffman and Ray Kunze, Linear Algebra, second ed., Prentice Hall, Englewood Cliffs,N.J., 1971. v
10. Steven Roman, Advanced Linear Algebra, second ed., Springer-Verlag, New York, 2005. v
11. Gilbert Strang, Linear Algebra and its Applications, second ed., Academic Press, New York, 1980. v
12. Eric W. Weisstein, MathWorld, A Wolfram Web Resource, http://mathworld.wolfram.com. v
13. Wikipedia, Wikipedia, The Free Encyclopedia, http://en.wikipedia.org. v

55
 Index

](x, y) (angle between x and y), 21 Laplace expansion, 14

At (transpose of A), 7 left
M  V (M is a subspace of V ), 33 inverse, 10
hx, yi (inner product of x and y), 21 length, 21
x · y (inner product of x and y), 21 Lie bracket, 7
M ⊕ N (direct sum of M and N ), 33 linear
combination, 41
action trivial, 41
of a matrix, 13 dependence, 41
additive independence, 41
inverses, 27
angle, 21 matrix
associative, 27 nonsingular, 27
symmetric, 7
basis, 49 transpose of a, 7
bounded function, 33 minor, 14

Cauchy-Schwarz inequality, 21 nonsingular, 27

cofactor, 14 norm, 21
combination
linear, 41 pivoting, 1
commutative, 27
commutator, 7 Riemann integrable function, 33
commute, 7 right
complementary subspaces, 33 inverse, 10
row
dependence equivalent matrices, 13
linear, 41 operations, 1
determinant, 14
det A (determinant of A), 14 scaling, 1
dimension, 49 Schwarz inequality, 21
dim V (dimension of V ), 49 space
direct sum, 33 vector, 27
span, 41
equivalent span A (the span of the set A), 41
row, 13 subspace, 33
expansion subspaces
Laplace, 14 complementary, 33
sum
finite dimension, 49 direct, 33
swapping, 1
identity symmetric
additive, 27 matrix, 7
independence
linear, 41 trace, 7
infinite dimensional, 49 tr A (trace of A), 7
integrable, 33 transpose
inverse of a matrix, 7
additive, 27 trivial

57
58 INDEX

linear combination, 41

vector
space, 27
complex, 27
real, 27