Study Pack MA100 Inside All

MA100 Mathematical Methods
General Information for 2013/14

Calculus Lecturer:
Dr. Ioannis Kouletsis
Linear Algebra Lecturer:
Dr Michele Harvey
Mathematics Departmental Oce:
oce: COL 2.05C

oce: COL 4.15
COL 4.01
Course Materials
Study Pack
That is what you are holding now, and formed of two parts: Directed Reading and Notes,
and the Exercise Sets. All this material, and more, including copies of the lecture slides,
will be made available on the MA100 Moodle page as well.
The lectures will broadly follow the material as presented in the notes.
Course Texts
All students are expected to have a copy of these books.
references to the relevant parts of the textbooks.
The notes contain detailed
Calculus: K. Binmore and J. Davies, Calculus, Concepts and Methods.

Linear Algebra: M. Anthony and M. Harvey, Linear Algebra, Concepts and Methods.
Reference Texts (for additional reading and exercises; find the texts in the library):
H. Anton, Elementary Linear Algebra (or Applications version by Anton and Rorres);
S. L. Salas and E. Hille, Calculus, One and Several Variables;
Schaum Outline Series: Mathematics for Economists, Linear Algebra, Advanced Calculus,
Dierential and Integral Calculus.
Teaching Arrangements and What You Need To Do!

Go to the lectures.
There is one Calculus lecture (Tuesday) and one Linear Algebra lecture (Friday) each week.
Links to the video recordings of the lectures will be published on the MA100 Moodle page,
but these videos are not a substitute for attending the lectures.
Read the texts.
Each week you should do the reading as directed in the study pack for that lecture, in
conjunction with any notes you have taken.
Work all exercises to the best of your ability before each class.
There is one set of exercises for you to complete each week. Start working the questions
as soon as the material is covered in a lecture. (You can start on questions 1, 2, 3, 9
of Exercise Set 1 immediately after the first lecture.) Starred and un-starred questions
are equally important. (The only dierence is that starred ones are marked by your class
teacher.) Also look at the reading and its connection to the exercises.
There are answers (but not full solutions) to the un-starred questions on the Moodle page so
you can check your work before class. Questions appearing with the symbol are optional
extras and will not be discussed in class unless time permits.
Solutions to all questions will be put on the MA100 Moodle page after the Exercise Set
has been discussed in the classes. Use these to check your work each week, as soon as your
solutions have been returned to you. Note that often other solutions may be possible.
Go to all classes.
Class attendance is obligatory. Each week complete the Exercise Set for that week. Prepare
and hand in solutions of the starred questions to your class teacher, either before or at
the next weeks class depending on your class teachers instructions. Your solutions will be
marked and returned to you at the following class.
The classes will be used to discuss the exercises and any questions you may have on the
course material.
See an MA100 class teacher or lecturer in their oce hours as soon as you have
diculty doing or understanding any part of the material: lectures, reading or exercises.
You can see any MA100 class teacher or lecturer during their Oce Hours. Oce hour times
can be found via the MA100 Moodle page and on the Mathematics Department website:
www2.lse.ac.uk/maths/Courses/Office Hours.aspx.
Attend Extra Example Sessions (optional).
The Extra Examples Sessions are optional interactive sessions designed to reinforce the
material covered in the lectures. There will be no new material presented in these sessions.
They will be held each week (starting Week 1): Thursday, 11.00-12.00, in the Old Theatre.
Maple.
Maple is a computer algebra program which will be used throughout the course for
visualising graphs and solutions to certain exercises, and to enhance your understanding
of the material.
Before the end of Week 3, work through the Maple tutorial found on the MA100 Moodle
page. If Maple doesnt start immediately when clicking on the file name, save the file, and
open it inside Maple.
Classes during Week 5 (only) will be held in one of the computer rooms (the room will be
listed on your LSEforYou timetable), where you will use Maple to work parts of Exercise
Set 4 using the Maple Instructions in your Study Pack.
A few Words of Advice

Each weeks material builds on the previous weeks work. Success in this course depends on
working the exercises and understanding the material covered week by week.
Do not fall behind!

If after the classes you still have questions or diculties with any of the material,
see a class teacher or one of the lecturers immediately!
MA100 Mathematical Methods

Background You Should Know and Exercises
This is background material for MA100 and is intended to be a review of your A-level
Mathematics course. Please work through it before term or in your spare time. (You do
not need and should not use a calculator.) If you have difficulties with anything, dont
panic but do make a note of any background you are missing and work on it, or see one
of the lecturers.
It is also recommended that you look through the review booklets, An Algebra Refresher
and A Calculus Refresher. These can be found at: www.maths.lse.ac.uk/Refreshers/.
References are made to these booklets (AlgR and CalcR) for corresponding material in the
exercises that follow.
Solutions to the exercises can be found on the MA100 Moodle page at the start of term.
1. The set of real numbers, denoted R, includes the following subsets:

N, the set of natural numbers: 1, 2, 3, 4, . . .
Z, the set of integers: . . . , 3, 2, 1, 0, 1, 2, 3, . . .
Q, the set of rational numbers: p/q with p, q Z, q 6= 0; such as, 25 , 29 , 41 =4.
the set of irrational numbers: real numbers which are not rational; for example, 2, .

a
if a 0
The absolute value (modulus) of a real number: |a| =
.
a if a 0
The absolute value satisfies |a| = a2 , and the Triangle Inequality: |a + b| |a| + |b|.
Intervals of real numbers, such as:
2. Polynomials of degree n:
How to graph:
(a, b) = {x | a < x < b}, [a, b] = {x | a x b},

(, b) = {x | x < b}
[a, ) = {x | a x}.
Pn (x) = a0 + a1 x + a2 x2 + . . . + an xn ,
where the ai are real constants, an 6= 0, and x is a real variable.
straight lines, P1 (x) = a0 + a1 x

quadratics,
P2 (x) = a0 + a1 x + a2 x2
cubics,
P3 (x) = a0 + a1 x + a2 x2 + a3 x3 .
The laws of indices:
xr xs = xr+s (xr )s = xrs
How to expand expressions like (1 + x)n , (x + y)n
How to factorize quadratics (such as x2 5x + 6 and 3x2 + 14x 5)
and simple cubics, (such as x3 1 and x3 2x2 5x + 6).
The Quadratic Formula.
How to sum: arithmetic series, a + (a + d) + ... + (a + (n 1)d)
geometric series,
a + ar + ar2 + ... + arn1
Exercises.
(AlgR, sections 23, 810)
(2.1) Sketch f (x) = 4 x2 and g(x) = 2x + 1 and find their points of intersection.
3
(2.2) Expand (2 + 5x)4 and x2 4/x .
(2.3) Find an expression for the profit function (q) as a function of q given that
(q) = pq C(q),
where
2p + 0.4q = 155 and
C(q) = q 2 10q + 300 .
(2.4) Find the complete solution of each of the following systems of equations, both
algebraically (by solving the equations simultaneously) and graphically (by illustrating
them as lines on a graph).

x + 2y = 4
x + 2y = 4
x + 2y = 4
(a)
(b)
(c)
2x y = 3
2x + 4y = 4
2x + 4y = 8
Note. (x, y) is a solution of a system of equations if it satisfies all equations simultaneously.
(2.5) Factorise P (x) = x3 7x + 6 and sketch the graph of P (x).
3. Be able to manipulate algebraic expressions accurately and efficiently.

Be able to manipulate indices.
Exercises.
(AlgR, sections 17)

a
(a) 6ab (b2 4bc)
b
(3.1) Simplify
(3.2) Solve for s:
(b)
49x2
4xy 2
35y
(2xy)3
5
2
=0.
3s + 1 s + 1
(3.3) Rewrite the simultaneous equations

x=
a c by
2b
y=
a c f bx
2b
as a system of linear equations in x and y. Solve the system for x and y (in terms of the
constants a, b, c, f ), and then show that
x+y =
1
(2a 2c f ) .
3b
4. Properties of the exponential (ex ) and logarithmic (ln x = loge x) functions and their
graphs:
ln xr = r ln x
ln ex = x
(er )s = ers
eln y = y
er+s = er es
ln(xy) = ln x + ln y
Exercises.
(4.1) Sketch the graphs of ex , ex , ln x .
(4.2) What is the value of e0 , ln 0, ln 1 ?
a
(4.3) Show that: ln
= ln a ln b (a > 0, b > 0).
b
(4.4) Simplify: ln x5 2 ln x + ln y 3 (x > 0, y > 0).
For which values of x and y is this expression positive?
5. Properties of the trigonometric functions (sin, tan, etc.) using radian measure of an angle:
Be able to sketch their graphs.
Know their values at multiples of , 2 , 4 , 3 , 6 .
(Know the triangles associated with these last values.)
Know the basic trigonometric identities, e.g.
sin2 x + cos2 x = 1
sin(a + b) = sin a cos b + cos a sin b ;
cos(a + b) = cos a cos b sin a sin b ;
sin x = sin(x) is an odd function;
cos x = cos(x) is an even function.
Exercises.
(5.1) Sketch the graphs of sin x, cos x, 2 cos 3x, tan x .
(5.2) Deduce the formulas for sin 2x, cos 2x from the identities above, and deduce that
sec2 x = tan2 x + 1 ;
sin x cos y =
tan 2x =
2 tan x
;
1 tan2 x
1
[sin(x + y) + sin(x y)] .
2
(5.3) Evaluate (without using a calculator):
11
sec 3 , cot( 6 ), sin( 5
4 ), cos( 4 ).
6. Rules of differentiation: Product, quotient and chain rules

Be able to differentiate polynomials.
Know the derivatives of the exponential, logarithmic and trigonometric functions.
Exercises.
(CalcR, sections 19)
(6.1) Calculate
d 3x
e
(a)
dx
d
(e)
dx
d
(b)
ln(5 2x)
dx
xex
2x2 + 1
(f)

d 3x
(c)
e
ln(5 2x)
dx
d p 2
3x 1
dx
(g)
d
tan 3x
dx
(h)
d
(d)
ln
dx
1
x2
d
sin(x2 5)
dx
7. Know that the equation of the tangent line to the curve y = f (x) at the point where
x = a is given by:
y f (a) = f (x)(x a)
Exercise.
(7.1) Find the equation of the tangent line to the curve y = 2x3 9x2 38x + 21
at the point where x = 1.
8. Find the stationary (turning) points of graphs of functions and determine which point is
a maximum and which is a minimum.
Exercises.
(8.1) Do this for the function y = 2x3 9x2 38x + 21.
(8.2) Sketch the graph of f (x) = x3 3x2 .
Find the maximum and the minimum value of f on the interval [0, 3],
and on the interval [0, 4].
9. Techniques of integration: partial fractions, integration by parts, substitution.

Exercise.
(CalcR, sections 10-17; AlgR, section 11)
(9.1) Calculate
(a)
sin 3x dx
(d)
(g)
(b)
x
dx
x1
Z
sec x dx
(c)
(e)
x(x 2) dx
x cos 2x dx
(h)
1
dx
x2 5x + 6
Z
(f)
(2x + 2)ex
x2 ex dx
+2x+3
dx
10. The technique of Completing the Square in a quadratic expression.

Exercise.
(AlgR, section 9)
(10.1) Complete the squares in the expressions: x2 + 6x + 11 and 2x2 4x + 7.

(10.2) Complete the square in ax2 + bx + c.
Use this to derive the quadratic formula to solve ax2 + bx + c = 0 for x.
MA 100
Mathematical Methods
Calculus
Directed Reading and Notes
Textbook
Ken Binmore and Joan Davies, Calculus, Concepts and Methods
Cambridge University Press, 2001. ISBN 978-0-521-77541-0
All students are expected to have a copy of this book.
The notes for this part of MA 100 consist mainly of detailed references to the
relevant parts of the textbook. Reading those parts is essential for understanding
the material in the course.
2013/14
c LSE 2013

Course Content CALCULUS

1
Vectors and general geometric concepts

Lines, planes, hyperplanes; their different types of equations. Flats. Curves and conics.
Scalar valued functions

One-variable functions f : R R. Constrained and unconstrained optimisation.
Inverse and local inverse functions.
Two-variable functions f : R2 R. Partial derivatives, tangent planes, gradients,
derivatives, and directional derivatives. Examples including linear and quadric surfaces.
Functions of several variables f : R n R. Gradients and derivatives, tangent hyperplanes. Taylor approximation. Convex and concave functions.
The Chain Rule for differentiation.
Stationary points and optimisation of functions of several variables

Calculation and classification of stationary points, including using the quadratic term
in the Taylor series expansion of functions. Local and global extrema.
Lagranges method for constrained optimisation.
Vector-valued functions
Functions f : R n R m . Derivatives and tangent flats.
Inverse functions and their derivatives. Local inverses and critical points.
Summation and integration

Methods for summation of series.
Techniques of integration.
Differential and difference equations

First-order differential equations. Solution methods for differential equations of the
following types : separable, exact, linear, homogeneous. Use of substitution.
Linear differential and difference equations with constant coefficients. Nonhomogeneous linear differential and difference equations.
General and particular solutions. Asymptotic behaviour of solutions.
Applications
Applications of all the topics above.
MA 100 Mathematical Methods Calculus
( By something like 1.4.1, we mean Section 1.4, Question 1, of the textbook. )
Lines : 1.8.1, 1.8.2, 1.8.4, 1.8.5, 1.8.6, 1.8.11, 1.8.12, 1.8.13, 1.8.14 and 1.8.15.
Combinations of functions (pages 55-56)
(If you have any difficulty with the reading, attend office hours.)
Vectors in R3 : 1.8.7 and 1.8.9;
Some elementary functions (pages 52-55)
Vectors in R2 : 1.4.1, 1.4.2 ( except part (vi) ), 1.4.3 and 1.4.4;
1.4 Optional extra exercises related to these topics
For this we follow Section 1.6 of the textbook.
1.3 Lines
1.2 Vectors in R3
For this we follow Section 1.3 of the textbook, until line 8 of page 19.
1.1 Vectors in R2
Vectors and Lines
Real valued functions of one real variable (pages 5152)
Intervals (page 51)
To begin, you should look through the first sections of Chapter 2:
working through the examples.
Read each section carefully,
Section and page number references refer to the calculus textbook.
Lecture 1
Curves and Conics
Planes
R2
2.15.7 and 2.15.10.
1.8.16, 1.8.17, 1.8.18, 1.8.19, 1.8.21, 1.8.22, 1.8.23, 1.8.24, 1.8.25 and 1.8.26;
Curves and conics :
Planes :
The equation
+ a y = b represents a parabola.
x 2 y2
The equation 2 + 2 = 1 represents an ellipse.
a
b
An ellipse with a = b can be rewritten as x2 + y2 = a2 , which is the equation of a circle.
y2
x2
An hyperbola is represented by the equation 2 2 = 1. Hyperbolas have two
a
b
asymptotes ( sometimes also called slant asymptotes ); see page 80 of the textbook.
x2
For conics have a look at the pictures in Section 2.14 of the textbook.
E XAMPLES
A line is represented by the equation a x + b y + c = 0, for some real numbers a, b, c.
The circle with centre (0, 0) and radius 1 is the set of points satisfying x2 + y2 1 = 0
( which we usually write as x2 + y2 = 1 ).
A curve is the set of points ( x, y)

satisfying an equation P( x, y) = 0. The equation
P( x, y) = 0 is said to represent the curve.
2.2 Curves and conics
For this we follow Section 1.7 of the textbook, until line 12 of page 31. You can skip
Example 8 on pages 29-30.
2.1 Planes
Lecture 2
3.2 Functions of one variable
2.11.1, 2.11.2, 2.11.9, 2.11.11, 2.11.13 and 2.11.15;
Stationary points and optimisation of one-variable functions : 4.5.4 and 4.5.5.
Taylor Series of one-variable functions : 2.15.1, 2.15.2, 2.15.4 and 2.15.5;
Derivatives :
3.6 Stationary points and optimisation of one-variable functions
3.5 Taylor series of one-variable functions
3.4 Higher order derivatives
For this we follow Sections 2.7, 2.8 and 2.10 of the textbook. The rules of differentiation
on page 67 are very important.
3.3 Derivatives
Taylor Series
Derivatives
Functions of One Variable
Stationary Points and Optimisation
Read Section 2.1 of the textbook yourself.
3.1 Intervals
Lecture 3
4.5.17;
3.7.1.
Functions of two variables :
Convex and concave functions :
Global and constrained optimisation of one-variable functions :

4.5.1, 4.5.2, 4.5.6, 4.5.8, 4.5.10, 4.5.11 and 4.5.12;
4.3 Functions of two variables
For this we follow page 126 and 127 of Section 4.1 of the textbook.
4.2 Convex and concave functions
For this we follow Sections 4.2 and 4.3 of the textbook.
f
f
+y
= n f ( x, y).
x
y
f
f
+y
= x (2 x + 2 y) + y (2 x ) = 2 ( x2 + 2 x y) = 2 f ( x, y).
x
y
Homogeneous functions and Eulers formula : 5.8.14;
Tangent planes : 3.7.5, 3.7.6 and 3.7.10.
Partial derivatives : 3.7.3 and 3.7.4;
E XAMPLE
For f ( x, y) = x2 + 2 x y we have
A homogeneous function f ( x, y) of degree n satisfies Eulers Formula :
If f ( x, y) is homogeneous of degree n > 1, then f has increasing returns to scale. If

f ( x, y) is homogeneous of degree n < 1, then f has decreasing returns to scale.
E CONOMIC INTERPRETATION
If x and y are thought of as inputs and f ( x, y) as the output, then a homogeneous function
of degree 1 has constant returns to scale : if we multiply all inputs by , then the output
is multiplied by .
E XAMPLE
The function f ( x, y) = x2 + 2 x y is homogeneous of degree 2, since f ( x, y) =
( x )2 + 2 ( x ) ( y) = 2 x2 + 2 2 x y = 2 ( x2 + 2 x y) = 2 f ( x, y).
f ( x, y) = n f ( x, y).
A function f : R2 R is called homogeneous of degree n if for all x, y and :
5.2 Homogeneous functions and Eulers formula
5.1 Partial derivatives and tangent planes
Homogeneous Functions
Functions of Two Variables
Partial Derivatives
Tangent Planes
Lecture 5
Convex and Concave Functions
Global and Constrained Optimisation
4.1 Global and constrained optimisation of one-variable functions
Lecture 4
3.7.7, 3.7.8, 3.7.9, 3.7.11 and 3.7.12.
Directional derivative :
Gradient and derivative of a function of two variables :
6.2 Directional derivative
3.7.2;
Directional Derivatives
Gradient and Derivative
6.1 Gradient and derivative of a function of two variables
Lecture 6
Vectors in n Dimensions
Stationary points : 4.9.1, 4.9.2, 4.9.3, 4.9.4, 4.9.5 and 4.9.6.
Functions of more than 2 variables; tangent hyperplanes; directional derivatives :

3.9.1, 3.9.2, 3.9.3, 3.9.4, 3.9.5, 3.9.6, 3.9.7, 3.9.8, 3.9.9, 3.9.10 and 3.9.12;
Flats in R n : 1.11.8, 1.11.9 and 1.11.10;
Vectors, lines and hyperplanes in R n : 1.11.1, 1.11.2, 1.11.3, 1.11.4 and 1.11.6;
For this we follow Sections 4.6, 4.7 and 4.8 of the textbook.
7.4 Stationary points
For this we follow Section 3.8 of the textbook ( including Subsections 3.8.1 and 3.8.2 ).
7.3 Functions of more than 2 variables; tangent hyperplanes; directional derivatives
7.2 Flats in R n
Stationary Points
Functions of more than 2 Variables
Lines, Hyperplanes and Flats
7.1 Vectors, lines and hyperplanes in R n
Lecture 7
Derivatives and tangent flats of vector functions :
Vector functions : 5.8.1, 5.8.2 and 5.8.3;
y f 2 ( ) = f 2 ( ) (t ).
5.8.4 and 5.8.5.
In other words, this tangent line is the intersection of the two planes in R3 :
(
x f 1 ( ) = f 1 ( ) (t ),
And the tangent flat of such a function at t = is a line

f 1 ( )
f 1 ( )
x
=
( t ).
y
f 2 ( )
f 2 ( )
The graph of a function f : R R2 is a 1-dimensional surface in R3 ; in other words :

a curve in R3 .

f 1 (t)
x
= f (t) =
We can write such a function as
.
y
f 2 (t)
The only exceptions are the cases n = m = 1; n = 1, m = 2; and n = 2, m = 1.
The graph of a vector function f : R n R m is some kind of n-dimensional surface

in R m+n . So in general there is no hope we can draw it !
8.4 Graphs of vector functions
8.3 Derivatives and tangent flats of vector functions
8.2 Affine functions
R : 5.8.10, 5.8.11 and 5.8.12.
5.8.8, 5.8.9 and 5.8.13;
Taylor series for functions f :
Rn
The chain rule for vector functions :
9.3 Taylor series for functions f : R n R
9.2 Second derivatives
9.1 The chain rule for vector functions
Taylor Series for Functions f : Rn R
Chain Rule for Vector Functions
Derivatives and Tangent Flats
Lecture 9
Second Order Derivatives of Vector Functions
Vector Functions
Affine Functions
8.1 Vector functions ( = vector valued functions )
Lecture 8
Classification of Stationary Points
Optimisation of Functions
f : Rn R
= 0, since we are looking at a stationary point ).
So now we need a good method to decide in which of the three cases we are. In other
words, we want a method to classify stationary points.
then is a saddle point.
if for some u : u f ( ) u > 0, and for some v: v f ( ) v < 0 :
so is a local maximum;
then for all u 6= 0 small : f ( + u) < f ( ),
if for all u 6= 0 we have u T f ( ) u < 0 :
so is a local minimum;
then for all u 6= 0 small : f ( + u) > f ( ),
if for all u 6= 0 we have u T f ( ) u > 0 :
More precisely, we have :
And since we want to know what the difference between f ( ) and f ( + u) for small u
is, it all seems to be decided by the term 12 u T f ( ) u.
f ( + u) f ( ) + 12 u T f ( ) u.
We are interested what happens close to , so lets write x = + u. This gives
( here we use that
f ( )
= f ( ) + 0 + 12 ( x )T f ( ) ( x ),
f ( x ) f ( ) + f ( ) ( x ) + 12 ( x )T f ( ) ( x )
To decide if a stationary point is a local maximum, a local minimum, or a saddle point,

we study the second order Taylor polynomial around :
But not every stationary point is a local maximum or local minimum. It can be a saddle
point as well.
A function f : R n R can have a local maximum or a local minimum in a stationary

point only ( a point for which f ( ) = 0 ).
10.1 Local maxima, local minima, and saddle points
Lecture 10
Lecture 10 Page 2
is a local maximum;
if > 0 and f xx < 0 :
The principal minors are the determinants of all the top left hand
of f ( ). In other words, they are the numbers

f x1 x1

f x1 x1 f x1 x2 f x1 x3
f
f x x f x x2

1
1
1
, f x x f x x f x x , . . . , x2. x1
f x1 x1 ,
2 2
2 3
2 1
..
f x2 x1 f x2 x2

f

x3 x1 f x3 x2 f x3 x3
fx x
n 1
f x1 x2
f x2 x2
..
..
.
.
f x n x2
n n

f x1 xn
f x2 xn
.. .
.
fx x
square submatrices
We can generalise the method from the previous part to functions f : R n R for any
dimension n. For this we use the second derivative
f x1 x1 ( ) f x1 x2 ( ) f x1 x n ( )
f x x ( ) f x x ( ) f x x ( )
2 1
2 2
2 n
f ( ) =
.
..
..
..
..
.
.
.
.
f x n x1 ( ) f x n x2 ( ) f x n x n ( )
10.3 Classification of stationary points for functions f : R n R
E XAMPLE
See Example 4 on pages 191 and 192 of the textbook.
In the case that = 0, it may well be possible to get a conclusion by other ways.
( It looks as if we forgot the possibility that > 0 and f xx = 0. But its easy to convince
yourself that if f xx = 0, then = 0 or < 0. )
we cant derive a conclusion with this method.
is a local minimum;
if > 0 and f xx > 0 :

if = 0 :
is a saddle point;
if < 0 :
Now we can show that the following holds :
where we use that f xy = f yx .
For functions f : R2 R, to classify stationary points we use the Hessian :

f ( ) f xy ( )
= f xx ( ) f yy ( ) f xy ( )2 ,
= det f ( ) = xx
f yx ( ) f yy ( )
10.2 Classification of stationary points for functions f : R2 R
Lecture 10 Page 3
is a local minimum;
Rn
Classification of stationary points for functions f :
6.5.1 and 6.5.3;
R : 6.5.2, 6.5.7 and 6.5.10.
Classification of stationary points for functions f : R2 R :
We will later see another way to define ( and test ) positive and negative definite matrices, in terms of their eigenvalues. That is also the approach that is taken in the textbook.
For the moment we will use the tests based on the principal minors.
And we have : A symmetric matrix A is negative definite if and only if a11 < 0 and the
principal minors of A are alternating in sign :

a11 a12 a13

a11 a12

> 0, a21 a22 a23 < 0, . . . .
a11 < 0,

a21 a22
a

31 a32 a33
A symmetric matrix A is negative definite if x T A x < 0 for all x 6= 0.
It can be shown that : A symmetric matrix A is positive definite if and only if all principal
minors of A are positive.
The analysis above is actually a special case of a more general kind of properties of
symmetric matrices. ( A matrix A is symmetric if it is square and if aij = a ji for all i, j. )
a11 a1n
..
.. are the determinants
..
The principal minors of an n n matrix A = .
.
.
an1 ann

a11 a12 a13
a11 a12

, a21 a22 a23 , . . . , | A|.
a11 ,

a21 a22
a

31 a32 a33
We say that a symmetric matrix A is positive definite if x T A x > 0 for all x 6= 0.
10.4 Positive and negative definite matrices
is a saddle point;
this method gives no conclusion.
E XAMPLE
See Example 5 on pages 192 and 193 of the textbook.
if det f ( ) = 0 :
if det f ( ) 6= 0 and neither of the above holds :
if f x1 x1 < 0 and the principal minors alternate in sign : is a local maximum;
if all principal minors are positive :
Suppose R n is a stationary point of a function f : R n R ( so f ( ) = 0 ). Then
And very similar to the classification of stationary points for 2-dimensional functions,
we have the following.
Convex and Concave Functions
Convex Sets
6.11.1 and 6.11.2.
6.5.4, 6.5.6 and 6.5.11;

Ad-hoc methods for constrained optimisation :
Convex sets; convex and concave functions :
11.2 Ad-hoc methods for constrained optimisation
If the function f : R n R is convex, then a local minimum is automatically a global

minimum.
If the function f : R n R is concave, then a local maximum is automatically a global

maximum.
The importance of convex and concave functions is given by the following :
Similarly, a function can be neither convex nor concave.
Note that is is not always the case that a symmetric matrix A is one of the two ( nonnegative definite or nonpositive definite ). It is well possible that there are x for which
x T A x < 0, and that there are y for which y T A y > 0.
Note that if a matrix is positive definite, then it certainly is nonnegative definite; and if it is
negative definite, then it certainly is nonpositive definite.
For this we follow mostly Section 6.4 of the textbook, except in the following definitions :
A symmetric matrix A is nonnegative definite ( also called positive semidefinite ) if
x T A x 0 for all x.
And the matrix is nonpositive definite ( or negative semidefinite ) if x T A x 0 for all x.
11.1 Convex sets; convex and concave functions
Introduction to Constrained Optimisation
Lecture 11
The Lagrange multiplier
Lagranges Method
for Constrained Optimisation
G
f
= 0
x
x
and
f
G
= 0.
y
y
for every b.
G dx
G dy
+
= 1.
x
db
y
db
(2)
We can differentiate this expression with respect to b, using the Chain Rule, to get
G ( x (b)) = b,
Now we can expect that x depends on b, i.e., x is a function of b. So lets write it as

x = x (b). Since we are only interested in points with G ( x, y) = b, this means we
always must have
(1)
and in particular find its stationary points. So lets assume that we find the optimum
of f at a point x = ( x , y ), with corresponding Lagrange multiplier . This means in
particular that those points satisfy the equations
L( x, y, ) = f ( x, y) + (b G ( x, y)),
Following the theory in the book, we should consider the Lagrangean
g( x, y) = b G ( x, y) = 0.
We can translate this into our standard form, by giving the constraint as
Suppose we want to optimise a function f ( x, y) subject to a constraint of the form

G ( x, y) = b ( with G : R2 R some function, and b a real number ). And we are
interested in what happens when we vary b.
This subsection is a simplified version of Section 8.5 of the textbook. The part below is
more than enough for our purposes.
12.2 The Lagrange multiplier
12.1 Lagranges method
Lecture 12
Lecture 12 Page 2
d
h ( ) ( x ).
dx
d
f ( x ( B)) (b B) = f ( x ( B)) + (b B).
db
L( x, y, ) = 250 x1/4 y3/4 + (1000 4 x 9 y).
The Lagrangean in the format we need is
maximise P( x, y) = 250 x1/4 y3/4 , for 4 x + 9 y = 1000.
According to the argumentation in the book, we know for sure that there is a maximum,
and that it will occur on the line 4 x + 9 y = 1000. So we need to solve
maximise P( x, y) = 250 x1/4 y3/4 , for 4 x + 9 y 1000.
E XAMPLE
We consider Example 15 on page 212 of the textbook. So we want to
In other words : Changing b by a small amount leads to a change in optimal value of approximately times .
f ( x ( B + )) f ( x ( B)) + .
If we write b = B + , then for close to 0 this gives
f ( x (b)) f ( x ( B)) +
Now lets use this for the function f ( x (b)) ( seen as a function R R in the one
variable b ). The optimal value f ( x (b)) for b close to some B behaves as
h( x ) h( ) +
We can interpret the previous part further. For a function h : R R around a point
we have the linear approximation
Weve shown that : The Lagrange multiplier is the marginal rate of change of f with respect
to b at the optimal point x .
f dx
f dy
G dx
G dx
d
f ( x (b)) =
+
=
+
db
x
db
y
db
x
db
y
db
G dx

G dx
=
+
= .
x db
y
db
Using the expressions in (1) ( evaluated in x , y , ) and (2), we can rewrite this as
d
f dx
f dy
f ( x (b)) =
+
.
db
x db
y
db
Now lets consider the function value in an optimum, in other words consider f ( x (b)).
Note that we can see this as a function depending on b, hence as a function R R
( for every b R we have a corresponding optimal function value f ( x (b)) ). To find
the derivative with respect to b of this function, we again use the Chain Rule :
Lecture 12 Page 3
Lagranges method :
6.11.3, 6.11.4, 6.11.6, 6.11.10, 6.11.11, 6.11.12, 6.11.13 and 6.11.14.
12.3 Optional extra exercises related to this topic
The Lagrange multiplier appears with respect to the constraint used in the Lagrangean
function. So any change should be seen as a change in the constraint used in the Lagrangean function.
As an example, suppose that instead of a total cost of 30000, we are allowed to use
100 dollar more. So the new constraint, in terms of total cost, becomes 120 x + 270 y =
30100. But this means that the constrained we used changes from 4 x + 9 y = 1000 to
4 x + 9 y = 1003 13 = 1000 + 3 31 . So the expected increase in the optimal production level
is approximately 3 13 64.63.
E XAMPLE ( continued )
In the example above we should take some care when trying to interpret the Lagrange
multiplier in terms of the original information, with the original constraint 120 x +
270 y = 30000. Although this constraint is equivalent to the one we used, 4 x + 9 y =
1000, the corresponding Lambda multipliers are not the same !
Suppose we change the constraint 4 x + 9 y = 1000 to 4 x + 9 y = 999 = 1000 1. Then

the change in the optimal production level is approximately (1) 19.39. I.e.,
the best possible production level goes down.
From this we can make conclusions such as the following :
1 3/4 3/4
1 1/4 1/4
= 250
(x )
(y )
= 250
( x ) (y )
19.39.
16
12
As in the book, we find that the solutions are x = 62 12 and y = 83 13 , with corresponding multiplier
250 14 x 3/4 y1/4 4 = 0;

250 43 x1/4 y3/4 9 = 0;
1000 4 x 9 y = 0.
( Notice that we need to reverse the order inside the brackets from the way it is done in
the book. ) The stationary points of the Lagrangean are found by solving
Derivatives of Inverse Functions
Exponential and Logarithmic Functions
Inverses of Functions f : R R
Derivatives of exponential and logarithmic functions, and of inverse functions :

2.11.3, 2.11.4, 2.11.6, 2.11.7 and 2.11.8.
13.2 Derivatives of inverse functions
For this we follow Sections 2.5 and 2.6 of the textbook. As a refresher, you should look
at Section 2.3.
13.1 Inverse functions; exponential and logarithmic functions
Lecture 13
7.6.6;
Local inverses of vector valued functions, and their derivatives :

7.6.7, 7.6.8, 7.6.9, 7.6.10, 7.6.11 and 7.6.13.
Local inverse functions, and their derivatives :
14.2 Local inverses of vector valued functions, and their derivatives
14.1 Local inverse functions, and their derivatives
Derivatives of Local Inverse Functions
Local Inverses of Functions

f : R R
Local Inverses of Functions f : R R
Lecture 14
Interest Compounding
(1 + r ) P + r (1 + r ) P =
after 2 years the total is :

..
.
(1 + r )t P.
(1 + r )2 P;
..
.
(1 + r ) P;
We also consider the case that the rate is negative, r < 0. That is often called depreciation
or deflation.
S = P ert .
We call this continuous compounding. So : if a principal P is invested for time t, at

annual interest rate r, and interest is compounded continuously, then the final sum is
So what would happen if we add interest more and more often ? In other words, what

r tm
happens for the final sum as m ? The easy answer is lim P 1 +
. But
m
m
what is that limit ?

1 m
= e, where e is the basis of the natural logarithm.
For this we use that lim 1 +
m
m
From this we can derive :

rt

r tm
1 m/r
lim P 1 +
= P ert .
= lim P 1 +
m
m
m/r
(m/r )
If instead of adding interest once a year, it is added m times per year, at an annual rate r,
r
for t years, then this is equivalent to compounding for t m periods, each at rate .
m
So the final sum in that instance is

r tm
.
S = P 1+
m
after t years the total is :
P+rP =
after 1 year the total is :
Suppose we have a principal P, compounded annually at interest rate r, for t years

( where t {1, 2, 3, . . .} ). To calculate the final total amount, we deduce :
Introduction to Integration
Arithmetic and Geometric Progressions
15.1 Interest compounding
Lecture 15
Lecture 15 Page 2
We find S = 1000 e0.163 = 1000 e0.48 1616.07.
re = e 1.
The process of determining the present value P of a future sum S is called discounting.
The rate of discount is the interest rate used when finding the present value.
A future sum of money is not worth as much as a present sum, since money available
now could have earned interest in the meantime.
15.3 Discounting and future value
Hence re = (1.01)12 1 0.127; which can be written as re 12.7 %.
We want to find the re so that

0.12 12t
.
P (1 + r e ) t = P 1 +
12
E XAMPLE
Find re for a nominal interest rate of 12 % compounded monthly.
if interest is compounded continuously :
When we know how often is compounded, we can calculate the effective annual interest rate re . That is the rate that would need to be given if we only compounded once a
year. Using the formulas above, it is easy to check that :

r m
if interest is compounded m times per year :
re = 1 +
1;
m
The rate r we used in the calculations above is called the nominal rate.
15.2 Nominal and effective interest
We need to find r so that 3 P = S = P (1 + r )16 . Hence we need to solve 3 = (1 + r )16 .

This gives 1 + r = 31/16 1.071, hence r 0.071. This can be written as r 7.1 %.
E XAMPLE
At what rate will money treble in 16 years under compounding annually ?
(iii)
This time we have m = 4, hence t m = 12, and so
(ii)

0.16 12
S = 1000 1 +
= 1000 (1.04)12 1601.03.
4
Now we have P = 1000, r = 0.16, and t = 3, so S = 1000 (1 + 0.16)3 1560.90.
(i)
E XAMPLE
Find the final value when a principal of 1000 is invested for three years at 16 % interest,
and compounded (i) annually; (ii) quarterly; or (iii) continuously.
Lecture 15 Page 3

0.10 24
= 750 (1.05)8 507.63.
P = 750 1 +
2
P = V ert = 100 e
t 0.08t
The present value P, when it would be sold at time t, is given by
0.08.
d2 P
=
dt2
1 1/2
2t
dP

0.08
+ P 41 t3/2 .
dt
But we still need to check that this stationary is really the maximum ( the best time to

dP
= P 21 t1/2 0.08 ,
sell the piece ). For this, we look at the second derivative. Since
dt
we find
t1/2
dP
= 0. This is equivalent to 12 t1/2 = 0.08, hence to
dt
= 0.16. So the stationary point is t = (0.16)2 39.06 years.
1 1/2
2t
For a stationary point we have
1 dP
=
P dt
Differentiating this with respect to t we get
Maximising P is the same as maximising
ln( P) = ln 100 + t 0.08 t.
(i)
How long should it be kept to maximise its present value, assuming continuous compounding, at a rate of (i) r = 0.08; and (ii) r = 0.12 ?
V = 100 e t .
E XAMPLE
A piece of cut class is currently worth 100. When it gets older, it is appreciating in
value according to the formula
Maximising natural exponential functions, which describe constant rates of continuous

growth, can be useful for calculating optimal timing.
We find
E XAMPLE
Find the present value of 750 to be paid in four years, when interest is compounded
semi-annually at 10 %.
When compounding annually we have S = P (1 + r )t , so P = S (1 + r )t .

r tm
When compounding m times a year we find P = S 1 +
.
m
rt
And when compounding continuously, we get P = S e .
dP
= 0 ( since we are at a stationary point ), which means that
dt
Lecture 15 Page 4

= (n + 1) a + 12 n b .
i =1
i =0
(0 + i 1 )

= (n + 1) 0 + 12 n 1 =
E XAMPLE
A special case is if a = 0 and b = 1. Then we get
1
2
n ( n + 1 ).
In other words : the sum is equal to the number of terms multiplied by the average of
the terms, where the average is the sum of the first and the last term divided by two.
i =0
( a + i b)
The total is
a + ( a + b ) + ( a + 2 b ) + ( a + 3 b ) + + ( a + n b ).
An arithmetic progression is a sum in which there is a fixed difference between consecutive

terms :
i =1
For sums a1 + a2 + + ak we use the notation ak .
For this you should read Section 10.1 of the textbook, together with the text below.
15.4 Arithmetic and geometric progression
The difference in the answers can be understood as follows : The larger the discount
rate, the less we value future earning, so the shorter we want to wait before selling it.
(ii) Repeating the calculations, we find that if r = 0.12, then P is maximised at 17 13

36
years.
1
years.
Hence the present value P is maximised when we sell in 39 16
1
is a
= 14 P t3/2 . This expression is always negative ( since P, t > 0 ). So t = 39 16
dt2
local maximum. And since it is the only local maximum of a one-dimensional function,
we can conclude it is a global maximum.
d2 P
1
, we have
At t = 39 16
Lecture 15 Page 5
i =0

n
1 x n +1
.
xi = a
1x
i =0
( for x 6= 1 ).
1 (1 + r ) n + 1
(1 + r ) n + 1 1
(1 + r ) n + 1 1
=
=
.
1 (1 + r )
(1 + r ) 1
r
1 x n +1
1x
(1 + r ) n + 1 1
1
r
i =0
(1 + r ) n + 1 1 r
(1 + r ) n + 1 (1 + r )
=
.
r
r
r
(1 + r )i (1 + r )0
(1 + r ) n + 1 (1 + r )
.
r
For the binomial coefficient and the binomial theorem, see the textbook.
Sn = P (1 + r ) + P (1 + r )2 + + P (1 + r ) n = P

S2 = P (1 + r ) + P (1 + r ) (1 + r ) = P (1 + r ) + P (1 + r )2 ,

S3 = P (1 + r ) + P (1 + r ) + P (1 + r )2 = P (1 + r ) + P (1 + r )2 + P (1 + r )3 ,
..
.
S1 = P (1 + r ),
Let Sn be the sum accumulated at the end of n years, for n = 1, 2, . . . , t. Then we get
that S1 = P (1 + r ). Moreover, for n 1, Sn+1 is formed by a term P (1 + r ) from
the deposit we made at the beginning of the year, plus a term Sn (1 + r ) from what we
already accumulated before that year ( both with interest earned ). So we get that
E XAMPLE
Suppose a deposit of P is made at the beginning of each year, in an account earning
interest at a rate r, compounded annually, and paid at the end of each year. Find the
sum accumulated at the end of t years.
i =1
(1 + r )i
But we should not forget that we should start at i = 1 and not at i = 0. But that is easy
to deal with, we just subtract the term for i = 0 :
i =0
(1 + r )i
Setting a = 1 and x = 1 + r, we see that
i =1
E XAMPLE
n
Determine (1 + r )i .
= 1 + x + x2 + x3 + + x n =
From this its easy to obtain that a xi = a
i =0
xi
The base form is a = 1, for which we have the outcome
a + x a + x2 a + x3 a + + x n a.
A geometric progression is a sum in which there is a fixed ratio between consecutive terms :
Lecture 15 Page 6
Introduction to integration :
10.10.6, 10.10.7, 10.10.8 and 10.10.12.
The table on page 303 should be known by all students.
OF THE
f (t) dt is a primitive of the function f .
F UNDAMENTAL T HEOREM OF C ALCULUS

Z b

b
If F is any primitive for f , then
f (t) dt = F( x ) a = F(b) F(a).
V ERSION 2
Z x
F UNDAMENTAL T HEOREM OF C ALCULUS

Z
d x
f (t) dt = f ( x ).
dx a
f (t) dt, then I ( x ) =
OF THE
Z x
In other words : The integral I ( x ) =
If I ( x ) =
V ERSION 1
If f is a function R R, then a primitive ( sometimes also called an antiderivative )

of f is a function F such that F ( x ) = f ( x ).
From Section 10.3 we only need to know some terminology and the two versions of the
Fundamental Theorem of Calculus given below. You should also study the example
on page 300.
For this we follow Sections 10.2, 10.4 and 10.5 of the textbook completely.
15.5 Introduction to integration
Integration by parts : 10.10.23, 10.10.24 and 10.10.25;
Integration by change of variable :

10.10.17, 10.10.18, 10.10.19, 10.10.20, 10.10.21 and 10.10.22;
Integration by partial fractions : 10.10.13, 10.10.14, 10.10.15 and 10.10.16;
Consumers and producers surplus : 10.10.9 and 10.10.10;
16.4 Integration by parts
16.3 Integration by change of variable
16.2 Integration by partial fractions
This topic can be found as Example 4 in Section 10.5.
Integration by Parts
Integration by Change of Variable
Integration by Partial Fractions
Application of Integration :
Consumers and Producers Surplus
16.1 Consumers and producers surplus
Lecture 16
2 f
N
M
=
=
=
.
y
yx
xy
x
2 f
In other words, a differential equation of the form M dx + N dy = 0 ( where M and N

are expressions in x and/or y ) is exact if and only if there is a function f ( x, y) such that
f
f
= M and
=N
x
y
But that holds in its turn only if and only if
Such a differential equation is called exact.
The differential of z = f ( x, y) is dz =
f
f
dx +
dy.
x
y
f
f
dx +
dy = 0 has solution f ( x, y) = c,
It follows that the differential equation
x
y
c R.
For this we follow Sections 12.5 and 12.11 of the textbook, together with the text below.
17.4 Exact differential equations
17.3 Partial differential equations
For this we follow Section 12.4 of the textbook, although you can skip Example 3 on
pages 379 and 380.
17.2 Separable differential equations
17.1 Introduction to differential equations
Linear Differential Equations of Order One
Exact Differential Equations
Partial Differential Equations
Lecture 17 Introduction to Differential Equations
and
f
= N.
y
x 2 + y2
dy
=
.
dx
2xy
M
N
= 2 y and
= 2 y. So
y
x
and
f
= N = 2 x y.
y
1
3
x 3 + x y2 + g ( y )
and
f ( x, y) = x y2 + h( x ),
x3 + x y2 = c,
1
3
( c R ).
x3 + x y2 + k, and hence the solution to the differential equa-
17.5 Linear differential equations of order one
1
3
So we find that f ( x, y) =
tion is
Since we must have 13 x3 + x y2 + g(y) = x y2 + h( x ), it follows that we must take g(y) =

k and h( x ) = 31 x3 + k, for some constant k R.
where g(y) is an arbitrary function in y, and h( x ) is an arbitrary function in x.
f ( x, y) =
Integrating both equations with respect to x and y, respectively, we find the general
solutions
f
= M = x 2 + y2
x
N
M
=
. Hence the equation is exact, and we need to find a function f so that
y
x
Setting M = x2 + y2 and N = 2 x y, we can check that
( x2 + y2 ) dx + 2 x y dy = 0.
We can rewrite this as 2 x y dy = ( x2 + y2 ) dx, hence as
Solve the equation
E XAMPLE
Once we have f , the solution to the exact equation is given by f ( x, y) = c, c R.
f
= M
x
M
N
=
.
y
x
f
f
And in an exact equation, M and N are the respective partial derivatives
and
of
x
y
a function f that is a common solution of the equations
Lecture 17 Page 2
We get that the differential equation M dx + N dy = 0 is exact if and only if
12.13.1 and 12.13.2;
Lecture 17 Page 3
Linear differential equations of order one : 12.13.4;
Exact differential equations : 12.13.7, 12.13.8, 12.13.17 and 12.13.18;
Separable differential equations : 12.13.3.
Introduction to differential equations :
Linear Differential Equations with

Constant Coefficients
Differential Equations and

Change of Variable
Homogeneous Differential Equations
12.13.14 and 12.13.15;
Linear differential equations with constant coefficients :

14.11.1, 14.11.2,, 14.11.4, 14.11.13 and 14.11.14.
Types of differential equations : 12.13.6, 12.13.9 and 12.13.10.
Solving differential equations by change of variable :
Homogeneous differential equations : 12.13.5;
You should also look at Section 14.3, although we will use the knowledge from the
Linear Algebra part of the course for most of that material.
18.4 Linear differential equations with constant coefficients
18.3 Types of differential equations
18.2 Solving differential equations by change of variable
18.1 Homogeneous differential equations
Lecture 18
Difference Equations
Convergence and Divergence of

Solutions of Differential Equations
A, B constants.
Difference equations; linear difference equations :
14.11.5, 14.11.6 and 14.11.12.
Convergence and divergence of solutions of differential equations :

14.11.3 and 14.11.11;
For this we follow Sections 14.2 and 14.6 and the final part of Section 14.8 of the textbook.
19.2 Difference equations; linear difference equations
Because of the factor ex/2 , we have that y oscillates infinitely as x ; except if A =

B = 0, in which case we have that y is identical 0.
E XAMPLE
Describe the behaviour as x of the general solution

A, B constants.
y = ex/2 A cos 12 x + B sin 12 x ,
The term C x e2x
, if C > 0;
, if C < 0.
If C = 0, the term B e2x becomes dominant, and similar arguments apply.
A, B, C constants.

is dominant. So as x we have y
y = A e x + B e2x + C x e2x ,
E XAMPLE
, if A > 0;
The term A ex is dominant. So as x we have y
, if A < 0;
0, if A = 0.
y = A ex + B e3x ,
E XAMPLE
For this we follow Section 14.8 of the textbook, together with the text below.
19.1 Convergence and divergence of solutions of differential equations
Lecture 19
Application : The Cobweb Model
Nonhomogeneous Differential
and Difference Equations
Nonhomogeneous difference equations : 14.11.9, 14.11.10 and 14.11.15.
Nonhomogeneous differential equations : 14.11.7 and 14.11.8;
For this we follow Section 14.14.1 of the textbook, but you can ignore the model III
from page 466 onwards.
20.3 The cobweb model
For this we follow Section 14.7.2 of the textbook.
20.2 Nonhomogeneous difference equations
For this we follow Section 14.7.1 of the textbook ( but dont worry about the physics part
at the beginning of Example 14 ).
20.1 Nonhomogeneous differential equations
Lecture 20
MA 100
Linear Algebra
Directed Reading and Notes
Textbook
Martin Anthony and Michele Harvey, Linear Algebra, Concepts and Methods
Cambridge University Press, 2012. ISBN 978-0-521-27948-2
All students are expected to have a copy of this book.
The notes for this part of MA 100 consist mainly of detailed references to the
relevant parts of the textbook. Reading those parts is essential for understanding
the material in the course.
2013/14
c LSE 2013

Course Content { LINEAR ALGEBRA

1
Matrices
Manipulation of matrices (A + B, AB, AT ), rules of matrix algebra.
For n n matrices, compute the determinant of A, |A|. Find A1 if it exists.
Write a system of equations in matrix form; coecient matrix.
Gaussian elimination
Solve a system of linear equations, Ax = b (Ax = 0) by putting the augmented
(coecient) matrix into reduced row echelon form using elementary row operations.
If A is n n and invertible: Cramers rule or A1 .
Vectors and vector spaces

Understand the basic concepts: vector space, subspace, span, linear independence, basis,
dimension. Apply these to Rn and subspaces of Rn .
Find the subspace spanned by a given set of vectors; nd a basis for a given subspace.
Apply the Gram-Schmidt method to obtain an orthonormal basis of a given vector space.
Subspaces associated with a matrix

Null space, range (column space), row space. The rank-nullity theorem for matrices.
Linear Transformations
Use a matrix to represent a linear transformation T : Rn Rm .
Null space and range. The rank-nullity theorem for linear transformations.
Change of basis
Use a matrix P to represent a change of basis, v = P[v]B from coordinates relative to a
new basis B to standard coordinates.
Diagonalisation and its applications

Find eigenvalues and corresponding eigenvectors of a square matrix A. Find an invertible
matrix P and a diagonal matrix D such that P1 AP = D.
Use this to: compute Ak for large k, solve a system of linear dierential equations y = Ay
by a change of basis y = Pz, solve a system of dierence equations xt = Axt1 , in
particular, Markov chains.
Orthogonal diagonalisation and applications

Find an orthogonal matrix P whose columns are eigenvectors of a symmetric matrix A (and
are an orthonormal basis of Rn ) and a diagonal matrix D such that PT AP = P1 AP = D.
Use this to sketch the graph of a conic section given by xT Ax = 1 where A is symmetric.
If A is symmetric, determine if the quadratic form xT Ax is positive denite, negative
denite or indenite, and the implications of this.
Complex numbers
Basic properties, z = a + ib = r(cos + i sin ) = rei .
DeMoivres theorem. Eulers formula: ei = cos + i sin .
In particular read
(If you have any difficulty with the reading, attend office hours.)
A little bit of logic (page 8).
Mathematical terminology (page 3)
Preliminaries: before we begin.
To begin, you should look through the introductory chapter:
working through the activities.
Read each section carefully,
Section and page number references refer to the linear algebra textbook.
Vectors and lines, as covered in Calculus Lecture 1, are also an important

part of Linear Algebra. See Sections 1.91.10 of the linear algebra text.
Exercises 1.11.7
Optional extra exercises on these topics.
Section 1.8
Vectors in Rn
What is the transpose of A, AT ? What is a symmetric matrix?

Section 1.7
Section 1.6
If A is an n n matrix and r Z, what is Ar ?
Section 1.5
The inverse of an invertible matrix
Section 1.4 (You do not need to know the proofs, but you should
understand why and how the rules work, and use them properly.)
Rules of matrix algebra
Section 1.3
Matrix multiplication. Matrix multiplication is not commutative.
Read Sections 1.11.2
What is a matrix? Equality, addition and scalar multiplication
This lecture assumes no previous knowledge of matrices. Below is a list of

the topics covered in the lecture with the appropriate reading.
Le ture 1 | Matri es
Planes in R3 , as covered in Calculus Lecture 2, are also part of Linear

Algebra. Have a look at Section 1.11 of the linear algebra text.
3. What is the definition of a matrix in row echelon form?

What is the definition of a matrix in reduced row echelon form?
Exercises 2.12.3
2. Write down the algorithm (6 steps) to put an augmented matrix into

row echelon form and then into reduced row echelon form.

Answer the following three questions:
1. What are the three elementary row operations which you can do to the
augmented matrix without changing the solution to the system?
Linear systems with free variables

Section 2.3.3
Section 2.3.2
Consistent and inconsistent systems
Sections 2.3, 2.3.1
Gaussian elimination, (an algorithmic method using elementary row

operations to solve a system of linear equations)
Section 2.2
Row operations
Systems of linear equations, matrix form, coefficient matrix,

augmented matrix
Section 2.1
Le ture 2 | Gaussian elimination
1 )?
Solving a system of linear equations using Gaussian elimination is one of the

basic skills you will need throughout this course.
With practice, you should be able to read all solutions directly

from the reduced row echelon form of the augmented matrix.
Write all solutions in vector form.
yes assign arbitrary parameters to the non-leading

variables. Then solve for the leading variables in terms
of these parameters. There are infinitely many solutions.
no read the unique solution from the reduced row echelon

form of the augmented matrix.
(3) Locate the leading ones. Are there non-leading variables?
yes stop the system is inconsistent (why?)

no continue to reduced row-echelon form
(2) Is there a row: ( 0
(1) Write down the augmented matrix and reduce it to

row-echelon form using elementary row operations.
Below is a summary of Gaussian elimination.
Exercises 2.42.6
Section 3.5
Leontieff input-output analysis
Section 4.1.1, page 132
The rank of a matrix
Section 2.4.2
Null space, principle of linearity
Section 2.4.1
Homogeneous systems
Section 2.3.4
Solution sets
Le ture 3 | Linear systems

Exercises 3.13.2
Section 3.1.5
Verifying an inverse
Section 3.1.4
Using row operations to find an inverse matrix
Section 3.1.3
The Main Theorem
Section 3.1.2
Row Equivalence
Section 3.1.1
Elementary matrices
Le ture 4 | Matrix inverse using row operations
Exercises 3.33.4
These are contained in Section 3.3, pages 104106 as Corollary 3.28,

Corollary 3.29 and Corollary 3.30.
Corollaries Results on determinants
Section 3.2.2
Elementary signed products
Section 3.2.1
Determinants using cofactors
Le ture 5 | Determinants I
Cramers Rule: To find xi ,

(1) replace column i of A by b,
(2) evaluate the determinant of the resultant matrix,
(3) divide by det A.
Note that row operations on a determinant can affect its value.
2 Methods to find A1
(1) Using cofactors and the adjoint matrix
(2) By elementary matrices: row reduction of (A|I) to (I|A1 )
3 Methods to solve Ax = b if A is n n and |A| =

6 0
(1) Gaussian Elimination
(2) Find A1 , then x = A1 b
(3) Cramers Rule
At this point in the course, make sure you know well the material we have
covered, in particular Gaussian elimination: finding and understanding the
solutions of linear systems of equations. In addition, you now have:
Exercises 3.53.9
Section 3.4.2
Cramers rule
Section 3.4.1
Matrix inverse using cofactors
Section 3.4.1, Theorem 3.39
Results on determinants
Section 3.3, Theorem 3.27. Section 3.2.2, Theorem 3.37
Section 3.3.1
Using row operations
Le ture 6 | Determinants II
There are many abstract ideas here, but they are not difficult once you get
used to them. You need to know what a vector space is, but you do not need
to memorise the 10 axioms. You need to understand that a subspace is a
vector space sitting inside a (possibly larger) vector space. We will usually
be dealing with subspaces of Rn .
to conclude that it is a subspace.
The linear span of a set of vectors is a subspace. So you could also show
that

1
W = Lin{v},
where v = 1
1
We cover all of Chapter 5 in this lecture, except for Sections 5.2.4 and 5.3.1,
which we defer to weeks 10 and 11 respectively.
To do this, use general vectors, u, v W (use letters to represent any

numbers), then use the fact that u and v satisfy the definition of W to show
that the sum u + v and the scalar product u also satisfy the definition of
W (and are therefore in W ).
You can take some examples of vectors in the subset to help you decide
whether or not adding two of them together or scalar multiplying them by
a real number will always get you another vector in the subset. If you think
it is a subspace, then you need to prove it. If you think it is not a subspace,
you only need to give one counterexample.
To prove the set W is a subspace, verify that it is non-empty, and closed
under addition and closed under scalar multiplication.
Is 0 in the subset? If not, the subset is not a subspace.
Look carefully at the definition. What is the requirement for a vector v Rn

to be in the given subset?
Extra notes (See Section 5.2.3.)

A subset of Rn can be defined either by giving a general vector or a condition
the vectors satisfy, such as

x

t

or U = y x + y + z = 2
W = t t R

z
t
To prove a set U is not a subspace find one counterexample, one

particular set of vectors (use numbers) for which the sum or the scalar
product does not satisfy the definition of U . (Here it is simple: 0
/ U .)
Exercises 5.15.2
Deciding if a subset is a subspace

Section 5.2.3
Section 5.3, page 160; Section 5.3.2
Linear span
Sections 5.2.15.2.2
Subspaces
Section 5.1.3; Section 5.6, Exercise 5.3(a)
Linear combinations
Vector spaces
Sections 5.1.15.1.2
Le ture 7 | Ve tor spa es
Write down the definitions of linear independence and span; that is, if
S = {v1 , v2 , . . . , vk } V , what does it mean to say that the set S is
linearly independent? What does it mean to say that the set S spans V ?
Exercises 5.3, 5.4, 6.1, 6.2, 6.4, 6.5
Section 6.1.4 Linear independence and span in Rn
Section 6.1.1 Uniqueness of linear combinations

Section 6.1.2 Testing for linear independence in Rn
Section 6.1 Linear independence

1
v2 = 2 ,
3
1
u = 2 ,
5

1
w = 2.
5
See Exercise 5.3 and Section 6.1.4.
The set {v1 , v2 } is linearly independent and spans a plane in R3 .
The set {v1 , v2 , u} does not span R3 and is not linearly independent.
The set {v1 , v2 , u, w} spans R3 but is not linearly independent.
Show that
The set {v1 , v2 , w} spans R3 and is linearly independent.
1
v1 = 0 ,
1
As in the lecture, consider the vectors
So to be linearly independent, we must have rank(A) = k, so then k n.
You should read
Section 5.3 Linear span (page 160), also look at Exercise 5.3.
The system of equations Ax = 0 will have a unique solution (the trivial

solution) if and only if the row echelon form of A has a leading one in every
column. (Can you explain why?)
So to span, we must have rank(A) = n, and in this case k n.
The system of equations Ax = b will be consistent for all b Rn if and

only if the row echelon form of A has a leading one in every row. (Make
sure you understand why.)
You do both by putting the matrix A into row echelon form and looking at
the leading ones.
To determine if a set of vectors {v1 , v2 , . . . , vk } Rn spans Rn , you need

to determine if the system of equations Ax = b has a solution for all b Rn ;
that is, if it is always consistent.
To determine if a set {v1 , v2 , . . . , vk } Rn is linearly independent, you
need to determine if the system of equations Ax = 0 has a unique solution;
that is, if Ax = 0 has only the trivial solution.
Extra notes
If {v1 , v2 , . . . , vk } Rn , the vector equation 1 v1 + 2 v2 + + k vk = b
is equivalent to the matrix equation Ax = b, where A is the n k matrix
whose columns are the vectors v1 , v2 , . . . , vk . (Make sure you know why.)
(2) Any vector in V can be written in at most one way as a linear combination
of the vectors v1 , v2 , . . . , vk .
This is the concept that a set of vectors v1 , v2 , . . . , vk V
is linearly independent.
(1) Any vector in V can be written in at least one way as a linear combination
of the vectors v1 , v2 , . . . , vk .
This is the concept that a set of vectors v1 , v2 , . . . , vk
spans a vector space V .
There are two concepts contained in the definition:
Our goal (which we will not achieve until next week) is the definition of a
basis of a vector space V , which is a set of vectors {v1 , v2 , . . . , vk } with
the property that every vector v V can be expressed as a unique linear
combination of v1 , v2 , . . . , vk . This means that we essentially know every
vector v V by just knowing a finite set v1 , v2 , . . . , vk of vectors in V .
In this lecture we develop the ideas of linear independence and span using
examples in a slightly different way than in the text, so you may wish to
print a copy of the lecture slides. However all the material you need is
contained in the text, in Sections 6.1 and 5.3.
Le ture 8 | Linear span { linear independen e
To be a basis of Rn , there must be precisely n vectors in the set. (See Section

6.1.4.) So to determine if a set {v1 , v2 , . . . , vn } is a basis of Rn , you can
write the vectors as the columns of an n n matrix A and use the main
theorem. If |A| =
6 0, then the set is a basis. Why? The main theorem states
that |A| =
6 0 if and only if the system of equations Ax = b has a unique
solution for every vector b in Rn , and this matrix equation is equivalent to
the vector equation 1 v1 + 2 v2 + + n vn = b. (See Exercise 6.4.)
To show a set of vectors is a basis of a subspace W , you need to show that

it spans W and it is linearly independent. (See Example 6.26)
Extra notes
Write down the definitions of basis and dimension; that is, if

S = {v1 , v2 , . . . , vk } V , what does it mean to say that the set S
is a basis of V ? What is the definition of dim(V )?
Exercises 6.86.10
Sections 6.4.1 and 6.4.2 (You do not need to know the proofs of the
theorems.)
Dimension
Section 6.3
Coordinates
Section 6.2
Basis
Sections 6.1.3 and Section 6.1.4
Linear independence and span
Le ture 9 | Basis and dimension
Section 2.4.2 (null space),

Sections 4.1.1, 4.1.2, (rank of a matrix)
Sections 4.2.1, 4.2.2 (rank and systems of linear equations),
Section 5.2.4 (If A is m n, the null space is a subspace of Rn .)
Optional extra exercises and review exercises.

Exercises 4.14.6 (do these!), Exercise 5.5 (optional)
Section 5.1.2, Example 5.7
Spaces of matrices
Section 5.1.1 (pages 149-150); Section 5.1.2, Example 5.5; Section 5.1.3,
Example 5.14
F , functions of real numbers
Section 5.1.2, Example 5.8; Section 6.8, Exercise 6.15
S, infinite sequences of real numbers
We then look at vector spaces which are not Euclidean spaces (not Rn ). Two
of these will be needed for the calculus part of the course: the vector space
F of functions from R to R, and the vector space S of infinite sequences.
Section 4.3
Section 5.2.4 (If A is m n, the range is a subspace of Rm .)
Range
Null space, rank
In this lecture we first look at vector spaces associated with a system of linear
equations; that is with the matrix A when the linear system is written in
matrix form as Ax = b. For this you need to read Chapter 4.
Le ture 10 | Linear spa es
If A
is an n n matrix, then the following statements are equivalent.

A1 exists.
Ax = b has a unique solution for any b Rn .
Ax = 0 has only the trivial solution, x = 0.
The reduced row echelon form of A is I.
|A| =
6 0.
The rank of A is n.
The column vectors of A are a basis of Rn .
The rows of A (written as vectors) are a basis of Rn .
Main Theorem (Theorem 6.28)
This is a good time to make sure you are familiar with all the material you
have seen in Chapters 16. Here is the main theorem for n n matrices.
Extra notes
Exercises 6.126.14
Section 6.5.2
The rank-nullity theorem, (Dimension theorem for matrices)
Section 6.5.1
Basis and dimension of row space, column space and null space
Section 5.3.1; (Review Sections 2.4.2 and 5.2.4)
Row space, column space and null space
Le ture 11 | Rank and nullity
Exponential form and Eulers formula

Section 13.1.6
Section 13.1.5
Polar form of complex numbers, DeMoivres theorem
Sections 13.1.4
The complex plane
Section 13.1.3
Roots of polynomials
Algebra of complex numbers

Section 13.1.2
Sections 13.1, 13.1.1
Complex numbers
This lecture covers the basics of complex numbers, Section 13.1 of the text.
You will need these later for calculus, for solutions of linear differential
equations and linear difference equations with constant coefficients.
Le ture 12 | Complex numbers
Exercise 7.2
Section 7.1.7, Theorem 7.20 only.
A linear transformation is determined by what it does to a basis.
Section 7.1.6
Inverse linear transformations
Section 7.1.5
Composition and combinations of linear transformations
Sections 7.1.4
Identity and zero linear transformations
Section 7.1.3
Linear transformations on R2 rotation
Section 7.1.2
Linear transformations and matrices
Section 7.1, 7.1.1
Linear transformations
Le ture 13 | Linear transformations
D(f ) = f
f : x f (x)
D : C (R) C (R),
f (x) C (R)
D(f ) = (f ) = f = D(f )
D(f + g) = (f + g) = f + g = D(f ) + D(g)
is a linear operator (linear transformation) on this space, since
Differentiation,
Examples. ex , sin x, x3 + 2x + 4 C (R)

C (R) is a subspace of the vector space of all real functions, f : R R
(1) addition
f + g : x 7 f (x) + g(x)
(2) scalar multiplication
f : x 7 f (x)
since these are again differentiable functions.
Let C (R) denote the set of continuously differentiable functions (functions

for which derivatives of all orders exist).
Linear transformations in calculus

This topic is not covered in the linear algebra text, so we include some notes.
It can be found in the calculus text Binmore/Davies, Section 14.3.
Exercises 7.1, 7.37.6
Section 7.2.2
Rank-nullity theorem for linear transformations

(Dimension theorem for linear transformations)
Section 7.2.1
Range and null space (kernel)
of a linear transformation
Le ture 14 | Range and null spa e (kernel)
(D 5D + 6)y = 0
e2x and e3x (check this), which are linearly independent
all solutions are a linear combination of these two functions.

The general solution is y(x) = ae2x + be3x , a, b R.
has solutions
functions.
d2 y
dy
+ 6y = 0
5
2
dx
dx
Example. The differential equation of order two,
By a theorem in calculus, this subspace has dimension n (where n is the

order of the highest derivative in the equation), so has a basis of n functions.
The set of all solutions of this homogeneous differential equation is the null
space of the linear transformation, and is therefore a subspace of C (R).
and is called a linear homogeneous differential equation with constant

coefficients.
A function y = f (x) which satisfies this equation is a
solution of the differential equation.
dn1 y
dy
dn y
+ an1 n1 + a1
+ a0 y = 0
n
dx
dx
dx
This can be rewritten as
(Dn + an1 Dn1 + + a1 D + a0 )y = 0
The null space of this map consists of all functions y(x) for which
is a linear operator on the vector space C (R).
Dn + an1 Dn1 + + a1 D + a0 I
= any polynomial expression in D,
(D2 3D)(f )) = f 3f
and so is any linear combination of these operators, such as
D2 (f ) = D(D(f )) = D(f ) = f
Moreover, so are compositions of this linear operator, such as
y0 = 0,
is a linear operator on this vector
(E 2 5E + 6)yt = 0
2t and 3t (check this), which are linearly independent
all solutions are a linear combination of these two functions.

The general solution is yt = a2t + b3t , a, b R.
has solutions
functions.
yt+2 5yt+1 + 6yt = 0
Example. The difference equation of order two
The fact that this null space is an n-dimensional subspace of S (so has a
basis of n sequences) is another calculus theorem.
yt+n + + 1 yt+1 + 0 yt = 0
which can be rewritten as,
(E n + an1 E n1 + + a1 E + a0 )yt = 0
is again a linear operator on S, and the null space of this linear transformation is a subspace consisting of all solutions to the linear homogeneous
difference equation of order n,
E n + an1 E n1 + + a1 E + a0
Example
yt = t 0, 1, 2, 3, . . .
E(yt ) = yt+1 = t + 1 1, 2, 3, 4, . . .
In the same way as with D in C (R), any linear combination,
The shift operator, E : E(xt ) = xt+1

space. (check this)
The set S of all sequences, xt , is a vector space:

(1) addition
xt + yt : t 7 xt + yt
(2) scalar multiplication
xt : t 7 xt
which are again sequences. The ten axioms of a vector space hold.
A sequence can also be given by a recursion formula and a starting value;

yt = yt1 + 1
y0 = 0 is precisely the same sequence, 0, 1, 2, . . ..
A sequence is a function, y : {0, 1, 2, 3, . . .} R.

Example. yt = t The first few terms of this sequence are
y1 = 1, y2 = 2, and so on.
Suppose B = {v1 , v2 , . . . , vn } is a basis of Rn and v Rn .

What is meant by the coordinate vector [v]B ?
Derive the relationship v = P[v]B . What is the matrix P?

Exercise 7.7
Why change basis?

Example 7.42, page 234 (Describe a linear transformation example)
Section 7.4.1 Change of basis and linear transformations

Section 7.4.2 Similarity
Similarity
Section 7.3.2 Change of basis as a linear transformation

Example 7.40, pages 232234 (Sketch an ellipse example)
Why change basis?
Section 7.3.3
Change of coordinates from basis B to B
Sections 7.3, 7.3.1
Change of coordinates
You should read all of Sections 7.3 and 7.4 (although you do not need to
look closely at the general case given in Section 7.4.1 as Theorem 7.36). You
can do this in the order of the text, or in the order in which the material is
covered in the lecture as given below
Le ture 15 | Change of basis
|
v2
|
|
vn .
|
P is also the matrix of the linear transformation, T : ei 7 vi ,
i = 1, . . . , n.
P1 is the transition matrix from standard coordinates

to coordinates in the basis B.
P is the transition matrix from B coordinates to standard coordinates.
Since the columns of P are a basis of Rn , P1 exists, [v]B = P1 v.
|
P = v1
|
v1 , v2 , . . . , vn are the columns of the matrix P,
The vectors v and v1 , v2 , . . . , vn are given as components in Euclidean space;

that is, in coordinates in the standard basis.
v is a any vector in Rn .
B = {v1 , v2 , . . . , vn } is a basis of Rn .
and understand how and why it works. You need to know exactly what is
meant by [v]B , which is the coordinate vector of v in the basis B (page 224).
v = P[v]B
To change basis in Rn , remember the relationship
Extra notes
Here is a summary of the main points of the lecture.
Write down the definitions of eigenvalue and eigenvector; that is, if A is

an n n matrix, what does it mean to say that the real number is an
eigenvalue of A, and that v is a corresponding eigenvector?
What does it mean to say that an n n matrix A can be diagonalised?
What do you need to do to diagonalise a matrix A?
Exercises 8.1, 8.2, 8.4, 8.68.8
Section 8.2.2 If and only if A has n linearly independent eigenvectors.

Section 8.3.2 If the matrix A has n distinct eigenvalues.
Section 8.3.3, Example 8.35 An example of a matrix without distinct
eigenvalues which can be diagonalised.
When is diagonalisation possible?
Section 8.2.1 Diagonalisation

Section 8.2.2 General method
Diagonalisation
Section 8.1.3 Eigenspaces
Section 8.1.2 Finding eigenvalues and eigenvectors
Section 8.1.1 Eigenvalues and eigenvectors
Eigenvalues and eigenvectors
This is perhaps the most important topic in the linear algebra part of this
course. It brings together everything we have done so far, from solving
systems of linear equations to linear transformations and change of basis.
In this lecture we do not exactly follow the order in which the material is
covered in the text, but in this and the next lecture we will cover most of
Chapter 8.
Le ture 16 | Diagonalisation
(5) If initial conditions are given, substitute these into the general solution
found in step (4) to obtain the unique solution satisfying the conditions.
(4) Substitute this solution z into the equation y = Pz to obtain

the general solution y = y(t).
(2) Set y = Pz (change coordinates). Then substitute for y in the equation

y = Ay to obtain the equivalent equation z = Dz (new coordinates).
(3) Find the general solution of the system z = Dz.
(1) Diagonalize the matrix A: find D (diagonal) and P (invertible),

such that D = P1 AP. Use P as a change of basis (to what basis?).
Below is a summary of the technique of using diagonalisation to solve a

linear system of differential equations of the form y = Ay, where y = y(t).
Extra notes
Exercises 8.3, 8.5, 8.10; Exercises 9.1, 9.79.9
Section 9.3
Application II Linear systems of differential equations
Section 9.1
Application I Powers of matrices
Section 8.2.4
Similar matrices
Section 8.2.3
Geometric interpretation
Section 8.3 (page 263), Section 8.3.1
When is diagonalisation not possible?
Le ture 17 | Appli ations of diagonalisation
kvk =
hv, vi
cos =
hv, wi
.
||v|| ||w||
page 342. Read the central paragraph:Consider . . . . . . of squares, where

it is shown: If PT AP = P1 AP = D and x = Pz, then xT Ax = zT Dz.
Section 11.2.4 Quadratic forms in R2 : conic sections. There may be a
typing error on page 355 in the expression for cos . This line should be
Application III - sketching conic sections in R2
Section 11.1.2 (A is orthogonally diagonalisable A is symmetric)

Section 11.1.3 Matrices with distinct eigenvalues
When is orthogonal diagonalisation possible?
Orthogonal diagonalisation of symmetric matrices

Sections 11.1, 11.1.1
Section 10.3.2 Orthonormal sets
Section 10.3.1 Orthogonal matrices
Section 10.2.3 Orthogonality and linear independence
Section 10.1.1 The inner product of vectors in Rn (review)
Orthonormal sets Orthogonal matrices
Quadratic forms
Section 11.2.1
In Lectures 18 and 19 we study orthogonal diagonalisation and some of its

applications. The reading is in Chapters 10 and 11 of the text, but these
contain material which is beyond the scope of this course, so we will not
cover all sections. For example you should omit Section 11.1.5 (the proof of
Theorem 11.5).
Le ture 18 | Orthogonal diagonalisation I
Make sure you understand how and why this works.
(4) Change coordinates: Set x = Pz.

Then substitute Pz for x in xT Ax to obtain xT Ax = zT Dz.
(5) Sketch the standard form conic zT Dz = 1 on the new X, Y axes.
(By step (3), the new X, Y axes are a rotation anticlockwise of the
standard x, y axes.)
(3) Set these eigenvectors as the columns of a matrix P so that the linear
transformation defined by P is a rotation anticlockwise.
P orthogonally diagonalises A: PT AP = P1 AP = D.
(1) Express the curve as xT Ax = 1, where A is a symmetric matrix.

(2) Find an orthonormal basis of R2 consisting of eigenvectors of A.
(Know that this is possible because A is symmetric.)
Extra notes Below is a summary of how to use orthogonal diagonalisation

to sketch a conic section ax2 + 2cxy + by 2 = 1 in the xy-plane.
If A is an n n symmetric matrix, what does it mean to orthogonally

diagonalise A? What do you need to do to orthogonally diagonalise A?
What does it mean to say that the set {v1 , v2 , . . . , vn } is orthonormal?

What is meant by: {v1 , v2 , . . . , vn } is an orthonormal basis of Rn ?
What do you know about the column vectors of an orthogonal matrix?
What is the definition of an orthogonal matrix?

Exercises 11.811.11
How do you determine if a matrix A is positive definite, negative definite

or indefinite? What does this tell you about the quadratic form xT Ax?
What can you deduce if |A| = 0?
Will the eigenvalues of A always tell you the definiteness of xT Ax? Why?
Exercises 10.6, 11.111.7
Section 11.2.2 (Omit Section 11.2.3, the proof of Theorem 11.32.)

Section 8.1.4, page 253. This section contains the result (Theorem 8.11)
that the determinant of an n n matrix equals the product of its
eigenvalues.
Application IV definiteness of quadratic forms
Section 11.1.4
Orthogonal diagonalisation when eigenvalues are not distinct
Section 10.4
Gram-Schmidt orthonormalisation process
Le ture 19 | Orthogonal diagonalisation II
Answer the following questions.

1. Given a linear system of difference equations xt+1 = Axt , state the two
properties the matrix A must satisfy for it to be the transition matrix of
a Markov chain.
2. What can you deduce about the eigenvalues of A, if A is the transition
matrix of a Markov chain?
3. Explain why a Markov chain has a long term distribution.
How do you find the long-term distribution?

Exercises 9.29.5
Markov chains
Section 9.2.6
Section 9.2.4 (optional)

This section and Section 9.2.5 are optional extra reading. The method of
solving linear systems of difference equations by change of variables is
analogous to the method we use for solving linear systems of differential
equations. However for difference equations, this method turns out to be
just another way of looking at the solution using matrix powers.)
Section 9.2.3 Solving using matrix powers
Solutions of linear systems of difference equations
Sections 9.2.1 and 9.2.2
Systems of difference equations
Markov chains are special examples of linear systems of difference equations.

You should be able to solve any system of difference equations of the form
xt+1 = Axt where the matrix A can be diagonalised, and you should know
the special properties that make such a system a Markov chain.
Le ture 20 | Markov Chains
MA 100
Exercises
Exercise sets contain starred questions to hand in and un-starred questions for
practice. Work all the exercises, both starred and un-starred, before your class.
Any difficulties you may have had with the solutions should be discussed in class.
Answers to the un-starred questions can be found on the MA 100 Moodle page so
that you can check your work. Only the answers are provided, not the solutions.
Full solutions of all questions (starred and un-starred) will appear on the MA 100
Moodle page after the questions have been discussed in the classes. It is very
important for you to check your own solutions to all questions for each exercise
set as soon as these appear.
Questions appearing with the symbol are optional extras. They will not be
discussed in class, unless time permits.
2013/14
c LSE 2013

Exercise Set 1
Hand in solutions to the starred questions to your class teacher.
1. Are the lines with equations

x
1
3
y = 2 + s5
z
2
7
and

x
7
6
y = 12 + t 10
z
16
14
parallel or are they coincident (meaning: are they the same line)?
2. Which of the following sets of points in R3 are collinear that is, lie on a line?
(a)
(2, 1, 4), (4, 4, 1), (6, 7, 6)
(b)
(1, 2, 3), (4, 2, 1), (1, 1, 2).
* 3. Let 1 be the line with equation (x, y, z)T = (1, 1, 2)T + t(1, 2, 3)T , 2 the line through
(5, 7, 4) and (8, 13, 3) and 3 the line through (1, 17, 6) parallel to the vector (7, 4, 3)T .
Show that
(a) 1 and 3 are skew, (b) 1 and 2 intersect (c) 2 and 3 intersect.
Find the points of intersection.
Determine whether each pair of intersecting lines is orthogonal. If not, find the angle
between them.
4. Given the matrices:
2 1
A = 1 1
0 3
1
b= 1
1
1 2
C = 3 0
4 1
1
1
1
0
D = 2
6
1
5
3
Which of the following matrix expressions are defined? Compute those which are defined.
(a) Ab
(b) CA
(c) A + Cb
(d) A + D
(f ) DAT + C
(g) bT b
(h) bbT
(i) Cb
(e) bT D
5. If a and b are both column matrices of the same size (n 1), what is the size of the
matrix product aT b? What is the relationship between aT b and bT a?
6. Using the definition of the inverse of a matrix and the rules of matrix algebra, prove that
if A and B are invertible matrices of the same order, then AB is invertible and
(AB)1 = B1 A1 .
(
* 7. (a) Let B =
)
(
)
2 5
x y
1
, and set B =
.
0 1
z w
Find B1 by solving the system of 4 equations given by the matrix equation BB1 = I:
(
2
0
5
1
)(
x y
z w
(
=
1
0
0
1
)
.
(b) Use the definition of inverse matrix to show that if A is a 2 2 matrix given by
(
)
a b
A=
with ad bc = 0, then
c d
)
(
1
d b
1
A =
.
(ad bc) c a
Note. The term (ad bc) is the determinant of A, denoted det(A) or |A|,

a b
= ad bc
|A| =
c d
8. Suppose A is an m n matrix, x Rn , and ci is the ith column vector of A. Show that

Ax = x1 c1 + x2 c2 + + xn cn .
(See Theorem 1.38 on page 26 of the linear algebra text.)
9. Look through Background You Should Know and Exercises (see the first two pages of
the Study Pack or look on the website). Work Exercise 3.3.
Objectives.
(i) Know concept of vectors in R2 and R3 , and find a vector equation of a line.
(ii) Know concept of a matrix, its inverse and transpose, and matrix operations.
Make sure you have gone through the Directed Reading and Notes for this week.
Exercise Set 2

x
1
y , 1 = 5
1. Find any points of intersection of the plane
z
1

x
1
1
x
1
1
and the lines

y = 2 +t 0
and
y = 2 + t 2.
z
3
5
z
3
1
2. The slope or gradient of a line in R3 is the change in the height (vertical direction) divided
by
in the horizontal direction. If denotes the angle between the
the
change

vectors
1
1
1
0 and 0 , find tan and deduce the slope of a line in the direction of 0 .
3
0
3

1/2
1/2
1/ 2
1/ 2 .
Repeat the exercise for the pair
and
5
0

1
1 and check your

Write down the slope of a line in the direction of the vector
2
answer by calculating it.

v1
x
= t v2
What condition must the vi s satisfy to make the slope of the line
y
v3
z
equal v3 ?
*3. Write down the general cartesian equation of a vertical plane (parallel to the z-axis), a
non-vertical plane and a horizontal plane (parallel to the x, y-plane) together with their
normal vectors.
Find the cartesian equations of the line of intersection of the plane
5x + y 2z = 6
()
and a general horizontal plane.

How is the normal to () related to the horizontal normal to the line of intersection?
*4. Sketch the graphs of the following conics, and in the case of the hyperbola show the
asymptotes. Find their points of intersection with the coordinate axes.
(i)
(ii)
(x 2)2 + y 2 = 4
(iii)
x2 + 4y 2 = 16
x2 = 15y
(iv)
y 2 2x2 = 1
*5. Write down the augmented matrix for each of the following systems of equations, and use
it to solve the system by reducing the augmented matrix to reduced row echelon form.
Express all solutions in vector form.
x+y+z =2
(i)
2y + z = 0
x + y z = 4
(iv)
x + y + 2z = 2
(ii)
2y + z = 0
x + y z = 0
3x y + z = 0
2x + 3y + 2z = 0
x + 2y + 3z = 0
(v)
x + y + 2z = 2
(iii)
2y + z = 0
x + y z = 2
x1 + x2 + x3 + x5 = 1
3x1 + 3x2 + 6x3 + 3x4 + 9x5 = 6

2x1 + 2x2 + 4x3 + x4 + 6x5 = 5
Interpret the solutions to (i)(iv) as intersections of planes, describing them geometrically.

6. A matrix A is said to be symmetric if A = AT . It is skew symmetric if A = AT .
Given that B is symmetric,
C is skew symmetric find the values of a, b, c, and
r, s, t, u, v, w:
r 3 s
3 a 1
C= t u v
B = 2 5 c
2 6 w
b 8 2
(
7. Suppose A =
a b
c d
)
is a 2 2 matrix such that AB = BA for all 2 2 matrices B.
Show that a = d, b = 0, c = 0.
Hint: If something is true for all 2 2 matrices,
(
) then it is true for any such matrix. Try
1 0
some simple choices for B such as B =
, and calculate AB = BA. Use the fact
0 0
that two matrices are equal if and only if corresponding entries are equal to derive the
conditions on a, b, c and d.
This illustrates that it is rather unusual for matrices to commute. Can you generalise the
result to 3 3 matrices? To n n matrices?
Objectives.
(i) Know 3 equations of a plane.
(ii) Know and sketch conics in standard form parabola, ellipse and hyperbola.
(iii) Express a system of linear equations in matrix form.
Know coecient matrix and augmented matrix.
(iv) Carry out Gaussian elimination accurately to reduced row echelon form (RREF).
Interpret the solution and express it in vector form.
Make sure you have done the reading for this week.
Exercise Set 3
1. (i) Dierentiate the following functions with respect to x:

(
(a) ln
x+
1+
3
x2
)
(b) e sin(x 1)
x
x2 + 1
(c) 2
x +x+1
(d)
x2
x
+x+1
Explain the relationship between the results for (c) and (d).
*(ii) Find the equation of the tangent line to the function
y = ex sin(x3 1)
at the point (1, 0) on the curve.
*2. Sketch the function

f (x) = x3 ex
by nding and classifying its stationary points. Find the quadratic Taylor polynomials
that approximate to the following functions
(1) f (x) = x3 ex about the point x = 3
(2) g(x) = x ln x about the point x = 4.

3. Derive the Taylor series expansions of
ex ,
sin x and cos x
about x = 0.
*4. Solve the following system of equations Ax = b by reducing the augmented matrix to
reduced row echelon form. Find the rank of A. What is the rank of the matrix (A|b)?
Show that your solution can be written in the form p + su1 + tu2 where u1 and u2 are
solutions of the associated homogeneous system. Verify Ap = b, Au1 = 0, Au2 = 0.
x1 x2 + x3 + x4 + 2x5 = 4
x1 + x2
+ x4 x5 = 3
x1 x2 + 2x3 + 3x4 + 4x5 = 7
5. Find the reduced row echelon form and the rank of the matrix
3 2
C = 2 1
6 2
1
1
4
3
0 .
6
(a) If C is the augmented matrix of a system of equations Ax = b, C = (A|b), what

can you say about the solutions? (What are they? What Euclidean space are they in?)
(b) If C is the coecient matrix of a system of homogeneous equations, Cx = 0, what
can you say about the solutions? (What are they? What Euclidean space are they in?)
6. Consider an economy with three industries,

i1 : water
i2 : electricity
i3 : gas
interlinked so that the corresponding consumption matrix is
0.2 0.3 0.2

C = 0.4 0.1 0.2
0
0 0.1
Each week the external demands for water, electricity and gas are respectively,
d1 = $40, 000
d2 = $100, 000
d3 = $72, 000
(a) How much water, electricity and gas is needed to produce $1 worth of electricity?
(b) What should be the weekly production of each industry in order to satisfy all demands
exactly?
7. Conditioning of matrices.
Some systems of linear equations lead to ill-conditioned matrices. These occur if a
small dierence in the coecients or constants yields a large dierence in the solution, in
particular when the numbers involved are decimal approximations of varying degree of
accuracy.
The following two systems of equations represent the same problem, the rst to 2 d.p.
accuracy, the second to 3 d.p. accuracy. Solve them using Gaussian elimination and note
the signicant dierence in the solutions you obtain.
{
{
x +
y
= 51.11
x +
y
= 51.106
x + 1.02y = 2.22
x + 1.016y = 2.218
(If a mathematical formulation of a problem leads to an ill-conditoned matrix, then a
dierent formulation needs to be found.)
Objectives.
(i) Dierentiate one variable functions.
(ii) Appreciate that each Taylor polynomial in question 2 is a quadratic polynomial which
approximates the function at the point, so has the same tangent line and shape (concavity),
indicating if the function has a local maximum or minimum at a stationary point.
(iii) Understand the general solution of Ax = b is x = p + w where p is one particular
solution, Ap = b and w is the general solution of Aw = 0.
(iv) Know null space and rank of a matrix.
(v) Become aware of the Leontief input-output model.
Exercise Set 4
This will be used to introduce you to Maple, a mathematics software program. Your class
during Week 5 will be held in a computer room. (Check your timetable for your room
assignment.)
Before attending the class, work this set of exercises, leaving the sections preceded by
= to do on the computer.
The instructions for the Maple class follow this exercise set.
Make sure you have both Exercise Set 4 and the Maple Instructions with you when you
attend your class.
*1. Find the local maxima and minima of the cost function
C(x) = x3 24x2 + 117x + 784
Find both the marginal cost function and the average cost function and investigate the
relationship between them.
Sketch the graphs of all three functions on a single set of axes.
= Use Maple to graph the three functions (in dierent colours) on the same axes.
*2. For each of the functions z = f (x, y), f : R2 R, given below, draw the contours
f (x, y) = c in the xy-plane for the specied values of z = c. Then also draw the intersection
of the surface z = f (x, y) with the xz plane (y = 0) and the intersection of the surface
with the yz plane (x = 0).
(a) f (x, y) = x 3y + 1
(b) f (x, y) = x2 + 4y 2
(c)
f (x, y) = cos y
(d)
f (x, y) = xy
c = 0, 1, 2, 1, 2
c = 1, 2, 3
1
c = 1, 0, , 1
2
c = 1, 2, 1, 2
For (d) also draw sections with the vertical planes y = x.

Hand in the sketches (label contours) together with the equations you used to make them.
= Use Maple to draw these surfaces in R3 and to look at the contours on the surfaces. Check
your sketches by plotting the level curves (contours) and the vertical sections using Maple.
3. The following fact is sometimes useful in solving optimisation problems. If f (x) is a convex
function, and f has a local minimum at , then is a global minimum of f .
Use this to show that f (x) = x2 cos x has a global minimum at x = 0.
= Use Maple to graph the function and observe the global minimum.
4. Use elementary row operations to nd any inverses of the following matrices.
1 2
*(a) A = 2 3
0 1
3
0
2
*(b)
1 2
B = 2 3
0 1
3
0
6
(c)
1
0
C=
0
0
0
1
0
0
4
0
0
1
0
0
1
0
Is C an elementary matrix? If the answer is yes, what operation does it perform? If

the answer is no, write it as the product of two elementary matrices.
5. Given a system of equations Ax = b, for several dierent values of b, it is often more

practical to nd A1 , if it exists, and to nd the solutions using x = A1 b.
Use this method to solve Ax = br for the matrix A given in the previous question, and
for the following vectors br , r = 1, 2, 3:

1
1
0
*(a) b1 = 0
(b) b2 = 1
(c) b3 = 1 .
3
1
0
Be certain your solution for A1 is correct. (check AA1 = I rst).
6. Consider an economy with three industries: coal, electricity, railways. To produce $1 of
coal requires $0.25 worth of electricity and $0.25 rail cost for transportation. To produce
$1 of electricity requires $0.65 worth of coal for fuel, $0.05 of electricity for the auxiliary
equipment, and $0.05 for transport. To provide $1 worth of transport, the railway requires
$0.55 coal for fuel and $0.10 electricity. Each week the external demand for coal is $50,000
and the external demand for electricity is $25,000. There is no external demand for the
railway. What should be the weekly production schedule for each industry?
Write down the consumption matrix C and the demand vector d for this problem and
formulate it in matrix form.
(Note that the consumption matrix C is most easily compiled by its columns, each column
j being the dierent outputs of all the industries needed by industry j to produce $1 of its
own output.)
= Use Maple to solve the system for the production vector.
Objectives.
(i) Understand concept of two variable function and visualise its graph. Sketch contours
and vertical sections with coordinate planes.
(ii) Use row operations to nd A1 by reducing (A|I) to (I|A1 ).
(iii) Use Maple.
Maple Instructions Computer class on Exercise Set 4

This week you are being introduced to the mathematics software package Maple. In
this class you will use Maple primarily to help you visualise surfaces in 3 dimensions.
The program is a powerful tool for graphics, and in finding real life solutions to
problems which can be solved by the methods covered in MA100. You will be expected
to use it throughout the course.
The current version of Maple is Maple 15. The program is available on all computers in
LSE computer class rooms. (You can buy a copy for your own computer for 10 pounds
from the IT Help Desk. See www2.lse.ac.uk/intranet/LSEServices/itservices/
facilities/software/softwareSales.aspx for more information.)
Before starting this computer class, you should have worked as much of Exercise 4 as
you were able to do.
Start Maple on your computer. Remember to choose Worksheet mode.
Question 2
Carry out the following instructions for the example f (x, y) = x2 + 2y 2 , which you
can then apply to the four given functions.
1. Make sure that Maple is in Math input mode, by clicking on the Math button
towards the top of the Maple window. TYPE
f:= (x,y)-> x^2 + 2*y^2
The arrow is typed as the hyphen - followed by the greater than sign, >. After
youve type ^2, use the right-arrow key to make sure you go back to the baseline
2
when typing the +, otherwise you get x2+2y instead of x2 + 2y 2 .
Take a minute to understand the code you have just typed. You have assigned (:=)
the name f to a function which takes a point (x, y) and maps it (->) to x2 + 2y 2 .
An asterisk * is useful to indicate multiplication. (Maple knows that 2x is the
same as 2*x. But you must type a*b to multiply a and b, as Maple thinks that ab
indicates a variable with name ab.)
Now press the return key. Maple will respond by showing that is has accepted your
definition of the function f :
(x, y) x2 + 2y 2
M1
2. To plot the function as a surface in 3-dimensional space, type and enter the following
input. Maple interprets f (x, y) as the function expression x2 + 2y 2 , and plots the
graph of the surface z = x2 + 2y 2 .
plot3d( f(x,y), x=-5..5, y=-5..5 )
(Maple will ignore spaces in the input; they have been put in so that you will notice
the precise way to type the commands.)
LOOK very carefully at the menu and tool bars at the top of your window. Now
CLICK on the picture of the surface with your mouse. A frame appears around the
surface and the lower toolbar changes into the plot toolbar.
PLACE your cursor (which will look like an arrow) on the surface and hold down the
left mouse button. MOVE the mouse slowly and OBSERVE what happens.
Now experiment by clicking on the different icons on the plot toolbar. Which view
shows the level curves on the surface? Which view allows you to see vertical intersections of the surface?
The plot commands can also be used with a function expression. So you could as well
have entered
plot3d( x^2 + 2*y^2, x=-5..5, y=-5..5 )
Since you need to graph several functions, the function notation is more efficient. (It
will become clear why as you continue.)
3. Before you can plot the contours in the xy plane, you need to instruct Maple to load
the plots package, which will give us extra commands.
Place your cursor in the next Maple input line, type and enter
with(plots)
You will see a list of additional commands which are now available. The one you are
interested in is the contourplot command. To find out for yourself how to use any of
these commands, highlight the name with your mouse and then click on Help in the
menubar. The box that appears will offer help on that command as the fifth option
from the top.
M2
To graph some level curves f (x, y) = c in the xy plane for the function f (x, y), use
contourplot( f(x,y), x=-5..5, y=-5..5 )
To graph a specific set of level curves, say at c = 0, 1, 2, 3, use
contourplot( f(x,y), x=-5..5, y=-5..5, contours=[0,1,2,3] )
(Explain why you see only 3 contours.)
Note the use of square brackets [0, 1, 2, 3] to denote a list in Maple.
The curves in the last graph are not as smooth as the first set. If you want a smoother
set, you have to instruct Maple to plot more points. Place your cursor in the input
line that you have just typed, just after the final square bracket of the list, and add
the option as shown below (on the same line, and dont forget the comma).
contourplot( f(x,y), x=-5..5, y=-5..5,
contours=[0,1,2,3], numpoints=2000 )
Is the scaling on the two axes different, distorting the shape of the curves? To fix this,
either click on the graph and then on the [1:1] icon in the toolbar, or add a scaling
option to your command line:
contourplot( f(x,y), x=-5..5, y=-5..5,contours=[0,1,2,3],
numpoints=2000, scaling=constrained )
Experiment: CLICK on the graph and look at the effects of clicking on different icons
in the plot bar. Experiment how many points you want Maple to compute to obtain
sufficiently smooth curves without wasting computational time and space.
4. To graph the vertical intersection with the xz plane (when y = 0) and with the
yz plane (when x = 0), use the following plot commands:
plot( f(x,0), x=-5..5)
plot( f(0,y), y=-5..5)
To compare the graphs, you might want to see them on the same set of axes, and
specify the colour of each graph.
plot( [f(t,0),f(0,t)], t=-5..5, color=[red,blue] )
Note the square brackets since you are now asking Maple to plot a list of functions.
M3
You are now ready to plot the graphs, contours and vertical intersections of the surfaces
in Question 2. SCROLL back to the beginning. Your Maple Worksheet should
contain the following commands:
(function definition)
f:= (x,y)-> x^2 + 2*y^2

plot3d(f(x,y), x=-5..5, y=-5..5 )
(graph of surface)
contourplot(f(x,y), x=-5..5, y=-5..5 )
(general contours)
contourplot(f(x,y), x=-5..5, y=-5..5, contours=[0,1,2,3],

numpoints=2000, scaling=constrained )
(specific contours)
(xz intersection, y = 0)
plot(f(x,0), x=-5..5 )
(yz intersection, x = 0)
plot(f(0,y), y=-5..5 )
Instead of re-typing these lines for each surface (or using cut and paste), CHANGE the
definition of the function f and then RE-EVALUATE the input lines. For example
to do part (b), place your cursor in the input line defining f and change it to
f:= (x,y)-> sqrt(x^2 + 4*y^2)
When you press return, Maple will replace the previous definition of the function f
with this new one.
Now place your cursor anywhere in the plot3d line and press return. Maple will
respond by graphing the new surface z = f (x, y).
EXPLORE this new surface as before. Rotate it to view it from different angles. Click
on the icons to view the surface with level curves and a set of axes through the origin.
CONTINUE to re-evaluate the other input lines. COMPARE the contours on the
surface with the graph of the general contourplot. ALTER the list in the specific
contours plot to match the set of contours required in the question.
Do the above for all functions in Question 2.
M4
Question 1
First define the cost function:
C := x -> x^3 - 24*x^2 + 117*x + 784
Define the marginal cost function (using the differential operator D which maps the
cost function into its derivative function),
mc := D(C)
Note Maples feedback.
Finally define the average cost function,
ac := x->C(x)/x
Each time that you enter an input line, verify that Maple has the correct definition.
To graph the functions on the same axes use a list,
plot( [C(x), mc(x), ac(x)], x=0.5..20 )
ENLARGE the graph: Click on it and a frame appears. Position the mouse over the
bottom right corner on the frame until it turns into a double arrow. Hold down the
left mouse button as you drag the frame down and right to enlarge it.
You can alter the input to specify the colours yourself, so that you can tell which graph
corresponds to which function, as you did for the vertical intersections of Question 2.
Note that the lowest x value graphed was 0.5 and not 0. What happens if you change
the lower x limit to 0? Why?
Were you able to find the minimum of the average cost function, the point at which
the marginal cost and the average cost are equal? If not, try
solve( mc(x)=ac(x) ,x )
Or look for the minimum value of ac(x):
solve( diff(ac(x),x)=0, x )
Note that Maple returns the real and the complex roots.
(Note the use of diff for the derivative of an expression. For more on this, see the
final section More Maple.)
M5
Question 3
You should be able to plot this function without any difficulty:
plot( x^2-cos(x), x=-3..3 )
Linear Algebra
To learn how to use Maple to solve systems of linear equations, Ax = b, we will work
examples from Question 5 of Exercise Set 2. The equations are
x1 + x2 + x3 = 2
2x2 + x3 = 0
(i)
x1 + x2 x3 = 4
x + x2 + 2x3 = 2
1
2x2 + x3 = 0
(ii)
x1 + x2 x3 = 0
x + x2 + 2x3 = 2
1
2x2 + x3 = 0
(iii)
x1 + x2 x3 = 2
First you need to load the linear algebra package,

with(LinearAlgebra)
Then input the vector b and the coefficient matrix A for (i):
b := Vector([2,0,-4])
A := Matrix( 3, 3, [1,1,1, 0,2,1, -1,1,-1] )
The first two numbers indicate the size of the matrix (3 3). Determine yourself how
the list of elements must be entered.
Obtain the augmented matrix and assign it the name Ab, then find the reduced row
echelon form:
Ab := <A|b>
ReducedRowEchelonForm(Ab)
from which you easily can obtain the solution.
M6
The solution x can also be found directly by

x := LinearSolve(A,b)
You can now modify the Maple definitions of A and b to work parts (ii) and (iii).
DO this to see how Maple handles systems which are inconsistent or which have
infinitely many solutions.
Questions 4, 5, 6
The inverse of a matrix is found as follows:
MatrixInverse(A)
Try to find out what happens if you ask for the inverse of a matrix which is not
invertible.
Matrix multiplication is denoted in Maple by a dot. Once you have as input a matrix
assigned the name A and a vector assigned the name b, the product matrix Ab is
given by
A.b
Use this method to find x = A1 b and to verify that Ax = b.
For Question 6 you need to create a 3 3 identity matrix; assign it the name Id.
Id := Matrix( 3, 3, [1,0,0, 0,1,0, 0,0,1] )
INPUT the consumption matrix C and the demand vector d, and then solve the
system:
LinearSolve(Id-C,d)
More Maple
Whatever question you ask in Maple, if you intend to use the result again, assign a
name to the commandline, using :=. For instance
sola := LinearSolve(A,b);
A.sola - b
M7
functions vs. expressions

To understand the difference, enter and OBSERVE the output of the following lines:
(a function)
f := x -> x^3 + 3*x
(an expression)
g := x^3 + 3*x
f(1); g(1)
(The semi-colon is there to separate the two questions, otherwise Maple thinks you
want to see the product of f (1) and g(1).)
If f is defined as a function, then f(x) is the corresponding expression.
f(x); g
diff( f(x),x );
diff( g,x )
Use diff( expression, x ) to differentiate an expression with respect to the variable x; its output is an expression.
To have the derivative as a function, you need to use the differential operator D; it
is applied to a function and its output is a function .
D(f)
Maple has excellent Help facilities. You can search for topics. If you know the command or topic, but want to know more how to use it, type a question mark followed
by the topic,
?plot, color
or highlight a word with your mouse, click on Help in the menubar, and click on the
fifth option Help on . . . .
If Maple does something unexpected or gives error messages, it is often because it is
using old definitions you have made. For instance, in the above part on Linear Algebra
you used x to denote the solution of Ax = b. This gives surprises if you later try to
do something with a completely different x; say want to solve x2 + x = 0.
To clear all definitions, assignments, etc., enter
restart
M8
Exercise Set 5
z
z
and
.
x
y
are homogeneous, and verify Eulers formula in each case.
1. For each of the three functions below, compute
(a)
x3 + y 3
z=
x+y
(
(b)
z=
1 1/5 4 1/5
x
+ y
5
5
Show that all the functions
)5
(c)
ex/y sin
x
y
+ ey/x cos
y
x
2. *(i) Find the equation of the tangent plane to the surface

z = f (x, y) = 4x2 + y 2 ,
at each of the points (1, 1) and (1, 1).
Find also a normal vector to the surface at each of these points.
At what points, if any, is the tangent plane to the surface z = f (x, y) horizontal?
(The next part of this question will be covered in the next calculus lecture, but you should
be able to answer it before then by thinking carefully about what is being asked.)
Find and sketch the contour 4x2 + y 2 = f (1, 1) = f (1, 1), together with the tangent
line and a normal vector to the contour at each of the points (1, 1) and (1, 1).
(ii) Do all of the above (part (i)) for the function z = g(x, y) = x2 y 2 , and the points
(2, 1) and (2, 1).
3. Determine whether the following permutations are even or odd, and find the total number
of inversions:
(i) (2 4 3 1)
(ii) (1 3 4 5 2).
4. Evaluate the following determinants using the cofactor expansion along an appropriate
row or column.

7 5 2 3
1 2 1 0
2 6 1

2 0 0 0
3 2 1 0

(i) 1 0 5
(ii)
(iii)

11 2 0 0
0 1 6 5
3 1 4

23 57 1 1
0 1 1 1

0

1
(iv)
0

0
1
0
0
0
0
0
1
0

0

0

0

1

0

0

0
(v)
0

0

1
0
0
0
0
6
3
0
0
0
1
9
4
0
0
2
0
8
2
0
3
9
7
7
9

1

2

3

4

5

6

3 t 2

(vi) 1 5 3
2 1 1
*For what values of t is the determinant (vi) equal to zero?
* 5. Use Maple to plot the functions f (x) and g(x) in question 2 of Exercise Set 3 about the
= respective points x = 3 and x = 4, together with the Taylor approximations you found.
Plot the functions in red and the Taylor approximations in blue. Hand in a printout of
the graphs (you can colour it by hand).
6. Let
a1
a = a2 ,
a3
b1
b = b2 ;
b3
e1 = 0 ,
0
e2 = 1 ,
0
e3 = 0
1
The vector product or cross product, a b of the two vectors is the vector defined by

e1 e2 e3

a b = a1 a2 a3 = (a2 b3 a3 b2 )e1 (a1 b3 a3 b1 )e2 + (a1 b2 a2 b1 )e3
b1 b2 b3
The vector a b is perpendicular to both a and b (see part (b)).

1
2
(a) Calculate w = u v for the vectors u = 2 and v = 5 .
3
4
Check that w is perpendicular to both u and v
(b) Show that for general vectors a, b, c R3
is given by

a1

a, b c = b1
c1
the scalar triple product, a, b c

a2
b2
c2

a3
b3
c3
()
Use this and properties of the determinant to show that the vector b c is perpendicular
to both b and c.
(c) Show that the vectors a, b, c are coplanar (lie in the same plane)
the determinant () is equal to zero.
2
t
3
Obtain the constant t if the vectors 1 , 5 , 3 are coplanar.
1
1
2
Objectives.
(i) Master partial dierentiation
(ii) Be able to check whether a function is homogeneous, know Eulers theorem.
(iii) Draw contours of surfaces, tangent lines and normals.
Find equations of tangent planes, contours and tangent lines, find normal vectors.
(iv) Evaluate a determinant by cofactor expansion.
Exercise Set 6
* 1. Roughly sketch a contour map of the function h : R2 R dened by
h(x, y) = 4000 0.001x2 0.004y 2
which represents the surface of a mountain. Suppose a mountain climber is at the point
(500, 300, 3390). In what direction should he move in order to ascend at the greatest rate?
What is this rate of ascent?
There is a path through the point in the direction (3, 4)T . Find its rate of ascent.
For each case, sketch a vector along the path of the climber.
2. A rm producing two goods x and y has the prot function

(x, y) = 32x 6x2 + 8xy + 16y 3y 2 20.
Find the level of output for each of the two goods for which x (x, y) = y (x, y) = 0.
* 3. The production function P (k, l) = 3k 1/3 l2/3 represents the output of a rm if it uses k
units of capital and l units of labour. Sketch some contours P (k, l) = c.
At present, k = 8 and l = 1. Find a normal vector to the contour P (k, l) = 6 at the point
(8, 1). At what rate would production increase if the rm increases k and l at the same
rate? In what ratio should it increase k and l in order to increase production most rapidly?
Find P for a general Cobb-Douglas function, P (k.l) = Ak l (for constants A, , ).
4. Evaluate the following determinants (use row

1
1 4 3 2

1
2 7 5 1

*(a)
(b) 2

2
6 0
1

2
2 10 14 4

1
(
5. For which values of is the matrix A =
operations to simplify the calculation):

4 1 3 0

3 3a 3a2
7 4 3 8

8 2 6 0
(c) 2 2b 2b2

1 c
0 5 5 7
c2

9 0 9 2
2
3
2
1
)
not invertible?
* 6. A is 3 3, det(A) = 5. Find det(3A), det(A2 ), det(2A1 ), det((2A)1 )).

7. Use the method of the adjoint matrix to nd
2 0 3
1
A= 0 3 1
B = 2
1 4 2
0
the inverse of
0 2
1
1 3
C = 0
1 1
2
2
1
1
0
1
1
8. Use Cramers rule to nd the value of x, y, z for system (a) and to nd the value of z
for system (b) (where a, b are constants, a = b, a = 2b).
x+y+z =8
ax ay + bz = a + b
(a)
2x + y z = 3
(b)
bx by + az = 0
x + 2y + z = 3
ax + 2by + 3z = a b
9. A portfolio is a row vector Y = ( y1 . . . ym ) in which yi is the number of units

of asset i held by the investor.
After a year, say, the value of the assets will increase
(or decrease) by a certain percentage. The change in each asset depends on states the
economy will assume, predicted by economists as a returns matrix, R = (rij ) where
rij is the factor by which investment i changes in one year if state j occurs.
Suppose an investor has assets in y1 = land,
1.05
R = 1.05
1.20
y2 = bonds and y3 = stocks, and that
0.95 1.0
1.05 1.05
1.26 1.23
Then the total values of the portfolio in one years time are given by YR where (YR)j
is the total value of the portfolio if state j occurs.
(a) Find the total values of the portfolio W = ( 5000
the possible states.
(b) Show that U = ( 600 8000
value in all states j.
2000
0 ) in one year for each of
1000 ) is a riskless portfolio, that is, it has the same
An arbitrage portfolio Y = ( y1 . . . ym ) is one which costs nothing (y1 + +ym = 0),

cannot lose ((YR)j 0), and in at least one state makes a prot ((YR)j > 0 some j).
(c) Show that Z = ( 1000 2000 1000 ) is an arbitrage portfolio. (The bond asset of
2000 indicates that this sum was borrowed from the bank).
Objectives.
(i) Understand the idea of the gradient and the derivative of a function of two variables,
and be able to evaluate these.
(ii) Understand the concept of directional derivative for a two variable function, know its
formula, f, u, and the direction and value of maximum increase/decrease.
(iii) Understand the eect of row operations on a determinant and use them to simplify
its evaluation.
(iv) Use |A| to determine if A is invertible or what solutions Ax = b might have.
(v) Find A1 by cofactor method and use Cramers rule for solving systems of n equations
in n unknowns.
Exercise Set 7
* 1. Let f (x, y, z) = 3x2 + 2y 2 z 2 .
Find
f
f f
,
and
.
x y
z
(a) What is the tangent hyperplane to the hypersurface u = 3x2 + 2y 2 z 2

where x = y = z = 1?
(b) Find the equation of the plane that is tangent to the surface
3x2 + 2y 2 z 2 = 13
at the point (2, 1, 1)
Write down its normal vector.

(Have a look at this surface on Maple. It is called a hyperboloid of one sheet.)
* 2. Find all the stationary points of the functions, f : R2 R dened by:
(a) f (x, y) = xy 2 + x2 y xy
(b) f (x, y) = x3 + y 3 3xy 8
3. Find all the stationary points of the function, f : R3 R given by
(
1)
f (x, y, z) = ln (x3 2y 3 3z 3 6xy + 4) 2
* 4. Which of the following are subspaces of R3 ?
x
x

S2= y x2 + y 2 + z 2 = 1
S1= y x + y + z = 0

z
z

S4 =
y xy = 0

S3= y x = 0
x
0
S5 =
y x = 0 and y = 0 =
0 zR
z
z
Provide proofs or counterexamples to justify your answers.
Describe the sets geometrically.
5. If A is an m n matrix, show that the null space, N (A), is a subspace of Rn .
2
1
* 6. Express each of the vectors below as a linear combination of u = 1 and v = 1
1
3
if possible:
0
7
3
(a) 2
(b) 0
(c) 5
4
0
7
7. (a) Solve the matrix equation to nd the coecients and by nding A1 :

(
2
5
)( )
) (
(
( )
1 1
1
1
= Ax
=
+
=
2 1
2
1
{( ) (
)}
1
1
(b) Show that lin{w1 , w2 } = lin
,
= R2 . That is, show any vector b R2
2
1
can be expressed as a linear combination of w1 and w2 by solving b = Ax for x:
(
b1
b2
) (
( )
(
1
1
1
+
=
=
1
2
2
1
1
)( )
( )
a
(c) Show, in general, that if v and w are non-zero vectors in R , v =
,
c
2
lin{v, w} = R
v = tw for any t R
( )
b
w=
d

a b

c d = 0
Objectives
(i) Extend concepts of three dimensional vectors to n dimensions.
(ii) Understand concept of a function of three or more variables; Find its
derivative, tangent hyperplanes to the hypersurface and contour, gradient and directional
derivatives. Find tangent plane to a contour of the graph of a 3-variable function.
(iii) Find stationary points of functions of two or more variables. (Systematically solve the
equations simultaneously; factorise expressions where possible and delete non-zero factors
before solving.)
(iv) Understand concept of a linear subspace: Decide whether a set is a subspace and prove it
is or use a counterexample to show that it isnt.
(v) Express a vector as a linear combination of given vectors, or show that it is not possible.
Exercise Set 8
1. The function f : R3 R2 with component functions f1 and f2 is dened by
u = f1 (x, y, z) = x2 + y 2 + z 2 ;
v = f2 (x, y, z) = x y.
Find all the points x = (x, y, z)T such that f (x) = (8, 0)T , and describe the curve consisting
of these points.
= Use Maple to verify your answer.
(
* 2. Let
A=
3 0
1 1
)
,
( )
1
c=
.
0
The ane function f : R2 R2 is dened by f (x) = Ax + c.

Write down the equations of two hyperplanes whose intersection is the graph of f .
3. A function f : R R2 with component functions f1 and f2 is dened by
}
}
x = f1 (t) = 2t
x = f1 (t) = sin t
(a)
*(b)
y = f2 (t) = 4t2
y = f2 (t) = ln t
For the function in (a) nd the set of (x, y) which correspond to the set of all real values of
t. Find the derivative at t = 1, then write down the equation of the tangent line at t = 1.
For the function in (b) nd the set of (x, y) which correspond to the set of all positive real
values of t. Find the derivative at t = /2, and the equation of the tangent line at t = /2.
4. Find the derivatives at the given point of the functions dened below. Write down the
equations for the corresponding tangent ats in each case.
(a) f : R2 R2 with component functions f1 and f2 dened by
}
u = f1 (x, y) = xy
at (x, y) = (1, 1).
v = f2 (x, y) = x y
*(b) f : R2 R3 with component functions f1 , f2 and f3 dened by
u = f1 (x, y) = x
v = f2 (x, y) = x + y
w = f3 (x, y) = xy
at (x, y) = (1, 1).
(c) f : R3 R2 with component functions f1 and f2 dened by

u = f1 (x, y, z) = xy
v = f2 (x, y, z) = yz
}
at (x, y, z) = (1, 1, 1).
* 5. Show that the set S1 spans R3 , but any vector v R3 can be written as a linear combination
of the vectors in S1 in innitely many ways. Show that S2 and S3 do not span R3 .

1
0
1
2
1
2
1
1
1
S1 =
2
0
1
1
S2 =
0
1
2
S3 =
1
0
3
1
1
0
1
3
9
1
1
6. Determine which of the following sets of vectors are linearly independent.

{( ) ( )}
{( ) ( ) ( )}
{( ) ( )}
1
1
1
1
2
0
1
L1 =
L2 =
L3 =
2
3
2
3
5
0
2

1
2
1
2
3
0
1
2
L4 = 2 7 5
L5 =

1
1
1
0
0
0
3
2
2

4
1
4
2
1
2
1
2 0 4
2 0 41
L6 =
L7 =

1
1
1
1
1
1
1
8
2
2
8
3
2
3
* 7. Determine which of the following sets of vectors spans R3 , and which of the sets is linearly
independent.

1
2
1
2
1
1
0
1
0
S5 =
2
1
S4 =
2
1
4
1
5
4
3
* 8. Let A be any matrix. Let v1 and v2 be 2 non-zero vectors and suppose that Av1 = 2v1
and Av2 = 5v2 . Prove that {v1 , v2 } is linearly independent.
(Hint: Assume 1 v1 + 2 v2 = 0; multiply this equation through by A to get a second
equation for v1 and v2 ; then solve the two equations simultaneously).
Can you generalize this result?
Objectives.
(i) Understand concept of a vector-valued function, f : Rn Rm , nd derivatives and
tangent ats. Know that a tangent at of a function f : R Rm is a line.
(ii) Understand the concept of an ane function.
(iii) Determine whether a set {v1 , , vr } is linearly independent or spans a given space.
Understand these concepts in terms of linear combinations of the vectors.
Exercise Set 9
1. The length and width of a rectangle decrease at the rate of 2 cm per minute and 3 cm per
minute respectively. When the length is 6m and the width is 3m how fast are the following
changing:
(a) the area
and
*(b) the diagonal?
2. Suppose that z = f (x, y) is a function of two variables x and y, and that y depends on x
dz
via the function y = g(x). Write down an expression for the derivative
.
dx
dz
Set z = xy 4 + 3y 2 + ln y and y = x2 + 3x + 10.
Find
.
dx
(Your answer can involve y.)
*3. Find the second derivative of the real-valued function f : R2 R defined by
f (x, y) = ex sin y
Write down the quadratic Taylor approximation at the point (x, y) = (1, 0).
4. Recall that a function f is homogeneous of degree n if

f (x, y) = n f (x, y).
Fix x and y and think of f as a function of . Use the Chain Rule to dierentiate the above
equation with respect to . Put = 1 to obtain Eulers Theorem:
xfx (x, y) + yfy (x, y) = nf (x, y).
*5. For each of the sets of vectors given below, find a basis of the vector space Lin(Si ) and state
its dimension. Describe any proper subspaces of R2 or R3 geometrically (give Cartesian
equations for any lines or planes).
{( ) ( )}
1
2
S1 =
2
3
{(
S2 =

2
1
1
S3 =
0
1
2
1
3
9
1
1
)( )(
)(
)}
0
2
3
0
2
3

1
2
4
1
2 0 41
S4 =

1
1
1
1
3
2
8
2
*6. Which of the following sets are a basis for R3 ? (State reasons for your answers.)

2
4
7
1
S1 = 2 1 1 2
3
0
0
1

1
3
2
S3 =
1
2
3
1
1
0
1
1
S2 = 0 1
1
1

1
3
2
S4 =
1
2
3
1
1
1
1
*7. Find the coordinates of the vector 2 with respect to each of the bases for R3 :
1

0
0
1
B1 = 0 1 0
0
0
1
1
2
1
B2 = 1 1 3
1
0
3
8. Find a basis for each of the following subspaces of R3 .

(a) The plane x 2y + z = 0
(b) The yz plane
9. If v1 , v2 Rn , explain the dierence between the sets A = {v1 , v2 } and B = Lin{v1 , v2 }.
Objectives.
(i) Know how chain rule formula applies to functions of several variables and use it to find
derivatives of composite functions.
(ii) Find the Taylor approximation at a point for a function of several variables.
(iii) Show that a set of vectors is a basis of a vector space V Rn .
(iv) Find a basis of any subspace V of Rn .
(v) Know that any vector in V is a unique linear combination of the basis vectors, and be
able to find the coordinates of any vector in the basis.
Exercise Set 10
1. The total cost of producing x units of a product and y units of another product is given by
the function
C(x, y) = 5x2 + 7y 2 2xy 46x 86y + 600
Find the production level that minimizes this cost and the corresponding cost of
production.
*2. For each of the following functions, find all the stationary points and classify them as local
maxima, local minima, or saddle points.
(a) f (x, y) = 1 y 3 3yx2 3y 2 3x2
(b) f (x, y) = 4xy x4 y 4
(c) f (x, y) = ex sin y
(d) f (x, y) = 4x2 ey 2x4 e4y
= Use Maple to plot the contours of the functions near their stationary points.
Print out the contours graph of (a) and mark the stationary points in red.
3. Find the general solution of each of the following systems of equations Ax = b in the form
x = p + 1 s1 + . . . + r sr where p is a particular solution of the system and {s1 , . . . , sr } is a
basis for the null space of A.
*(a) x1 + 2x2 + 3x4 + x5 = 4,
*(b)
x1 + x2 + x 3 + x4 = 4
2x1 + x3 x4 = 2
2x2 + x3 + 3x4 = 6
x3 + 2x4 + x6 = 1,
(c)
x7 = 2
x1 + 2x2 x3 x4 = 3
x1 x2 2x3 x4 = 1
2x1 + x2 x3 = 3
*4. For each of the following systems of equations Ax = b, determine the set of all vectors b for
which the system is consistent. In each case the set will be a subspace of R3 ; describe what
that subspace is.
(b)
x1 + x 2 + x3 + x4 = b 1
2x1 + x3 x4 = b2
2x2 + x3 + 3x4 = b3
(c)
x1 + 2x2 x3 x4 = b1
x1 x2 2x3 x4 = b2
2x1 + x2 x3 = b3
5. (a) Convince yourself that F(R) is a vector space. Which of these sets are subspaces:
S1 = {f F(R) | f (0) = 1}
S2 = {f F(R) | f (1) = 0}
(b) Show S3 = {f C 1 (R) | f f = 0} is a subspace.

6. Which of the following subsets are subspaces of M2 (R)?
{(
)
}
{(
)
}
a 0
a 1
W1 =
| a, b R
W2 =
| a, b R
0 b
1 b
{(
W3 =
a2
0
0
b2
}
| a, b R
Objectives.
(i) Classify stationary points of functions of two or more variables.
(ii) Recognise the subspaces associated with a system of equations Ax = b:
the null space of A, N (A) = {x | Ax = 0} Rn ,
the range of A, R(A) = {Ax | x Rn } = {b | Ax = b is consistent } = CS(A) Rm .
(iii) Know Ax = b consistent general solution is x = p + 1 s1 + 2 s2 + + k sk
where Ap = b and {s1 , s2 . . . , sk } is a basis of the null space of A.
(iv) Understand functions can be treated as vectors, e.g. C (R) is a vector space.
Exercise Set 11
*1. Show that the function

f (x, y, z) = x2 + 5y 2 + z 2 + 2xy + Ayz
is convex if the positive constant A is less than 4, but not convex if A > 4.
Identify one stationary point of f , and show that this is the global minimum of f if A < 4.
For A > 4, evaluate f (t, t, 2t) for any t, and deduce that f has no global minimum.
2. (a) Find the production levels x and y which minimize the cost function C : R2 R
dened by
C(x, y) = 4x2 + 4y 2 2xy 40x 140y + 1800
for a rm producing two goods x and y.
(b) Find the production levels x and y which minimize this cost function subject to each
of the following production requirements.
(i)
x + y 25
(ii) x + y 35
For the following three questions consider the matrices
1
A = 1
0
0 1
1 1
1 0
1 2
B = 0 1
1 3
1
1
2
3
1
0
0
1
1
3. (a) Find a basis of the row space of A, RS(A), a basis for the column space of A, CS(A),
and nd the rank of A, (A).
(b) Show that the RS(A) and CS(A) are each planes in R3 . Find Cartesian equations
for these planes and hence show that they are two dierent subspaces.
(c) Find the null space of A, N (A), and verify the dimension theorem. Show that the
basis vectors of the null space are orthogonal to the basis vectors of the row space of A.
*4. (a) Find a basis of the row space of B.

(b) Find a basis of the column space of B.
(c) Find a basis of the null space of B.
(d) Find the rank of B and verify the dimension theorem for matrices. Show that the basis
vectors of the null space are orthogonal to the basis vectors of the row space of B.
*5. (a) Using the information you obtained in questions 3 and 4, and without solving the
system, determine for each of the matrices A and B above if the system of equations
Ax = b or Bx = b is consistent, for each of the given values of b:
1
b1 = 1
2

2
b2 = 1
3
(b) If the system is consistent, then nd the general solution in each case. Hence or
otherwise express b as a linear combination of the column vectors of the given matrices.
def
6. A matrix A has full column rank the columns of A are linearly independent.
If A is an m k matrix with full column rank, show that:
(1) AT A is a symmetric k k matrix,
(2) AT A is invertible.
Then verify the above results for the matrix
1 0
M = 0 2 .
1 1
Objectives.
(i) Optimise a several variable function constrained either within, or on the boundary of
a region.
(ii) Understand convex and concave function and investigate if a function is one of these.
Understand the connection with global optimisation.
(iii) Know coecients xi in b = x1 v1 + + xn vn = components of x in Ax = b,
where columns of A are the vectors vi .
(iv) Know rank(A) = number of leading ones in RREF(A) = dim(RS(A)) = dim(CS(A)).
Know CS(A) = R(A) (Column space and range are the same subspace.)
(v) Find a basis of RS(A), CS(A) and N (A) from RREF(A).
(vi) Understand the rank-nullity theorem: rank(A)+nullity(A) = n, where n is the
number of columns of A, rank(A) = dim(R(A)) and nullity(A) = dim(N (A)).
Exercise Set 12
*1. The Cobb-Douglas production function for a particular manufacturer is given by

P (x, y) = 100x3/5 y 2/5
where x represents the units of labour (at 150 per unit) and y represents the units of
capital (at 250 per unit).
Using the method of Lagrange multipliers,
(i) nd the maximum production level if the total cost of labour and capital is limited to
100,000, and
(ii) nd the minimum cost of producing 20,000 product units.
Roughly sketch appropriate contours P and C to justify your use of Lagranges method
and to establish whether you have minimised or maximised the function.
In each case nd and interpret the value of the Lagrangian multiplier, .
3 i and w = 1 + i as points in the complex plane. Express them in

( 3 i)6
exponential form and evaluate q =
. Express your answer in the form a + ib.
(1 + i)10
2. Plot z =
*3. Write each of the following complex numbers in the form a + ib
ei 2
ei
3
2
ei
3
4
ei
11
3
e1
e1+i
e3+i 2
4ei
7
6
4. Show that the functions y(x) = eix and y(x) = eix each satisfy the dierential equation
d2 y
+ y = 0.
dx2
*5. Find the roots w and w

of the equation x2 4x + 7 = 0
For these values of w and w,
nd the real and imaginary parts of the following functions,
f (t) = ewt
tR,
g(t) = wt
t Z+
*6. The function y(t) can be written in the form

y(t) = Aet + Bet
where A and B are non-zero complex constants and = a + ib.
Show that y(t) can be written in the alternative forms
y(t) = eat (Aeibt + Beibt )
b cos bt + B
b sin bt)
= eat (A
b = A + B and B
b = i(A B).
where A
b and B
b are real?
How can A and B be chosen so that A
b and B
b real, show that y(t) can be written as
For A
b cos bt + B
b sin bt) = Ceat cos(bt )
y(t) = eat (A
b2 + B
b 2 = 2 AB and satises tan = (B/
b A).
b
where C = A
7. Show that for any z C, the expressions
are both real.
ezt + ezt , t R
8. (a) Find the four complex roots of the equation z 4 = 1.

in the complex plane.
and
z t + zt , t Z+
Illustrate the roots as points
Find the roots of z 3 = 1 and illustrate them on another graph of the complex plane.
Without actually solving the equations, illustrate the roots of
x5 = 1
and
x6 = 64
as points in the complex plane.

(b) Use the roots you found in part (a) to write z 4 + 1 as a product of quadratic factors
with real coecients.
Objectives.
(i) Sketch and use the contours of a function of several variables and its constraint to
dene a Lagrangian function.
(ii) Deduce constrained maximum and minimum points from the Lagrangian function.
Interpret the value of the Lagrangian multiplier, .
(iii) Know: a complex number is z = a + ib where i2 = 1.
All roots of a polynomial with real coecients exist as complex numbers.
(iv) Know Eulers formula:
ei = cos + i sin .
(v) Change between forms of a complex number:
a + ib = rei = r cos + ir sin .
Exercise Set 13
1. For the following, you may use the laws of exponents, but do not assume any property of
the logarithm other than its definition as the inverse of the exponential function.
(a) For a > 0, show that
loga y = ln y/ ln a
(b) Find the derivative of f (x) = ax .
2. Use the chain rule to show that, for 0 < x < ,
d
(arcsin(cos x)) = 1 .
dx
What can you deduce about h(x) = arcsin(cos x) for x in this range? Can you explain
this?
*3. Let p be the price per unit required to sell x units of a commodity. Verify that each of the
following demand functions:
(i) x = 85e6p
and
(ii) x = 90 2p
is invertible and sketch the graphs of the functions and their inverses. Find the revenue
functions and the outputs x and prices p which maximize them.
For (i) express the price elasticity of demand given by
=
dx/x
p/x
=
dp/p
dp/dx
as a function of x and verify that || = 1 when the revenue function is maximised.

If the average cost function for (ii) is
AC(x) = x2 8x + 157 +
2
x
find the output x which minimises marginal costs.

4. Two linear transformations T and S : R2 R2 are defined by
( ) (
)
( ) (
)
x
x
x
y
T
=
and
S
=
y
y
y
x
(a) Sketch the eects of T and S on the standard basis, and hence on the unit square with
sides e1 , e2 . Describe T and S in words.
(b) Find the matrices for T and S.
(c) What are the linear transformations ST and S 2 T ? First do this directly and describe
them in words, then check by multiplying matrices. Is ST = T S?
*5. T and S are linear transformations with respective matrices:

(
AT =
1
2
1
2
12
(
AS =
1
2
1 0
0 1
(a) Describe T and S in words.

(b) Illustrate ST and T S using the unit square. Then calculate their matrices to check
that ST = T S.

1
1
0
*6. v1 = 0 , v2 = 1 , v3 = 0
1
3
1
(a) Show that {v1 , v2 , v3 } is a basis of R3 .
1
1
0
w1 = 1 , w2 = 1 , w3 = 1
1
0
1
Show that {w1 , w2 , w3 } is a basis of R3 .
(b) Write down the matrix AT of the linear transformation T given

T (e1 ) = v1 , T (e2 ) = v2 , T (e3 ) = v3 where {e1 , e2 , e3 } R3 is the standard basis.

x
Express T y as a vector in R3 (in terms of x, y and z).
z
by
(c) Write down the matrix AS of the linear transformation S given by

S(v1 ) = e1 , S(v2 ) = e2 , S(v3 ) = e3
(Hint: What is the relationship between S
and T ?)
(d) Write down the matrix AR of
R(e1 ) = w1 , R(e2 ) = w2 , R(e3 ) = w3 .
the
linear
transformation
given
by
(e) Is RS defined? What does this linear transformation do to v1 , v2 and v3 ? Find the
matrix ARS and use it to check your answer.
Objectives.
(i) Understand concept of an inverse of a one variable function and know when it exists.
(ii) Find a formula for an inverse function and its derivative. Find inverses of elementary
functions.
(iii) Know and apply the definition of a linear transformation.
(iv) Understand why and how a linear transformation T : Rn Rm can be given as
multiplication by a matrix AT . Columns of AT are the images of a basis of Rn .
(v) Recognise the order of composed mappings. ST is first T and then S.
Exercise Set 14
1. = Use Maple to plot the function f : R R defined by
y = x4 6x2 + 12 .
Find a formula for a local inverse to the function at each of the points x = 1, 1, 2. Sketch
the graph of each local inverse function.
Calculate the derivative of the local inverse at x = 2 directly from your formula. Evaluate
it at the point y = 4 and verify that at this point
dx
=
dy
dy
dx
)1
Find any points at which no local inverse exists.

*2. Functions f : R2 R2 are defined by
(a)
( ) (
)
u
2x y
=
v
x+y
(b)
( ) (
)
u
y x2
=
x+y
v
Find the derivatives of these functions.

Find where the functions have a local
inverse, and find the derivative of each local inverse at points where it exists.
Find also a formula for each local inverse, when it exists, at each of the points (x, y) = (1, 1).
and (x, y) = ( 21 , 0).
*3. For each of the following linear transformations, find a basis for the null space of T , N (T ),
and the range of T , R(T ). Verify the rank-nullity theorem (dimension theorem) in each
case. If any of the linear transformations are invertible, find the inverse, T 1 .
( )
x + 2y
x
(a) T : R2 R3 by T
= 0
y
0

x+y+z
x
(b) T : R3 R3 by T y = y + z
z
z

x
1 1 0
x
3
3
x= y
(c) T : R R
by T (x) =
0 1 1
y
z
z
1 0 1
4. Is it possible to construct a linear transformation T : R3 R3 where

N (T ) is the line with Cartesian equation x = y = z,
and R(T ) is the plane 2x + y z = 0?
If not, explain why. If yes, then write down a matrix for such a linear transformation.
5. F = D2 2D + 1 is a linear operator on C (R), the vector space of functions on R
with continuous derivatives of all orders, where D is the dierential operator, D(f ) = f .
Find F (sin x) and F (3xex ).
6. S is the set of sequences, yt : {0, 1, 2, 3, . . .} R. Convince yourself that S is a vector
space. (Are the axioms satisfied? What is the 0 sequence? What is the additive inverse?)
E is the shift operator, E(yt ) = yt+1 .
Show that E is a linear operator on S.
(a) Let yt be the sequence yt = ( 12 )t . Then E(yt ) = yt+1 = zt , where zt = ( 12 )t+1 .

Write out the first four terms of the sequences yt and zt to see the eect of E.
(b) Is 2E 1 also a linear operator on S? Justify your answer.
(c) The null space of the linear transformation 2E 1 is therefore a subspace of S.
Show that yt = ( 21 )t is in the null space by showing it is a solution of the dierence
equation
2yt+1 yt = 0
Given that the null space has dimension 1, explain why any solution can be written as
yt = ( 12 )t for some R.
Objectives.
(i) Understand concept of a local inverse for a function at a point and find it.
(ii) Understand the concept of a local inverse of a vector-valued function of several variables
at a point. Be able to find critical points, and to find formulae for a local inverse and
its derivative when these exist.
(iii) Understand null space (kernel) and range (image) of a linear transformation, and
their correspondence to null space and range of an associated matrix.
(iv) Know the rank-nullity theorem (the dimension theorem) for linear transformations.
Exercise Set 15
1.*(a) Find the eective annual rate of interest on 1000 at 8% compounded
(i) quarterly and (ii) continuously.
*(b) Determine the interest rate needed to have money double in 8 years when compounded
semiannually.
(c) Find the present value of 10,000 in 5 years (that is, nd how much to invest now in
order to obtain 10,000 in 5 years time) at 6% interest, when interest is compounded
(i) annually and (ii) continuously.
*(d) The estimated value of a collection of antiques bought for investment is increasing
according to the formula
2/5
V = 325, 000 (1.95)t
The discount rate under continuous compounding is 6.8%. How long should the collection
be held to maximize the present value?
2. *(a) A deposit of P is made each month for t years in an account that is compounded
monthly at an interest rate r. Find the sum accumulated after t years.
(b) A deposit of 50 is made each month in an account that is compounded monthly at
12% interest. Find the balance in the account after 10 years.
*3. In question 6 of exercise set 13, you showed that each of the sets is a basis of R3 :

0
1
0
1
1
1
b = 1 1 1
and
B
B = 010
1
1
0
1
1
3
(a) Write down the transition matrix P from B coordinates to standard.
b coordinates to standard.
Write down the transition matrix Q from B
2
3
(b) If v = 1 , nd [v]B
(c) Given [x]Bb = 1
nd [x]B
2
3 Bb
b coordinates to B coordinates.
(d) Write down the transition matrix from B
b to standard to B).
(Note you will need to change coordinates from B

clockwise:
6
(a) Write down the matrix of the linear transformation which accomplishes this rotation.
4. Change the basis in R2 by a rotation of the axes through an angle of
(b) Write down the new basis vectors, v1 and v2 (which are the images of e1 and e2 ).
Denote the new basis by B, B = {v1 , v2 }.
Write down the transition matrix P from B coordinates to standard coordinates.
(c) A curve C is given in standard coordinates by the equation 3x2 + 2 3xy + 5y 2 = 6.

Find the equation of the curve in the new B coordinates, (X, Y ).
(d) Hence sketch the curve in the x, y plane.
*5. Let
{( ) (
)}
2
1
M=
,
1
1
(a) Show that M is a basis of R2 .

coordinates to standard coordinates.
(
v=
1
2
( ) (
)
x
7x 2y
T
=
y
x + 8y
Write down the transition matrix from M
(b) Find [v]M , the M coordinates of the vector v.

On one graph, sketch the vector v and show how to obtain it geometrically:
(1) as a linear combination of the standard basis vectors and
(2) as a linear combination of the M basis vectors.
(c) Write down the matrix of the linear transformation T , T : R2 R2 with respect to
the standard basis. Call it A.
(d) Find the matrix of T in M coordinates. Call it B.
Describe geometrically the eect of the transformation T as a map from R2 R2 .
(e) Find the image of [v]M using B.
On the same graph as in part (b), use the M basis vectors (and the information you
obtained about the linear transformation in part (d)) to sketch T (v) as a linear combination of the M basis vectors.
Check your answer using standard coordinates.
Objectives.
(i) Know formulae for interest compounding and discounting and their uses.
(ii) Find the transition matrix P from coordinates with respect to a new basis B to
standard coordinates and use P to change coordinates. Know v = P[v]B .
(iii) Understand changing coordinates does not change the position of the vector v.
Exercise Set 16
*1. Given the inverse demand function yd = 116 + 2x x2 and inverse supply function
ys = (x + 2)2 , where x represents quantity and yd and ys each represents price, find the
consumers surplus and the producers surplus.
2. Use partial fractions to find the integrals
(a)
3. (i) Find
(a)
(ii)
dx
(1 x2 )2
x2 dx
x2 + x 2
(b)
x3 x2 + x + 1
dx
(x2 + 1)2
(b)
cos8 x sin3 x dx
Find, using appropriate changes of variable
(a)
6 3
(x + 2) x dx
(b)
sin x cos x
dx
(1 + sin x)3
4. Use integration by parts to evaluate
arcsin x dx
0
5. A firm has fixed costs of 125. Its marginal revenue function is R (x) = 170 0.4x, and its
marginal cost function is C (x) = 10 + 2x, where x represents quantity. Find the revenue
function, R(x) and the cost function C(x). Find also the value of x which maximises profit.
(
*6. Let
A=
1
3
)
4
.
2
(a) Find the eigenvalues and the corresponding eigenvectors for A.

(b) Hence, find an invertible matrix P such that P1 AP is diagonal.
(c) Calculate P1 AP directly to check your answer.
(
7. Diagonalize the matrix
*8. Let
B= 5
1
A=
)
7 2
.
1 8
1 2
3 5 .
1 2
(a) Find the characteristic equation of B. Find the eigenvalues and the corresponding
eigenvectors for B.
(b) Hence, find an invertible matrix P such that P1 BP is diagonal, and check your
answer for P by multiplying out P1 BP.
Objectives.
(i) Know the concepts of consumer and producer surplus and how to calculate them.
(ii) Know the method of partial fractions.
(iii) Convert integrals to standard form by techniques of completing the square,
substitution and integration by parts.
(iv) Know definition of an eigenvector and an eigenvalue of a matrix.
Find eigenvalues and corresponding eigenvectors for a matrix.
(v) Diagonalize a matrix A: use eigenvalues and corresponding eigenvectors to create (if
possible) an invertible matrix P and a diagonal matrix D such that P1 AP = D.
Know that eigenvectors and eigenvalues are in corresponding columns, and that the
eigenvectors form a basis of Rn .
(vi) Check eigenvectors and eigenvalues by computing AP.
Exercise Set 17
1. Classify each of the dierential equations below in one or more of the following categories:
(i) linear equation (ii) separable equation (iii) exact equation
Find the general solutions of the equations by any appropriate method.
*(a) (x2 + 6x) dy = y 2 dx 12 dy
(b) 2x
dy
=2 x+y+2
dx
*(c) (cos y + y cos x) dx + (sin x x sin y) dy = 0

(d)
dy
= yx2 + 2x2
dx
(e) xy = e(3x
+5)
dy
dx
*(f) x dy 4y dx = (x + 1) dx
*2. Find the general solution of
2f
1
= ex sin y + ln(xey )
xy
x
*3. Find the eigenvalues of the matrices
1 1 1
A = 0 1 1
1 0 2
2 1
B = 1 0
2 1
2
1
2
and show that neither matrix can be diagonalized over R. Why must each of these matrices
have at least one real eigenvalue?
*4. (a) Find the general solution of the following system of linear dierential equations:
{
y1 (t) = y1 (t) + 4y2 (t);
.
y2 (t) = 3y1 (t) + 2y2 (t)
(b) Then nd the unique solution satisfying the initial conditions y1 (0) = 1 and y2 (0) = 0.
For this solution nd the values of y1 (0) and y2 (0).
(
5. Find A5 if A =
)
1 4
.
3 2
*6. (a) Diagonalize the matrix
1 3 0
A = 0 2 0
3 3 2
(b) Find the general solution of the system of linear dierential equations y = Ay.
Objectives.
(i) Understand concepts of a dierential equation and its solution.
(ii) Solve a dierential equation of order one by separating variables.
Know: solution of equation (D )y = 0 is y = cex .
(iii) Recognise and solve exact and linear dierential equations of order one.
(iv) When possible, solve a partial dierential equation by partial integration.
(v) Recognise if a matrix can be diagonalized or know why it cannot.
(vi) Calculate An (n N) using diagonalization.
(vii) Solve systems of linear dierential equations y = Ay using diagonalization
(by changing to a basis of eigenvectors).
Exercise Set 18
1. Find the general solution of each of the equations by any appropriate method.
(a) xy
dy
y 2 = 3x2 e2y/x
dx
(b) x2
dy
+ 2xy = 6x2
dx
2. Solve the following dierential equations with the help of the given substitutions:
dy
+ y = x3 y 6
dx
(Let y 5 = u)
(i)
(ii)
(1 + y 2 ) = (arctan y x)
dy
dx
(Let arctan y = u)
*3. Write down a linear dierential equation P (D)y = 0 whose solutions include the function
e2x and another such equation whose solutions include the function xex .
Hence write down a homogeneous linear dierential equation whose solutions include both
e2x and xex together with its general solution.
*4. Find the general solutions of the following dierential equations:

(i)
d4 y
d3 y
d2 y
+
2
+
=0
dx4
dx3
dx2
(ii)
d2 y
dy
2
+ 10y = 0
2
dx
dx
Find and sketch the particular solution of equation (ii) which satises the conditions
y(0) = 1 and y
( )
6
=0
Record your answers to this question for later use.

(
*5. (a) Orthogonally diagonalize the matrix A =
)
1 2
.
2 1
(b) Use this to sketch the curve xT Ax = 3 in the xy plane.

Find the points of intersection of the curve with the old and new axes.
*6. Let C be the curve dened by
3x2 + 2 3xy + 5y 2 = 6.
(a) Find a symmetric matrix A such that C is given by xT Ax = 6.

(b) Orthogonally diagonalize the matrix A. (Find an orthogonal matrix P and a diagonal
matrix D such that P1 AP = PT AP = D).
(c) Use P as a change of coordinates to sketch the curve in the xy plane, showing the old
and new axes on your diagram.
(d) How does this relate to question 4 of Exercise Set 15? Why would a rotation
anti-clockwise by
also have been an appropriate choice there?
3
7. (a) Show that if P is an orthogonal matrix, then det(P) = 1.
(b) Show that if A is orthogonally diagonalizable, then A is symmetric.
(c) If P is an orthogonal matrix and x = Pz, show that x = z.
Objectives.
(i) Recognise and solve homogeneous dierential equations.
(ii) Recognise when a change of variable can transform a dierential equation of order one
to one of the types: separable, homogeneous, exact or linear.
(iii) Solve linear constant coecient higher order dierential equations P (D)y = 0 when
the roots of P (D) are distinct, repeated or complex.
(iv) Know matrix P is orthogonal PT = P1 its columns are an orthonormal
basis of Rn .
(v) Orthogonally diagonalize a symmetric matrix A, PT AP = D.
In the 2 2 case create P from eigenvectors so P is a rotation anticlockwise and use
this to change coordinates to sketch a curve xT Ax = k, where k is a constant.
Find points of intersection of the curve with new and old axes.
(vi) Know A is orthogonally diagonalizable A is symmetric.
Exercise Set 19
*1. Describe the behaviour as x of the general solutions of the equations in question 4
of Exercise Set 18.
2. Find the general solutions of the following dierence equations; describing their behaviour
as x
(i) yx+2 yx+1 2yx = 0
(ii)
yx+3 3yx+2 + 9yx+1 + 13yx = 0
*Find and sketch the particular solution of equation (i) which satises the conditions
y1 = 1 and y2 = 5
Record your answers to question 2 for use next week.
1 4 2
A = 4 1 2
2 2 2
*3. Let
(a) Find the eigenvalues of A and for each eigenvalue nd an orthonormal basis for the
corresponding eigenspace.
(b) Hence nd an orthogonal matrix P such that PT AP = P1 AP = D.
(c) Write down D and check that PT AP = D.
(d) Express f (x, y, z) = xT Ax as a function of the variables x, y and z. Write down a
matrix Q and an expression zT Dz = 1 X 2 + 2 Y 2 + 3 Z 2 where 1 2 3 such
that f (x, y, z) = xT Ax = zT Dz if x = Qz.
Is the quadratic form f (x, y, z) positive denite, negative denite or indenite?
Evaluate f (x, y, z) at one unit eigenvector corresponding to each eigenvalue.
*4. (a) Find an orthonormal basis for the subspace of R3 given by

V = y 5x y + 2z = 0 .
z
(b) Extend this to an orthonormal basis of R3 .
5. Beginning with the vector v1 , use Gram-Schmidt to obtain an orthonormal basis for the
subspace of R4 spanned by the vectors,

1
0
v1 =
1
0

3
0
v2 =
2
0
2
1
v3 =
1
3
*6. (See Calculus Lecture 10 and Algebra Lecture 19)

(a) Let q = (x )T f ( )(x ).
For x near , , an eigenvalue of the matrix f ( ),
q > 0 all s > 0 = is a local minimum.

q < 0 all s < 0 = is a local maximum.
q diers in sign the s dier in sign is a saddle point.
Classify each stationary point of the function in question 2(a) of Exercise Set 10 by
nding the signs of the s.
(b) For all , and x near
q 0 all s 0 f (x) is convex.
q 0 all s 0 f (x) is concave.
Use the signs of the s to examine where, if at all, the function
convex or concave.
f (x, y) = y 3 x2
is
7. Is F (x, y) = 3x2 8xy + 3y 2 or G(x, y) = 43x2 48xy + 57y 2 positive denite?

Find, if possible, (ai , bi ) (i = 1, 2, 3, 4) such that F (a1 , b1 ) > 0, F (a2 , b2 ) < 0,
G(a3 , b3 ) > 0, G(a4 , b4 ) < 0.
8. If A is an m k matrix of rank k, show that the matrix AT A is positive denite.
Objectives.
(i) Describe the behaviour of solutions of dierential equations as x .
(ii) Know the dierence equation (E )yx = 0 has solution yx = c()x . Solve linear,
constant coecient, higher order dierence equations. Describe behaviour as x .
(iii) Sketch particular solutions of dierence equations.
(iv) Use Gram-Schmidt orthogonalization to obtain an orthonormal basis of a vector space.
(v) Obtain information about a quadratic form xT Ax from the eigenvalues of A.
Exercise Set 20
*1. Find the general solutions of the following dierential equations where x R+ :
(i)
d4 y
d3 y
d2 y
+
2
+
= sin x
dx4
dx3
dx2
d3 y
d2 y
dy
3
+
9
+ 13y = x + ex
dx3
dx2
dx
(ii)
2. Find the general solutions of the following dierence equations where x N:

(i)
yx+2 yx+1 2yx = (1)x

(iii)
(ii)
yx+2 + 2yx+1 + 4yx = 12
yx+3 3yx+2 + 9yx+1 + 13yx = 2x
*3. Find the general solution of the dierential equation

2f
2f
=x
x2
y 2
x = u + v, y = u v.
by making the change of variable
*4. A Markov process satises the dierence equation:

(
xk = Axk1
where
A=
0.7
0.3
0.6
0.4
(
x0 =
0.6
0.4
Solve the equation and nd an expression for xk as a linear combination of the eigenvectors
of A. Use this to predict the long term distribution; that is, nd lim xk as k .
5. The population of osprey eagles at a certain lake is dying out. Each year the new population
is only 60% of the previous years population. Conservationists introduce a new species of
trout into the lake and nd that the populations satisfy the following system of dierence
equations, where xt is the number of osprey in year t and yt is the number of trout in
year t.
(
xt = Axt1
where
xt =
xt
yt
(
A=
.6
.2
.25 1.2
(
and
x0 =
20
100
(a)
Give a reason why this system is not a Markov process.
(b) Describe in words how each of the populations depends on the previous years populations.
(c) Solve the system (as you would a Markov process).
Show that the solution xt
satises xt = At x0 .
Diagonalize the matrix A.
Hence nd an expression for xt as a
linear combination of the eigenvectors of A.
(d) Show that the situation is not stable, that according to this model both osprey and
trout will increase without bound as t .
What will be the eventual ratio of osprey to trout?
In order to have the populations of osprey and trout achieve a steady state, they decide to
allow an amount of shing each year, based on the number of osprey in the previous year.
The new equations are
(
xt = Bxt1
where
xt =
xt
yt
(
B=
.6
.2
1.2
(
and
x0 =
20
100
where > 0 is a constant to be determined.

(e) What property of the transition matrix of a Markov process determines that there is a
(nite, non-zero) long-term distribution?
Deduce a condition on the eigenvalues of the matrix B to produce the same eect.
Then nd the value of which satises this condition.
(f ) Show that for this value of the population now reaches a steady state as t and
determine what this stable population of osprey and trout will be.
Objectives.
(i) Find a particular solution of a nonhomogeneous linear dierential/dierence equation
with RHS itself the solution of a linear dierential/dierence equation.
(ii) Recognise dening properties of a Markov process, xk = Axk1 :
A has non-negative entries and the entries of each column sum to 1.
Know that A always has an eigenvalue = 1 with other eigenvalues |i | 1.
(iii) Find the solution of a Markov process xk = Axk1 . Express it as a linear combination
of eigenvectors of A. Use it to predict lim xk as k .
Apply the same method to solve any system of linear dierence equations, xk = Axk1 .
Some Review Exercises

(optional)
1. Consider a matrix, for example the 3 3 matrix A given below, and think about . . .
what this matrix can represent . . . what you can ask about the matrix (e.g. invertible?
diagonalizable?) . . . what subspaces of Rn are associated with the matrix?
7 6
A = 5 4
2 2
9
5
4
(i) Solve Ax = 0.
(ii) Describe the linear transformation T (x) = Ax by finding the images of e1 , e2 , e3
(iii) Can this matrix be used as a change of coordinates in R3 ?
(iv) Find A1 if it exists.
(v) Diagonalise the matrix A. (Find an invertible matrix P and a diagonal matrix D such
that P1 AP = D, and check your work.)
(vi) Use the diagonalisation to give a description of the eect of the linear transformation
T (x) = Ax on R3 .
(vii) Use the diagonalisation to solve the system of dierential equations y = Ay
(viii) Find the row space, column space and nullspace of A. Find the kernel and the
range of the linear transformation T (x) = Ax.
Describe these spaces geometrically,
giving Cartesian equations for any planes and vector equations for any lines.
2. Show that the function f (x, y, z) = 5x2 +5y 2 +9z 2 6xz 12yz
points and determine their nature.
has a line of stationary
Find a symmetric matrix A such that f (x, y, z) = xT Ax.

Is A positive (or negative) definite (or semidefinite)?
What information does this give
about the function f ? About the stationary points of f ?
How is A related to the Hessian of f ?
3. Let A be an m k matrix with full column rank (i.e. the columns of A are linearly
independent). Show that:
(1) AT A is a symmetric k k matrix,
(2) AT A is invertible,
(3) AT A is positive definite.
4. Functions f, g : R3 R are defined by

f (x, y, z) = x2 + 4y 2 + 2z 2
g(x, y, z) = x2 + y 2 2z 2
(a) Show that the point (3, 2, 1) lies on both of the following level surfaces and hence
on their curve of intersection:
f (x, y, z) = 27
and
g(x, y, z) = 11
Describe these surfaces, sketching their intersections with the coordinate planes.
(b) Find normal vectors and the tangent planes to both surfaces at (3, 2, 1).
surfaces touch tangentially at this point?
Do the
(c) Using the normal vectors found in (b) write down Cartesian equations for the tangent
line to the curve of intersection of the level surfaces at (3, 2, 1)
5. The function f (x, y) is given by
f (x, y) = (xa y + xy a ) ln
x2 + y 2
where a is some constant. Show that f is homogeneous and verify that Eulers equation
holds.
6. A consumer has utility function u(x, y) = (x + 1)2 (y + 2)3 for two goods, X and Y . (Here,
x denotes the amount of X and y the amount of Y .) Each unit of x costs p dollars and
each unit of Y costs q dollars, and the consumer has a budget for X and Y of M dollars.
Use the Lagrange multiplier method to find the quantities of X and Y the consumer will
consume in order to maximise his utility subject to the budget constraint. (Your answers
should be expressions depending on p, q and M .)
7. Let V and W be the subspaces

1
1
V = Lin 0 , 1 R3 ,
0
1
1
2
3 5
4
W = Lin ,
R
1
0
0
1
Consider the possibility of linear transformations T or S such that

T : Ra Rb
with
N (T ) = V
and
S : Rc Rd
with
N (S) = W
and R(S) = V
R(T ) = W
Show that one of these, S or T cannot exist. Then find a matrix AS or AT representing the
other linear transformation (with respect to the standard basis in each Euclidean space).
Check your answer by row reducing the matrix and finding the null space and range of the
linear transformation.

Study Pack MA100 Inside All

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Study Pack MA100 Inside All

Încărcat de

Drepturi de autor:

Formate disponibile

MA100 Mathematical Methods

General Information for 2013/14

oce: COL 2.05C

The notes contain detailed

Calculus: K. Binmore and J. Davies, Calculus, Concepts and Methods.

Teaching Arrangements and What You Need To Do!

A few Words of Advice

Do not fall behind!

MA100 Mathematical Methods

1. The set of real numbers, denoted R, includes the following subsets:

(a, b) = {x | a < x < b}, [a, b] = {x | a x b},

straight lines, P1 (x) = a0 + a1 x

(AlgR, sections 23, 810)

2p + 0.4q = 155 and

C(q) = q 2 10q + 300 .

3. Be able to manipulate algebraic expressions accurately and efficiently.

(AlgR, sections 17)

(3.3) Rewrite the simultaneous equations

(5.3) Evaluate (without using a calculator):

6. Rules of differentiation: Product, quotient and chain rules

(CalcR, sections 19)

9. Techniques of integration: partial fractions, integration by parts, substitution.

(CalcR, sections 10-17; AlgR, section 11)

10. The technique of Completing the Square in a quadratic expression.

(10.1) Complete the squares in the expressions: x2 + 6x + 11 and 2x2 4x + 7.

Course Content CALCULUS

Vectors and general geometric concepts

Scalar valued functions

Stationary points and optimisation of functions of several variables

Summation and integration

Differential and difference equations

( By something like 1.4.1, we mean Section 1.4, Question 1, of the textbook. )

MA 100 Mathematical Methods Calculus

MA 100 Mathematical Methods Calculus

Combinations of functions (pages 55-56)

Vectors in R3 : 1.8.7 and 1.8.9;

Some elementary functions (pages 52-55)

Vectors in R2 : 1.4.1, 1.4.2 ( except part (vi) ), 1.4.3 and 1.4.4;

1.4 Optional extra exercises related to these topics

For this we follow Section 1.6 of the textbook.

For this we follow Section 1.5 of the textbook.

Vectors and Lines

Real valued functions of one real variable (pages 5152)

Intervals (page 51)

To begin, you should look through the first sections of Chapter 2:

working through the examples.

Read each section carefully,

Section and page number references refer to the calculus textbook.

Curves and Conics

MA 100 Mathematical Methods Calculus

2.15.7 and 2.15.10.

Curves and conics :

2.3 Optional extra exercises related to these topics

A curve is the set of points ( x, y)

2.2 Curves and conics

For this we follow Section 2.2 of the textbook.

3.2 Functions of one variable

2.11.1, 2.11.2, 2.11.9, 2.11.11, 2.11.13 and 2.11.15;

MA 100 Mathematical Methods Calculus

Stationary points and optimisation of one-variable functions : 4.5.4 and 4.5.5.

Taylor Series of one-variable functions : 2.15.1, 2.15.2, 2.15.4 and 2.15.5;

3.7 Optional extra exercises related to these topics

For this we follow Section 4.1 of the textbook.