You are on page 1of 102

Mathematical Tripos: IA Algebra & Geometry (Part I

)

Contents
0 Introduction i
0.1 Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
0.2 Lectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
0.3 Printed Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
0.4 Example Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
0.5 Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
0.6 Revision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

This is a supervisor’s copy of the notes. It is not to be distributed to students.
1 Complex Numbers 1
1.0 Why Study This? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Real numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 The general solution of a quadratic equation . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.3 Complex numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Algebraic Manipulation of Complex Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Functions of Complex Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.1 Complex conjugate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.2 Modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 The Argand Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 Polar (Modulus/Argument) Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5.1 Geometric interpretation of multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 The Exponential Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6.1 The real exponential function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6.2 The complex exponential function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6.3 The complex trigonometric functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6.4 Relation to modulus/argument form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.6.5 Modulus/argument expression for 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.7 Roots of Unity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.8 De Moivre’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.9 Logarithms and Complex Powers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.9.1 Complex powers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.10 Lines and Circles in the Complex Plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.10.1 Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.10.2 Circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.11 Möbius Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.11.1 Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Mathematical Tripos: IA Algebra & Geometry (Part I) a
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

1.11.2 Inverse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.11.3 Basic Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.11.4 The general Möbius map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2 Vector Algebra 15
2.0 Why Study This? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1 Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.1 Geometric representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Properties of Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2.1 Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2.2 Multiplication by a scalar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

This is a supervisor’s copy of the notes. It is not to be distributed to students.
2.2.3 Example: the midpoints of the sides of any quadrilateral form a parallelogram . . . . . . . 17
2.3 Scalar Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.1 Properties of the scalar product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.2 Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3.3 Another property of the scalar product . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3.4 Example: the cosine rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Vector Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4.1 Properties of the vector product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.2 Vector area of a triangle/parallelogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5 Triple Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5.1 Properties of the scalar triple product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6 Bases and Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6.1 Two dimensional space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6.2 Three dimensional space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.6.3 Higher dimensional spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.7 Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.7.1 The Cartesian or standard basis in 3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.7.2 Direction cosines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.8 Vector Component Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.9 Polar Co-ordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.9.1 2D plane polar co-ordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.9.2 Cylindrical polar co-ordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.9.3 Spherical polar co-ordinates (cf. height, latitude and longitude) . . . . . . . . . . . . . . . 28
2.10 Suffix Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.10.1 Dyadic and suffix equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.10.2 Summation convention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.10.3 Kronecker delta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.10.4 More on basis vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.10.5 The alternating tensor or Levi-Civita symbol . . . . . . . . . . . . . . . . . . . . . . . . . 33

Mathematical Tripos: IA Algebra & Geometry (Part I) b
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

2.10.6 The vector product in suffix notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.10.7 An identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.10.8 Scalar triple product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.10.9 Vector triple product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.11 Vector Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.12 Lines and Planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.12.1 Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.12.2 Planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.13 Cones and Conic Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.14 Maps: Isometries and Inversions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

This is a supervisor’s copy of the notes. It is not to be distributed to students.
2.14.1 Isometries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.15 Inversion in a Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3 Vector Spaces 42
3.0 Why Study This? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.1 What is a Vector Space? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.1.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.1.2 Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.1.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.2 Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3 Spanning Sets, Dimension and Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.1 Linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.2 Spanning sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.4 Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.5 Intersection and Sum of Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.5.1 dim(U + W ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.5.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.6 Scalar Products (a.k.a. Inner Products) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.6.1 Definition of a scalar product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.6.2 Schwarz’s inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.6.3 Triangle inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
n
3.6.4 The scalar product for R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.6.5 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Mathematical Tripos: IA Algebra & Geometry (Part I) c
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

4 Linear Maps and Matrices 55
4.0 Why Study This? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.1 General Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.2 Linear Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.2.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3 Rank, Kernel and Nullity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.3.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.4 Composition of Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.4.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.5 Bases and the Matrix Description of Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

This is a supervisor’s copy of the notes. It is not to be distributed to students.
4.5.1 Matrix notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.5.2 Examples (including some important new definitions of maps) . . . . . . . . . . . . . . . . 62
4.5.3 dim(domain) 6= dim(range) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.6 Algebra of Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.6.1 Multiplication by a scalar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.6.2 Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.6.3 Matrix multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.6.4 Transpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.6.5 Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.6.6 Trace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6.7 The unit or identity matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6.8 The inverse of a matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.6.9 Determinants (for 3 × 3 and 2 × 2 matrices) . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.7 Orthogonal Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.8 Change of basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.8.1 Transformation matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.8.2 Properties of transformation matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.8.3 Transformation law for vector components . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.8.4 Transformation law for matrices representing linear maps . . . . . . . . . . . . . . . . . . 74

5 Determinants, Matrix Inverses and Linear Equations 76
5.0 Why Study This? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.1 Solution of Two Linear Equations in Two Unknowns . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.2 Determinants for 3 × 3 Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.2.1 Properties of 3 × 3 determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.3 The Inverse of a 3 × 3 Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.3.1 Minors and cofactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.3.2 Construction of the inverse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.4 Solving Linear Equations: Gaussian Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.5 Solving linear equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

Mathematical Tripos: IA Algebra & Geometry (Part I) d
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

5.5.1 Inhomogeneous and homogeneous problems . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.5.2 Geometrical view of Ax = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.5.3 Linear mapping view of Ax = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.5.4 Implications for the solution of the inhomogeneous equation Ax = d . . . . . . . . . . . . 85

6 Complex Vector Spaces 86
6.0 Why Study This? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.1 Vector Spaces Over The Complex Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.1.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

This is a supervisor’s copy of the notes. It is not to be distributed to students.
6.1.2 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.1.3 Cn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.2 Scalar Products for Complex Vector Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.2.2 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.2.3 Scalar Products in Terms of Components . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.3 Linear Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Mathematical Tripos: IA Algebra & Geometry (Part I) e
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

0 Introduction

0.1 Schedule

This is a copy from the booklet of schedules.1 Schedules are minimal for lecturing and maximal for
examining; that is to say, all the material in the schedules will be lectured and only material in the
schedules will be examined. The numbers in square brackets at the end of paragraphs of the schedules
indicate roughly the number of lectures that will be devoted to the material in the paragraph.

ALGEBRA AND GEOMETRY 48 lectures, Michaelmas term

Review of complex numbers, modulus, argument and de Moivre’s theorem. Informal treatment of complex

This is a supervisor’s copy of the notes. It is not to be distributed to students.
logarithm, n-th roots and complex powers. Equations of circles and straight lines. Examples of Möbius
transformations. [3]
Vectors in R3 . Elementary algebra of scalars and vectors. Scalar and vector products, including triple
products. Geometrical interpretation. Cartesian coordinates; plane, spherical and cylindrical polar coor-
dinates. Suffix notation: including summation convention and δij , ijk . [5]
Vector equations. Lines, planes, spheres, cones and conic sections. Maps: isometries and inversions.
[2]
Introduction to Rn , scalar product, Cauchy–Schwarz inequality and distance. Subspaces, brief introduc-
tion to spanning sets and dimension. [4]
Linear maps from Rm to Rn with emphasis on m, n 6 3. Examples of geometrical actions (reflections,
dilations, shears, rotations). Composition of linear maps. Bases, representation of linear maps by matrices,
the algebra of matrices. [5]
Determinants, non-singular matrices and inverses. Solution and geometric interpretation of simultaneous
linear equations (3 equations in 3 unknowns). Gaussian Elimination. [3]
Discussion of Cn , linear maps and matrices. Eigenvalues, the fundamental theorem of algebra (statement
only), and its implication for the existence of eigenvalues. Eigenvectors, geometric significance as invariant
lines. [3]
Discussion of diagonalization, examples of matrices that cannot be diagonalized. A real 3 × 3 orthogonal
matrix has a real eigenvalue. Real symmetric matrices, proof that eigenvalues are real, and that distinct
eigenvalues give orthogonal basis of eigenvectors. Brief discussion of quadratic forms, conics and their
classification. Canonical forms for 2 × 2 matrices; discussion of relation between eigenvalues of a matrix
and fixed points of the corresponding Möbius map. [5]
Axioms for groups; subgroups and group actions. Orbits, stabilizers, cosets and conjugate subgroups.
Orbit-stabilizer theorem. Lagrange’s theorem. Examples from geometry, including the Euclidean groups,
symmetry groups of regular polygons, cube and tetrahedron. The Möbius group; cross-ratios, preservation
of circles, informal treatment of the point at infinity. [11]
Isomorphisms and homomorphisms of abstract groups, the kernel of a homomorphism. Examples. Intro-
duction to normal subgroups, quotient groups and the isomorphism theorem. Permutations, cycles and
transpositions. The sign of a permutation. [5]
Examples (only) of matrix groups; for example, the general and special linear groups, the orthogonal
and special orthogonal groups, unitary groups, the Lorentz groups, quaternions and Pauli spin matrices.
[2]

Appropriate books
M.A. Armstrong Groups and Symmetry. Springer–Verlag 1988 (£33.00 hardback).

Alan F Beardon Algebra and Geometry. CUP 2005 (£21.99 paperback, £48 hardback).

1 See http://www.maths.cam.ac.uk/undergrad/schedules/.

Mathematical Tripos: IA Algebra & Geometry (Part I) i
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

Michaelmas 2006 . Bloom Linear Algebra and Geometry. . Nelson Thornes 1992 (£30. • I will stay around for a few minutes at the front after lectures in order to answer questions.C.Cowley@damtp.P.E. I will occasionally use straightforward or similarly to last time. • Finally. This is a supervisor’s copy of the notes. • The course is somewhere between applied and pure mathematics. at some point. Chapman and Hall/CRC 1988 (£38. but please do not discuss. research suggests that within the first 20 minutes I will. . J.95 paperback). 0. unclear or just plain wrong then please shout out.J. • I will endeavour to have a 2 minute break in the middle of the lecture for a rest and/or jokes and/or politics and/or paper aeroplanes2 .D. if it is not. . If you feel that the course is going over your head. or you are spending more than 20-24 hours a week on it.3 • I will aim to finish by 11:55.3 Printed Notes • I hope that printed notes will be handed out for the course . I will pick up the first one to stay in the air for 5 seconds. D. 0. • I aim to avoid the words trivial. otherwise you will get hopelessly lost (especially at 6 lectures a week).cam. I was in your position 32 years ago and nearly gave up the Tripos.M.cam. a Path to Geometry. • If anyone is colour blind please come and tell me which colour pens you cannot read. but am not going to stop dead in the middle of a long proof/explanation. D. Sernesi Linear Algebra: A Geometric Approach.75 paperback). Let me know if I fail.uk) or catch me at the end of the next lecture. Please be on time since it is distracting to have people walking in late. Bourne and P.ac. Smart Linear Algebra and Geometry. • By all means chat to each other quietly if I am unclear. email me (S.99 paperback). Cambridge University Press 1988 (out of print). If it is not in the notes or on the example sheets it should not be in the exam. E. have lost the attention of all of you.uk. R. obvious and yes 4 . easy.J. . Such chatting is a distraction. If I am inaudible. 2 If you throw paper aeroplanes please pick them up.99 paperback). say. last night’s football results. CRC Press 1993 (£38. Under such circumstances I will have to plough on as a result of time constraints. • Sometimes I may confuse both you and myself (I am not infallible). so that you can listen to me rather than having to scribble things down (but that depends on my typing keeping ahead of my lecturing). I will do my best to be clear but you must read through and understand your notes before the next lecture . Green Sets and Groups: a first course in Algebra. Hence do not always expect pure mathematical levels of rigour. Mathematical Tripos: IA Algebra & Geometry (Part I) ii c S. .2 Lectures • Lectures will start at 11:05 promptly with a summary of the last lecture. Cambridge University Press 1987 (£20. An understanding of your notes will not diffuse into you just because you have carried your notes around for a week . It is not to be distributed to students. 3 Having said that.Cowley@damtp. Burn Groups. Cambridge University Press 1979 (out of print). however I will clear up any problems at the beginning of the next lecture. • I want you to learn. • I welcome constructive heckling. or put them under your pillow.A. 4 But I will fail miserably in the case of yes. come and chat.ac. Kendall Vector Analysis and Cartesian Tensors. having said that all the outline/sketch ‘proofs’ could in principle be tightened up given sufficient time. illegible. . students seem to find that the break makes it easier to concentrate throughout the lecture. or who did (or did not) get drunk and/or laid. and may not be able to extract myself in the middle of a lecture.

ac.Cowley@damtp. 0. from which you may conclude that I will not have copies of last time’s notes (so do not ask). . • With one or two exceptions figures/diagrams are deliberately omitted from the notes. but if they are ill then please take copies. In the past I have been tempted to not include these because I was worried that students would be unhappy with material in the notes that was not lectured. athletes do press-ups.4 Example Sheets • There will be four main example sheets. • I do not keep back-copies (otherwise my office would be an even worse mess) . they are an adjunct to lectures and are not meant to be used independently. The following notes were adapted (i.J. • Please email me corrections to the notes and example sheets (S. stolen) from those of Peter Haynes. I was taught to do this at my teaching course on How To Lecture . Mathematical Tripos: IA Algebra & Geometry (Part I) iii c S. then come and see me.uk). but bring your sick-note. 5 With the exception of the first three lectures for the pedants. • Any notes will only be available in lectures and only once for each set of notes.cam. They will be available on the WWW at about the same time as I hand them out (see http://damtp. my esteemed Head of Depart- ment.cam.J. .ac. pianists do scales.ac. • Please do not take copies for your absent friends unless they are ill. • There is some repetition on the sheets by design. This is a supervisor’s copy of the notes. 0. . • If you do not want to attend lectures then there are a number of excellent textbooks that you can use in place of my notes. Michaelmas 2006 . • There are a number of unlectured worked examples in the notes.uk.uk/user/examples/).e. There will also be two supplementary ‘study’ sheets. • Your supervisors might like to know that the example sheets will be the same as last year (but that they will not be the same next year). It is not to be distributed to students. 6 If you really have been ill and cannot find a copy of the notes.5 We are going to fell a forest as it is. .Cowley@damtp. a vote in one of my previous lecture courses was overwhelming in favour of including unlectured worked examples. • There will only be approximately as many copies of the notes as there were students at the lecture on the previous Saturday. or there- abouts. Personally I suggest that you do not have your first supervision before the middle of week 2 of lectures. • You should be able to do example sheets 1/2/3/4 after lectures 6/12/18/23 respectively. However.cam. the aim being that it might help you to stay awake if you have to write something down from time to time. and I have no desire to be even more environmentally unsound. mathe- maticians do algebra/manipulation.5 Acknowledgements.6 • The notes are deliberately not available on the WWW. Please bear this in mind when arranging supervisions.

The Greek alphabet. The exponential function. (0.e. (0. (0. H η eta T τ tau Θ θ theta Υ υ upsilon I ι iota Φ φ phi K κ kappa X χ chi Λ λ lambda Ψ ψ psi M µ mu Ω ω omega There are also typographic variations on epsilon (i.1b) exp(1) = e ≈ 2. (0. %).1c) exp(x) = ex . It is not to be distributed to students.e. Michaelmas 2006 . is defined by the series ∞ X xn exp(x) = .0. (0. (0.2e) The sine and cosine functions. (0. i.J. exp(x). (0. and rho (i.71828183 .1a) n=0 n! It has the following properties exp(0) = 1. You should check that you recall the following. ϕ).Cowley@damtp.3b) n=0 2n! Mathematical Tripos: IA Algebra & Geometry (Part I) iv c S. log x. (0.1d) ex+y = ex ey .2d) log(xy) = log x + log y .2b) log(e) = 1. A α alpha N ν nu B β beta Ξ ξ xi Γ γ gamma O o omicron ∆ δ delta Π π pi E  epsilon P ρ rho Z ζ zeta Σ σ sigma This is a supervisor’s copy of the notes.e. (0.2a) It has the following properties log(1) = 0. ε). The logarithm of x > 0. phi (i.1e) The logarithm. (0.6 Revision.2c) log(exp(x)) = x. The sine and cosine functions are defined by the series ∞ X (−)n x2n+1 sin(x) = .ac.e. (0.3a) n=0 (2n + 1)! ∞ X (−)n x2n cos(x) = . The exponential function.cam.uk. is defined as the unique solution y of the equation exp(y) = x .

β and γ respectively. y). (0. b and c respec- tively. The equation of a circle.6b) where t is the parametric variable and a is an arbitrary real number. y0 ) is given by y − y0 = m(x − x0 ) . (0. q) is given by (x − p)2 + (y − q)2 = r2 .J. (0. B and C be a. (0.4b) tan(x) ± tan(y) tan(x ± y) = .5c) The equation of a line. 2 2     x+y x−y sin(x) − sin(y) = 2 cos sin . (0. Let ABC be a triangle. (0. y = y0 + a m t .Certain trigonometric identities. Further suppose that the angles subtended at A. (x. Michaelmas 2006 . (0.4e) 2 2     x+y x−y cos(x) − cos(y) = −2 sin sin .ac.4d) 2 2     x+y x−y sin(x) + sin(y) = 2 sin cos . In 2D Cartesian co-ordinates.5b) c2 = a2 + b2 − 2ab cos γ .4a) cos(x ± y) = cos(x) cos(y) ∓ sin(x) sin(y) .uk.4g) 2 2 The cosine rule. B and C are α.cam.6a) In parametric form the equation of this line is given by x = x0 + a t . y). the equation of a circle of radius r and centre (p.5a) b2 = a2 + c2 − 2ac cos β . You should recall the following sin(x ± y) = sin(x) cos(y) ± cos(x) sin(y) . (0. (0. (0. It is not to be distributed to students.4c) 1 ∓ tan(x) tan(y)     x+y x−y cos(x) + cos(y) = 2 cos cos . Then the cosine rule (also known as the cosine formula or law of cosines) states that a2 = b2 + c2 − 2bc cos α . (0. (0. the equation of a line with slope m which passes through (x0 . (0.7) Mathematical Tripos: IA Algebra & Geometry (Part I) v c S.4f) This is a supervisor’s copy of the notes.Cowley@damtp. (x. In 2D Cartesian co-ordinates. Let the lengths of the sides opposite vertices A.

6. or reflection and dilatation). log(xy) = log(x) + log(y). using some of the Imperial examples? 3. multiplication (using suffix notation). Include a complex numbers study sheet. Add proof of rank-nullity theorem.J. 4. 8. 3. Use the following alternative proof of Schwarz’s inequality as an example: 2 2 2 kxk kyk − |x · y| = xi xi yj yj − xi yi xj yj = 12 xi xi yj yj + 12 xj xj yi yi − xi yi xj yj = 12 (xi yj − xj yi )(xi yj − xj yi ) > 0. and simplify notes on Gaussian elimination. Examples.g.Cowley@damtp. inverses. Introduce a sheet 0 covering revision.ac. Mathematical Tripos: IA Algebra & Geometry (Part I) vi c S. and geometric interpretation. n × n determinants. rotated reflection matrix. It is not to be distributed to students. e. 7.Suggestions. 5. e. Deliberate mistake a day (with mars bar). Add construction of the inverse matrix when doing Gaussian elimination. This is a supervisor’s copy of the notes. Change of basis: add pictures and an example of diagonalisation (e. Do (Shear Matrix)n .g.g. Michaelmas 2006 .cam. Additions/Subtractions? 1. and Gaussian elimination. 2. Add a section on diagonal matrices. 1. Add more grey summation signs when introducing the summation convention. 2.uk.

1. to be a number with the form z = a + ib. If β 2 < 4αγ then the square root is not equal to any real number.1) can now be rewritten p p β 4αγ − β 2 β 4αγ − β 2 z1 = − +i and z2 = − −i . where ∈ means ‘belongs to’. . We sometimes visualise real numbers as lying on a line (e. 1. 2. Note that i is sometimes denoted by j by engineers.1. denoted by Q.1. Michaelmas 2006 . q are integers√ (q 6= 0) 2 irrationals. rationals. Subject to us being happy with the introduction and existence of i. Mathematical Tripos: IA Algebra & Geometry (Part I) 1 c S. This has two roots p p β + β 2 − 4αγ β − β 2 − 4αγ z1 = − and z2 = − .1 Introduction This is a supervisor’s copy of the notes. where a. say z.J. α 6= 0 .0 Why Study This? For the same reason as we study real numbers: because they are useful and occur throughout mathematics. (1.2) Remark. (1. 1. If β 2 < 4αγ. For z = a + ib. e. . We say that z ∈ C. 0. We define a complex number.3) are of this form. denoted by Z.4) and i is as defined in (1. 2. 1. (1. OCR FP1 and FP3 this will be new.1 Real numbers The real numbers are denoted by R and consist of: integers. b = Im (z) : the imaginary part of z. . for a few of you who have not done. we sometimes write a = Re (z) : the real part of z. we can now always solve a quadratic equation. instead note the speed that new concepts are being introduced.uk. b ∈ R. − 3. π.1.g.2 The general solution of a quadratic equation Consider the quadratic equation αz 2 + βz + γ = 0 : α.3) 2α 2α 2α 2α where the square roots are now real [numbers]. −2. Note that z1 and z2 in (1.2). γ ∈ R .Cowley@damtp.ac. say. In order that we can always solve a quadratic equation. For many of you this section is revision. 1.cam. .3 Complex numbers Complex numbers are denoted by C. (1. we introduce √ i = −1 . p/q where p. (1. e. between any two points on a line there is another point.1 Complex Numbers 1. β. . For those for which it is revision. −1.g.1) 2α 2α If β 2 > 4αγ then the roots are real (there is a repeated root if β 2 = 4αγ). π . the rest of the reals. and between any two real numbers there is always another real number). It is not to be distributed to students. . please do not fall asleep.

ac.Cowley@damtp.8) Mathematical Tripos: IA Algebra & Geometry (Part I) 2 c S.0 ∈ C.3.0. Extending the number system from real (R) to complex (C) allows a number of important general- isations in addition to always being able to solve a quadratic equation. Proof. Hence a = c and b = d. 1. check that z1 z1−1 = 1 + i.2. is defined as a − ib. 3. 2.uk.g. c.2 Algebraic Manipulation of Complex Numbers In order to manipulate complex numbers simply follow the rules for reals. b. All the above operations on elements of C result in new elements of C.e.1 Complex conjugate The complex conjugate of z = a + ib. 1. A complex number 0 + i. b. but adding the rule i2 = −1. (1. It is not to be distributed to students.Remarks. Complex numbers were first used by Tartaglia (1500-1557) and Cardano (1501-1576). unique. The terms real and imaginary were first introduced by Descartes (1596-1650). z2 ∈ C. c. Michaelmas 2006 . (1. C contains all real numbers since if a ∈ R then a + i.cam. If z1 = z2 where z1 . This is described as closure: C is closed under addition and multiplication. where a. Hence for z1 = a + ib and z2 = c + id. d ∈ R. 4. Theorem 1. is zero. (1. (1.6) 1 1 a − ib a ib inverse : z1−1 = = = 2 − 2 .7) z a + ib a − ib a + b2 a + b2 Exercise: For z1−1 as defined in (1.b is said to be pure imaginary. if z = a + ib then z = a − ib. 1. d ∈ R such that z = a + ib = c + id. But the only number greater than or equal to zero that is equal to a number that is less than or equal to zero. Assume ∃ a. 2 2 Then a − c = i (d − b). then Re (z1 ) = Re (z2 ) and Im (z1 ) = Im (z2 ). i. Corollary 1. The representation of a complex number z in terms of its real and imaginary parts is This is a supervisor’s copy of the notes. Remark. A complex-valued function f is one that takes a complex number as ‘input’ and defines a new complex number f (z) as ‘output’.3 Functions of Complex Numbers We may extend the idea of functions to complex numbers.J. which is usually written as z but sometimes as z ∗ .1.5) multiplication : z1 z2 = (a + ib) (c + id) = ac + ibc + ida + (ib) (id) = (ac − bd) + i (bc + ad) . 1.7). it makes solving certain differential equations much easier. e. we have that addition/subtraction : z1 + z2 = (a + ib) ± (c + id) = (a ± c) + i (b ± d) . and so (a − c) = − (d − b) .

uk. r ∈ C then by using (1. is defined as 1/2 |z| = a2 + b2 . the real and imaginary parts of z are viewed as co- ordinates in an xy plot. Michaelmas 2006 .9c) f (z) ≡ f (z) = pz 2 + qz + r = p z 2 + q z + r. (1.12a) 2. (1. We label the 2D vector be- → tween the origin and (x.9c) 4. (z −1 ) = (z)−1 .9d) This is a supervisor’s copy of the notes. which is written as |z|.9b) 3. and re-invented by Jean-Robert Argand (1806). (1.Cowley@damtp.J.9b) and (1. The xy plane is referred to as the complex plane. Exercises. Then we can rep- resent each z = x + iy ∈ C by the point (x. (1. Show that 1. (1.2 Modulus The modulus of z = a + ib. Show that 1.12b) |z|2 1/03 1. and the y-axis as the imaginary axis.10) Example. Let f (z) = pz 2 + qz + r with p. Hence f (z) = p z 2 + q z + r. Remarks. the complex conjugate function f is defined by f (z) = f (z). by the complex number z.9a) f (z) = f (z). z z −1 = . The Argand diagram was invented by Caspar Wessel (1797).ac.cam. y). (1. z 1 z2 = z1 z2 . 2. |z|2 = z z.e. y).11) Exercises. and hence from (1. Given a complex-valued function f . q. Definition. 1. z1 ± z2 = z1 ± z2 .9a) 2. Mathematical Tripos: IA Algebra & Geometry (Part I) 3 c S. z = z. (1.3.4 The Argand Diagram Consider the set of points in two dimensional (2D) space referred to Cartesian axes. (1. say OP . We refer to the x-axis as the real axis. It is not to be distributed to students. i. 1. Such a plot is called an Argand diagram.

where P 0 is the point (x. Then z3 = z1 + z2 = (x1 + x2 ) + i (y1 + y2 ) . −y).13a) . If z1 . Addition. This is a supervisor’s copy of the notes. In terms of vector addition → → → OP3 =OP1 + OP2 . then OP 0 represents z. (1. z2 ∈ C then |z1 + z2 | 6 |z1 | + |z2 | . Let z1 = x1 + iy1 be associated with P1 .e. is associated with the point P3 that is obtained by completing the parallelogram P1 O P2 P3 .3. Modulus. The modulus of z corresponds to the mag- → nitude of the vector OP since 1/2 |z| = x2 + y 2 . and z2 = x2 + iy2 be associated with P2 . i. It is not to be distributed to students. which is sometimes called the triangle law. → → Complex conjugate. P 0 is P reflected in the x-axis. 1/02 Theorem 1. If OP represents z.

.

|z1 − z2 | > .

|z1 | − |z2 | .

ac.13a) implies that |z10 | 6 |z10 − z20 | + |z20 | . Then (1.Cowley@damtp. Michaelmas 2006 .13a) is known as the triangle inequality (and is in fact one of many).uk. Interchanging z10 and z20 we also have that |z20 − z10 | = |z10 − z20 | > |z20 | − |z10 | . Proof. Result (1. . (1. and hence that |z10 − z20 | > |z10 | − |z20 | .13a).13b) Remark.13b) follows. (1. By the cosine rule |z1 + z2 |2 = |z1 |2 + |z2 |2 − 2|z1 | |z2 | cos ψ 6 |z1 |2 + |z2 |2 + 2|z1 | |z2 | 2 = (|z1 | + |z2 |) . (1. Mathematical Tripos: IA Algebra & Geometry (Part I) 4 c S. Let z10 = z1 + z2 and z20 = z2 .13b) follows from (1. so that z1 = z10 − z20 .cam.J.

1 Geometric interpretation of multiplication Consider z1 . Michaelmas 2006 . cos θ2 − sin θ1 .5 Polar (Modulus/Argument) Representation Another helpful representation of complex numbers is obtained by using plane polar co-ordinates to rep- resent position in Argand diagram. Let x = r cos θ and y = r sin θ. (1. 1/06 The pair (r. Find equivalent results for z1 /z2 . then z = x + iy = r cos θ + ir sin θ = r (cos θ + i sin θ) . θ) uniquely. Remark. z2 written in modulus argument form: z1 = r1 (cos θ1 + i sin θ1 ) . sin θ2 +i (sin θ1 .5.17b) In words: multiplication of z1 by z2 scales z1 by | z2 | and rotates z1 by arg(z2 ). Then z1 z 2 = r1 r2 (cos θ1 . Mathematical Tripos: IA Algebra & Geometry (Part I) 5 c S. z does not specify (r.14) Note that 1/2 |z| = x2 + y 2 = r.uk. It is not to be distributed to students.17a) arg (z1 z2 ) = arg (z1 ) + arg (z2 ) (+2nπ with n an arbitrary integer). the integers) does not change z. cos θ1 )) = r1 r2 (cos (θ1 + θ2 ) + i sin (θ1 + θ2 )) . (1. In order to get a unique value of the argument it is sometimes more convenient to restrict θ to 0 6 θ < 2π 1.Cowley@damtp. For each z there is a unique value of the argument θ such that −π < θ 6 π.ac. θ) specifies z uniquely. (1. • Hence r is the modulus of z (mod(z) for short). z2 = r2 (cos θ2 + i sin θ2 ) .J. • The expression for z in terms of r and θ is called the modulus/argument form. (1. since adding 2nπ to θ (n ∈ Z. However. cos θ2 + sin θ2 . Exercise. • θ is called the argument of z (arg (z) for short). sometimes called the principal value of the argument.cam.15) This is a supervisor’s copy of the notes. (1. 1. i.16) Hence |z1 z2 | = |z1 | |z2 | .e.

p.ac.6. For n = 1 there is nothing to prove. Next note from applying (1.1 The real exponential function The real exponential function.18) 2! n=0 n! This series converges for all x ∈ R (see the Analysis I course). Show for x. From the power series definition (1. q Mathematical Tripos: IA Algebra & Geometry (Part I) 6 c S. It is not to be distributed to students. and thence exp(−1) = = e−1 . is defined by the power series ∞ x2 X xn exp(x) = exp x = 1 + x + ··· = .6 The Exponential Function 1. Michaelmas 2006 .1. and using (1.19). q Thence on taking the positive qth root   p p exp = eq . q Solution.J. where without loss of generality (wlog) q > 0. y ∈ R that (exp x) (exp y) = exp (x + y) .cam.19) we have that 1 exp(−1) exp(1) = exp(0) . q ∈ Z. Worked exercise. e For n 6 −2 proceed by induction as above. and thence by induction exp(n) = en .uk. ∞ ∞ X xn X y m exp x exp y = n=0 n! m=0 m! ∞ X r X xr−m y m = for n = r − m r=0 m=0 (r − m)! m! ∞ r X 1 X r! = xr−m y m r=0 r! m=0 (r − m)!m! ∞ X (x + y)r = by the binomial theorem r=0 r! = exp(x + y) .19) q times that   q p exp = exp (p) = ep . (1. For n > 2. Definition. Solution.18) with n = 0: exp(0) = 1 = e0 . exp(n) = exp(1) exp(n − 1) = e exp(n − 1) . (1. Show for n. that:   n p p exp(n) = e and exp = eq .Cowley@damtp. Also from (1. Worked exercise. We write exp(1) = e . exp(x).19) This is a supervisor’s copy of the notes.

For z ∈ C and z 6∈ R we define ez = exp(z) .22) n=0 (2n)! n=0 (2n + 1)! The result (1. This series converges for all finite |z| (again see the Analysis I course). Mathematical Tripos: IA Algebra & Geometry (Part I) 7 c S.4.. Next.J. For w ∈ C exp (iw) ≡ eiw = cos w + i sin w .21) then follows for w complex. Michaelmas 2006 . then it is consistent to write exp(y) = ey . (1.21) Proof. the complex exponential is defined by ∞ X zn exp(z) = . 1. Definition. for w ∈ C define the complex trigonometric functions by7 ∞ ∞ X w2nn X n w2n+1 cos w = (−1) and sin w = (−1) ...cam. (1.ac. 2! 4! 3! 5! ∞ 2n ∞ 2n+1 n w n w X X = (−1) +i (−1) n=0 (2n)! n=0 (2n + 1)! = cos w + i sin w .Cowley@damtp. we have that ∞ n X (iw) w2 w3 exp (iw) = = 1 + iw − −i . n=0 n! 2 3! w2 w4 w3 w5     = 1− + . For z ∈ C. From the above it follows that if y ∈ R.6. with the proof essentially as for (1. Remarks. For irrational x. This definition is consistent with the earlier results and definitions for z ∈ R.20) n=0 n! This is a supervisor’s copy of the notes. 1.. which is as required (as long as we do not mind living dangerously and re-ordering infinite series). Remark.. (1.6. When z ∈ R this definition is consistent with (1. Then from using the power series definitions for cosine and sine when their arguments are real. 7 Again note that these definitions are consistent with the definitions when the arguments are real. define ex = exp(x) . z2 ∈ C.3 The complex trigonometric functions Theorem 1. For z1 . First consider w real. (exp z1 ) (exp z2 ) = exp (z1 + z2 ) .uk.Definition. It is not to be distributed to students.19)..18).2 The complex exponential function Definition. + i w − + .

21) eiθ = cos θ + i sin θ . One solution is z = 1.6. From taking the complex conjugate of (1.17a) and (1.Cowley@damtp. for k ∈ Z. with z ∈ C and n a positive integer. Remarks.5. (1. θ ∈ R.27) and hence from §1. there are n ‘nth roots of unity’) Proof.cam. rn = 1 and n θ = 2kπ with k ∈ Z.17b). 1. (1.uk. 2/03 1.e.J. It is not to be distributed to students. Let w = θ where θ ∈ R. Michaelmas 2006 . Since by definition r.   cos w = 2 and sin w = (1. We conclude that within the requirement that 0 6 θ < 2π. Then from (1. (1. it follows that r = 1. exp (−iw) ≡ e−iw = cos w − i sin w .21).28) n Mathematical Tripos: IA Algebra & Geometry (Part I) 8 c S.26) with (again) r = |z| and θ = arg z.6.21) and (1. n − 1.4 Relation to modulus/argument form This is a supervisor’s copy of the notes.24) 2i 2/02 1. 1.23) it follows that 1 iw 1 eiw + e−iw . From (1. ei θ = cos θ + i sin θ = 1 . . Then n r ei θ = rn einθ = 1 . (1. (1. and thence that cos θ = 1 and sin θ = 0.7 Roots of Unity A root of unity is a solution of z n = 1. e − e−iw . Theorem 1.   confirming (1. . Seek more general solutions of the form z = r ei θ with the restriction 0 6 θ < 2π so that θ is not multi-valued.14) that z = r (cos θ + i sin θ) = reiθ . or otherwise. .5. We deduce that θ = 2kπ . there are n distinct roots given by 2kπ θ= with k = 0.25) It follows from the polar representation (1. 1. In this representation the multiplication of two complex numbers is rather elegant: z1 z2 = r1 ei θ1 r2 ei θ2 = r1 r2 ei(θ1 +θ2 ) .5 Modulus/argument expression for 1 Consider solutions of z = rei θ = 1 .6. There are n solutions of z n = 1 (i.ac. .23) 2.

cos pθ − sin θ. If we write ω = e2π i/n . and so holds for all n > 0. θ = 2πk/5 with k = 0. Then. ω n−1 . Put z = ei θ . 4. Larger (or smaller) values of k yield no new roots. There are thus five distinct roots given by This is a supervisor’s copy of the notes. .25) cos nθ + i sin nθ = ei (nθ) n = ei θ n = (cos θ + i sin θ) Remark. From (1. De Moivre’s theorem states that for θ ∈ R and n ∈ Z n cos nθ + i sin nθ = (cos θ + i sin θ) . If we write ω = e2π i/5 . . ω 3 . 2. sin pθ) = cos (p + 1) θ + i sin (p + 1) θ . Each root corresponds to a vertex of a pen- tagon. 3.Cowley@damtp. . (1. Further.30) holds for θ. (1. say n = −p. Michaelmas 2006 . ω.Remark. then we require that e5i θ = e2πki for k ∈ Z. Example.cam. Now consider n < 0. n ∈ C in the sense n that (cos nθ + i sin nθ) (single valued) is one value of (cos θ + i sin θ) (multivalued). sin pθ + i (sin θ. using the proved result for p > 0. (1. Alternative proof (unlectured). i.30) Proof. It is not to be distributed to students. 1.J. Mathematical Tripos: IA Algebra & Geometry (Part I) 9 c S.ac. and 1 + ω + ω2 + ω3 + ω4 = 0 . cos pθ + cos θ. Now argue by induction. ω 2 . Hence the result is true for n = p + 1.e.30) is true for n = 0.6. ω 2 . then the roots are 1. 2/06 1. for n > 2 n−1 X 1 − ωn 1 + ω + · · · + ω n−1 = ωk = = 0. Although De Moivre’s theorem requires θ ∈ R and n ∈ Z. Then p+1 p (cos θ + i sin θ) = (cos θ + i sin θ) (cos θ + i sin θ) = (cos θ + i sin θ) (cos pθ + i sin pθ) = cos θ. assume that (cos θ + i sin θ) = cos pθ + i sin pθ. Solve z 5 = 1. (1. ω 4 . .8 De Moivre’s Theorem Theorem 1. Solution. then the roots of z n = 1 are 1.uk. n −p (cos θ + i sin θ) = (cos θ + i sin θ) 1 = p (cos θ + i sin θ) 1 = cos p θ + i sin pθ = cos pθ − i sin p θ = cos n θ + i sin nθ Hence De Moivre’s theorem is true ∀ n ∈ Z.29) 1−ω k=0 because ω n = 1. p Assume true for n = p > 0. ω.

32a) Remark. This is a supervisor’s copy of the notes. z. z w is only defined upto an arbitrary multiple of e2kiπw . Example. The value of ii is given by ii = ei log i = ei(log |i|+i arg i) = ei(log 1+2ki π+iπ/2) 1 “ ” − 2k+ 2 π = e for any k ∈ Z (which is real). (1. the com- plex equation ey = x has a unique real solution. Since arg z is a multi-valued function. Definition. For z 6= 0. It is not to be distributed to students.9 Logarithms and Complex Powers We know already that if x ∈ R and x > 0.cam. Then eu+iv = z = reiθ .32b) Example. v ∈ R. Definition. w ∈ C. Michaelmas 2006 .1 Complex powers Recall the definition of xa . Since log z is multi-valued so is z w .uk. then log z = log | −x | +i arg (−x) = log | x | +(2k + 1)iπ for any k ∈ Z . (1. for x.e. (1.31) To understand the nature of the complex logarithm let w = u + iv with u.9. (1.33) Remark. i. 1. x > 0 and a irrational. namely xa = ea log x = exp (a log x) .Cowley@damtp.J. The principal value of log(−x) is log |x| + iπ. v = arg z = θ + 2kπ for any k ∈ Z . Mathematical Tripos: IA Algebra & Geometry (Part I) 10 c S. so is log z. a ∈ R. and hence eu = |z| = r . Thus w = log z = log |z| + i arg z .ac. We define log z for z ∈ C as ‘the’ solu- tion w of ew = z . for any k ∈ Z. The principal value of log z is such that −π < arg z = Im(log z) 6 π. namely y = log x (or ln x if you prefer). Definition. If z = −x with x ∈ R and x > 0.1. define z w by z w = ew log z .

Show that z0 w̄ − z¯0 w = 0 if and only if (iff) the line (1. which is the equation for a circle with centre (p.Cowley@damtp. d ∈ C are constants.34b). Solution. • Since |z − v|2 = (z̄ − v̄) (z − v). b.10.2 Circles In the Argand diagram. (1.uk. c. Remark.36) cz + d where a. w w̄ Thus z w̄ − z̄w = z0 w̄ − z̄0 w (1. Show that if z w̄ − z̄w = 0.cam.35b) 3/03 1. Worked exercise. Michaelmas 2006 . then z = γw for some γ ∈ R. We require that Mathematical Tripos: IA Algebra & Geometry (Part I) 11 c S. q) and radius r in Cartesian coordinates. w ∈ C with w 6= 0. 1. and the result follows.11 Möbius Transformations Consider a map of C → C (‘C into C’) defined by az + b z 7→ z 0 = f (z) = . z − z0 z̄ − z̄0 = .e.35a) i. 1. and varying λ ∈ R. Remarks. This is satisfied by z = 0.10 Lines and Circles in the Complex Plane 1.1 Lines For fixed z0 . the set of complex numbers z such that |z − v| = r. It is not to be distributed to students.34a) represents in the Argand diagram (complex plane) points on straight line through z0 and parallel to w. Since λ = (z − z0 )/w ∈ R. (1. a circle of radius r 6= 0 and centre v (r ∈ R.10. and hence that This is a supervisor’s copy of the notes. v ∈ C) is given by S = {z ∈ C : | z − v |= r} .J. (1.34b) is an alternative representation of the line. • If z = x + iy and v = p + iq then 2 2 |z − v|2 = (x − p) + (y − q) = r2 . and hence the line passes through the origin.ac. it follows that λ = λ̄. Exercise. If the line passes through the origin then put z = 0 in (1. then the equation of the line is z w̄ − z̄w = 0. If z0 w̄ − z¯0 w = 0.34a) passes through the origin. the equation z = z0 + λw (1. an alternative equation for the circle is 3/02 |z|2 − v̄z − vz̄ + |v|2 = r2 .

Those who have taken Further Mathematics A-level should then conclude something. except z = −d/c.38) can be deduced from (1.Cowley@damtp.uk. while (1. (1.36).38) maps C \{a/c} to C \{−d/c}. δ ∈ C. • f (z) maps every point of the complex plane. β = b. i.e. (ii) different points map to different points.e. • Adding the ‘point at infinity’ makes f complete. 1. 1. Then from (1. cz1 + d cz2 + d Remarks.e.36) by formal manipulation.38) cz 0 − a i. γz 0 + δ Then the combined map z 7→ z 00 is also a Möbius transformation: z 00 = g (z 0 ) = g (f (z))   α az+b cz+d + β =   γ az+b cz+d + δ α (az + b) + β (cz + d) = γ (az + b) + δ (cz + d) (αa + βc) z + (αb + βd) = . It is not to be distributed to students. The inverse (1. (f g)h = f (gh). For Möbius maps f . i. we require that az1 + b az2 + b 6= .11. Hence the set of all Möbius maps is therefore closed under composition. 2. so that the map is finite (except at z = −d/c). b. or equivalently (ad − bc)(z1 − z2 ) 6= 0 .2 Inverse For the a. c. hence we need only require that (ad − bc) 6= 0. d ∈ R as in (1. γ. α = −d.38) is the inverse to (1. β. We conclude that (1. consider the Möbius map −dz 0 + b z 00 = . and αδ − βγ 6= 0 . to infinity).cam. g and h demonstrate that the maps are associative. if z1 6= z2 then z10 6= z20 .36). Mathematical Tripos: IA Algebra & Geometry (Part I) 12 c S.37).1 Composition Consider a second Möbius transformation αz 0 + β z 0 7→ z 00 = g (z 0 ) = where α.ac. i. • Condition (i) is a subset of condition (ii).37) (γa + δc) z + (γb + δd) where we note that (αa + βc) (γb + δd) − (αb + βd) (γa + δc) = (ad − bc)(αδ − βγ) 6= 0. Remarks.11. Michaelmas 2006 . and vice versa. (1. Exercise. (i) c and d are not both zero. z 00 = z. i.e. γ = c and δ = −a. (ad − bc) 6= 0 .e.J. (1. 1.36) maps C \{−d/c} to C \{a/c}. into another (z = −d/c is mapped This is a supervisor’s copy of the notes.

(1. e. becomes z 0 = az0 + λaw = z00 + λw0 . c = 0 and d = 1 so that z 0 = az = |a||z| ei(arg a+arg z) . while circles map to circles of the same radius but with a centre offset by b. Next put b = 0. which is just another line.g lines map to parallel lines. This map scales z by |a| and rotates z by arg a about the origin O.11.39a) This map represents a translation.1. It is not to be distributed to students. c = 0 and d = 1 to obtain z0 = z + b . where z00 = az0 and w0 = aw.39b) This is a supervisor’s copy of the notes. 3/06 Dilatation and rotation. (1. Put a = 1.3 Basic Maps Translation. The line z = z0 + λw. where λ ∈ R and w ∈ C. The circle |z − v| = r becomes .

0 .

.

z |z 0 − v 0 | = r0 . .

.

− v .

= r or equivalently .

a .

z0 z̄ By multiplying by |z 0 |2 . becomes w̄ w − 0 = z0 w̄ − z̄0 w . or equivalently (see (1. this equation can be rewrit- ten successively as z¯0 w̄ − z 0 w = (z0 w̄ − z̄0 w) z 0 z¯0 z¯0 w̄ z0w z 0 z¯0 − − =0 z0 w̄ − z¯0 w z¯0 w − z0 w̄ .34b)) z w̄ − z̄w = z0 w̄ − z̄0 w . b = 1. Thus if z = reiθ then 1 −iθ z0 = e . etc. r Hence this map represents inversion in the unit circle cen- tred on the origin O. The line z = z0 + λw. Inversion and reflection.e. i. Now put a = 0. so that z 0 = z1 .. which is just another circle. and reflection in the real axis. |z 0 | = |z|−1 and arg z 0 = − arg z. c = 1 and d = 0. where v 0 = av and r0 = |a|r.

.

2 .

.

2 .

0 .

z − w̄ .

.

=.

.

w̄ .

.

.

z0 w̄ − z¯0 w .

.

z0 w̄ − z¯0 w .

35a) this is a circle (which passes . From (1.

through.

the w̄ w̄ origin) with centre z0 w̄−z¯0 w and radius .

z0 w̄−z¯0 w .

The ..

.

uk.ac. Michaelmas 2006 .J.Cowley@damtp. z¯0 w̄ − z 0 w = 0. and the mapped curve. is also a straight line through the origin. exception is when z0 w̄−z̄0 w = 0. in which case the original line passes through the origin. Mathematical Tripos: IA Algebra & Geometry (Part I) 13 c S.cam.

Further. under the map z 0 = 1 z the circle | z − v |= r becomes .

.

.

1 .

− v .

.e. i. |1 − vz 0 | = r|z 0 | . = r .

.

z0 .

cz + d (ii) Construct a similar sequence if c = 0 and d 6= 0. Summary. Hence (1 − vz 0 ) 1 − v̄ z¯0 = r2 z¯0 z 0 .  or equivalently z 0 z¯0 |v|2 − r2 − vz 0 − v̄ z¯0 + 1  = 0. Mathematical Tripos: IA Algebra & Geometry (Part I) 14 c S. or equivalently This is a supervisor’s copy of the notes. The above implies that a general Möbius map transforms circles and straight lines to circles and straight lines (since each constituent transformation does so). while circles and straight lines that do pass through origin map to straight lines. and the map reduces to vz 0 + v̄ z¯0 = 1 .35b) this is the equation for a circle with centre v̄/(|v|2 − r2 ) and radius r/(|v|2 − r2 ).cam.Cowley@damtp. dilatation and rotation.11. v v̄ |v|2 r2 z 0 z¯0 − z 0 − ¯0 + z 2 = 2 . Under inversion and reflection. in which case the original circle passed through the origin.uk. The exception is if |v|2 = r2 . 1. It is not to be distributed to students.4 The general Möbius map The reason for introducing the basic maps above is that the general Möbius map can be generated by composition of translation. (|v|2 − r2 ) (|v|2 − r2 ) (|v|2 − r2 ) (|v|2 − r2 ) From (1. To see this consider the sequence: dilatation and rotation z 7→ z1 = cz (c 6= 0) translation z1 7→ z2 = z1 + d inversion and reflection z2 7→ z3 = 1/z2   bc − ad dilatation and rotation z3 7→ z4 = z3 (bc 6= ad) c translation z4 7→ z5 = z4 + a/c (c 6= 0) Exercises. which is the equation of a straight line. (i) Show that az + b z5 = .ac. Michaelmas 2006 . and inversion and reflection. circles and straight lines which do not pass through the origin map to circles.J.

subtraction. with length |v| and direction that of v (the direc- tion/sense of v is from A to B).Cowley@damtp. the position of a particle. y. Remarks. that is a function of position (x. e.2 Vector Algebra 2.e. It is not to be distributed to students. Remarks.g.g. 2. multiplication and. concentration.1. • Often the position vector is represented by x rather than r. the direction of propagation of a wave. However other quantities have both a magnitude and a direction. y. in the case of our example we write F ≡ F(x. You have learnt how to manipulate such scalars (e. that is a function of position (x. This is a supervisor’s copy of the notes. temperature.g. y. • A vector. but even then the length (i. |u| = |v|. Mathematical Tripos: IA Algebra & Geometry (Part I) 15 c S. Michaelmas 2006 . Examples. 2. A scalar.e. from some chosen origin O. temperature T . v. differentiation) if you are to be able to describe them mathematically. On the over- head/blackboard I will put a squiggle under the v . time.g. with r =OP and r = OP = |r|.1 Geometric representation → We represent a vector v as a line segment. • For the purpose of this course the notes will represent vectors in bold. an electric field. • The position vector is an example of a vector field. u is parallel to v and in both vectors are in the same direction/sense.0 Why Study This? Many scientific quantities just have a magnitude.e. z) is referred to as a scalar field . We refer to such numbers as scalars. by addition.J. You need to know how to manipulate these quantities (e. z).g. 8 Sophisticated mathematicians use neither bold nor squiggles.1 Vectors Definition. say AB. and they are in the same direction. force F. differentiation) since your first day in school (or possibly before that). i.8 ∼ • The magnitude of a vector v is written |v|. e. by addition. a magnetic field. and hence to the po- sition vector of that point. e. i. subtraction. a force. density. the velocity of a particle.uk. multiplication. (i) Every point P in 3D (or 2D) space has a position → vector. r.g.g. in the case of our example we write T ≡ T (x. A quantity that is specified by a [positive] magnitude and a direction in space is called a vector. z) is referred to as a vector field . next term. z). • Two vectors u and v are equal if they have the same magnitude. y. 4/03 magnitude) is usually represented by r. e.cam. (ii) Every complex number corresponds to a unique point in the complex plane. e.ac. Such quantities can be completely specified by a single number.

i. i. but I will not always. Depending on context 0 will sometimes mean 0 and sometimes 0. and it has the same direction/sense as a if λ > 0. Mathematical Tripos: IA Algebra & Geometry (Part I) 16 c S.3) (iii) Addition is associative.J.ac.2 Properties of Vectors 2.7b) 2. However.2.e.8a) λ(a + b) = λa + λb . 2. but to have the opposite direction/sense (so that it is anti-parallel ). |a + b| 6 |a| + |b| . A number of properties follow from the above definition.6) (vi) Define the vector −a to be parallel to a. I will try and get it right. (2. (2. a + b = b + a.1 Addition Vectors add according to the parallelogram rule: a + b = c.2. is parallel to a. (2. (2. write a = 0.2 Multiplication by a scalar 4/02 If λ ∈ R then λa has magnitude |λ||a|.cam. (2. This is a supervisor’s copy of the notes. (2. on the overhead/blackboard you will need to be more ‘sophisticated’.5) This is proved in an analogous manner to (1. and from (2. (2. µ ∈ R.7a) Define subtraction of vectors by b − a ≡ b + (−a) . i. a + (b + c) = (a + b) + c . (2.2) (ii) Addition is commutative.1a) or equivalently → → → OA + OB = OC . it is also true that → → → OA + AC=OC . (2.1b) where OACB is a parallelogram. (2. but the opposite direction/sense as a if λ < 0 (see below for λ = 0).8b) 9 I have attempted to always write 0 in bold in the notes. (2. Remarks.13a).4) (iv) There is a triangle inequality analogous to (1. Distributive law: (λ + µ)a = λa + µa .uk. (v) If |a| = 0.9 For all vectors b b + 0 = b. if I have got it wrong somewhere then please let me know. It is not to be distributed to students.Cowley@damtp.3) 0 + b = b .13a) using the cosine rule. In what follows λ. This is called the negative of a and is such (−a) + a = 0 .e. → → (i) Since OB=AC.e. where 0 is the null vector or zero vector. Michaelmas 2006 . to have the same magnitude as a.

(2.10b) (−1) a = −a since | − 1| |a| = |a| and −1 < 0. Unit vectors.6).7b).10b) since This is a supervisor’s copy of the notes.8a). 1 and -1: 0 a = 0 since 0 |a| = 0. (2. Definition. Suppose a 6= 0. (2. The vector c = λa + µb is described as a linear combination of a and b.9) Multiplication by 0. (2.10a) 1a = a.11) |a| â is termed a unit vector since . (2. (−1)a = (−1)a + 0 = (−1)a + (a − a) = ((−1)a + 1 a) − a = (−1 + 1)a − a = 0−a = −a .4). then define a â = . (2. (2. (2.Associative law: λ(µa) = (λµ)a .7a). (2. (2.10a) and (2.10c) We can also recover the final result without appealing to geometry by using (2. It is not to be distributed to students.

.

.

1 .

1 |â| = .

.

.

.

1). (2.7.Cowley@damtp. → RS = 12 (c + d) = − 21 (a + b) → = − PQ.2. from using (2. Let a.ac. BC and CD. |a| |a| A ˆ is often used to indicate a unit vector. Since P Q and SR have equal magnitude and are parallel.uk. c and d represent the sides DA.12). and let P . Denote the vertices of the quadrilateral by A. Q. 2. see §2. Then since the quadrilateral is closed a + b + c + d = 0. but note that this is a convention that is often broken (e. Michaelmas 2006 . R and S denote the respective midpoints. 04/06 Mathematical Tripos: IA Algebra & Geometry (Part I) 17 c S.cam.12) Further → → → P Q=P A + AQ= 12 a + 12 b . C → and D. B. → → → AB. P QSR is a parallelogram. Similarly. → → and thus SR=P Q. b.J.3 Example: the midpoints of the sides of any quadrilateral form a parallelogram This an example of the fact that the rules of vector manipulation and their geometric interpretation can be used to prove geometric theorems.g. |a| = |a| = 1 .

If instead λ < 0 then a · (λb) = |a||λb| cos(π − θ) = −|λ||a||b| cos θ = −|λ| a · b = λa · b. 2.15) (iii) If 0 6 θ < 12 π.1 Properties of the scalar product (i) The scalar product of a vector with itself is the square of its modulus: a · a = |a|2 = a2 . (v) Suppose λ ∈ R. It is not to be distributed to students. In summary a · (λb) = (λa) · b = λ a · b . (2. The scalar product is also referred to as the dot product. then a · b < 0.ac. b b a0 = |a0 | = |a| cos θ . (2.14) (ii) The scalar product is commutative: a · b = b · a.uk.Cowley@damtp.3 Scalar Product Definition.15).e. This is a supervisor’s copy of the notes. Thus since a0 is parallel to b.J. From geometry. then a and b must be orthogonal (i. The scalar product of two vectors a and b is defined to be the real (scalar) number a · b = |a||b| cos θ . Similarly.17a) |b| |b| Mathematical Tripos: IA Algebra & Geometry (Part I) 18 c S. If λ > 0 then a · (λb) = |a||λb| cos θ = |λ||a||b| cos θ = |λ| a · b = λa · b. (λa) · b = λ a · b.13) where 0 6 θ 6 π is the angle between a and b mea- sured in the positive (anti-clockwise) sense once they have been placed ‘tail to tail’ or ‘head to head’. The pro- jection of a onto b is that part of a that is parallel to b (which here we will denote by a0 ).2.3. Remark. we need to discuss projections. θ = 12 π). (2. Michaelmas 2006 .16) 2. (2.2 Projections Before deducing one more property of the scalar product. (2. |a0 | = |a| cos θ (assume for the time being that cos θ > 0). or by using (2. then a · b > 0. (iv) If a 6= 0 and b 6= 0 and a · b = 0.3. while if 21 π < θ 6 π.cam.

(iii) a × b has the sense/direction defined by the ‘right-hand rule’. the second finger in the direction of b. (2.3 Another property of the scalar product We wish to show that a · (b + c) = a · b + a · c .J.4 Example: the cosine rule → → → BC 2 ≡ | BC |2 = | BA + AC |2     → → → → = BA + AC · BA + AC → → → → → → → → = BA · BA + BA · AC + AC · BA + AC · AC → → = BA2 + 2 BA · AC +AC 2 = BA2 + 2 BA AC cos θ + AC 2 = BA2 − 2 BA AC cos α + AC 2 . and then a × b is in the direction of the thumb.17b) (after exchanging a for b. and so λ + µ = γ). etc.17b) |a||b| |b| |b|2 where b̂ is the unit vector in the direction of b. Michaelmas 2006 .19) with 0 6 θ 6 π defined as before. Mathematical Tripos: IA Algebra & Geometry (Part I) 19 c S. (ii) a × b is perpendicular/orthogonal to both a and b (if a × b 6= 0). i.13) a·b b a·b a0 = |a| = b = (a · b̂) b̂ . and b or c or (b + c) for a. point the index finger in the direction of a. 5/03 2. if λa + µa = γa. so henceforth assume a 6= 0. (2. Show that (2. take a right hand.e. It is not to be distributed to students.ac. (2. 2. |a|2 Now use (2. Then from (2.uk. hence (λ + µ)|a|2 = γ|a|2 . The vector product a × b of an ordered pair a. 2.3.Exercise.cam. then (λ + µ)a = γa.18) The result is [clearly] true if a = 0. b is a vector such that (i) |a × b| = |a||b| sin θ .e.8a) on the LHS before ‘dotting’ both sides with a to obtain the result (i.4 Vector Product Definition. This is a supervisor’s copy of the notes. Hence from (2.17a) remains true if cos θ < 0.3.Cowley@damtp.) a·b a·c 2 a+ a = {projection of b onto a} + {projection of c onto a} |a| |a|2 = {projection of (b + c) onto a} (by geometry) a · (b + c) = a.

then θ = 0 or θ = π. It is not to be distributed to students. a and b are parallel (or equivalently there exists λ ∈ R such that a = λb). (2.20d) by first noting by geometry that {projection of b onto plane ⊥ to a} + {projection of c onto plane ⊥ to a} = {projection of (b + c) onto plane ⊥ to a} . Mathematical Tripos: IA Algebra & Geometry (Part I) 20 c S.e. (vi) We can use this geometric interpretation to show that a × (b + c) = a × b + a × c .20c) (v) Consider b00 = â × b. (2.cam. First project b onto a plane orthogonal to â to generate the vector b0 .ac. (i) The vector product is also referred to as the cross product. Further.4. (ii) The vector product of a vector with itself is zero: a × a = 0. and b00 has the same magnitude as b0 . and thus has the correct magnitude and sense/direction to be equated to â × b. Remarks. (ii) An alternative notation (that is falling out of favour except on my overhead/blackboard) is a ∧ b.20b) (iii) If a 6= 0 and b 6= 0 and a × b = 0. Michaelmas 2006 .uk. by construc- tion b00 is orthogonal to both a and b. then rotate b0 about â by π2 in an ‘anti-clockwise’ direction (‘anti-clockwise’ when looking in the opposite direction to â).e. i. (2. a a b θ b’’ π 2 b’ b’ b’ is the projection of b onto b’’ is the result of rotating the vector b’ the plane perpendicular to a through an angle π 2 anti−clockwise about a (looking in the opposite direction to a ) By geometry |b0 | = |b| sin θ = |â × b|.J.20a) This is a supervisor’s copy of the notes. 0 b0 + c0 = (b + c) . 5/02 2. (2.Cowley@damtp. (iv) It follows from the definition of the vector product a × (λb) = λ (a × b) . This vector can be constructed by two operations.1 Properties of the vector product (i) The vector product is not commutative (from the right-hand rule): a × b = −b × a . i.

0
and by then rotating b0 , c0 and (b + c) by π
2 ‘anti-
clockwise’ about a to show that
00
b00 + c00 = (b + c) .

This is a supervisor’s copy of the notes. It is not to be distributed to students.
2.4.2 Vector area of a triangle/parallelogram

Let O, A and B denote the vertices of a triangle, and let N B
→ →
be the altitude through B. Denote OA and OB by a and b
respectively. Then
1 1
Area of triangle = 2 OA . N B = 2 OA . OB sin θ = 21 |a × b| .
1
2a × b is referred to as the vector area of the triangle. It has
the same magnitude as the area of the triangle, and is normal
to OAB, i.e. normal to the plane containing a and b.

→ →
Let O, A, C and B denote the vertices of a parallelogram, with OA and OB as before. Then

Area of parallelogram = |a × b| ,

and the vector area is a × b.

2.5 Triple Products

Given the scalar (‘dot’) product and the vector (‘cross’) product, we can form two triple products.

Scalar triple product:
(a × b) · c = c · (a × b) = − (b × a) · c , (2.21)
from using (2.15) and (2.20a).
Vector triple product:

(a × b) × c = −c × (a × b) = − (b × a) × c = c × (b × a) , (2.22)

from using (2.20a).
5/06

2.5.1 Properties of the scalar triple product
Volume of parallelepipeds. The volume of a parallelepiped
(or parallelipiped or parallelopiped or parallelopi-
pede or parallelopipedon) with edges a, b and c is

Volume = Base Area × Height
= |a × b| |c| cos φ
= | (a × b) · c| (2.23a)
= (a × b) · c > 0 , (2.23b)

if a, b and c have the sense of the right-hand rule.

Mathematical Tripos: IA Algebra & Geometry (Part I) 21
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

Identities. Assume that the ordered triple (a, b, c) has the sense of the right-hand rule. Then so do the
ordered triples (b, c, a), and (c, a, b). Since the ordered scalar triple products will all equal the
volume of the same parallelepiped it follows that

(a × b) · c = (b × c) · a = (c × a) · b . (2.24a)

Further the ordered triples (a, c, b), (b, a, c) and (c, b, a) all have the sense of the left-hand rule,
and so their scalar triple products must all equal the ‘negative volume’ of the parallelepiped; thus

(a × c) · b = (b × a) · c = (c × b) · a = − (a × b) · c . (2.24b)

It also follows from (2.15) and (2.24a) that

(a × b) · c = a · (b × c) , (2.24c)

This is a supervisor’s copy of the notes. It is not to be distributed to students.
and hence the order of the ‘cross’ and ‘dot’ is inconsequential.10 For this reason we sometimes use
the notation
[a, b, c] = (a × b) · c . (2.24d)

Coplanar vectors. If a, b and c are coplanar then

[a, b, c] = 0 ,

since the volume of the parallelepiped is zero. Conversely if non-zero a, b and c are such that
[a, b, c] = 0, then a, b and c are coplanar.
6/03

2.6 Bases and Components

2.6.1 Two dimensional space

First consider 2D space, an origin O, and two non-zero and non-parallel vectors a and b. Then the
position vector r of any point P in the plane can be expressed as

r =OP = λa + µb , (2.25)

for suitable and unique real scalars λ and µ.
Geometric construction. Draw a line through P par-
→ →
allel to OA= a to intersect OB= b (or its ex-
tension) at N (all non-parallel lines intersect).
Then there exist λ, µ ∈ R such that
→ →
ON = µb and N P = λa ,

and hence

r =OP = λa + µb .

Definition. We say that the set {a, b} spans the set of vectors lying in the plane.

Uniqueness. Suppose that λ and µ are not-unique, and that there exists λ, λ0 , µ, µ0 ∈ R such that

r = λa + µb = λ0 a + µ0 b .

Hence
(λ − λ0 )a = (µ0 − µ)b ,
and so λ − λ0 = µ − µ0 = 0, since a and b are not parallel (or ‘cross’ first with a and then b).

10 What is important is the order of a, b and c.

Mathematical Tripos: IA Algebra & Geometry (Part I) 22
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

Definition. If for two vectors a and b and α, β ∈ R,

αa + βb = 0 ⇒ α = β = 0, (2.26)

then we say that a and b are linearly independent.

Definition. We say that the set {a, b} is a basis for the set of vectors lying the in plane if it is a spanning
set and a and b are linearly independent.

Remark. {a, b} do not have to be orthogonal to be a basis.

This is a supervisor’s copy of the notes. It is not to be distributed to students.
2.6.2 Three dimensional space

Next consider 3D space, an origin O, and three non-zero and non-coplanar vectors a, b and c (i.e.
[a, b, c] 6= 0). Then the position vector r of any point P in space can be expressed as

r =OP = λa + µb + νc , (2.27)

for suitable and unique real scalars λ, µ and ν.
Geometric construction. Let Πab be the plane con-
taining a and b. Draw a line through P parallel

to OC= c. This line cannot be parallel to Πab
because a, b and c are not coplanar. Hence it
will intersect Πab , say at N , and there will ex-

ist ν ∈ R such that N P = νc. Further, since

ON lies in the plane Πab , from §2.6.1 there

exists λ, µ ∈ R such that ON = λa + µb. It
follows that
→ → →
r = OP = ON + N P
= λa + µb + νc ,

for some λ, µ, ν ∈ R.

We conclude that if [a, b, c] 6= 0 then the set {a, b, c} spans 3D space. 6/02

Uniqueness. We can show that λ, µ and ν are unique by construction. Suppose that r is given by (2.27)
and consider

r · (b × c) = (λa + µb + νc) · (b × c)
= λa · (b × c) + µb · (b × c) + νc · (b × c)
= λa · (b × c) ,

since b · (b × c) = c · (b × c) = 0. Hence, and similarly or by permutation,

[r, b, c] [r, c, a] [r, a, b]
λ= , µ= , ν= . (2.28)
[a, b, c] [a, b, c] [a, b, c]

The uniqueness of λ, µ and ν implies that the set {a, b, c} is linearly independent, and thus since the set
also spans 3D space, we conclude that {a, b, c} is a basis for 3D space.

Remark. {a, b, c} do not have to be mutually orthogonal to be a basis.

Mathematical Tripos: IA Algebra & Geometry (Part I) 23
c S.J.Cowley@damtp.cam.ac.uk, Michaelmas 2006

j and k respectively. Definition.J. However. Let OX. k] = 1 . (2. vy . ν) the components of r with respect to {a. If for a vector v and a Cartesian basis {i. and if for a vector r. j. k}. 0. 2. 0) .ac. (2. j.Cowley@damtp. k = (0. (2. (i) Assuming that we know the basis vectors (and remember that there are an uncountably infinite number of Cartesian axes). It is not to be distributed to students. we often write v = (vx . vz ) .29b) i × j = k.29a) i · j = j · k = k · i = 0. j be the unit vector along OY .30) with i.29d) Definition. we deduce from (2.29a) and (2. we define (vx . in n-dimensional space we would find that the basis had n vectors. c} is a basis for 3D space. 0. vz = v · k . 2. However. j = (0. in which case they are said to define a orthonormal basis.3 Higher dimensional spaces We can ‘boot-strap’ to higher dimensional spaces.34) Mathematical Tripos: IA Algebra & Geometry (Part I) 24 c S. vy = v · j .31) Hence for all 3D vectors v v = (v · i) i + (v · j) j + (v · k) k . vy . (2. (2. c}. k} form a basis for 3D space satisfying i · i = j · j = k · k = 1. 1) .7. j. Then {i. v = vx i + vy j + vz k .1 The Cartesian or standard basis in 3D We have noted that {a. Michaelmas 2006 . OY .29c) [i. b. k}. Let i be the unit vector along OX . In fact we define the dimension of a space as the numbers of [different] vectors in the basis. k × i = j. b.uk. vz ∈ R. vy . It is also conventional to order them so that they are right-handed.32) Remarks.30) where vx . OZ be a right-handed set of Cartesian axes. (2. j. c} do not have to be mutually orthogonal (or right-handed) to be a basis. If {a. then we call (λ. matters are simplified if the basis vectors are mutually orthogonal and have unit magnitude. 0) . this is not necessarily the best way of looking at higher dimensional spaces. By ‘dotting’ (2. j × k = i. vz ) to be the Cartesian components of v with respect to {i.cam. This is a supervisor’s copy of the notes. b. (2.6.29b) that vx = v · i . r = λa + µb + νc .33) In terms of this notation i = (1. (2. where it is not conventional to add a ˆ. 1. k be the unit vector along OZ . (2. µ.2.7 Components Definition.

. (2. . ty . We wish to show that a × (b × c) = (a · c)b − (a · b)c . 2. z) . tx = t · i = |t| |i| cos α .29c) a × b = (ax i + ay j + az k) × (bx i + by j + bz k) = ax i × bx i + ax i × by j + ax i × bz k + . the direction cosines of t are defined by t = (cos α. (2.20c). b = bx i + by j + bz k and c = cx i + cy j + cz k .e. (2. cos β. . = ax bx + ay by + az bz . (2.42) Mathematical Tripos: IA Algebra & Geometry (Part I) 25 c S. (2. = ax bx i × i + ax by i × j + ax bz i × k + . From repeated application of (2.36a) 6/06 where α is the angle between t and i.40) Scalar triple product.20a). z). (2. .J. Hence if β and γ are the angles between t and j. cos γ) .36b) 2. (2.9) λa + µb = (λax + µbx )i + (λay + µby )j + (λaz + µbz )k .18).7. (ii) If the point P has the Cartesian co-ordinates (x. Michaelmas 2006 . j. (2.37) Then we can deduce a number of vector identities for components (and one true vector identity).2 Direction cosines If t is a unit vector with components (tx . (2.Cowley@damtp.29a) and (2.29b) a·b = (ax i + ay j + az k) · (bx i + by j + bz k) = ax i · bx i + ax i · by j + ax i · bz k + . tz ) with respect to the basis {i.8 Vector Component Identities Suppose that a = ax i + ay j + az k . .38) Scalar product. (2. (2.ac. From (2.20b). (2. . . . (2.9). y. then This is a supervisor’s copy of the notes.41) Vector triple product.20d). From repeated application of (2. (2.9). Addition.cam. From repeated application of (2. It is not to be distributed to students.40) (a × b) · c = ((ay bz − az by )i + (az bx − ax bz )j + (ax by − ay bx )k) · (cx i + cy j + cz k) = ax by cz + ay bz cx + az bx cy − ax bz cy − ay bx cz − az by cx . (2. i. (2. so we often write R2 /R3 for 2D/3D space.uk. then the position vector → OP = r = x i + y j + z k . (2. r = (x.39) and (2. and t and k. = (ay bz − az by )i + (az bx − ax bz )j + (ax by − ay bx )k .8b) and (2. (2. respectively. = ax bx i · i + ax by i · j + ax bz i · k + . y. (2.16).39) Vector product.35) (iii) Every vector in 2D/3D space may be uniquely represented by two/three real numbers.15). k}.8a).

The curves of constant r and curves of constant θ intersect at right angles. Michaelmas 2006 . as a basis in 2D Cartesian co-ordinates.ac. er · eθ = 0 . or note that if its true for one component it must be true for all components because of the arbitrary choice of axes. but a key difference in the case of polar co-ordinates is that the unit vectors are position dependent.J. Then from (2. y) so that x = r cos θ .45b) Exercise. Similarly we can use the unit vectors in the directions of increasing r and θ respectively as ‘a’ basis in 2D plane polar co-ordinates.46) Mathematical Tripos: IA Algebra & Geometry (Part I) 26 c S.42) has no component in the direction a.1 2D plane polar co-ordinates Define 2D plane polar co-ordinates (r. It is not to be distributed to students. are orthogonal. π < θ < 2π if y < 0. To prove this. Now proceed similarly for the y and z components. (2. Identity (2. (2. y = r sin θ .cam.44b) Equivalently i = cos θ er − sin θ eθ .43a) where 0 6 r < ∞ and 0 6 θ < 2π.44a) eθ = − sin θ i + cos θ j . Define er as unit vectors orthogonal to lines of con- stant r in the direction of increasing r. define eθ as unit vectors orthogonal to lines of con- stant θ in the direction of increasing θ.e. Similarly. (2.Cowley@damtp.9 Polar Co-ordinates Polar co-ordinates are alternatives to Cartesian co-ordinates systems for describing positions in space. θ = arctan . begin with the x-component of the left-hand side of (2. eθ · eθ = 1 . They naturally lead to alternative sets of orthonormal basis vectors. (2. θ = 0 if x > 0 and y = 0. orthogonal unit vectors in the directions of increasing x and y respectively.e. (2. From inverting (2. 2.43b) x where the choice of arctan should be such that 0 < θ < π if y > 0. (2.Remark.9. and θ = π if x < 0 and y = 0. no component in the direction of the vector outside the parentheses.uk. i. θ) in terms of 2D Cartesian co-ordinates (x. This is a supervisor’s copy of the notes. We use i and j. Remark. Confirm that er · er = 1 .45a) j = sin θ er + cos θ eθ .42). 7/03 2.40)   a × (b × c) x ≡ (a × (b × c) · i = ay (b × c)z − az (b × c)y = ay (bx cy − by cx ) − az (bz cx − bx cz ) = (ay cy + az cz )bx + ax bx cx − ax bx cx − (ay by + az bz )cx = (a · c)bx − (a · b)cx   = (a · c)b − (a · b)c · i ≡ (a · c)b − (a · b)c x .43a) it follows that 1 y r = x2 + y 2 2 . (2. i. Then er = cos θ i + sin θ j .

π < φ < 2π if y < 0.49a) it follows that 1 y ρ = x2 + y 2 2 .9. eθ } vary with θ. z = z.ac.J. (2. → (ii) The polar co-ordinates of a point P are (r.Given the components of a vector v with respect to a basis {i. in 3D there is a good reason for not using r and θ as we shall see once we get to spherical polar co-ordinates. It is not to be distributed to students. eθ } are (r. This means that even constant vectors have com- ponents that vary with position. e. eθ }: v = vx i + vy j = vx (cos θ er − sin θ eθ ) + vy (sin θ er + cos θ eθ ) = (vx cos θ + vy sin θ) er + (−vx sin θ + vy cos θ) eθ .45a) the components of i with respect to the basis {er . Mathematical Tripos: IA Algebra & Geometry (Part I) 27 c S.47) Example. while the components of r =OP with respect to the basis {er . from (2. (2. From inverting (2. j} we can deduce the components with respect to the basis {er . (i) It is crucial to note that {er . y. (2.49a) where 0 6 ρ < ∞. θ) is (ρ. and φ = π if x < 0 and y = 0. y = ρ sin φ . φ). sin θ). (2.g.48) Remarks. 2. 0 6 φ < 2π and −∞ < z < ∞. IMHO there is also a good reason for not using r and θ in plane polar co-ordinates.43a) and (2.Cowley@damtp. This is a supervisor’s copy of the notes. eθ } are (cos θ.uk. z) so that x = ρ cos φ . φ. z ez eφ P eρ r z k O y i j φ eφ ρ N eρ x 11 Note that ρ and φ are effectively aliases for the r and θ of plane polar co-ordinates. φ = arctan . which as shall become clear has its advantages. 0) (in the case of components the θ dependence is ‘hiding’ in the basis). However. Michaelmas 2006 .49b) x where the choice of arctan should be such that 0 < φ < π if y > 0.47) that r = (x cos θ + y sin θ) er + (−x sin θ + y cos θ) eθ = rer . z)11 in terms of 3D Cartesian co-ordinates (x. (iii) An alternative notation to (r.2 Cylindrical polar co-ordinates Define 3D cylindrical polar co-ordinates (ρ. φ = 0 if x > 0 and y = 0. θ). For the case of the position vector r = xi + yj it follows from using (2.cam.

e ρ · ez = 0 . y.51a) j = sin φ eρ + cos φ eφ . (2.9. φ = constant and z = constant. are mutually orthogonal. show that (cf. z) so that x = r sin θ cos φ . eφ . k}. 0 6 θ 6 π and 0 6 φ < 2π. (2. want to describe fluid flow down a long [straight] cylindrical pipe (e. (2. eφ . (2. e φ · ez = 0 . (ii) At any point P on the surface.50c) Equivalently i = cos φ eρ − sin φ eφ . With respect to {eρ . the vectors orthogonal to the surfaces ρ = constant. we can show that analogous vector component identities to (2. However. Michaelmas 2006 . (2. ez } form a basis we can write any vector v in component form as v = vρ eρ + vφ eφ + vz ez .29d). eφ .52d) Remark.Cowley@damtp. (2. z) . 0.29c) and (2.J.51c) Exercise. say. Show that {eρ . (2.55b) z x where again some care is needed in the choice of arctan. ez ] = 1 .40) and (2.uk. θ = arctan   (2. eφ and ez as unit vectors orthogonal to lines of constant ρ.29a). y = r sin θ sin φ .29c) and (2.53) Example.54b) 7/06 2.52c) [eρ . Since {eρ . (2. eφ × ez = eρ . (2.29d)) e ρ · eρ = 1 . (2. ez · ez = 1 . Since {eρ . (2. φ and z respectively. (2. ez } are a right-handed triad of mutually orthogonal unit vectors.55a) it follows that  2 1  2 2 1 x + y  .cam. (2. Remarks. φ is the φ of cylindrical polar co-ordinates. (2. latitude and longitude) Define 3D spherical polar co-ordinates (r. θ. i. (2.ac.   r = x2 + y 2 + z 2 2 . φ)12 in terms of 3D Cartesian co-ordinates (x. the normals. eφ . φ = arctan y . Component form.52b) e ρ × eφ = ez .38). eφ .54a) = (ρ. ez } satisfy analogous relations to those satisfied by {i. eφ · eφ = 1 . ez = k.41) also hold. (i) Cylindrical polar co-ordinates are helpful if you. height.e. (2.29b). (2.e. z = r cos θ .51b) k = ez . Then eρ = cos φ i + sin φ j .50b) This is a supervisor’s copy of the notes.e. From inverting (2.55a) where 0 6 r < ∞. (2. i. ez × eρ = eφ . j.52a) eρ · eφ = 0 . Mathematical Tripos: IA Algebra & Geometry (Part I) 28 c S. φ and z in the direction of increasing ρ. (2.50a) eφ = − sin φ i + cos φ j .3 Spherical polar co-ordinates (cf. ez } the position vector r can be expressed as → → r = ON + N P = ρ eρ + z ez (2.39). i. 7/02 Define eρ . (2.29a).29b).g. like the new gas pipe between Norway and the UK). It is not to be distributed to students. (2. 12 Note that r and θ here are not the r and θ of plane polar co-ordinates.

(2. (2. (2. eθ } would form a left-handed triad.e.Cowley@damtp. (2.29a).57a) j = sin φ (sin θ er + cos θ eθ ) + cos φ eφ . (2. (i) Spherical polar co-ordinates are helpful if you.56b) eφ = − sin φ i + cos φ j . eφ · eθ = 0 .41) also hold. eθ .29c) and (2. eθ · eθ = 1 . It is not to be distributed to students. (2.58d) Remark. (2. show that (cf. eφ ] = 1 . eθ × eφ = er . (2. (2. θ = constant and φ = constant at any point P are mutually orthogonal.29d)) er · er = 1 . Since {er . θ and φ in the direction of increasing r. eφ .57b) k = cos θ er − sin θ eθ .ac.cam.Remarks.e. (2. Michaelmas 2006 . (2.40) and (2.29b). eθ . in order to understand global warming. eφ }13 satisfy analogous relations to those satisfied by {i. say. eθ .29d). or both) motion on the earth. (2.J.58a) er · eθ = 0 . we can show that analogous vector component identities to (2. Mathematical Tripos: IA Algebra & Geometry (Part I) 29 c S. eφ } are a right-handed triad of mutually orthogonal unit vectors. (2.58c) [er .g. i. eθ .39).38). eθ and eφ as unit vectors orthogonal to lines of constant r. 13 The ordering of er . (2.58b) er × eθ = eφ .56a) eθ = cos θ cos φ i + cos θ sin φ j − sin θ k . k}. er eφ P r eθ θ z k O y i j φ eφ ρ N x Define er . Then er = sin θ cos φ i + sin θ sin φ j + cos θ k . eφ × er = eθ . (2. want to describe atmospheric (or oceanic. (2.56c) Equivalently i = cos φ (sin θ er + cos θ eθ ) − sin φ eφ . Show that {er . e r · eφ = 0 . e φ · eφ = 1 . z This is a supervisor’s copy of the notes. (2.29a).29b).29c) and (2. (2. j. θ and φ respectively.uk.57c) Exercise. and eφ is important since {er . e. i. (ii) The normals to the surfaces r = constant.

2.J. Similarly c = λa + µb ⇔ ci = λai + µbi ⇔ cj = λaj + µbj ⇔ cα = λaα + µbα ⇔ cU = λaU + µbU . Once familiar with suffix notation. (2. x3 ) = {xi } . vy . α and U. 3 . (2. 2. v2 . range through (1. it is understood that the one free suffix i ranges through 1.ac. eφ } form a basis we can write any vector v in component form as v = vr er + vθ eθ + vφ eφ . (2. 3 . it is generally easier to manipulate vectors This is a supervisor’s copy of the notes. when we omit the for i = 1.Cowley@damtp.14 In (2. (2. y.60a) = (r. Refer to v as {vi }. 2.63a) or equivalently in component form a1 = b1 .15 14 Although there are dissenters to that view.62) Remark. z) = (x1 . i is then termed a free suffix. 2. v3 ) (2. respectively. Write the position vector r as r = (x.Component form. It is not to be distributed to students. Since {er . With respect to {er .1 Dyadic and suffix equivalents If two vectors a and b are equal.10 Suffix Notation So far we have used dyadic notation for vectors. eθ . 3 understood. (2.63e) This is a vector equation where.uk. Example: the position vector. with the i = 1. (2. rather than r. Mathematical Tripos: IA Algebra & Geometry (Part I) 30 c S. x2 .10. (2. Michaelmas 2006 .30) we introduced the notation v = vx i + vy j + vz k = (vx . Henceforth we will use x and r interchangeably. The use of x. j. we write a = b.60b) 2. 0.63c) a3 = b3 . 2. where is is assumed that i. 0) . 3). 2. vz ) .63d) In suffix notation we express this equality as ai = bi for i = 1. (2.61b) Suffix notation.63b) a2 = b2 . for the position vector in dyadic notation possibly seems more understandable given the above expression for the position vector in suffix notation. 3.59) Example.cam. using suffix notation. An alternative is to write v = v1 i + v2 j + v3 k = (v1 . Suffix notation is an alternative means of expressing vectors (and tensors).61a) = {vi } for i = 1. 15 In higher dimensions the suffices would be assumed to range through the number of dimensions. 2. eφ } the position vector r can be expressed as → r = OP = r er (2. eθ . (2. 3 so as to give three component equations.

etc. This is a supervisor’s copy of the notes. and i is the free suffix that is assumed to range through (1. In suffix notation this becomes 3 X 3 X (ak bk ) ci = ak bk ci = di . 2. are referred to as dummy suffices since they are ‘summed out’ of the equation. In particular • if a suffix appears once it is taken to be a free suffix and ranged through. it is essential that the dummy suffices are different.. Under Einstein’s summation the explicit sum. . k. especially after the rearrangement. Remark. but the above rules must be followed.g.ac. is chosen for the free suffix. • if a suffix appears more than twice in one term of an equation. etc. • if a suffix appears twice it is taken to be a dummy suffix and summed over. 2. 2. but it must be the same in each term. Mathematical Tripos: IA Algebra & Geometry (Part I) 31 c S. 3) without the need to explicitly say so. Similarly X3 a·b=λ ⇔ aα bα = λ .Remark.cam. Further examples. It is not to be distributed to students. 3).uk. and remember to check your answers (e. k=1 where the i. This notation is powerful because it is highly abbreviated (and so aids calculation. something has gone wrong (unless there is an explicit sum). It is essential that we used different symbols for both the dummy and free suffices! (ii) In suffix notation the expression (a · b)(c · d) becomes 3 ! 3  X X (a · b)(c · d) = ai bi  cj dj  i=1 j=1 3 X X 3 = ai bi cj dj . (i) As another example consider the equation (a · b)c = d.2 Summation convention In the case of free suffices we are assuming that they range P through (1. i=1 j=1 where.64) k=1 k=1 where k is the dummy suffix.10. can be omitted for dummy suffices. In suffix notation the scalar product becomes a · b = a1 b1 + a2 b2 + a3 b3 X3 = ai bi i=1 3 X = ak bk . Michaelmas 2006 . Dummy suffices. the free suffices should be identical on each side of an equation). (2.Cowley@damtp. It does not matter what letter. especially in examinations).J. or symbol. α=1 where we note that the equivalent equation on the right hand side has no free suffices since the dummy suffix (in this case α) has again been summed out.

Michaelmas 2006 .66e) 2. Using the definition of the delta function: 3 X ai δi1 = ai δi1 i=1 = a1 δ11 + a2 δ21 + a3 δ31 = a1 . An alternative notation is e(1) . (2. The following equations make no sense ak = bj because the free suffices are different ak bk ck = dk because k is repeated more than twice on the left-hand side.66b) 2. δ33 = 1 .65a) δij = 0 if i 6= j . δij . (2. Mathematical Tripos: IA Algebra & Geometry (Part I) 32 c S. (2.J.65c) δ31 δ32 δ33 0 0 1 Properties. δ22 = 1 .Cowley@damtp. 3 X δii = δii = δ11 + δ22 + δ33 = 3 . The Kronecker delta.65b) This can be written as a matrix equation:     δ11 δ12 δ13 1 0 0 δ21 δ22 δ23  = 0 1 0 .4 More on basis vectors Now that we have introduced suffix notation.66c) j=1 3.ac.uk. e2 and e3 for the Cartesian unit vectors i. ap δpq bq = ap bp = aq bq = a · b .66a) Similarly ai δij = aj .3 Kronecker delta This is a supervisor’s copy of the notes. is a set of nine numbers defined by δ11 = 1 . 3. 2. (2. Under suffix notation and the summation convection a + b = c is written as ai + bi = ci (a · b)c = d is written as ai bi cj = dj .10. 8/03 2. i. It is not to be distributed to students.66d) i=1 4. 3 X δij δjk = δij δjk = δik . (2. j = 1. j and k. (2. Examples. where the use of superscripts may help emphasise that the 1. 2 and 3 are labels rather than components. (2. 1.10. e(2) and e(3) .cam. (2. it is more convenient to write e1 .

10. (2.cam. 2. For a symmetric tensor sij .67d) Or equivalently (ej )i = (ei )j = δij . 3) to be the set of 27 quantities such that εijk = 1 if i j k is an even permutation of 1 2 3. (2.68b) = 0 otherwise. Do we need to do more? Example. in this case 1 2 3. (2.69a) ε132 = ε213 = ε321 = −1 (2.67e) This is a supervisor’s copy of the notes. It is not to be distributed to students. Since ijk sij = jik sji = −ijk sij . k = 1. is even/odd. (2. (2.67a) (i) a·e = ai .40). (2. Check.68a) = −1 if i j k is an odd permutation of 1 2 3. 8/06 2. (2. (a × b)1 = ε123 a2 b3 + ε132 a3 b2 = a2 b3 − a3 b2 .Then in terms of the superscript notation e(i) · e(j) = δij . e(j) × e(k) = εilm e(j) l e(k) m    i = εilm δjl δkm = εijk . (2. We define εijk (i.uk.71) j=1 k=1 where we note that there is one free suffix and two dummy suffices.10.69b) Further εijk = εjki = εkij = −εikj = −εkji = −εjik .69c) 8/02 Example.67b) and hence e(j) · e(i) e(j) i  = (the ith component of e(j) ) (2.72) Mathematical Tripos: IA Algebra & Geometry (Part I) 33 c S. 2.J.ac.68c) An ordered sequence is an even/odd permutation if the number of pairwise swaps (or exchanges or transpositions) necessary to recover the original ordering.70) we conclude that ijk sij = 0. (2. Hence the non-zero components of εijk are given by ε123 = ε231 = ε312 = 1 (2.5 The alternating tensor or Levi-Civita symbol Definition. i. j = 1. 2. 3. as required from (2. such that sij = sji evalulate ijk sij .Cowley@damtp. Michaelmas 2006 . (2.6 The vector product in suffix notation We claim that 3 X X 3 (a × b)i = εijk aj bk = εijk aj bk . j.  = (2.67c) e(i) j = δij .

0 otherwise  Similarly whenever j 6= k.9 Vector triple product Using suffix notation for the vector triple product we recover  a × (b × c) i = εijk aj (b × c)k = εijk aj εklm bl cm = −εkji εklm aj bl cm = −(δjl δim − δjm δil ) aj bl cm = aj bi cj − aj bj ci  = (a · c)b − (a · b)c i . If j = k = 1.10.73) Remark. (2.73) as an example of a repeated suffix.10.74) 2.2.7 An identity Theorem 2.75) 2. then LHS = εi12 εipq = ε312 ε3pq   1 if p = 1. There are four free suffices/indices on each side. Next suppose j = 1 and k = 2. (2. then LHS = εi11 εipq = 0 .ac. Similarly whenever j = k (or p = q). Mathematical Tripos: IA Algebra & Geometry (Part I) 34 c S.8 Scalar triple product In suffix notation the scalar triple product is given by a · (b × c) = ai (b × c)i = εijk ai bj ck . RHS = δ1p δ1q − δ1q δ1p = 0 .J. say.1. Hence (2.Cowley@damtp. q = 1 . (2. q = 2 = −1 if p = 2. This is a supervisor’s copy of the notes. with i as a dummy suffix on the left-hand side. q = 2 = −1 if p = 2. 0 otherwise  while RHS = δ1p δ2q − δ1q δ2p   1 if p = 1. εijk εipq = δjp δkq − δjq δkp . Take j = p in (2. q = 1 .42). It is not to be distributed to students. in agreement with (2.uk. Michaelmas 2006 . Example.73) represents 34 equations.10. Proof. say. then εipk εipq = δpp δkq − δpq δkp = 3δkq − δkq = 2δkq .cam.

way forward is to use vector manipulation to make progress. For given a.uk.76) Solution.Cowley@damtp. e. For given a. c ∈ R3 find solutions x ∈ R3 to This is a supervisor’s copy of the notes. then substitute this result into the previous equation to obtain: x(1 + a · b) = c + a(b · c) . the multiplicity of the solutions is represented by a single arbitrary scalar. (2. First expand the vector triple product using (2. It is not to be distributed to students. now rearrange to deduce that c + a(b · c) x= .77b) This is an equivalent equation for the line since the solutions to (2.12 Lines and Planes Certain geometrical objects can be described by vector equations.11 Vector Equations When presented with a vector equation one approach might be to write out the equation in components. Michaelmas 2006 .cam.42): x − a(b · x) + x(a · b) = c .77b) has many solutions. Equation (2. 2. (1 + a · b) Remark. x − (x × a) × b = c .12. x2 .81) below). b. then dot this with b: b·x=b·c . Remark. (2. and let P be a point on the line. For the case when a and c are not parallel. An alternative. Worked Exercise.g. (2. c and a × c as a basis.77b) for t 6= 0 are either x = a or 9/03 (x − a) parallel to t.J. x3 ) ∈ R3 . We may eliminate λ from the equation by noting that x − a = λt. Mathematical Tripos: IA Algebra & Geometry (Part I) 35 c S. and hence we might expect two arbitrary parameters in the solution (as we shall see is the case in (2.1 Lines Consider the line through a point A parallel to a vector t. (x − a) · n = 0 would become x1 n1 + x2 n2 + x3 n3 = a1 n1 + a2 n2 + a3 n3 . 2. n ∈ R3 this is a single equation for three unknowns x = (x1 . we could have alternatively sought a solution using a.ac. and hence (x − a) × t = 0 .77a) for some λ ∈ R. Then the vector equation for a point on the line is given by → → → OP = OA + AP or equivalently x = a + λt . 2. and often better.

a straight line in direction t through (t × u)/|t|2 . (2.e.78) with t to obtain t × u = t × (x × t) = (t · t)x − (t · x)t . (2. then d is the distance of the plane to O. Hence t × u (t · x)t This is a supervisor’s copy of the notes. i. (2. Further.78) so is x + µt for any µ ∈ R. 9/02 2.J. First ‘dot’ (2.2 Planes Consider a plane that goes through a point A and that is orthogonal to a unit vector n.80a) that (dn − a) · n = 0 .e.ac. Then → AP ·n = 0.80a) → Let Q be the point in the plane such that OQ is → parallel to n. The equation of the plane is thus x · n = a · n = d. Let P be any point in the plane.78) with t to obtain t · u = t · (x × t) = 0 . since Q is in the plane. t ∈ R3 find solutions x ∈ R3 to u = x × t. it follows from (2. |t|2 |t|2 Finally observe that if x is a solution to (2. x= + . µ ∈ R. (2. The arbitrariness in the two independent arbitrary scalars λ and µ means that the equation has [uncountably] many solutions. 2. Hence the general solution to (2. (2. then any point x in the plane may be written as x = a + λl + µm .78).78) can only be found up to an arbitrary multiple of t.cam.80b) Remarks. → →  AO + OP · n = 0. Thus there are no solutions unless t · u = 0. Michaelmas 2006 .81) is a solution to equation (2. For given u. is t×u x= + µt . n is the normal to the plane. It is not to be distributed to students.78) Solution.12. (x − a) · n = 0.uk. (2. 1. Suppose that OQ= dn.Cowley@damtp. If l and m are two linearly independent vectors such that l · n = 0 and m · n = 0 (so that both vectors lie in the plane). solutions of (2.79) |t|2 i. and hence a · n = dn2 = d .81) where λ.80b). Worked Exercise. assuming that t · u = 0. Next ‘cross’ (2. Mathematical Tripos: IA Algebra & Geometry (Part I) 36 c S.

If the latter case. we deduce that x is a point on both L1 and L2 (as required). Let x = a + λt = b − µu . with normal t × u. It is not to be distributed to students.77a). then from the equation of a line. (2.uk.ac.83) Further.J.9/06 Worked exercise. either L2 intersects Π nowhere (in which case L1 does not intersect L2 ). then b lies in Π and we deduce that a necessary condition for the lines to intersect is that (b − a) · (t × u) = 0 .84b) to the equation for a cone with a vertex at a general point a: 2 (x − a) · n = (x − a)2 cos2 α . Michaelmas 2006 . say α (0 < α < 12 π). we can show that (2. (2. Under what conditions do the two lines L1 : (x − a) × t = 0 and L2 : (x − b) × u = 0 intersect? Solution. Because L2 is parallel to L02 and thence Π.cam.84b) where by squaring the equation we have included the ‘reverse’ cone. or L2 lies in Π (in which case L1 in- tersects L2 ). This plane is spanned by t and u. let L02 be the line passing through a parallel to u.  (2. For if (2. Then from the above description x · n = |x| cos α . By means of a translation we can now generalise (2.83) holds.84a) The vector equation for a cone with its vertex at the origin is thus 2 (x · n) = x2 cos2 α . then (b − a) must lie in the plane through the origin that is normal to (t × u). µ ∈ R such that b − a = λt + µu . (2.Cowley@damtp.84c) Mathematical Tripos: IA Algebra & Geometry (Part I) 37 c S. with a given axis that passes through O. 2. and hence there exists λ. Let Π be the plane containing L1 and L02 . The point O is referred to as the vertex of the cone. If the lines are to intersect they cannot be paral- lel.82) This is a supervisor’s copy of the notes.80a) the equation specifying points on the plane Π is Π: (x − a) · (t × u) = 0 .83) is also a sufficient condition for the lines to intersect. (2.13 Cones and Conic Sections A right circular cone is a surface on which every point P is such that OP makes a fixed angle. Hence from (2. L1 passes through a. hence t and u must be linearly independent. (2. Let n be a unit vector parallel to the axis.

b. cos2 α − sin2 β so that (2. In that case l = 0 and we can express n in component form as n = (l. Next suppose that π β+α > 2 i. Michaelmas 2006 . suppose that.e. n = (l.87) cos2 α − sin2 β 10/02 There are now three cases that need to be considered: sin2 β < cos2 α. m.e. Mathematical Tripos: IA Algebra & Geometry (Part I) 38 c S. wlog.Component form. y. n) . n) = (0. To see why this is. In this case the intersection of the cone with the z = 0 plane will yield a closed curve. 2 cos2 α − sin2 β cos2 α − sin2 β which can be simplified to  c2 sin2 α cos2 α X 2 cos2 α + Y 2 cos2 α − sin2 β = . we choose Cartesian axes so that the axis of the cone is in the yz plane.uk.Cowley@damtp. sin2 β > cos2 α and sin2 β = cos2 α.   (2. sin β > cos α .J. c) . (2. 1 β < 2π −α i. m.84c) becomes 2  (x − a)l + (y − b)m + (z − c)n = (x − a)2 + (y − b)2 + (z − c)2 cos2 α . 1  sin β < sin 2π − α = cos α .85) 10/03 Intersection of a cone and a plane.86) This is a curve defined by a quadratic polynomial in x and y. In order to simplify the algebra This is a supervisor’s copy of the notes. It is not to be distributed to students. translate the axes by the transformation c sin β cos β X = x − a. In that plane 2  (x − a)l + (y − b)m − cn = (x − a)2 + (y − b)2 + c2 cos2 α . and for definiteness suppose that 0 6 β 6 π2 .e. Further. then (2. sin β. Y =y−b+ . z) . consider graphs of the intersection of the cone with the X = 0 plane.ac. First suppose that π β+α < 2 i.86) becomes   2   2  c sin β cos β 2 c sin β cos β Y − sin β − c cos β = X + Y − + c cos2 α .   (2. a = (a. cos β) . In this case the intersection of the cone with the z = 0 plane will yield two open curves. Let us consider the intersection of this cone with the plane z = 0. while if sin β = cos α there will be one open curve. Suppose that in terms of a standard Cartesian basis x = (x.cam.

(i) The above three curves are known collectively as conic sections. An isometry of Rn is a mapping from Rn to Rn such that distances are preserved (where n = 2 or n = 3 for the time being). Then |x01 − x02 | = |(x1 + b) − (x2 + b)| = |x1 − x2 | .J. In this case (2. x2 ∈ Rn are mapped to x01 .cam.89b) A2 B2 This is the equation of a hyperbola. (2. suppose x1 . i. In this case (2.14 Maps: Isometries and Inversions 2.88) | cos2 α − sin2 β| 2 2 cos α − sin β sin β < cos α.Cowley@damtp. where B is one half of the distance between the two vertices.14.uk. and that x 7→ x0 = x + b . In particular. Suppose that b ∈ Rn . (ii) The identification of the general quadratic polynomial as representing one of these three types will be discussed later in the Tripos.89a) A2 B2 This is the equation of an ellipse with semi-minor and semi-major axes of lengths A and B respec- tively (from (2. x02 ∈ Rn .e. sin β = cos α. It is not to be distributed to students.ac. x1 7→ x01 and x2 7→ x02 . (2. sin β > cos α. then |x1 − x2 | = |x01 − x02 | .90) Examples. In this case X2 Y2 − + = 1. Remarks.1 Isometries Definition. (2. (2. Translation is an isometry.89d) This is the equation of a parabola.87) becomes X2 Y2 + = 1. 2. This is a supervisor’s copy of the notes.88) it follows that A < B). (2. Define c2 sin2 α c2 sin2 α cos2 α = A2 and 2 2 = B .89c) where X = x − a and Y = y − b − c cot 2β . Mathematical Tripos: IA Algebra & Geometry (Part I) 39 c S.86) becomes X 2 = −2c cot β Y . Translation. (2. Michaelmas 2006 .

e.92) OP → → Let OP = x and OP 0 = x0 . then x k2 x k2 x0 = |x0 | = = x. i.Cowley@damtp. Having derived the mapping. For a point P . OP 0 =OP + P N + N P 0 =OP −2 N P .e. two points close to the origin map to far from the sphere. |x012 |2 = |x1 − 2(x1 · n)n − x2 + 2(x2 · n)n|2 = |x12 − 2(x12 · n)n|2 = x12 x12 − 4(x12 · n)(x12 · n) + 4(x12 · n)2 n2 = |x12 |2 . x0 is linear in x. and so → → → → → → This is a supervisor’s copy of the notes. 2) . Let x12 = x1 − x2 and x012 = x01 − x02 . Suppose xj 7→ x0j = xj − 2(xj · n)n (j = 1. (2. we now need to show that distances are preserved by the map- ping. (2. Then Σ represents a sphere with centre O and radius k.15 Inversion in a Sphere Let Σ = {x ∈ R3 : |x| = k ∈ R(k > 0)}.91) Remark. Show that P 0 has inverse point P .e. and far from each other. i.cam. as required for isometry.g. → Hence N P = (x · n)n. Remark. and → OP 0 = x0 = x − 2(x · n)n . x · n > 0 NP= → −|N P |n if N P has the opposite sense as n. But |N P | = |x · n| and ( → → |N P |n if N P has the same sense as n. 10/06 Mathematical Tripos: IA Algebra & Geometry (Part I) 40 c S. Sup- pose also that → → x =OP 7→ x0 =OP 0 .93) |x| |x| |x| |x|2 Exercise. where Π = {x ∈ R3 : x · n = 0} and n is a constant unit vector. e. x · n < 0 . Then → → → N P 0 =P N =−N P . Michaelmas 2006 . let N be the foot of the perpendicular from P to the plane. the mapping is a linear function of the components of x. (2. 2.J. then since n2 = 1. the inverse point P 0 with respect to Σ lies on OP and satisfies k2 OP 0 = . Definition.uk. Consider reflection in a plane Π. It is not to be distributed to students. Inversion in a sphere is not an isometry. For each point P . i.ac. Reflection.

k 4 − 2a · x0 k 2 + (a2 − r2 )|x0 |2 = 0. It is not to be distributed to students. Suppose that the sphere is given by |x − a| = r . (2. x2 − 2a · x + a2 − r2 = 0 .94) can thus be rewritten k4 0 k 2 − 2a · x + a2 − r2 = 0. |x|2 |x0 |2 (2.94) From (2. i. Example: inversion of a sphere.e. or equivalently if a2 6= r2 we can complete the square to obtain 2 . |x0 |2 |x0 |2 or equivalently This is a supervisor’s copy of the notes.93) k4 k2 0 |x0 |2 = and so x = x .

2 r2 k4 .

.

x − k a .

= .

0 .

. .

a2 − r2 .

ac.cam. (a2 − r2 )2 which is the equation of another sphere.uk. Michaelmas 2006 . Exercise.J.Cowley@damtp. Hence inversion of a sphere in a sphere gives another sphere (if a2 6= r2 ). What happens if a2 = r2 ? 11/02 Mathematical Tripos: IA Algebra & Geometry (Part I) 41 c S.

J. etc. This is a supervisor’s copy of the notes. z ∈ V x + (y + z) = (x + y) + z .1 What is a Vector Space? We will consider vector spaces over the real numbers.Cowley@damtp. in the case of real numbers. a set is a collection of objects considered as a whole. identifying certain key properties on which to focus (e. etc. they can also be sets of polynomials. engineering. i.0 Why Study This? One of the strengths of mathematics.3 Vector Spaces 3. i. The objects of a set are called elements or members. associativity.cam. for all x. Conventionally a set is listed by placing its elements between braces.1.uk. real numbers). {x : x ∈ R} is the set of real numbers.e. 16 Pedants may feel they have reached nirvana at the start of this section.ac. y. identity and the existence of an inverse under addition or multiplication).g. 17 The first four mean that V is an abelian group under addition. There are generalisations to vector spaces over the complex numbers.. Often this is done by taking something that we are familiar with (e.g. indeed one of the ways that mathematical progress is made.16 Sets. Up to this point you have thought of vectors as a set of ‘arrows’ in 3D (or 2D) space. functions. where x + y ∈ V so that there is closure under vector addition. called the null or zero vector. matrices. { } or ∅. To recap. normal service will be resumed towards the end of the section. for all x. However. Rings and Modules in Part IB for the definition of a field). It is not to be distributed to students. vector addition has an identity element. and then studying all mathematical objects with those properties (groups in the example just given). together with two binary operations • vector addition denoted for x. say. or ‘vectors’. We have referred to sets already. where ax ∈ V so that there is closure under scalar multiplication. Michaelmas 2006 . It is a subset of every set. 3. finance.1c) i.1b) A(iii) there exists an element 0 ∈ V .1 Definition A vector space over the real numbers is a set V of elements. 3.1a) A(ii) addition is commutative.e. e.e. closure. y ∈ V x + y = y + x. Mathematical Tripos: IA Algebra & Geometry (Part I) 42 c S. Vector spaces occur throughout mathematics. The empty set. (3..g. The aim of this section is to ‘abstract’ the previous section. or indeed over any field (see Groups. is the unique set which contains no elements. (3. is by linking what seem to be disparate subjects/results. satisfying the following eight axioms or rules:17 A(i) addition is associative. science. y ∈ V by x + y. (3. The empty set. such that for all x ∈ V x + 0 = x. • scalar multiplication denoted for a ∈ R and x ∈ V by ax. and I hope that you have covered them in Numbers and Sets.

e. (3. i. λµ). (3.1d) B(v) scalar multiplication of vectors is ‘associative’. i. It is not to be distributed to students. for suppose that both y and z are additive inverses of x then y = y+0 = y + (x + z) = (y + x) + z = 0+z = z. (3. (3. since there are two different operations in question.uk.g. for all λ. for all λ ∈ R and x. µx) and multiplication of real numbers (e.1h) 11/03 3. y ∈ V This is a supervisor’s copy of the notes.1f) where 1 is the multiplicative identity in R.18 i.g.2) (iv) Scalar multiplication by 0 yields the zero vector. µ ∈ R and x ∈ V λ(µx) = (λµ)x . (iii) The existence of a unique negative/inverse vector allows us to subtract as well as add vectors.ac. (3.3) since 0x = 0x + 0 = 0x + (x + (−x)) = (0x + x) + (−x) = (0x + 1x) + (−x) = (0 + 1)x + (−x) = x + (−x) = 0.e.e.2 Properties.1c) 0 + x = x and x + 00 = x for all x ∈ V . B(vii) scalar multiplication is distributive over vector addition.1.1e) B(vi) scalar multiplication has an identity element. (3. and hence 00 = 0 + 00 = 0 . (ii) The additive inverse of a vector x is unique. for all λ. for all x ∈ V 1 x = x.1g) B(viii) scalar multiplication is distributive over scalar addition. i. for all x ∈ V . We denote the unique inverse of x by −x.J. i. λ(x + y) = λx + λy . by defining y − x ≡ y + (−x) . 18Strictly we are not asserting the associativity of an operation. Michaelmas 2006 .e. A(iv) for all x ∈ V there exists an additive negative or inverse vector x0 ∈ V such that x + x0 = 0 . Mathematical Tripos: IA Algebra & Geometry (Part I) 43 c S. (i) The zero vector 0 is unique because if 0 and 00 are both zero vectors in V then from (3. namely scalar multiplication of vectors (e.cam.e. (3. 0x = 0 . µ ∈ R and x ∈ V (λ + µ)x = λx + µx .1b) and (3.Cowley@damtp.

. . . λxn ) . . . −xn ) .3 Examples (i) Let Rn be the set of all n−tuples {x = (x1 . . yn ). (3.7c) −x = (−x1 . . i. A(ii). for all x ∈ V . One possibility is that λ = 0. where λ ∈ R and x ∈ V . B(v). .4) since (−1)x = (−1)x + 0 = (−1)x + (x + (−x)) = ((−1)x + x) + (−x) = (−1 + 1)x + (−x) = 0x + (−x) = 0 + (−x) = −x . . y2 . −x2 . (vi) Scalar multiplication with the zero vector yields the zero vector. (3.1. 2. . . . If x. define x+y = (x1 + y1 . in which case there exists λ−1 such that λ−1 λ = 1. . It is not to be distributed to students. . n}. xn ) : xj ∈ R with j = 1. (−1)x = −x . (3.cam. (3.7d) It is straightforward to check that A(i).uk. for all λ ∈ R. This is a supervisor’s copy of the notes. So if λx = 0 then either λ = 0 or x = 0. Mathematical Tripos: IA Algebra & Geometry (Part I) 44 c S. and then appeal to the fact that the zero vector is unique to conclude that λ0 = 0 . (viii) Negation commutes freely since for all λ ∈ R and x ∈ V (−λ)x = (λ(−1))x = λ((−1)x) = λ(−x) . xn + yn ) .5) (vii) Suppose that λx = 0. B(vi).6b) 3. . To see this we first observe that λ0 is a zero vector since λ0 + λx = λ(0 + x) = λx . 0. y ∈ Rn . where n is any strictly positive integer. Hence Rn is a vector space over R. Then we conclude that x = 1x = (λ−1 λ)x = λ−1 (λx) = λ−1 0 = 0.e. . However suppose that λ 6= 0. x2 + y2 . .e. (v) Scalar multiplication by −1 yields the additive inverse of the vector. 0) . B(vii) and B(viii) are satisfied. .7b) 0 = (0. λx2 . . .Cowley@damtp. . (3.ac. with x as above and y = (y1 . . (3. . (3.6a) and (−λ)x = ((−1)λ)x = (−1)(λx) = −(λx) . . (3. A(iii) A(iv). .J. λ0 = 0. Michaelmas 2006 . x2 .7a) λx = (λx1 . . . . i.

(ii) Consider the set F of real-valued functions f (x) of a real variable x ∈ [a.1 Examples (i) For n > 2 let U = {x : x = (x1 . y ∈ U and λ. y ∈ U and λ. . For f. . B(vi). A subset U of a vector space V is a subspace of V if and only if under operations defined on V (i) for each x. (3. B(vii) and B(viii) hold since the elements of U are also elements of V . A(ii). y2 . it follows from (ii) that 0x ∈ U . (ii) for each x ∈ U and λ ∈ R. Proof.Cowley@damtp.3) that 0x = 0. B(vii) and B(viii) are satisfied. 0) with xj ∈ R and j = 1. Hence −x ∈ U . b].1. B(vi). For each x ∈ U . . Then U is a subspace of Rn since. For each x ∈ U . A proper subspace is a subspace of V that is not V or {0}. λx + µy ∈ U . µ ∈ R.e. A subset U of the elements of a vector space V is called a subspace of V if U is a vector space under the same operations (i. . y ∈ U .cam. . (3. Again it is straightforward to check that A(i). . We need demonstrate that A(iii) (i.e.2. if and only if U is closed under vector addition and scalar multiplication. λxn−1 + µyn−1 . (3. xn−1 . λx2 + µy2 . (3. 0) = (λx1 + µy1 . x + y ∈ U . but since also x ∈ V it follows from (3. where a. vector addition and scalar multiplication) as are used to define V . . µ ∈ R. . Remark.ac. . Strictly V and {0} (i.e. This is a supervisor’s copy of the notes. b ∈ R and a < b. . 0) + µ(y1 . Thus U is closed under vector addition and scalar multiplication. . Michaelmas 2006 . If. A(iii) A(iv). We can combine (i) and (ii) as the single condition for each x. . . the every element has an inverse in U ) hold. . . . Hence 0 ∈ U . 3. 2.e. i. g ∈ F define (f + g)(x) = f (x) + g(x) . .J. Theorem 3. and A(iv) (i. Mathematical Tripos: IA Algebra & Geometry (Part I) 45 c S.2 Subspaces Definition. If U is a subspace then it is a vector space. . A(iv).8c) (−f )(x) = −f (x) .uk.4) that (−1)x = −x. and hence that F is a vector space (where each 11/06 vector element is a function). x2 . it follows from (ii) that (−1)x ∈ U . for x. xn−1 . Proper subspaces. 3.e.8a) (λf )(x) = λf (x) . 0) . It is not to be distributed to students. and hence (i) and (ii) hold.8b) O(x) = 0 . . λ(x1 . It is straightforward to show that A(i). . A(ii). A(iii). x2 . Only if. 0 is an element of U ). n − 1}. B(v). B(v). λx ∈ U . and is hence a subspace. but since also x ∈ V it follows from (3. yn−1 .8d) where the function O is the zero ‘vector’. the set containing the zero vector only) are subspaces of V .

. .3. i=1 12/03 Remark. n). for x.10) Remark. j = 1. n). 2. . . i=1 i=1 i=1 Thus W is closed under vector addition and scalar multiplication. n . .uk.cam. 2. Thus W f is not closed under vector addition. λxn + µyn ) . (3. .3 Spanning Sets. and we say that S spans U . . . λj ∈ R. Definition. 2. . n).24)). . 2. . . n X λi vi = 0 ⇒ λi = 0 for i = 1. λx2 + µy2 . the vectors are said to be linearly dependent since there exist scalars λj ∈ R (j = 1. A subset of S = {u1 . n) i=1 not all of which are zero (wlog α1 6= 0. n hyper-plane that does not pass through the origin. u2 .ac. i=1 n 12/02 and so W cannot be a subspace of R . such that v = λ1 u1 + λ2 u2 + . .9) i=1 Otherwise. . 3. . . Mathematical Tripos: IA Algebra & Geometry (Part I) 46 c S. or consider x ∈ W f such that x = (α1−1 . 0) . . . . . which is a This is a supervisor’s copy of the notes. v2 .Cowley@damtp. . . is not a subspace of R . vj ∈ V (j = 1. . y ∈ W and λ. . (iii) For n > 2 consider the set Wf = {x ∈ Rn : Pn αi xi = 1} for given scalars αj ∈ R (j = 1. and hence is a subspace of V . vn }. un } of vectors in V is a spanning set for V if for every vector v ∈ V . n). . 2. . . . . f 3. Dimension and Bases 3. (3. . is linearly independent if for all scalars λj ∈ R (j = 1. 2. . . . . λx + µy = (λx1 + µy1 . . Definition. . . . With this definition {0} is a linearly dependent set. The vector space U is called the span of S (written U = span S). A set of n vectors {v1 . there exist scalars λj ∈ R (j = 1. and is hence a subspace.2 Spanning sets Definition. . Michaelmas 2006 . Pn (ii) For n > 2 consider the set W = {x ∈ Rn : i=1 αi xi = 0} for given scalars αj ∈ R (j = 1. 0. . . . . 2. Further. n). n). . W is a subspace of V since. The λj are not necessarily unique (but see below for when the spanning set is a basis). not all of which are zero.J. . W f . . Then x ∈ W f but x+x 6∈ Wf since Pn αi (xi +xi ) = 2. Later we shall see that W is a hyper-plane through the origin (see (3. + λn un . . µ ∈ R. we are implicitly assuming here (and in much that follows) that the vector space V can be spanned by a finite number of vectors (this is not always the case). . .3. i. The set of all linear combinations of the uj (j = 1.e. .1 Linear independence Extend ideas from vectors in R3 . n i=1 is closed under vector addition and scalar multiplication. . ( n ) X U = u:u= λi ui . It is not to be distributed to students. To see this either note that 0 6∈ W f . such that Xn λi vi = 0 . and n X n X n X αi (λxi + µyi ) = λ αi xi + µ αi yi = 0 . 2. 2. if not reorder the numbering of the axes). .

3 X xi ui = (x1 . 0. • The set {u1 .ac. not all zero such that n X λi ui = 0 . x2 .11) • The set {u1 . then λ1 = λ2 = λ4 = 0. i=1 λn Since S spans V . i=1 This is a supervisor’s copy of the notes. the set Sn−1 = {u1 . un−1 .uk. This set spans R3 since for any x = (x1 . u3 . up }. u2 .g. It is not to be distributed to students. x2 .Cowley@damtp. . 1) . λ2 + λ4 . x2 . i=1 then λ1 = λ2 = λ3 = 0. . u3 . This set spans R3 . . u2 . If Sn−1 is linearly independent then we are done. . Then n−1 X λi un = − ui . A basis for a vector space V is a linearly independent spanning set of vectors in V . u2 . 1. 1. u5 . If not repeat until Sp = {u1 . u4 . u4 } and {u2 . −1) cannot be expressed as a linear combination of u1 . u4 }. . i=1 Suppose that λn 6= 0 (if not reorder the vectors). This set spans R3 (as above). Proof. (x1 − x3 )u1 + (x2 − x3 )u2 + x3 u4 = (x1 . x3 ) ∈ R3 . .2. u4 }. This set does not span R3 . u3 . x3 ) . n. u4 . • The set {u1 . 1. . e. 1. • The set {u1 .11). j = 1. . the vector (0. . If a set S = {u1 . {u1 . e. u4 . . u3 }. . which spans V .) Theorem 3. u4 } is linearly independent since if λ1 u1 + λ2 u2 + λ4 u4 = (λ1 + λ4 . u4 and u5 . un } is linearly dependent. λ4 ) = 0 . λ2 . u5 } is linearly dependent since u1 + u2 + u3 − u4 = 0 . for uj as defined in (3. 1) . . u5 = (0. u2 . . Example. . 0. u2 . (The set does span the plane containing u1 . . Michaelmas 2006 . λ3 ) = 0 . but is linearly dependent since u4 = u1 + u2 + u3 . u5 } is linearly dependent since u1 − u4 + u5 = 0 . is linearly independent. . u2 = (0. and let u1 = (1. and spans the vector space V . . (3. we can reduce S to a linearly independent set also spanning V . . . u4 = (1. x3 ) ∈ R3 .J. If S is linear dependent then there exists λj . un−1 } spans V . 0) . 12/06 Mathematical Tripos: IA Algebra & Geometry (Part I) 47 c S. u3 } is linearly independent since if 3 X λi ui = (λ1 . u2 . . {u1 . u3 = (0. u3 . u4 } all span R3 and are linearly independent.g. This set spans R3 since for any x = (x1 . Example. Definition. and un can be expressed in terms of u1 . the sets {u1 . 1) . Consider R3 .cam. x2 . Consider the set {u1 . 0) . There are a number of ways by which this set can be reduced to a linearly independent one that still spans R3 . x3 ) . u2 .

. . i = 1. Mathematical Tripos: IA Algebra & Geometry (Part I) 48 c S.cam. .12) that This is a supervisor’s copy of the notes. . w2 . . un } spans V . Let V be a finite-dimensional vector space. . w2 . we deduce that νi = 0 for i = 1. . . n. . Continue to deduce that {w1 . . Assume µ2 6= 0 (otherwise reorder the ui . un } spans V there exists µi such that n X x= µi ui . Michaelmas 2006 . . . . . un } forms a basis. If m > n this would mean that {w1 . . . . . . We can then deduce that {w1 .3. Hence. . . ν1 λi + νi = 0 . . at least one of the µi . there exists µi .12) n X λ1 ν1 u1 + (ν1 λi + νi )ui = 0 . wn } forms a basis. Since λ1 6= 0. Proof. . . i = 1. i=2 But {u1 . . suppose that n X ν1 w1 + νi ui = 0 . n must be non-zero. i=2 Moreover. .4. . i = 2. . . . . and hence spans V . w2 } would be linearly dependent and so violate the hypothesis that the set {wi . Suppose that {u1 . . un } is a basis for a vector space V . and hence that {w1 . n . Further. .Lemma 3. It is not to be distributed to students. otherwise the subset {w1 . The set {w1 . . .uk. . . un } is a basis so λ1 ν1 = 0 . n ! n n   µ1 X X µ1 X λi µ1 x= w1 − λi ui + µi ui = w1 + µi − ui . . Since the set {ui . . n (at least one of which is non-zero) such that n X w2 = µ1 w1 + µi ui . The number of vectors in a basis of V is the dimension of V . . . . . n). . Let x ∈ V . . . . i=2 Then from (3. Remark. m} forms a basis. i = 2. w2 . . . . . contradicting the original assumption. . un } for n X w1 = λi ui (3. . . . then every basis has the same number of elements. . . . Proof. wm } are two bases for V with m > n (wlog). . Theorem 3. . . u2 . i = 1. then so is {w1 . .12) i=1 if λ1 6= 0. .ac. . . . i = 1. . n). .J. . . . Hence m = n. u2 . . i = 1. λ1 i=2 i=2 λ1 i=2 λ1 and hence that {w1 . u2 . n such that n X w1 = λi ui i=1 with λ1 6= 0 (if not reorder the ui . un } is an alternative basis. . u3 . . . An analogous argument can be used to show that no linearly independent set of vectors can have more members than a basis. . If {u1 . .Cowley@damtp. i = 2. un } and {w1 . then since {u1 . . wm } was linearly dependent. . . . i=1 It follows from (3. . n} is a basis there exist λi . . u2 . Definition. . un } is linearly independent. written dim V .

. (0. (Unlectured. 0. 1)}. 0) + x3 (0. 0. .uk. . But dim R3 = 3. b]). spans V (which must be possible since we are only studying finite dimensional vector spaces). • The vector space {0} is said to have dimension 0. . x3 ∈ R = x1 (1. y} = R3 . 0). . If it spans V . the vector space of real functions defined on [a. . −1. 1). Let V = R3 and suppose U = {(x1 . with e1 = (1. −1. . and {(1. 0). y}. It is not to be distributed to students. 0. (0. of V . e. x3 ) x1 . . 1). Examples. Then {(1. . x) with x ∈ R is spanned by e = (1. Mathematical Tripos: IA Algebra & Geometry (Part I) 49 c S. u` } be a basis of U (dim U = `). . 1)}. u` . Then the set {u1 .5. 1)}. of V . . . w2 } is linearly independent. w2 6∈ U1 . (0. Exercise. Remarks. Choose w1 ∈ V . . 1. . . (0. −1. α3 = 0). Hence a basis for U is {(1. (i) The set {e1 .) Let {u1 . 0. {e} is a basis and thus dim U = 1. Show that any set of dim V linearly independent vectors of V is a basis. e2 . . 0. . .cam. 0). Then the set {u1 . Then each v ∈ V can be expressed as a linear combination of the ei . . w1 . • We shall restrict ourselves to vector spaces of finite dimension (although we have already encoun- tered one vector space of infinite dimension. . Since x1 + x2 = 0 implies that x2 = −x1 with no restriction on x3 . Now repeat until {u1 . y} is linearly independent and forms a basis for span {(1. is a basis for Rn . 0). 0). . . 3. . . i. Michaelmas 2006 . . . u2 . . . u2 . If it spans V we are done. (0. 0. hence span {(1. . y 6∈ U .4 Components Theorem 3. 1) . 0). . (ii) The subspace U ⊂ Rn consisting of vectors (x. and thus dim Rn = n.J. for x ∈ U we can write x = (x1 . Since {e} is linearly independent. . 0. x. 0).6. n X v= vi ei .13) i=1 where the vi are unique for each v. i. . U is a subspace (see example (ii) on page 46 with α1 = α2 = 1. 1. −1. any basis of U can be extended into a basis for V . en }. −1. i. .Cowley@damtp. Now choose w2 ∈ V . . . 1. . 0). w1 . Example. . w1 6∈ U . . or (0. 0). . which has dimension three. This is a supervisor’s copy of the notes. 1). 13/02 Theorem 3. . e2 = (0. If U is a proper subspace of V then there are vectors in V but not in U . Let V be a vector space and let S = {e1 . . y} is an extension of a basis of U to a basis of R3 . x2 . en = (0. say U1 . (dim V = ` + 1) we are done. . for some m. Proof. or (1. 1. u` . w1 } is linearly independent. .g. u2 .e. . −1. say U2 .e.ac. x3 ) ∈ R3 : x1 + x2 = 0}. . u` . 1). . . (3. u2 . y = (1. . wm }. Now choose any y such that y ∈ R3 . en } be a basis of V . 0). If U is a proper subspace of V . . if not it spans a proper subspace. . −x1 . .e. 0. if not it spans a proper subspace. 0. .

λxn + µyn ) ∈ Rn . Proof.uk. such that n X v= vi0 ei . . . 1)}. while U + W is the 2D-plane.5 Intersection and Sum of Subspaces Let U and W be subspaces of a vector space V over R. and thus U + W is a subspace. µ ∈ R. . For a given basis. If S is a basis then. i=1 Suppose also there exist vi0 ∈ R. i=1 i=1 i=1 But the ei are linearly independent. then x. It is not to be distributed to students. . . . hence (vi − vi0 ) = 0. Hence λv1 + µv2 = λ(u1 + w1 ) + µ(u2 + w2 ) = (λu1 + µu2 ) + (λw1 + µw2 ) ∈ U + W . U + W is a subspace of V . λx + µy ∈ U and λx + µy ∈ W (since both U and W are subspaces). If U = span{(1. . If x. then U ∪ W is the abscissa and the ordinate. w2 ∈ W such that v1 = u1 + w1 and v2 = u2 + w2 . .7.J. To be more precise. and hence the vi are unique. . . Proof. y ∈ U ∩ W . Proof. i = 1. Remark. . Theorem 3. . Remark. u2 ∈ U and w1 . We call the vi the components of v with respect to the basis S. 13/06 Definition. . . then there exist u1 . and U ∩ W is a subspace of V . . Since U ⊆ U + W and W ⊆ U + W . . Remark. .8. Theorem 3. xn ) ∈ Rn . if y is associated with components y = (y1 . For any λ. n. A little more work then demonstrates that that all dimension n vector spaces over the real numbers are the ‘same as’ Rn . . U ∩ W is also a subspace of U and a subspace of W . U ∩ W contains at least 0 so is not empty.cam. . This is a supervisor’s copy of the notes. U ∩ W is the intersection of U and W and consists of all vectors v such that v ∈ U and v ∈ W. we conclude that each vector x in a vector space V of dimension n is associated with a unique set of real numbers. Michaelmas 2006 . for all v ∈ V there exist vi ∈ R. i = 1. Thus Rn is the 13/03 prototypical example of a real n-dimensional vector space. Further. µ ∈ R. U + W is called the sum of subspaces U and W . Definition. namely its components x = (x1 . Remark. such that Xn v= vi ei .Cowley@damtp. Mathematical Tripos: IA Algebra & Geometry (Part I) 50 c S. 3. Remark. Definition. y ∈ W . i=1 Then n X n X n X vi ei − vi0 ei = (vi − vi0 )ei = 0 . n. it is possible to show that every real n-dimensional vector space V is isomorphic to Rn . . Hence λx + µy ∈ U ∩ W . . both U and W are subspaces of U + W . n. λu1 + µu2 ∈ U and λw1 + µw2 ∈ W . and consists of all vectors u + w ∈ V where u ∈ U and w ∈ W . hence for any λ. U + W is not the same as U ∪ W . i = 1. If v1 . v2 ∈ U + W . yn ) ∈ Rn then (λx + µy) is associated with components (λx + µy) = (λx1 + µy1 . 0)} and W = span{(0. y ∈ U and x.ac. from the definition of a basis. U ∩ W is a subspace of V . . . .

i = 1. . er .cam. . and so r X v= αi ei for some αi ∈ R. . . and hence x1 = x2 = 0 . 1). . . Then. i = 1. . On the other hand U +W = span{(0. s. Then {e1 . 3. . . fs . 1). If x ∈ U ∩ W then x1 = 0 and x1 + 2x2 = 0 . s and νi ∈ R. . −1)} ⇒ dim W = 3 . (3. . Similarly µi = 0. . . 0.15) i=1 i=1 i=1 Define v by s X r X t X v= µi fi = − λ i ei − νi gi . r. . 1. i = 1. . µi ∈ R. Then U + W = span{u1 . 0. 0. 0.1 dim(U + W ) Let {u1 . . we can say more than this. . 1). (0.15) since {e1 . .9. . It is not to be distributed to students.J. 1. . 0. . g1 . 0). 0. r from (3. 0. . . . . . 0. 1). 1. . 14/02 3. . . er } be a basis for U ∩ W (OK if empty). 1. gt } spans U + W . However. 1. . 1. . . 1. dim(U + W ) + dim(U ∩ W ) = dim U + dim W . and hence dim(U + W ) 6 dim U + dim W . . . gt } for W . r. . gt } is linearly independent and is thus a basis for U + W . (0. . U = span{(0. 0). . 0. . . i = 1. f1 . 0. . . (0. . . i = 1. Mathematical Tripos: IA Algebra & Geometry (Part I) 51 c S. . . er . Further λi = 0. 1.5. 0. 0.5. f1 . 1. i=1 i=1 But {e1 . 0). (3. w1 . g1 . seek solutions to This is a supervisor’s copy of the notes. . −1)} = span{(0. . (−2. i = 1. . 0. . (0. . r X s X t X λ i ei + µi fi + νi gi = 0 . . . t. 1. 0). . 1. It follows that {e1 . . . .2 Examples (i) Suppose that U = {x ∈ R4 : x1 = 0} and W = {x ∈ R4 : x1 + 2x2 = 0} (both are subspaces of R4 from example (ii) on page 46). . Extend this to a basis {e1 . and so νi = 0. fs } for U and a basis {e1 . gt } is a basis (for W ) and hence linearly independent.ac. i=1 i=1 i=1 where the first sum is a vector in U .14) Proof. 0). 1). g1 . Let {e1 . 1. . 1. . 0). . . (0. . . 0. . 0). . 0. (0. . Theorem 3. 0. . . . Hence v ∈ U ∩ W . . er . . . (0. Thus U ∩ W = span{(0. er . g1 . wm } . 0. Thus dim(U + W ) = r + s + t = (r + s) + (r + t) − r = dim(U ) + dim(W ) − dim(U ∩ W ) . 0. . 0. 1)} ⇒ dim U = 3 . (0. . . 0. . . 0. . er . . 0)} . −1)} and dim U ∩ W = 2 . 1. . (0. . . . . . (0. i = 1. . For λi ∈ R. (0. i=1 We deduce that r X t X (αi + λi )ei + νi gi = 0 . . 0. . 1. Michaelmas 2006 . . while the RHS is a vector in W . . (0. (−2. W = span{(−2.Cowley@damtp. .uk. . . . u` } and {w1 . 1)} = span{(0. 0. . . 1. wm } be bases for U and W respectively. t. 0). . u` . er } is a basis (for U ∩ W ). . f1 . 0). fs . say. . . 0.

e1 (2) = 0 and e1 (3) = 0. i. i.18a) 1 kvk ≡ |v| = (v · v) . define e2 and e3 similarly (where we have deliberately not used bold notation).e.e. Properties (3. i. 2 (3.17) Alternative notation.1 Definition of a scalar product The three-dimensional linear vector space V = R3 has the additional property that any two vectors u and v can be combined to form a scalar u · v. (3.16b) (iii) Non-negativity.16a) (ii) Linearity in the second argument. as in (3. Define addition. a scalar product of a vector with itself should be positive. scalar multiplication.6. (3. u · v = v · u. 2.8d). or inner product.J.18b) Mathematical Tripos: IA Algebra & Geometry (Part I) 52 c S. f (2). f (2).16d) Remark.6 Scalar Products (a. i. Inner Products) 3. f (3)) ∈ R3 with f (1) = f (2) = 0}. for λ. in agreement with (3.8c) and (3. (ii) Let V be the vector space of real-valued functions on {1.8b). so dim U ∩ W = 1 .k. for λ. u · v ∈ R with the following properties. i. Then {ei . a scalar product. and dim V = 3.e. i. dim(U ∩ W ) + dim(U +W ) = 2 + 4 = 6 .ac. Hence dim U + W = 4.e. µ ∈ R u · (λv1 + µv2 ) = λ u · v1 + µ u · v2 .14). where the real positive number |v| is a norm (cf.e. µ ∈ R (λu1 + µu2 ) · v = v · (λu1 + µu2 ) = λv · u1 + µv · u2 = λu1 · v + µu2 · v . Then U +W = V so dim(U +W ) = 3 . (3.e. again verifying (3. |v| = 0 ⇒ v = 0. after reducing the linearly dependent spanning set to a linearly independent spanning set. f (3)) ∈ R3 with f (1) = 0}. the only vector of zero norm should be the zero vector. W = {f : (f (1).16c) This allows us to write v · v = |v|2 . (3. f (3)) ∈ R3 with f (2) = 0}.cam. This can be generalised to an n-dimensional vector space V over the reals by assigning. Hence. for every pair of vectors u. 2.e. then dim W = 2 . 14/03 3. Let e1 be the function such that e1 (1) = 1.16a) and (3.e. length) of the vector v. (i) Symmetry. This is a supervisor’s copy of the notes. i = 1. Let U = {f : (f (1). f (2). (3. then dim U = 2 . (3. Michaelmas 2006 . (3.a.uk. v ∈ V . f (3)) ∈ R3 }. i. etc. (3. It is not to be distributed to students. i.8a). dim U + dim W = 3 + 3 = 6 . f (2).16b) imply linearity in the first argument. (iv) Non-degeneracy. 3} is a basis for V .14). 3}. v · v > 0. U ∩W = {f : (f (1). An alternative notation for scalar products and norms is hu |v i ≡ u · v. V = {f : (f (1).Cowley@damtp.

This follows from taking square roots of the following inequality: ku + vk2 = kuk2 + 2 u · v + kvk2 from above with λ = 1 2 2 6 kuk + 2|u · v| + kvk 6 kuk2 + 2kuk kvk + kvk2 from (3. It is not to be distributed to students. 3. 3. has at most one real root. 2 (2u · v) 6 4kuk2 kvk2 . with equality only if u = −λv for some λ.13a) can be extended to (1.16a). suppose that v 6= 0. For λ ∈ R consider 0 6 ku + λvk2 = (u + λv) · (u + λv) from (3.18b) 2 = u · u + λu · v + λv · u + λ v · v from (3.2 Schwarz’s inequality Schwarz’s inequality states that |u · v| 6 kuk kvk .3 Triangle inequality This triangle inequality is a generalisation of (1. Second.e.5) and states that ku + vk 6 kuk + kvk .13b) we can similarly deduce that . In the same way that (1. Then. in which case (3. since 0 = 0+0 from (3. A more direct way of proving that u · 0 = 0 is to set λ = µ = 0 in (3. We have two cases to consider: v = 0 and v 6= 0.20) Proof. (3.1c) and 0 = 0vj from (3.19) is satisfied as an equality. Proof.19) with equality only when u is a scalar multiple of v. This is a supervisor’s copy of the notes. since it is not negative.16b). (3. so that kvk = 0. Hence ‘b2 6 4ac’. First. Remark and exercise. The right-hand-side then simplifies from a quadratic in λ to an expression that is linear in λ. Hence we conclude that u · v = 0 if v = 0 .6.19) 2 6 (kuk + kvk) . i. suppose that v = 0.13a) and (2.16b) and (3. u · 0 = u · (0v1 + 0v2 ) = 0 (u · v1 ) + 0 (u · v2 ) = 0 .17) 2 2 2 = kuk + 2λu · v + λ kvk from (3. The right-hand-side is then a quadratic in λ that.6. Schwarz’s inequality follows on taking the positive square root. Remark.3). If u · v 6= 0 we have a contradiction since for certain choices of λ this simplified expression can be negative.

.

14/06 ku − vk > .

kuk − kvk .

21) Mathematical Tripos: IA Algebra & Geometry (Part I) 53 c S.ac.uk. Michaelmas 2006 .cam.Cowley@damtp.J. (3. .

a ∈ Rn }. denoted by Rn (and referred to as real coordinate space). orthonormal bases always exist. 0.22).13) has a more complicated form than (3. of a vector x ∈ Rn is defined as n !1 X 2 kxk = x2i . y.Cowley@damtp. 0. This is a supervisor’s copy of the notes. x·y = y · x.16c) and (3. The good news is that for any scalar product defined 15/02 on a vector space over R.22) is consistent with the scalar product for R3 defined in (2. It is important to appreciate that the scalar product for Rn as defined by (3. en = (0.6. . . a non-orthogonal basis).24) n Remark.6. x · (λy + µz) = λx · y + µx · z.16a). Remarks. • The length. 0). 2 kxk = x · x > 0. r ∈ R.g. (3.cam. Hyper-plane in Rn .22) satisfies (3. or Euclidean norm. If b · n = 0 so that Π = {x ∈ Rn : i=1 xi ni = 0}. . .e.e.22) (and for which we need matrices). The (hyper-)sphere in Rn with centre a ∈ Rn and radius r ∈ R is given by Σ = {x ∈ Rn : kx − ak = r > 0.uk.16b).1). . To this end we might view (3.39)) n X x·y = xi yi = x1 y1 + x2 y2 + .16d). For the case when the xi and yi are components with respect to a non-orthonormal basis (e. 3. µ ∈ R. (3. . is given by Π = {x ∈ Rn : (x − b) · n = 0.22) i=1 Exercise. z ∈ Rn and λ. (3. Michaelmas 2006 . the scalar product in Rn equivalent to (2. i=1 while the interior angle θ between two vectors x and y is defined to be   x·y θ = arccos . (3. 1). + xn yn . .4 The scalar product for Rn In example (i) on page 44 we saw that the space of all n-tuples of real numbers forms an n-dimensional vector space over R. i. b. The (hyper-)plane in Rn that passes through b ∈ Rn and has normal n ∈ Rn . (3. .22) as the scalar product with respect to the standard basis e1 = (1. .13) only when the xi and yi are components with respect to an orthonormal basis. . then Π is a subspace of dimension (n − 1) (see example (ii) of §3. n ∈ Rn with knk = 1}.J. It is not to be distributed to students. .5 Examples Hyper-sphere in Rn .ac. so that the hyper-plane passes P through the origin. kxk = 0 ⇒ x = 0 . (2. .23) Remark. .2. . 3. kxkkyk • We need to be a little careful with the definition (3. . . . 0). . e2 = (0. We define the scalar product on Rn as (cf. 1. Σ is not a subspace of Rn . for x. Confirm that (3. i. Mathematical Tripos: IA Algebra & Geometry (Part I) 54 c S.

f (A) ⊆ B. It is not to be distributed to students.1b) or equivalently if T (λa + µb) = λT (a) + µT (b) for all a. f (A) is the image of A under f . Definition. or codomain. n ∈ Z+ . f (x) = x0 is the image of x under f . This is not unduly restrictive since every real n-dimensional vector space V is isomorphic to Rn (recall that any element of a vector space V with dim V = n corresponds to a unique element of Rn ). W be vector spaces over R. focussing on V = Rn and W = Rm . (4. Michaelmas 2006 . Remark. µ ∈ R . but there may be elements of B that are not images of any x ∈ A. (4. We write f : A → B and/or x 7→ x0 = f (x) Examples.uk. Definition. the set of all image points x0 ∈ B of x ∈ A.1a) (ii) T (λa) = λT (a) for all a ∈ V and λ ∈ R .g. 4. b ∈ V . A is the domain of f .Cowley@damtp.J. B is the range. e. inversion with respect to a sphere. Remark. The map T : V → W is a linear map or linear transfor- mation if (i) T (a + b) = T (a) + T (b) for all a. Definition. Moreover.ac. (4. b ∈ V and λ.2 Linear Maps We shall consider maps from a vector space V to a vector space W . 4. B be sets. Let A. Let V.2) Mathematical Tripos: IA Algebra & Geometry (Part I) 55 c S. The aim of this section is to construct a general framework for viewing such linear problems.e. Möbius maps of the complex plane.1 General Definitions This is a supervisor’s copy of the notes. Definition. Definition. A map f of A into B is a rule that assigns to each x ∈ A a unique x0 ∈ B. i.cam. and almost all sounds you hear are ‘linear’ (exceptions being sonic booms). of f . Definition. for m. many computational approaches to solving nonlinear problems involve ‘linearisations’ at some point in the process (since com- puters are good at solving linear problems).4 Linear Maps and Matrices 4. translation. electromagnetic waves satisfy linear equations.0 Why Study This? Many problems in the real world are linear.

µ ∈ R HΠ (λx1 + µx2 ) = (λx1 + µx2 ) − 2 ((λx1 + µx2 ) · n) n = λ(x1 − 2(x1 · n)n) + µ(x2 − 2(x2 · n)n) = λHΠ (x1 ) + µHΠ (x2 ) . However. Hence for x1 .1 on page 45. it follows that 0 ∈ T (V ). (iii) As another example consider projection onto a line with direction t ∈ R3 as defined by (cf. 15/03 This is a supervisor’s copy of the notes. T (b) ∈ T (V ) λT (a) + µT (b) = T (λa + µb) ∈ T (V ) for all λ. and so HΠ is a linear map.Cowley@damtp.17b)) P : R3 → R3 x 7→ x0 = P (x) = (x · t)t where t · t = 1. T (b) = 0 ∈ W does not imply b = 0 ∈ V .1a). n ∈ R3 and |n| = 1. Set b = 0 ∈ V in (4. (ii) A projection is not an isometry. we can say more than that. since for T (a). and so for all λ. Michaelmas 2006 .J.3) Now apply Theorem 3. which is a 1-dimensional subspace of R3 . Since T (V ) is a subspace. T (V ) is a subspace of W . Property. (4. Remarks. (ii) Next consider the isometric map HΠ : R3 → R3 consisting of reflections in the plane Π : x.e.91) x 7→ x0 = HΠ (x) = x − 2(x · n)n .2) is satisfied. It is not to be distributed to students. Remark. x2 7→ x02 = x2 − 2(x2 · n)n = HΠ (x2 ) . The zero element.1 Examples (i) Consider translation in R3 (an isometry). i. Thus (4. then from (2.2. (i) The range of P is given by P (R3 ) = {x ∈ R3 : x = λt for λ ∈ R}. 4. where a ∈ R3 and a 6= 0 . to deduce that T (a) = T (a + 0) = T (a) + T (0) for all a ∈ V . From the observation that P (λx1 + µx2 ) = ((λx1 + µx2 ) · t)t = λ(x1 · t)t + µ(x2 · t)t = λP (x1 ) + µP (x2 ) we conclude that this is a linear map. This is not a linear map by the strict definition of a linear map since T (x) + T (y) = x + a + y + a = T (x + y) + a 6= T (x + y) .cam. Thus from the uniqueness of the zero element it follows that T (0) = 0 ∈ W .ac. x2 ∈ R3 under the map HΠ x1 7→ x01 = x1 − 2(x1 · n)n = HΠ (x1 ) . µ ∈ R. Mathematical Tripos: IA Algebra & Geometry (Part I) 56 c S. consider x 7→ x0 = T (x) = x + a .n = 0 where x. (2.uk.

We observe that T is a linear map since T (λx1 + µx2 ) = ((λx1 + µx2 ) + (λy1 + µy2 ). T (e3 ) = (0. e1 = (1. T (e2 ) = T ((0. T is a linear map since T (λx1 + µx2 ) = ((λx1 + µx2 ) + (λy1 + µy2 ). 1. z) 7→ (u. 0). 0. 2(λx1 + µx2 ) − (λz1 + µz2 )) = λ(x1 + y1 . Remarks. 0). 2x1 − z1 ) + µ(x2 + y2 . y) 7→ (s. v) = T (x) = (x + y. 0) and e3 = (0. x2 . i. Remarks. 1.e. λy1 + µy2 ) = λ(x1 + y1 . y) = (x + y.cam. u. y) . 1)) = (1. y − 3x.uk. Michaelmas 2006 . Thus the whole of the subspace of R3 spanned by {e1 − e2 + 2e3 }. 2)  T (e2 ) = (1. 0. x. the 1-dimensional line specified by λ(e1 − e2 + 2e3 ). v) = T (x.Cowley@damtp. • The subspace T (R2 ) = span{(1. −1)  Hence T (R3 ) = R2 . 1.e. • The standard basis vectors for R3 . 2x − z) . −3. 0)) = (1. This is a supervisor’s copy of the notes. T (e3 )}. 1. for all λ ∈ R. 1)} of R4 is a 2-D hyper-plane through the origin. 0) which are linearly independent. x1 . is mapped onto 0 ∈ R2 . • We observe that R3 = span{e1 . T (e2 ). (1. y1 − 3x1 . 0.ac. It is not to be distributed to students. λy1 + µy2 − 3(λx1 + µx2 ). y2 ) = λT (x1 ) + µT (x2 ) . t. • In this case we observe that the standard basis vectors of R2 are mapped to the vectors  T (e1 ) = T ((1. y. e2 = (0. 0) which are linearly dependent and span R2 . y2 − 3x2 . −3. 1) are mapped to the vectors  T (e1 ) = (1. y1 ) + µ(x2 + y2 . 1. • We also observe that T (e1 ) − T (e2 ) + 2T (e3 ) = 0 ∈ R2 which means that T (λ(e1 − e2 + 2e3 )) = 0 ∈ R2 . i. Mathematical Tripos: IA Algebra & Geometry (Part I) 57 c S. 1) and which form a basis for T (R2 ). λx1 + µx2 . 0. (v) As an example of a map where the domain of T has a lower dimension than the range of T consider T : R2 → R4 where (x.J. 2x2 − z2 ) = λT (x1 ) + µT (x2 ) .(iv) As another example of a map where the domain of T has a higher dimension than the range of T consider T : R3 → R2 where (x. e2 . e3 } and T (R3 ) = R2 = span{T (e1 ).

r(T ) = dim T (V ) . Mathematical Tripos: IA Algebra & Geometry (Part I) 58 c S. Hence T (R2 ) is the line x = λ(1. Recall that T (V ) is the image of V under T and that T (V ) is a subspace of W . r(T ) = 2.cam. i. The nullity of T is defined to be the dimension of the kernel. 2. It is not to be distributed to students. Proof. K(T ) is a subspace of V . of T . In both example (iv) and example (v) of §4. (4. 4. i. and hence λu + µv ∈ K(T ). Michaelmas 2006 .2) T (λu + µv) = λT (u) + µT (v) = λ0 + µ0 = 0. In example (iv) on page 57 n(T ) = 1.1.3. µ ∈ R then from (4. 0 ∈ K(T ) ⊆ V . v ∈ K(T ) and λ. Kernel and Nullity Let T : V → W be a linear map.1. Definition. Definition.Cowley@damtp. y) 7→ T (x. while in example (v) on page 57 n(T ) = 0. or null space. Since T (0) = 0 ∈ W . Suppose that u.ac. and so the rank of the map is such that r(T ) = dim T (R2 ) = 1 .5) 15/06 Theorem 4.e.J.1 on page 45. 2. 4.uk. i.e.3 Rank. 4x + 6y.e. (4. −2x − 3y) = (2x + 3y)(1.1 Examples (i) Consider the map T : R2 → R3 such that (x. −1) ∈ R3 . K(T ) = {v ∈ V : T (v) = 0 ∈ W } . y) = (2x + 3y. The subset of V that maps to the zero element in W is call the kernel. Remark. Definition. The proof then follows from invoking Theorem 3. −1) .2. n(T ) = dim K(T ) . (4. This is a supervisor’s copy of the notes.4) Examples.6) Examples. The rank of T is the dimension of the image. so K(T ) contains at least 0.

For future reference we note that r(T ) + n(T ) = 2 = dimension of domain. i. R2 . For the map to be well-defined the domain of T must include the 16/03 image of S.8) Proof. 4. then T .4.1 Examples (i) Let HΠ be reflection in a plane H Π : R3 → R3 . where as in (2. Thus n(T ) = dim K(T ) = 1 . then r(T ) + n(T ) = dim V . thus n(P ) = 2. we may apply map twice. (ii) Next consider projection onto a line. (4. Then P (R3 ) = {x ∈ R3 : x = λn. Further.cam.7) Theorem 4.Cowley@damtp. 4. Mathematical Tripos: IA Algebra & Geometry (Part I) 59 c S.17b) x 7→ x0 = P (x) = (x · n)n . so K(T ) = {x = (−3s. thus r(P ) = 1.e. the kernel is given by K(P ) = {x ∈ R3 : x · n = 0} . Then by geometry (or exercise and algebra) it follows that 2 HΠ HΠ = HΠ =I. It is not to be distributed to students. (4.e. See Linear Algebra in Part IB. Michaelmas 2006 . Since the range ⊆ domain. λ ∈ R} which is a line in R3 .10) where I is the identity map. Further.uk.J. Again we note that r(P ) + n(P ) = 3 = dimension of domain.ac. y) ∈ K(T ) if 2x + 3y = 0. which is a line in R2 . (4. Let V and W be real vector spaces and let T : V → W be a linear map. which is a plane in R3 .4 Composition of Maps Suppose that S : U → V and T : V → W are linear maps such that u 7→ v = S(u) .9) Definition. i. The composite or product map T S is the map T S : U → W such that u 7→ w = T (S(u)) . v 7→ w = T (v) . 2s) : s ∈ R} .2 (The Rank-Nullity Theorem). where we note that S acts first. (4. where n is a fixed unit vector. x = (x. This is a supervisor’s copy of the notes. P : R3 → R3 . R3 .

cam. (e0j )i is the ith component of e0j with respect to the basis {ei } (i = 1.14) It follows that for general x ∈ R3 3 X x 7→ x0 = M(x) = xj M(ej ) j=1 3 X 3 X = xj Mij ei j=1 i=1   3 X X3 =  Mij xj  ei . 3). v) 7→ T (−v. (4.5 Bases and the Matrix Description of Maps Let ei . u + v) and This is a supervisor’s copy of the notes. Then 3 X 3 X M(ej ) ≡ e0j = (e0j )i ei = Mij ei .2). From the definition of a linear map. Let M : R3 → R3 be a linear map (where for the time being we take the domain and range to be the same). Again we may apply the map twice. y. Then T S : R2 → R is such that (u. u. It is not to be distributed to students. 3. 2. (4. and we have set Mij = (e0j )i = (M(ej ))i . 3. u + v) = 2u . 0. Using Theorem 3.6 on page 49 we note that any x ∈ R3 has a unique expansion in terms of this basis. v) = (−v.ac. j = 1. namely X3 x= xj ej .Cowley@damtp. (4.12) j=1 j=1 Now consider the action of M on a basis vector. (x. and in this case we can show (by geometry or algebra) that P P = P2 = P .15) i=1 j=1 Mathematical Tripos: IA Algebra & Geometry (Part I) 60 c S. 16/02 4. (4. be a basis for R3 (it may help to think of it as the standard orthonormal basis e1 = (1. (4. etc.J. j=1 where the xj are the components of x with respect to the given basis. (4. (ii) Let P be projection onto a line P : R3 → R3 .11) (iii) Let S : R2 → R3 and T : R3 → R be linear maps such that (u. 2. it follows that   X3 X 3 M(x) = M  xj ej  = xj M(ej ) . Michaelmas 2006 . i = 1. 2. j = 1. z) = x + y + z . respectively.uk. z) 7→ T (x. 3.13) i=1 i=1 where e0j is the image of ej . u. 0). y. say ej . Remark. ST not well defined because the range of T is not the domain of S. v) 7→ S(u. 2.).

• Sometimes we write (a) M = {Mij } and (b) Mij = (M)ij .17a) x3 x03 respectively.16b).5. j = 1. x2 . and sans serif x denoting a column matrix of components.    0 x1 x1 x = x2  and x0 = x02  (4. Hence we now have   x1 x = x2  = (x1 . or equivalently (4. The explicit relation between x0i . (4. 4. completely specified by the 9 quantities Mij . x01 = M11 x1 + M12 x2 + M13 x3 . 3. x3 where we draw attention to the commas on the RHS.17b) M31 M32 M33 Remarks. the mapping M : R3 → R3 is.16a) i=1 j=1 Alternatively. Equation (4.e. 3 . i. 2. 2. x. 3.Cowley@damtp. i = 1. in terms of the suffix notation and summation convention introduced earlier x0i = Mij xj .Thus in component form 3 X 3 X x0 = x0i ei where x0i = Mij xj . i. once the standard basis (or any other basis) has been chosen. i = 1. italic xi denoting a component of a vector.18b) j=1 Mathematical Tripos: IA Algebra & Geometry (Part I) 61 c S. In other words. and xj . • We now have bold x denoting a vector. while the second suffix j is the column number. an accepted convention is to use the standard notation for a vector. 3 the elements of the matrix M. 2.e.J. x02 = M21 x1 + M22 x2 + M23 x3 . Let x and x0 be the column matrices. what this means is that once we know the Mij we can calculate the results of the mapping M : R3 → R3 for all elements of R3 . a column vector. • The first suffix i is the row number.1 Matrix notation The above equations can be written in a more convenient form by using matrix notation. or column vectors. It is not to be distributed to students.16b) Since x was an arbitrary vector.16a). in the case of a column matrix of vector components.cam. (4. (4. However. 3. 3 is This is a supervisor’s copy of the notes. i. x. x3 ) . Michaelmas 2006 . 2. j = 1. and let M be the 3 × 3 square matrix   M11 M12 M13 M = M21 M22 M23  . i = 1. j = 1. x03 = M31 x1 + M32 x2 + M33 x3 . (4. • To try and avoid confusion we have introduced for a short while a specific notation for a column matrix. i. 2.ac. • We call the Mij . 2. can now be expressed in matrix notation as x0 = Mx or equivalently x0 = Mx .uk.18a) where a matrix multiplication rule has been defined in terms of matrix elements as 3 X x0i = Mij xj . (4.e.

for i = 1. it is necessary to give the basis with respect to which it has been constructed.e. that represents HΠ .uk. b3 . 0) = (0. b2 . Hence (H)ij = Hij = δij − 2ni nj . (i) For i = 1. 3. for i = 1.21) In order to construct the map’s matrix P with respect to the standard basis. 4.91) we have that x 7→ HΠ (x) = x0 = x − 2(x · n)n . 2. where the e0i on the RHS are to be interpreted as column vectors. Similarly we obtain 1 − 2n21 −2n1 n2   −2n1 n3 H = −2n1 n2 1 − 2n22  −2n2 n3  .cam.2 Examples (including some important new definitions of maps) (i) Reflection.19) (e01 )3 (e02 )3 (e03 )3 This is a supervisor’s copy of the notes. (ii) From (4. 2. Then in terms of the scalar product for vectors ith row of x0 = x0i = ri · x .19) and similar expressions for Pb (e2 ) and Pb (e3 ) to deduce that   0 −b3 b2 Pb =  b3 0 −b1  .Cowley@damtp. We wish to construct matrix H. To this end consider the action of HΠ on each member of the standard basis. It is not to be distributed to students.ac. (4.14) (e01 )1 (e02 )1 (e03 )1   M = (e01 )2 (e02 )2 (e03 )2  = e01 e02 e03 . From (2.Remarks. recalling that ej · n = nj .5. (iii) The elements of M depend on the choice of basis. Michaelmas 2006 . 2. Hence when specifying a matrix M associated with a map M.J. i. 0 n3 −2n1 n3 This is the first column of H. 3. −b2 ) .22a) −b2 b1 0 Mathematical Tripos: IA Algebra & Geometry (Part I) 62 c S.  (4. (4. 3 let ri be the vector with components equal to the elements of the ith row of M. Mi3 ) . first note that Pb (e1 ) = (b1 . 3. Then. with respect to the standard basis. j = 1. Consider reflection in the plane Π = {x ∈ R3 : x · n = 0 and |n| = 1}. Now use formula (4. (4.20b) (ii) Consider the map Pb : R3 → R3 defined by x 7→ x0 = Pb (x) = b × x . b3 ) × (1. 0. (4. i.91) 16/06 (HΠ (x))i = xi − 2xj nj ni = δij xj − 2xj nj ni = (δij − 2ni nj )xj ≡ Hij xj . 2.20a) −2n1 n3 −2n2 n3 1 − 2n23 Alternatively we can obtain this same result using suffix notation since from (2. it follows that 1 − 2n21       1 n1 HΠ (e1 ) = e01 = e1 − 2n1 n = 0 − 2n1 n2  = −2n1 n2  . Mi2 . ri = (Mi1 .

the x1 direction. Hence.24) 0 0 ν The effect on the unit cube 0 6 x1 6 1. 0 6 x2 6 1.22a). Under such a rotation e1 →7 e01 = e1 cos θ + e2 sin θ .23) 0 0 1 (iv) Dilatation. It is not to be distributed to students. (4. (4. 0 6 x03 6 ν. by an amount proportional to the distance in that plane from. Consider the mapping R3 → R3 defined by x 7→ x0 where x01 = λx1 . x03 = νx3 where λ.g. where λ ∈ R. is given by   cos θ − sin θ 0 R(θ) =  sin θ cos θ 0 . e2 7→ e02 This is a supervisor’s copy of the notes. the x1 -axis. ν ∈ R and λ. Pij = εikj bk = −εijk bk . in agreement with (4.21) Pij xj = x0i = εijk bj xk = (εikj bk )xj . From (4. If λ = µ = ν then the transformation is called a pure dilatation. e.16b) and (4. e03 = νe3 .ac.g. (4. Under this transformation e1 7→ e01 = e1 . e3 → 7 e03 = e3 . ν > 0. to a cuboid that has been stretched or contracted by different factors along the different Cartesian axes. the x1 x2 -plane. say D. Michaelmas 2006 . 0 6 x3 6 1.Cowley@damtp. µ. i. The elements {Pij } of Pb could also be derived as follows. and so the map’s matrix with respect to the standard basis.22b) (iii) Rotation. that displaces points in one direction.e. Consider rotation by an angle θ about the x3 axis. µ. = −e1 sin θ + e2 cos θ .J. of this map is to send it to 0 6 x01 6 λ. Then e01 = λe1 . A simple shear is a transformation in the plane. e02 = µe2 . e. 17/03 (v) Shear. Thus the rotation matrix. R(θ). (4. is given by   λ 0 0 D = 0 µ 0 . x02 = µx2 . 0 6 x02 6 µ. say. e3 7→ e03 = e3 .uk. e2 7→ e02 = e2 + λe1 .cam.25) Mathematical Tripos: IA Algebra & Geometry (Part I) 63 c S.

26) 0 0 1 17/02 4.. .   . . . m. . (4. (4. .29a)        .e. . n ∈ Z+ . . .28c) k=1 Using the same rules of multiplication as before we write  0     x1 N11 N12 . .. n) such that m X N (ek ) = Njk fj . This is a supervisor’s copy of the notes.  0 xm Nm1 Nm2 .. (4.e. (4.J. N1n x1  x02   N21 N22 . For instance. It is not to be distributed to students. . we have found that a map M : R3 → R3 leads to a 3 × 3 matrix M. . say Sλ . (4. Let {ek } be a basis of Rn .uk.Cowley@damtp.   . . Consider now a linear map N : Rn → Rm where m. x0 = Nx where N = {Nij } . . . (4. .   .28a) j=1 Hence   n X m X x0 = N (x) = xk  Njk fj  k=1 j=1 m n ! X X = Njk xk fj . Michaelmas 2006 .ac.  . i.cam. where x ∈ Rn and x0 ∈ Rm . .3 dim(domain) 6= dim(range) So far we have considered matrix representations for maps where the domain and the range are the same. so n X x = xk ek .   . .27b) k=1 Let {fj } be a basis of Rm . For this example the map’s matrix (with respect to the standard basis).) ..5. .c. (4.27a) k=1 and n X N (x) = xk N (ek ) . is given by   1 λ 0 Sλ = 0 1 0 .29b) Mathematical Tripos: IA Algebra & Geometry (Part I) 64 c S. (4. then there exist Njk ∈ R (j = 1.28b) j=1 k=1 and thus n X x0j = (N (x))j = Njk xk = Njk xk (s. Nmn xn | {z } | {z } | {z } m × 1 matrix m × n matrix n × 1 matrix (column vector (m rows. k = 1. N2n   x2  =  .. . n columns) (column vector with m rows) with n rows) i. a map x 7→ x0 = N (x)..

28b)   Xm Xn (λA)(x) = λA(x) = λ  ajk xk fj  This is a supervisor’s copy of the notes. (4. Then from (4.c. Now consider the composite map W = T S : Rn → R` . Michaelmas 2006 .cam.3 Matrix multiplication Let S : Rn → Rm and T : Rm → R` be linear maps. and using the summation convention. x0j = Sjk xk and x00i = Tij x0j (s.) .33a) then from (4.34b) and hence because (4.uk. In words.c. with associated ` × n matrix W = {Wij }.35) We interpret (4.6. j=1 k=1 Hence. (4.Cowley@damtp. (4. then for consistency we define A + B = {aij + bij } .4. Mathematical Tripos: IA Algebra & Geometry (Part I) 65 c S. Let S = {Sij } be the m × n matrix of S with respect to a given basis.34a) However.). x00 = T S(x) = W(x) and x00i = Wik xk (s. (4. T = {Tij } be the ` × m matrix of T with respect to a given basis.6 Algebra of Matrices 4. or a given basis of Rn if m = n). (4. It is not to be distributed to students. (4. for consistency the matrix of λA must be λA = {λaij } . (4.32) 4.6. 4.ac. if A = {aij } and B = {bij } are both m × n matrices associated with maps Rn → Rm . (4. Let A = {aij } be the matrix of A (with respect to given bases of Rn and Rm .6.).c.30) This is also a linear map.2 Addition Similarly.34a) and (4. Then for given λ ∈ R define (λA) such that (λA)(x) = λ(A(x)) .31) which we use as the definition of a matrix multiplied by a scalar.34b) must identical for arbitrary x.c.28c). the ik th element of TS is equal to the scalar product of the ith row of T with the k th column of S. j=1 k=1 m X X n = (λajk )xk fj .35) as defining the elements of the matrix product TS.). it follows that Wik = Tij Sjk (s.J.33b) and thus x00i = Tij (Sjk xk ) = (Tij Sjk )xk (s. If x0 = S(x) and x00 = T (x0 ) .1 Multiplication by a scalar Let A : Rn → Rm be a linear map.

Remarks.18b).J.uk. then its transpose AT is defined to be a n × m matrix with elements (AT )ij = (A)ji = aji . then AB exists only if q = r. 18/03 4. i.cam. then A(BC) = (AB)C .ac. and is then a r × q matrix. the special case when S is a column matrix (or column vector). If A = {aij } is a m × n matrix. AB 6= BA in general. s t u e f ep + f s eq + f t er + f u (iv) Even if p = q = r = s. s t u sa + tc + ue sb + td + uf e f while     a b   ap + bs aq + bt ar + bu c d p q r =  cp + ds cq + dt cr + du  . Michaelmas 2006 .36) For instance      0 1 1 0 0 −1 = . 0 −1 1 0 −1 0 Lemma 4. (4. (iii) If A is a p × q matrix.38b) Mathematical Tripos: IA Algebra & Geometry (Part I) 66 c S. so that both AB and BA exist and have the same number of rows and columns. this is the case above since T is a ` × m matrix. while S is a m × n matrix. (ii) The above definition of matrix multiplication is consistent with the n = 1 special case considered in (4. i.37) Proof. (4.4 Transpose Definition. the number of columns of T must equal the number of rows of S. It is not to be distributed to students. ((AB)C)ij = (AB)ik ckj = ai` b`k ckj = ai£ b£U cUj .18a) and (4.Cowley@damtp. This is a supervisor’s copy of the notes. if A = {aij }. In terms of suffix notation (and the summation convention) (A(BC))ij = aik (BC)kj = aik bk` c`j = ai£ b£U cUj .6.38a) Remark.e. BA exists only if s = p.3. (i) For the scalar product to be well defined. and is then a p × s matrix. (AT )T = A .e. The multiplication of matrices is associative. 1 0 0 −1 1 0 while      1 0 0 1 0 1 = . (4. B = {bij } and C = {cij } are matrices such that AB and BC exist. and B is a r × s matrix. For instance     a b   p q r  pa + qc + re pb + qd + rf c d  = . (4.

2 4 6 5 6 (ii)   x1  x2  If x =  .uk. Then xT Ay = xi aij yj = x£ a£U yU . Similarly we deduce that all the diagonal elements of an antisymmetric matrix are zero.4. = ann = 0 . i. then (AB)T = BT AT . Examples. a scalar.e. Example. a11 = −a11 .40) 18/02 Definition.  xn T  x = x1 x2 . For an antisymmetric matrix. .  xn This is a supervisor’s copy of the notes.5 Symmetry Definition.6.e. If A = {aij } and B = {bij } are matrices such that AB exists. . Michaelmas 2006 . Recall that commas are sometimes important:   x1  x2  x =  . xT = x1  x2 . It follows that (xT Ay)T = yT AT x = yi aji xj = yU a£U x£ = xT Ay . . (i)  T 1 2   3 1 3 5 4 =  .39) Proof. Let x and y be 3 × 1 column vectors.. A square n × n matrix A = {aij } is antisymmetric if AT = −A .  = (x1 . (4. aji = aij .e.41a) Remark. (4. Lemma 4.    . 4.. A square n × n matrix A = {aij } is symmetric if AT = A . is a 1 × 1 matrix. (4. .e. i.  is a column vector.41b) Mathematical Tripos: IA Algebra & Geometry (Part I) 67 c S. Remark. aji = −aij . xn ) .    . . a11 = 0. ((AB)T )ij = (AB)ji = ajk bki = (B)ki (A)jk = (BT )ik (AT )kj = (BT AT )ij .. xn is a row vector. i. and let A = {aij } be a 3 × 3 matrix. . It is not to be distributed to students.J.ac. . . xn . i.cam.Cowley@damtp. i. a11 = a22 = . x2 ..e. (4.

Examples. b −c 0 i. it has six independent elements. The unit or identity n × n matrix is defined to be   1 0 . and hence T r(BC) = T r(CB) (even if m 6= n so that the matrices are of different sizes).7 The unit or identity matrix Definition.J.  .. .44)    .6. but are not usually equal (even if m = n). . (ii) An antisymmetric 3 × 3 matrix A has the form   0 a −b This is a supervisor’s copy of the notes.) . T r(CB) = (CB)ii = cij bji = bij cji . .cam. 0 0 1 . (4. b = v2 and c = v1 . ..e. 0 0 .c.e.Cowley@damtp.. 4. Remark.e.42) Thus each antisymmetric 3 × 3 matrix corresponds to a unique vector v in R3 .6 Trace Definition. .. it has three independent elements. all the elements are 0 except for the diagonal elements that are 1. (4. Michaelmas 2006 . The 3 × 3 identity matrix is given by   1 0 0 I = 0 1 0 = {δij } . (4. It is not to be distributed to students. then (cf.ac. (4.43) Remark. T r(A) = aii (s. Let a = v3 . Example. 1 i. However T r(BC) = (BC)ii = bij cji . . The trace of a square n × n matrix A = {aij } is equal to the sum of the diagonal elements..6.22b)) A = {aij } = {εijk vk } ..22a) and (4. i. 0 I = . Let B = {bij } be a m × n matrix and C = {cij } be a n × m matrix. c e f i.uk.e. .. 4. (4. (i) A symmetric 3 × 3 matrix S has the form   a b c S = b d e . A = −a 0 c. .. then BC and CB both exist.45) 0 0 1 Mathematical Tripos: IA Algebra & Geometry (Part I) 68 c S.

From (4. (4.cam. Property.e. in the case of a square matrix. A is said to be invertible if there exists a n × n matrix B such that BA = AB = I .5.47) This is a supervisor’s copy of the notes. In general. is unique (see above) and is denoted by A−1 (see above). Let A be a n × n matrix. Let A be a square n × n matrix. Then (AB)−1 = B−1 A−1 . 4. B is a left inverse of A if BA = I. . the existence of a left inverse does not necessarily guarantee the existence of a right inverse. then (IA)ij = δik akj = aij . If B is a left inverse of A and C is a right inverse of A then B = C and we write B = C = A−1 .47). n.8 The inverse of a matrix Definition.6. Remark.uk. .48) it follows that B−1 A−1 (AB) = B−1 (A−1 A)B = B−1 IB = B−1 B = I . Michaelmas 2006 . (4.48) it follows that A = B−1 (in addition to B = A−1 ). Hence A = (A−1 )−1 . (4. j = 1. BA = I and AC = I it follows that B = BI = B(AC) = (BA)C = IC = C .48) The matrix B is called the inverse of A.J. From using (4. . Define the Kronecker delta in Rn such that  1 if i = j δij = for i. the existence of a left inverse does imply the existence of a right inverse. Proof. From (4. or vice versa.37).46) 0 if i 6= j Let A = {aij } be a n × n matrix. (4.ac. The lemma is based on the premise that both a left inverse and right inverse exist.49) Lemma 4.6.50) Proof. IA = AI = A . (AI)ij = aik δkj = aij . Mathematical Tripos: IA Algebra & Geometry (Part I) 69 c S. (4. The above lemma then implies that they are the same matrix.37). C is a right inverse of A if AC = I. Lemma 4. and vice versa (see Part IB Linear Algebra for a general proof). 2.Property. (AB)B−1 A−1 = A(BB−1 )A−1 = AIA−1 = AA−1 = I . Definition. However. (4.Cowley@damtp. i. Suppose that A and B are both invertible n × n matrices.47) and (4. (4. It is not to be distributed to students. .

(4.52b) = a11 (a22 a33 − a23 a32 ) + a12 (a23 a31 − a21 a33 ) + a13 (a21 a32 − a22 a31 ) (4. on the volume of the unit cube defined by standard orthonormal basis vectors ei .52c) = εijk a1i a2j a3k (4. 4. Consider the effect of a linear map. A : R3 → R3 .52d) = a11 a22 a33 + a12 a23 a31 + a13 a21 a32 − a11 a23 a32 − a12 a21 a33 − a13 a22 a31 .6.9 Determinants (for 3 × 3 and 2 × 2 matrices) Recall that the signed volume of the R3 parallelepiped defined by a. It is not to be distributed to students. b and c are right-handed. (4. Definition.52a) = a11 (a22 a33 − a32 a23 ) + a21 (a32 a13 − a12 a33 ) + a31 (a12 a23 − a22 a13 ) (4. b and c is a · (b × c) (positive if a. given by e01 · e02 × e03 = εijk (e01 )i (e02 )j (e03 )k = εijk ai` (e1 )` ajm (e2 )m akn (e3 )n = εijk ai` δ1` ajm δ2m akm δ3n = εijk ai1 aj2 ak3 . negative if left-handed).67e) and (4. then the volume of the mapped cube is. Alternative notations for the determinant of the matrix A include . Let A = {aij } be the matrix associated with A.51) This is a supervisor’s copy of the notes.28c). The determinant of a 3 × 3 matrix A is given by det A = εijk ai1 aj2 ak3 (4.52e) Alternative notation. with the aid of (2.

.

.

a11 a12 a13 .

.

.

a22 a23 .

.

= .

e01 e02 e03 .

. .

.

det A = | A | = .

.

a21 (4.53) .

a31 a32 a33 .

• A linear map R3 → R3 is volume preserving if and only if the determinant of its matrix with respect to an orthonormal basis is ±1 (strictly ‘an’ should be replaced by ‘any’. Remarks. 0 • If . but we need some extra machinery before we can prove that).

0{ei }0 is a0 .

standard right-handed orthonormal .

0 .

basis then the set {ei } is right-handed if 0 0 .

e1 e2 e3 .

> 0. and left-handed if .

e1 e2 e3 .

(i) Show that the determinant of the rotation matrix R defined in (4. (ii) Show that the determinant of the reflection matrix H defined in (4. (4. or equivalently (4. < 0.52c)) .23)).20a). A map R3 → R3 is effectively two-dimensional if a33 = 1 and a13 = a31 = a23 = a32 = 0 (cf. is −1 (since reflection sends a right-handed set of vectors to a left-handed set). Hence for a 2 × 2 matrix A we define the determinant to be given by (see (4.23) is +1. 18/06 Exercises.52b) or (4.20b). 2 × 2 matrices.

.

.

a a12 .

.

det A = | A | = .

.

11 = a11 a22 − a12 a21 . (4.54) a21 a22 .

If you are into the abuse of notation show that . A map R2 → R2 is area preserving if det A = ±1. This is not for the faint-hearted. An abuse of notation.

.

.

i j k .

.

.

a × b = .

.

a1 a2 a3 .

.

55) . . (4.

b1 b2 b3 .

ac.Cowley@damtp.J. as ‘components’. Michaelmas 2006 .cam.uk. j and k. where we are treating the vectors i. 19/03 Mathematical Tripos: IA Algebra & Geometry (Part I) 70 c S.

it follows that H2 = I . since AT A = I. e3 7→ Ae3 the third column of A. (4. Similarly.uk.58a) Moreover H is symmetric. and thus RRT = RT R = I .57) From (4. Thus if A is an orthogonal matrix the {ei } transform to an orthonormal set (which may be right- handed or left-handed depending on the sign of det A).J. Then from (4. (4. Examples.7 Orthogonal Matrices Definition. (4. (AT )ik (A)kj = aki akj = δij .56a) i.20b) the matrix of HΠ with respect to a standard basis is specified by (H)ij = {δij − 2ni nj } . (4.cam. or a little manipulation. This This is a supervisor’s copy of the notes.ac.23))   cos θ − sin θ 0 R =  sin θ cos θ 0 . implies that the rows of A form an orthonormal set. (4.19) we recall that e1 7 → Ae1 the first column of A. and so H2 = HHT = HT H = I .e. Property: orthogonal rows and columns.59b) 19/02 Mathematical Tripos: IA Algebra & Geometry (Part I) 71 c S. It is not to be distributed to students. i. 2 HΠ =I. e2 7→ Ae2 the second column of A. hence H = HT .56c) and so the columns of A also form an orthonormal set.e. Michaelmas 2006 . (i) We have already seen that the application of a reflection map HΠ twice results in the identity map. In components (4. (ii) With respect to the standard basis. if A is invertible and A−1 = AT . (4. (4.56a) becomes (A)ik (AT )kj = aik ajk = δij . 4. Suppose that the map A : R3 → R3 has a matrix A with respect to an orthonormal basis. rotation by an angle θ about the x3 axis has the matrix (see (4.57).59a) 0 0 1 R is orthogonal since both the rows and the columns are orthogonal vectors. From (4. Property: map of orthonormal basis.56b) Thus the scalar product of the ith and j th rows of A is zero unless i = j in which case it is 1. An n × n matrix A = {aij } is orthogonal if AAT = I = AT A .Cowley@damtp. (4.58b) Thus H is orthogonal.

where Aij is the ith component of the vector e Aij can be represented by a square n × n transformation matrix A   A11 A12 · · · A1n  A21 A22 · · · A2n   A= . .  . . . For suppose that.1 Transformation matrices Let {ei : i = 1. . (4.8 Change of basis We wish to consider a change of basis. . .  = e1 e2 . |x0 − y0 |2 = (x0 − y0 ) · (x0 − y0 ) = (x − y) · (x − y) = |x − y|2 . . i.. Isometric maps. .  An1 An2 ··· Ann where in the final matrix the e ej are to be interpreted as column vectors of the components of the e ej in the {ei } basis (cf. we will work in Rn and will not assume that either basis is orthonormal (unless stated otherwise). 4. However. Under a map represented by an orthogonal matrix with respect to an orthonormal basis. . 4..cam. . ..J.62) k=1 for some numbers Bki . . Similarly.uk. If a linear map is represented by an orthogonal matrix A with respect to an orthonormal basis.  . in component form.61)   . where Bki is the kth component of the vector ei in the basis {e ek }. Again the Bki can be viewed as the entries of a matrix B   B11 B12 · · · B1n  B21 B22 · · · B2n   B= . . then the map is an isometry since This is a supervisor’s copy of the notes. Hence |x0 − y0 | = |x − y|. The only length preserving maps of R3 are translations (which are not linear maps) and reflections and rotations (which we have already seen are associated with orthogonal matrices). The numbers for some numbers Aij . . the individual basis vectors of the basis {e ei } can be written as n X ej = e ei Aij (j = 1. . (4. en . . (4. Remark. Michaelmas 2006 . It is not to be distributed to students.e.. x 7→ x0 = Ax and y 7→ y0 = Ay (note the use of sans serif). een . . (4.8. To fix ideas it may help to think of a change of basis from the standard orthonormal basis in R3 to a new basis which is not necessarily orthonormal. n) . . a scalar product is preserved. . .Cowley@damtp. 2.63)   . n} and {e ei : i = 1.  = ee1 ee2 . .. . .  Bn1 Bn2 ··· Bnn Mathematical Tripos: IA Algebra & Geometry (Part I) 72 c S. (4. .. n) . .19)). . lengths are preserved. . . the individual basis vectors of the basis {ei } can be written as n X ei = ek Bki e (i = 1. . then x0 · y0 = x0T y0 (since the basis is orthonormal) = (xT AT )(Ay) = xT I y = xT y = x·y (since the basis is orthonormal) . . Since the {ei } is a basis.ac. . n} be two sets of basis vectors for an n-dimensional real vector space V . since the {e ei } is a basis.Preservation of the scalar product.60) i=1 ej in the basis {ei }. .

4. that v = Ae v. (4.J.67) j=1 Xn n X = vej ei Aij from (4. i=1 j=1 Since a basis representation is unique it follows that n X vi = Aij vej .3 Transformation law for vector components Consider a vector v.2 Properties of transformation matrices From substituting (4.68c) Equations (4.cam.68c) relate the components of v with respect to the {ei } basis to the components with respect to the {e ei } basis. the set {e ej } is a basis and so linearly independent. (4. it follows that n X Bki Aij = δkj . substituting (4. Michaelmas 2006 .64) i=1 k=1 k=1 i=1 However.68a) j=1 i. By applying A−1 to either side of (4.Cowley@damtp.13) that n X v= vi ei .68b) and (4. from noting that n This is a supervisor’s copy of the notes. where I is the identity matrix.8. Conversely.62) into (4.60) we have that Xn X n  n X n X  ej = e ek Bki Aij = e ek e Bki Aij . i=1 Similarly.uk.60) j=1 i=1 Xn Xn = ei (Aij vej ) swap summation order.8. BA = I. e (4. It is not to be distributed to students.68b) it follows that v = A−1 v .66) 4. Thus. (4. (4.ac. X ej = e ek δkj .e. Thus B = A−1 .60) into (4.where in the final matrix the ej are to be interpreted as column vectors of the components of the ej in the {e ei } basis. (4. e k=1 or otherwise.62) leads to the conclusion that AB = I (alternatively argue by a relabeling symmetry).65) i=1 Hence in matrix notation. then in the {ei } basis it follows from (3. Mathematical Tripos: IA Algebra & Geometry (Part I) 73 c S.68b) where we have deliberately used a sans serif font to indicate the column matrices of components (since otherwise there is serious ambiguity). in the {e ei } basis we can write n X v= vej e ej (4.

62) that ! 1 1 B = A−1 = 1 2 . Verify the transformation law for the components of an arbitrary vector v in the two coordinate systems. Answer. it follows from (4. Then it follows from from (4. −1 1 Moreover A−1 A = AA−1 = I. 1 1 Similarly.e. e2 = (−1.Cowley@damtp. 1)} and {e e1 = (1. 1) + (−1.uk. 1).e. 1) ) = 12 (e e1 − e e2 ) . We have that e1 = ( 1. 1) − (−1. with respect to the basis {ei }. since e1 = (1. 1) = −e1 + e2 . e e2 = (−1. Now consider an arbitrary vector v.68c).   1 −1 A= . It is not to be distributed to students. v0 and e Let e v be the component column matrices of v0 and v with respect to the alternative basis {e ei }. Find the transformation matrix A that connects them. e Hence from comparison with (4.cam. −1 1 4. 1) = 12 ( (1. i. 1) ) = 12 (e e1 + e e2 ) . Michaelmas 2006 .60) A11 = 1 . 1) = e1 + e2 . 0) + (0. We deduce that the matrix of M with respect to the alternative basis {e ei } is given by e = A−1 MA . v0 = (A−1 MA) e i. Let {e1 = (1. Then v = v1 e1 + v2 e2 = 12 v1 (e e2 ) + 12 v2 (e e1 − e e1 + e e2 ) = 12 (v1 + v2 ) e e1 − 12 (v1 − v2 ) e e2 .4 Transformation law for matrices representing linear maps Now consider a linear map M : Rn → Rn under which v 7→ v0 = M(v) and (in terms of column vectors) v0 = Mv . we deduce that (as above) and thus from (4. 0) + (0. 1) = e (1. 1)} be two sets of basis vectors in R2 . 0). i. e2 = (0.Worked Example. This is a supervisor’s copy of the notes. respectively. A21 = 1 . v = A−1 v. 0) = 12 ( (1. e2 = (0. Thus ve1 = 12 (v1 + v2 ) and ve2 = − 21 (v1 − v2 ) . A12 = −1 and A22 = 1 .8.e. e ! 1 1 A−1 = 21 . 1) = −1 (1.69) Mathematical Tripos: IA Algebra & Geometry (Part I) 74 c S. e v.ac.J. M (4.68b) that v0 = MAe Ae v. where v0 and v are the component column matrices of v0 and v. and M is the matrix of M with respect to this basis.

x2 ) plane. fm . and thus   cos θ − sin θ 0 A =  sin θ cos θ 0 . Then from (4. e3 e = e3 . and N is matrix of N with respect to these two bases. ei } of Rn and {e Now consider new bases {e fi } of Rm . where e v0 are component column matrices of v and v0 with respect to bases {e v and e ei } and {e fi } respectively.uk.59a) and (4. see (4.. 0 0 1 The matrix of the shear map with respect to new basis is thus given by Sγ e = A−1 Sγ A    cos θ sin θ 0 cos θ + γ sin θ − sin θ + γ cos θ 0 = − sin θ cos θ 0  sin θ cos θ 0 0 0 1 0 0 1 γ cos2 θ   1 + γ sin θ cos θ 0 =  −γ sin2 θ 1 − γ sin θ cos θ 0 . Michaelmas 2006 .. Then v = Aev . en e and C = e f1 . Hence v0 = NAe Ce v . Suppose {ei } is a basis of Rn . 0 0 1 Let {e e} be the basis obtained by rotating the stan- dard basis by an angle θ about the x3 axis. and let   e1 A= e . i. A similar approach may be used to deduce the matrix of the map N : Rn → Rm (where m 6= n) with respect to new bases of both Rn and Rm . N (4. and v0 = Ce v0 . Maps from Rn to Rm (unlectured).Cowley@damtp. and hence   cos θ sin θ 0 A−1 =  − sin θ cos θ 0  .61)) and C is a m × m matrix of components. It follows that C−1 NA is map’s matrix with respect to the new bases {e ei } and {e fi }. 0 0 1 We have already deduced that rotation matrices are orthogonal.29b) v 7→ v0 = N v where v and v0 are component column matrices of v and v0 with respect to bases {ei } and {fi } respectively. 0 0 1 20/03 Mathematical Tripos: IA Algebra & Geometry (Part I) 75 c S. and so e v0 = C−1 NAe v. e2 e = − sin θ e1 + cos θ e2 .J. then from (4.70) 19/06 Example. {fi } is a basis of Rm . Then e1 e = cos θ e1 + sin θ e2 .26) the matrix of this map with respect to the standard basis {ei } is   1 γ 0 Sγ =  0 1 0  .e.ac.59b). This is a supervisor’s copy of the notes. It is not to be distributed to students.cam.. e = C−1 NA .. Consider a simple shear with magnitude γ in the x1 direction within the (x1 . e where A is a n × n matrix of components (see (4.

a22 × (5. It is not to be distributed to students. We will concentrate on the case of three linear equations in three unknowns. The ‘real’ world can often be described by equations of various sorts. and A = {aij } (a 2 × 2 matrix).1a) − a12 × (5. From (4.54) we have that . However. d= .0 Why Study This? This section continues our study of linear mathematics.1 Solution of Two Linear Equations in Two Unknowns This is a supervisor’s copy of the notes. the solution of these more complicated models often involves the solution of linear equations. Matrix Inverses and Linear Equations 5. (a21 a12 − a22 a11 )x2 = a21 d1 − a11 d2 .2a) where     x1 d1 x= .1b) or equivalently Ax = d (5. but the systems in the case of the real world are normally much larger (often by many orders of magnitude). 5.1a) a21 x1 + a22 x2 = d2 (5. (5.5 Determinants. Some models result in linear equations of the type studied here.2b) x2 d2 Now solve by forming suitable linear combinations of the two equations (e.1b)) (a11 a22 − a21 a12 )x1 = a22 d1 − a12 d2 . even when the real world results in more complicated models.g. Consider two linear equations in two unknowns: a11 x1 + a12 x2 = d1 (5.

.

.

a a12 .

.

(a11 a22 − a21 a12 ) = det A = .

.

11 . a21 a22 .

3a) det A x2−a21 a11 d2 However. x2 = (−a21 d1 + a11 d2 )/ det A . the equations have a unique solution x1 = (a22 d1 − a12 d2 )/ det A . if det A 6= 0. Check that AA−1 = A−1 A = I.2a) by A−1 (if it exists) we have that x = A−1 d . (5. (5. i.      1 x1a22 −a12 d1 = . Thus. 5.4) det A −a21 a11 Exercise.e. from left multiplication of (5.3b) We therefore conclude that   −1 1 a22 −a12 A = . (5.2 Determinants for 3 × 3 Matrices For a 3 × 3 matrix A we already have that (see (4.52a) and following) .

.

.

a11 a12 a13 .

.

.

det A ≡ | A | ≡ .

.

a21 a22 a23 .

.

(5.5a) .

a31 a32 a33 .

uk. = εijk ai1 aj2 ak3 (5.J. Michaelmas 2006 .ac. (5.Cowley@damtp.5b) = εijk a1i a2j a3k (5.5e) Mathematical Tripos: IA Algebra & Geometry (Part I) 76 c S.cam.5d) = a11 (a22 a33 − a23 a32 ) − a21 (a12 a33 − a13 a32 ) + a31 (a12 a23 − a22 a13 ) .5c) = a11 (a22 a33 − a23 a32 ) − a12 (a21 a33 − a23 a31 ) + a13 (a21 a32 − a22 a31 ) (5.

We observe that the last line.e. (5. i. can be rewritten .5e).

.

.

.

.

.

.

a a23 .

.

.

a12 a13 det A = a11 .

.

22 .

+ a31 .

a12 a13 .

.

.

.

− a21 .

.

.6) a32 a33 . (5.

a32 a33 .

.

a22 a23 .

This observation can be used to generalise the definition (using recursion). . or otherwise. and 20/02 evaluation. • (5.2. Remarks.6).1 Properties of 3 × 3 determinants (i) Since εijk ai1 aj2 ak3 = εijk a1i a2j a3k det A = det AT . • Note the sign pattern in (5. of determinants to larger (n × n) matrices (but is left undone) This is a supervisor’s copy of the notes. 5.6) in terms of the elements of the first row of A and determinants of 2 × 2 sub-matrices also exists.6) is an expansion of det A in terms of elements of the first column of A and determinants of 2 × 2 sub-matrices. (5. Hence from (5.5d).7) This means that an expansion equivalent to (5. It is not to be distributed to students.

.

.

.

.

.

.

a22 a23 .

.

a21 a23 .

.

a21 a22 .

det A = a11 .

.

.

− a12 .

.

.

+ a13 .

.

.

(5..8) a32 a33 .

a31 a33 .

a31 a32 .

ak3 = γk . or from (5.7) if and only there is linear dependence between the rows of A. β and γ are linearly dependent. Remark.5b)   α1 β1 γ1 det  α2 β2 γ2  = ijk αi βj γk = α · (β × γ) . This property generalises to n × n determinants (but is left unproved). e by adding to a given column of A linear combinations of the other columns.7) generalises to n × n determinants (but is left unproved). Remark. aj2 = βj . This follows from the fact that . if α. Similarly det A = 0 if and only if there is linear dependence between the columns of A. Remark. Hence if we interchange any two columns of A we change the sign of det A. α3 β3 γ3 Now α · (β × γ) = 0 if and only if α. i. (iv) Suppose that we construct a matrix.e. (ii) Let ai1 = αi . (iii) If we interchange any two of α. (5. This property generalises to n × n determinants (but is left unproved). Then det A e = det A . say A. then from (5. β and γ are coplanar. β and γ we change the sign of α · (β × γ). similarly if we interchange any two rows of A we change the sign of det A.

.

.

α1 + λβ1 + µγ1 β1 γ1 .

.

.

.

α2 + λβ2 + µγ2 β2 .

γ2 .

.

= (α + λβ + µγ) · (β × γ) .

α3 + λβ3 + µγ3 β3 γ3 .

From invoking (5. = α · (β × γ) + λβ · (β × γ) + µγ · (β × γ) = α · (β × γ) .cam. • This property generalises to n × n determinants (but is left unproved).7) we can deduce that a similar result applies to rows.uk.ac. Mathematical Tripos: IA Algebra & Geometry (Part I) 77 c S. Michaelmas 2006 .J.Cowley@damtp. Remarks.

From (5. (vi) If we multiply the matrix A by λ.uk.10b) det AB = ijk (AB)i1 (AB)j2 (AB)k3 = ijk aip bp1 ajq bq2 akr br3 = ijk aip ajq akr bp1 bq2 br3 = pqr det A bp1 bq2 br3 = det A det B . Then the left-hand side is zero. then det AB = (det A)(det B) . so the sign of right-hand side also reverses. since (λα) · (β × γ) = λ(α · (β × γ)) . Similarly for swaps of p and r.10a) is just (5.10b) starting from (5. if a11 6= 0 add multiples of the first row to subsequent rows to make ai1 = 0 for i = 2. and suppose that p = 1. Wlog take p = q = 1. This property generalises for n × n determinants to det(λA) = λn det A .10a) is true.J. Michaelmas 2006 . This theorem generalises to n × n determinants (but is left unproved).9b) but is left unproved. then This is a supervisor’s copy of the notes. then pqr det A = ijk api aqj ark .10a). Theorem 5. or q and r. then det A = a11 ∆11 (see (5.10a) are equal. Remark.1. Remark. . (5.11) Proof. If A = {aij } is 3 × 3. q = 2.5b) and (5. It is not to be distributed to students. • This property is useful in evaluating determinants. It follows that (5.10a) holds for any {pqr} that is a permutation of {123}. which is also zero. Next suppose that p and q are swapped. (5. while the right-hand side is ijk a1i a1j ark = jik a1j a1i ark = −ijk a1i a1j ark . We prove the result only for 3 × 3 matrices. Start with (5. .5c). n.10a) reverses. to give A. Then the sign of the left-hand side of (5.2. q and r in (5.10b) Proof. while the right-hand side becomes ijk aqi apj ark = jik aqj api ark = −ijk api aqj ark . This property generalises to n × n determinants (but is left unproved). say. For instance. r = 3. .Cowley@damtp. Mathematical Tripos: IA Algebra & Geometry (Part I) 78 c S. Similarly for (5. (5. (v) If we multiply any single row or column of A by λ.ac.10a) pqr det A = ijk aip ajq akr . 20/06 Theorem 5. (5.5b) Remark. (5. Having covered all cases we conclude that (5. .9a) since λα · (λβ × λγ) = λ3 (α · (β × γ)) .13b)).cam. If A and B are both square matrices. det(λA) = λ3 det A . Then (5. Suppose now that two (or more) of p. b then det A b = λ det A .

.. 21/03 Remark.13b) The above definitions apply for n × n matrices.11) that (det A)2 = (det A)(det AT ) = det(AAT ) = det I = 1 . . . i. It follows from (5. (5.13b) . . If A is orthogonal then AAT = I. . a(i−1)n   A =  .. a (i+1)(j−1) a (i+1)(j+1) .. a (i+1)n   . (5. but henceforth assume that A is a 3 × 3 matrix.  an1 . Then from (5. .12) Proof.7) and (5. Hence   a11 .  . . . of the ij th element of square matrix A to be the determinant of the square matrix obtained by eliminating the ith row and the j th column of A... . . 5.. Hence det A = ±1.1 Minors and cofactors For a square n × n matrix A = {aij }.  ij  a(i−1)1 . . . a1(j−1) a1(j+1) . . .3. . a1n  . an(j−1) an(j+1) . . ..   .3 The Inverse of a 3 × 3 Matrix This is a supervisor’s copy of the notes. define Aij to be the square matrix obtained by eliminating the ith row and the j th column of A.. . . a  (i+1)1 . ..   . . . . Define the cofactor ∆ij of the ij th element of square matrix A as ∆ij = (−1)i+j Mij = (−1)i+j det Aij .. . (5. . ann Definition..e. 5. Define the minor. You have already verified this for some reflection and rotation matrices.. a(i−1)(j−1) a(i−1)(j+1) . .. It is not to be distributed to students.13a) Definition.3. If A is orthogonal then det A = ±1 ..... . Theorem 5. Mij = det Aij .6) and (5. .. Mij ..

.

.

.

.

.

.

a22 a23 .

.

a12 a13 .

.

a12 a13 .

det A = a11 .

.

.

− a21 .

.

.

+ a31 .

.

.

a32 a33 .

a32 a33 .

a22 a23 .

14a) Similarly. after noting from an interchanges of columns that . (5. = a11 ∆11 + a21 ∆21 + a31 ∆31 = aj1 ∆j1 .

.

.

.

.

a11 a12 a13 .

.

.

a12 a11 a13 .

.

.

.

det A = .

.

a21 a22 a23 .

.

= − .

.

a22 a21 a23 .

.

. .

a31 a32 a33 .

.

a32 a31 a33 .

we have that .

.

.

.

.

.

.

a a23 .

.

.

a11 a13 .

.

a11 a13 .

.

det A = −a12 .

.

21 + a22 .

.

.

− a32 .

.

a31 a33 .

a31 a33 .

a21 a23 .

cam. Michaelmas 2006 .ac.uk.J.14b) Mathematical Tripos: IA Algebra & Geometry (Part I) 79 c S. = a12 ∆12 + a22 ∆22 + a32 ∆32 = aj2 ∆j2 .Cowley@damtp. (5.

(5. (5.14c) Similarly.8) and subsequently interchanging rows. det A = a1j ∆1j = a2j ∆2j = a3j ∆3j .Analogously (or by a relabelling symmetry) det A = aj3 ∆j3 . but starting from (5. .15) Next we wish to show that aij ∆ik = 0 if j 6= k. It is not to be distributed to students.16a) To this end consider j = 2 and k = 1. (5. then ai2 ∆i1 = a12 ∆11 + a22 ∆21 + a32 ∆31 This is a supervisor’s copy of the notes.

.

.

.

.

.

.

a22 a23 .

.

a12 a13 .

.

a a13 .

.

= a12 .

.

.

− a22 .

.

.

+ a32 .

.

12 a32 a33 .

a32 a33 .

a22 a23 .

.

.

.

a12 a12 a13 .

.

.

= .

.

a22 a22 a23 .

.

.

a32 a32 a33 .

15).cam. It follows that B = A−1 and A is invertible. (5.17a) aji ∆ki = δjk det A . Mathematical Tripos: IA Algebra & Geometry (Part I) 80 c S. we can also show that aji ∆ki = 0 if j 6= k. We can combine all the above results in a lemma. (5.18a) det A then AB = BA = I . From (5. (AB)ij = aik (B)kj aik ∆jk = det A δij det A = det A = δij . (5. Lemma 5.17a) and (5. aij ∆ik = δjk det A . (5. Proceed similarly for other choices of j and k to obtain (5. = 0. Hence AB = I.14a). Further. Similarly from (5. Michaelmas 2006 .4.17b) Proof. 5.uk.17b) and (5.16a) and (5.18a).ac. See (5.18b) Proof.2 Construction of the inverse Theorem 5. (5. (5.Cowley@damtp. Given a 3 × 3 matrix A with det A 6= 0.16b) since two rows turn out to be linearly dependent. (5. (5.16b).5. (5. BA = I.3.16a).14b).J. define B by 1 (B)ij = ∆ji .18a).14c). since two columns are linearly dependent (actually equal).

21a) and (5.20b) Hence if we wished to solve (5. and then form A−1 d. Hence     ∆11 ∆21 ∆31 1 −γ 0 S−1 γ = ∆12 ∆22 ∆32  = 0 1 0 .e. ∆31 = 0 . i.21c) Assume a11 = 6 0.19). (5.21a) a21 x1 + a22 x2 + a23 x3 = d2 .20a) where A is a n × n matrix. then one method would be to calculate A−1 using (5. (5.J. Consider the simple shear   1 γ 0 Sγ = 0 1 0 . x is a n × 1 column vector of unknowns. this has the formal solution x = A−1 d .21b) − × (5. A better method is Gaussian elimination. Assuming that det A 6= 0.21a) to eliminate x1 by forming a21 a31 (5. However. Hence suppose we wish to solve a11 x1 + a12 x2 + a13 x3 = d1 .20a) numerically. ∆33 = 1 .21a) .cam. (5. 0 0 1 Then det Sγ = 1. and after a little manipulation This is a supervisor’s copy of the notes. ∆23 = 0 . Michaelmas 2006 . ∆11 = 1 . and if that is not possible then stop (since there is then no unique solution). this is actually very inefficient.Cowley@damtp.uk. (5.22a) a11 a11 a11     a31 a31 a31 a32 − a12 x2 + a33 − a13 x3 = d3 − d1 . ∆22 = 1 . It is not to be distributed to students. which we will illustrate for 3 × 3 matrices.21c) − × (5. Remark. Example.4 Solving Linear Equations: Gaussian Elimination Suppose that we wish to solve the system of equations Ax = d . (5.ac. Now use (5. (5. 21/02 5. otherwise re-order the equations so that a11 6= 0. The formula for the elements of the inverse of A.21b) a31 x1 + a32 x2 + a33 x3 = d3 .19) det A holds for n × n matrices (including 2 × 2 matrices) for suitably defined cofactors. ∆12 = 0 . a11 a11 so as to obtain     a21 a21 a21 a22 − a12 x2 + a23 − a13 x3 = d2 − d1 . 1 (A−1 )ij = ∆ji . ∆32 = 0 . ∆13 = 0 . and d is a given n × 1 column vector. (5. (5.22b) a11 a11 a11 Mathematical Tripos: IA Algebra & Geometry (Part I) 81 c S. ∆21 = −γ . ∆13 ∆23 ∆33 0 0 1 which makes physical sense in that the effect of a shear γ is reversed by changing the sign of γ.

where as usual we specialise to the case where A is a 3 × 3 matrix.24a) gives x3 .21a) gives x1 .23a) a032 x2 + a033 x3 = d03 . and if that is not possible then stop This is a supervisor’s copy of the notes.24b) a22 (5. otherwise re-order the equations so that a22 6= 0. Definition.e. In order to simplify notation let     a21 a21 a21 a022 = a22 − a12 . Mathematical Tripos: IA Algebra & Geometry (Part I) 82 c S.23b) Assume a22 = 6 0. and finally (5. (5.23a) gives x2 . It is possible to show that this method fails only if A is not invertible. Now use (5. only if det A = 0. i. Definition.22b) become a022 x2 + a023 x3 = d02 . If d 6= 0 then the system of equations Ax = d (5. a11 a11 a11 so that (5.ac. Michaelmas 2006 .5 Solving linear equations 5. a033 = a33 − a13 . (since there is then no unique solution).cam.5. 5. It is not to be distributed to students.25b) 21/06 is said to be a system of homogeneous equations. Remark.uk.J. (5. In such circumstances it follows that the system of equations Ax = d has a unique solution x = A−1 d.23b) − × (5.23a) .5 on page 80 that A is invertible and A−1 exists. d03 = d3 − d1 . The system of equations Ax = 0 (5. providing that a0   a033 − 032 a023 6= 0 . a022 so as to obtain a032 0 a032 0   a033 − 0 a 0 23 x3 = d3 − 0 d2 . a11 a11 a11     a31 a31 a31 a032 = a32 − a12 . If a0   a033 − 032 a023 = 0 . If det A 6= 0 then the unique solution is A−1 0 = 0.22a) and (5. (5.Cowley@damtp.24a) a22 a22 Now. then (5. d02 = d2 − d1 .23a) to eliminate x2 by forming a032 (5. a023 = a23 − a13 . This section is about what can be deduced if det A = 0.4) and Theorem 5.1 Inhomogeneous and homogeneous problems If det A 6= 0 (and A is a 2 × 2 or a 3 × 3 matrix) then we know from (5. a22 there is no unique solution (if there is a solution at all). (5.25a) is said to be a system of inhomogeneous equations.

r2 · x = 0 and r3 · x = 0 all imply each other. r2 and r3 must then all be parallel. r3 } = R3 . (iii) Finally we need to consider the case when the dimension of span{r1 . (5. 3) . 2. Mathematical Tripos: IA Algebra & Geometry (Part I) 83 c S. all points in this line satisfy r3 · x = 0. The first two equations of (5. There are three possibilities: (i) the three planes intersect only at O. and thus the solution space. first we consider the case when it is 2.27b) the three planes each pass through the origin. (iii) the three planes coincide.e. solutions to (5.J. r3 } being equal to 2 or 1. µ ∈ R} . r2 . in which case  T r1  T A = r2  .2 Geometrical view of Ax = 0 For i = 1. r3 } is 1.ac.e. For the homogeneous equations (5. by {x ∈ R3 : x = λa + µb where λ. 2. r2 . The final condition r3 · x = 0 then implies that λ = 0 (since we have assumed that r1 · (r2 × r3 ) 6= 0). (ii) Next suppose that det A = 0. i.e. The solution space thus has dimension one. 22/03 We consider each of these cases in turn. (ii) the three planes have a common line (including O). Then as above the first two equations of (5. t = r1 × r2 }. Michaelmas 2006 .25b) may then be expressed as rT i x = ri · x = di (i = 1. 3 let ri be the vector with components equal to the elements of the ith row of A. x must lie on the line {x ∈ R3 : x = λt.28) The solution space thus has dimension two. respectively. hence span{r1 .27b) again imply that x must lie on the line {x ∈ R3 : x = λt. then we may specify the plane.27b) imply that x must lie on the intersection of the planes r1 · x = 0 and r2 · x = 0.27b) lie on a plane.cam. 2. r2 . t = r1 × r2 }.27a) T ri x = ri · x = 0 (i = 1. the solution for x is a line. The three row vectors r1 . i. O. the three planes intersect only at the origin. (i) If det A 6= 0 then r1 · (r2 × r3 ) 6= 0 and the set {r1 . and hence x = 0. i. In this case the set {r1 . r2 . r2 .26) rT 3 Equations (5. Assume wlog that r1 and r2 are the two linearly independent vectors. the solution of each set of 3 equations is the intersection of 3 planes. 5. It is not to be distributed to students. λ ∈ R. Thus the intersection of the three planes is a plane.5.25a) and (5. i.Cowley@damtp. The solution space thus has zero dimension. r3 } consists of three linearly independent vectors. (5. (5. Since each of these individual equations represents a plane in R3 .e. 3) .uk. Since r1 · (r2 × r3 ) = 0. r3 } is linearly dependent with the dimension of span{r1 . (5. If a and b are any two linearly independent vectors such that a · r1 = b · r1 = 0. λ ∈ R. This means that r1 · x = 0.27b) This is a supervisor’s copy of the notes. Hence the intersection of the three planes is a line.

w} to be the standard basis {e1 . (ii) If n(A) = 1 choose non-zero u ∈ K(A). r(A) = 2 (since the dimension of span{A(u). r3 } = number of linearly independent rows of A (row rank ). A(e2 ). A(v).uk.5. Since Au = Av = 0 and Aw 6= 0. then the image of A is spanned by {A(u). We do not need to consider the case n(A) = 3 as long as we exclude the map with A = 0.ac. and that the rank of A is two.29) K(A) is the solution space of Ax = 0. (b) In each case we also have from comparison of the results in this section with those in §5. A(w)} is a linearly independent set since {λA(u) + µA(v) + νA(w) = 0} ⇔ {A(λu + µv + νw) = 0} ⇔ {λu + µv + νw = 0} .5. for some α ∈ R. It follows that the rank of A is three. It is not to be distributed to students. such that x 7→ x0 = Ax. Remark. i. We claim that {A(v).2 on page 59) r(A) + n(A) = 3 . A(v).5. (5.3 Linear mapping view of Ax = 0 Consider the linear map A : R3 → R3 . v.Cowley@damtp. Michaelmas 2006 . u is a basis for the proper subspace K(A). (a) In each of cases (i). A(w)}. We now consider the three different possibilities for the value of the nullity in turn. From our earlier definition (4. Next recall that if {u. Next choose w 6∈ K(A) to extend this basis to form a basis of R3 . Suppose now we choose {u. Ae3 } = number of linearly independent columns of A (column rank ). We conclude that A(v) and A(w) are linearly independent. A(w)} = r(A) = 1 . w} is basis. the kernel of A is given by K(A) = {x ∈ R3 : Ax = 0} . it follows that dim span{A(u).2 correspond to (i) n(A) = 0.5. i. i. v. (i) If n(A) = 0 then {A(u).5). A(v). and so λ = µ = ν = 0 since {u. Remarks. e3 }. A(w)} is two).e. A(e3 )}. with a dimension denoted by n(A). To see this note that {µA(v) + νA(w) = 0} ⇔ {A(µv + νw) = 0} ⇔ {µv + νw = αu} . Next choose v. r(A) = dim span{r1 . (ii) n(A) = 1.e.19) that r(A) = dim span{Ae1 . (iii) If n(A) = 2 choose non-zero u. r2 . w} is a basis for R3 . Hence −αu + µv + νw = 0. r(A) = 3. Then the image of A is spanned by {A(e1 ). A(R3 ) = span{A(u). w 6∈ K(A) to extend this basis to form a basis of R3 (recall from Theorem 3. A(v). It follows from (4. A(v). Ae2 . e2 .e. v} is a basis for K(A). v ∈ K(A) such that the set {u. where A is the matrix of A with respect to the standard basis. w} is basis. and the number of linearly independent vectors in this set must equal r(A). The three cases studied in our geometric view of §5. Mathematical Tripos: IA Algebra & Geometry (Part I) 84 c S. v. and so α = µ = ν = 0 since {u.2. A(w)} .cam. A(w)} are linearly independent.14) and (4. (ii) and (iii) we have in accordance of the rank-nullity Theorem (see Theorem 4. (iii) n(A) = 2.J. This is a supervisor’s copy of the notes.5 on page 49 that this is always possible). v.

then det A = 0. 0 −1 22/06 where λ ∈ R. 1 −1 and so r(A) = 1 and n(A) = 1.Cowley@damtp. then y = 0 and the solution is unique. i.e. (5.uk. Whether there is a solution now depends on the value of b.31) 1 − a −a 1 b If a = 1.ac. If d ∈ I(A) then the general solution to Ax = d can be written as x = x0 + y where x0 This is a supervisor’s copy of the notes.J. from §5. (5. The latter case is described by Theorem 5. if a 6= 1 then det A 6= 0 and A−1 exists and is unique. • or d ∈ I(A). Example. It is not to be distributed to students. i. then y = λt and x = x0 + λt (representing a line). and thus A(x0 + y) = d + 0 = d . 0 The general solution is then x = x0 + y. and the unique solution is x = A .5. Mathematical Tripos: IA Algebra & Geometry (Part I) 85 c S.3. First we note that x = x0 + y is a solution since Ax0 = d and Ay = 0. Then • either d ∈ / I(A). where y is any vector in K(A). (iii) n(A) = 2 and r(A) = 1. b   1 • If b = 1 then ∈ I(A) and solutions exist (the equations are consistent).e. Theorem 5. is a particular fixed solution of Ax = d and y is the general solution of Ax = 0.   1 • If b 6= 1 then ∈ / I(A). A particular b solution is   1 x0 = .4 Implications for the solution of the inhomogeneous equation Ax = d If det A 6= 0 then r(A) = 3 and I(A) = R3 (where I(A) is notation for the image of A). in which case there is at least one solution and the equations are consistent. x = A−1 d exists and unique. (ii) n(A) = 1 and r(A) = 2. in which there are no solutions and the equations are inconsistent.     1 1 x= +λ .cam. there must exist x ∈ R3 for which d is the image under A.6 below.30) a 1 x2 b Since det A = (1 − a). then y = λa + µb and x = x0 + λa + µb (representing a plane). Further. If det A = 0 then r(A) < 3 and I(A) is a proper subspace of R3 .2 and §5. Michaelmas 2006 .5. Consider the (2 × 2) inhomogeneous case of Ax = d where      1 1 x1 1 = . Proof.5. Since d ∈ R3 . if (i) n(A) = 0 and r(A) = 3. x1 + x2 Hence     1 1 I(A) = span and K(A) = span . 5.6. and   x1 + x2 Ax = . and there are no solutions because the equations are inconsistent. Specifically     −1 1 1 −1 −1 1 A = .

1c) i.ac. (6. Hence a vector space over the complex numbers is a set V of elements. vector addition has an identity element.1a) A(ii) addition is commutative. i. Inter alia they are important in quantum mechanics. for all x. such that for all x ∈ V x + 0 = x. y ∈ V x + y = y + x. for all λ. (6.J. 6. y ∈ V λ(x + y) = λx + λy .1 Vector Spaces Over The Complex Numbers We have considered vector spaces with real scalars. We now generalise to vector spaces with complex This is a supervisor’s copy of the notes. (6. 6.1 Definition We just adapt the definition for a vector space over the real numbers by exchanging ‘real’ for ‘complex’ and ‘R’ for ‘C’.e. and allow us to solve more problems. i. It is not to be distributed to students. A(iv) for all x ∈ V there exists an additive negative or inverse vector x0 ∈ V such that x + x0 = 0 . (6. i. (6. satisfying the following eight axioms or rules: A(i) addition is associative. i. y ∈ V by x + y. i. i.1g) B(viii) scalar multiplication is distributive over scalar addition. µ ∈ C and x ∈ V λ(µx) = (λµ)x . for all x ∈ V 1 x = x.6 Complex Vector Spaces 6.e. (6. together with two binary operations • vector addition denoted for x. (6. µ ∈ C and x ∈ V (λ + µ)x = λx + µx .1e) B(vi) scalar multiplication has an identity element.1.e.1f) where 1 is the multiplicative identity in C.cam.uk.e.Cowley@damtp. z ∈ V x + (y + z) = (x + y) + z . where x + y ∈ V so that there is closure under vector addition.1h) Mathematical Tripos: IA Algebra & Geometry (Part I) 86 c S. for all x.0 Why Study This? In the same way that complex numbers are a natural extension of real numbers. (6. for all λ ∈ C and x.1b) A(iii) there exists an element 0 ∈ V .e. or ‘vectors’.e. Michaelmas 2006 . and allow us to solve more problems.1d) B(v) scalar multiplication of vectors is ‘associative’.e. y. scalars. for all λ. where ax ∈ V so that there is closure under scalar multiplication. called the null or zero vector. complex vector spaces are a natural extension of real numbers. B(vii) scalar multiplication is distributive over vector addition. • scalar multiplication denoted for a ∈ C and x ∈ V by ax.

vn ). (6. . For λ ∈ C and complex vectors u. . (ii) the additive inverse of a vector is unique. . zn ) of complex numbers zi ∈ C. . . en = (0. . A(ii). 0. (ii) The standard basis for Rn .2. (i) the zero vector 0 is unique. i = 1. .ac.Cowley@damtp. (6. . since Rn is not closed under multiplication by an arbitrary complex number. we have to be careful with scalar products.2b) Lemma 6.1. . Michaelmas 2006 . 6. . .cam. .3b) i=1 6.J. (6.3a) also serves as a standard basis for Cn .g. . 6.1. define Cn to be the set of n-tuples (z1 . i. . A(iii).1 are satisfied. A(iv). where u = (u1 . It is not to be distributed to students. 2 Remarks.uk. . . z2 . . un + vn ) ∈ Cn . . v ∈ V by hu. define vector addition and scalar multiplication by u + v = (u1 + v1 .1 Definition Let V be a n-dimensional vector space over the complex numbers. . e1 = (1. For fixed positive integer n. B(vii) and B(viii) of §6. vi ∈ C. . 1). 0. This is a supervisor’s copy of the notes. but it is not a subspace of the vector space Cn . . . . . (6. .2a) λu = (λu1 . . . . 0). Check that A(i).e. or inner product.2 Scalar Products for Complex Vector Spaces In generalising from real vector spaces to complex vector spaces. . . B(v). . Further we can express any z ∈ Cn in terms of components as n X z = (z1 .2 Properties Most of the properties and theorems of §3 follow through much as before. (6. (−1)x = −x and λ0 = 0. (iii) if x ∈ V and λ ∈ C then 0x = 0. .1. Hence Cn has dimension n when viewed as a vector space over C. e. . We will denote a scalar product. Cn is a vector space over C Proof by exercise.6. B(vi). λun ) ∈ Cn . . . . zn ) = zi e i .1. (iv) Theorem 3. v ∈ Cn .1 on when a subset U of a vector space V is a subspace of V still applies if we exchange ‘real’ for ‘complex’ and ‘R’ for ‘C’. . n. of the ordered pair of vectors u.4) Mathematical Tripos: IA Algebra & Geometry (Part I) 87 c S. .3 Cn Definition. un ) and v = (v1 . (i) Rn is a subset of Cn . . .

Properties (6.2. (i) Conjugate symmetry. (iv) Non-degeneracy. We can again show that hu.3 Scalar Products in Terms of Components Suppose that we have a scalar product defined on a complex vector space with a given basis {ei }.ac. Further.5b) i. Michaelmas 2006 . (6. i = 1.5a) and (6.where alternative notations are u · v and h u | v i. hv. . 0i = h0.5c) (iii) Non-negativity.6) Anti-linearity in the first argument. (λv1 + µv2 ) i = λ h u . . for any two vectors n X n X v= vi ei and w = wj ej . where the real positive number kvk is the norm of the vector v. ej i (i. v i| 6 kuk kvk . h u .e. v1 i + µ h u . vi = hv. v2 i . then (6. a scalar product of a vector with itself should be positive. vi = hv. for λ.e. v i = h v .e. ui . i. . if we let u = v. (6. v i . . v i = kvk . (6. µ ∈ C This is a supervisor’s copy of the notes. .uk. ui = 0. vi > 0. 6.8a) ku + vk 6 kuk + kvk .e. vi . (6.Cowley@damtp. µ ∈ C ∗ h (λu1 + µu2 ) . v i + µ∗ h u2 . (ii) Linearity in the second argument. i.e. To see this first define the complex numbers Gij by Gij = h ei .cam. (6.2. We claim that the scalar product is determined for all pairs of vectors by its values for all pairs of basis vectors. i.e. u2 i = λ∗ h u1 . i.5d) 2 This allows us to write h v . (6. (6. v i is real. h v .5a) ∗ ∗ where a is an alternative notation for a complex conjugate (we shall swap between ¯ and freely). It is not to be distributed to students. (λu1 + µu2 ) i ∗ ∗ = λ∗ h v . (6. i.9) Then. It is again true that |h u .e.2 Properties Scalar product with 0. i.8b) with equality only when u is a scalar multiple of v. 2 kvk ≡ h v . n.5e) 6. v i = 0 ⇒ v = 0. A scalar product of a vector space over the complex numbers must have the following properties. . (6.10) i=1 j=1 Mathematical Tripos: IA Algebra & Geometry (Part I) 88 c S. ∗ hu. . j = 1. u1 i + µ∗ h v . the only vector of zero norm should be the zero vector. Implicit in this equation is the assumption that for a complex vector space the ordering of the vectors in the scalar product is important (whereas for Rn this is not important).J.7) Schwarz’s inequality and the triangle inequality.e. for λ. (6. (6.5c) imply so-called ‘anti-linearity’ in the first argument. i.5a) implies that ∗ hv. . n) .

.15) where G is the matrix. are such that Gij = h ei . wj ej i=1 j=1 n X X n = vi∗ wj h ei . with entries Gij . i.J. as before. Definition. (6.11) i=1 j=1 We can simplify this expression (which determines the scalar product in terms of the Gij ).13a) Also. It is not to be distributed to students. (6.11) can be written as h v . denotes a transpose.5a). Confirm that the scalar product given by (6.16b) satisfies the required properties of a scalar product. v† is the row matrix v† ≡ (v∗ )T = v1∗ v2∗ . (6. This is a supervisor’s copy of the notes. wi = vi∗ wi . . namely (6. w i = v† G w .5c) and (6. Hermitian conjugates. (6. from (6. . T∗ A†† = A∗T = A.5d) and (6.14b) Then in terms of this notation the scalar product (6. but first it helps to have a definition. and so (AB)† = B† A† .16b) i=1 Exercise.we have that * n n + X  X  hv. (6.   w1  w2  w= .cam. ej i = δij .uk.12). (6.e.5e).Cowley@damtp. (6. a∗11 a∗21     a11 a12 If A = then A† = . reduces to (cf. Remark. ej i from (6. (6. is defined to be A† = (AT )∗ = (A∗ )T .e. a21 a22 a∗12 a∗22 Properties. or metric.ac.22)) n X hv. i. wi = vi ei . Example. Mathematical Tripos: IA Algebra & Geometry (Part I) 89 c S.. If the {ei } form an orthonormal basis. Hence (AB)T∗ = BT∗ AT∗ .12) T where.5c). Michaelmas 2006 . (6.16a) then (6. or equivalently (6.11).  (6.7) i=1 j=1 Xn X n = vi∗ Gij wj .15). where aij ∈ C. vn∗ . (6. For matrices A and B recall that (AB)T = BT AT .13b) Let w be the column matrix of components. (3. The Hermitian conjugate or conjugate transpose or adjoint of a matrix A = {aij }.  wn and let v† be the adjoint of the column matrix v.14a)    .

ac.6. As before this defines a matrix. We have observed that the standard basis. y y0 sin θ cos θ y y Since diagonal matrices have desirable properties (e.e.g. If N is a real linear transformation so that N is real. {ei }. Extend T to a map Tb : Cn → Cm .cam. Example. N will in general be a complex (m × n) matrix. det A −a21 a11 a11 sin θ + a21 cos θ a12 sin θ + a22 cos θ and so 1 (A−1 RA)12 = (a22 (a12 cos θ − a22 sin θ) − a12 (a12 sin θ + a22 cos θ)) det A sin θ 2 a12 + a222 .  = − det A and Mathematical Tripos: IA Algebra & Geometry (Part I) 90 c S. Let {ei } be a basis of Cn and {fi } be a basis of Cm . i Further real components transform as before. N = {Nij }.18) to obtain    1 a22 −a12 a11 cos θ − a21 sin θ a12 cos θ − a22 sin θ A−1 RA = . they are straightforward to invert) we might ask whether there is a change of basis under which e = A−1 RA R (6. Consider a linear map T : Rn → Rm . i=1 where Nij ∈ C.uk. This is a supervisor’s copy of the notes. with respect to bases {ei } of Cn and {fi } of Cm . e = C−1 NA . Change of bases. it is not necessarily true that N e is real. Example. this will not be the case if we transform from standard bases to bases consisting of complex vectors.70) follows through for a linear map N : Cn → Cm . If ei → e0i . Maps such as Tb are referred to as real linear transformations of Cn → Cm . Consider the map R : R2 → R2 consisting of a rotation by θl. It is not to be distributed to students.23)    0      x x cos θ − sin θ x x 7→ = = R(θ) .Cowley@damtp. Then under N .3 Linear Maps Similarly. where Tb has the same effect as T on real vectors. from (4. is both a basis for Rn and Cn . then the components of v with respect to the standard basis transform to components of v0 with respect to the standard basis according to v0 = Tv . Xm ej → e0j = N ej = Nij fi .17) Remark. One way (but emphatically not the best way) to proceed would be to [partially] expand out the right-hand side of (6.g.J. T is a real matrix. Thus if v ∈ Cn . Under a change of bases {ei } to {e ei } and {fi } to {e fi } the transformation law (4. i. then as before b ij = (e0 )j (T) and hence T b = T.18) is a diagonal matrix. Real linear transformations. e. Michaelmas 2006 . we can extend the theory of §4 on linear maps and matrices to vector spaces over complex numbers. N (6. Consider the linear map N : Cn → Cm . and let T be the associated matrix with respect to the standard bases of both domain and range. but complex components are now also allowed.

(6.61) it follows that     1 1 1 −i e1 = √ . as can all normal matrices. ei . As mentioned above. i. h e ej i = δij . as required. as well as Hermitian matrices. (6.cam. For the time being we state that Hermitian matrices can always be diagonalized. where we note from using (6. RRT = RT R = I.12) ∗ (G† )ij = G∗ji = h ej .22b) Definition. from (6. It is not to be distributed to students. matrices such that A† A = A A† . i. diagonal matrices have desirable properties. e i. ei i = h ei .59b) that real rotation matrices are orthogonal. Michaelmas 2006 . Definition. e he e2 i = 1 . if A† = A . A complex square matrix U is said to be unitary if its Hermitian conjugate is equal to its inverse. ej i = (G)ij . (6. it is not true that R eReT = R eTR e = I. e he e2 i = 0 .  = det A Hence R e is a diagonal matrix if a12 = ±ia22 and a21 = ±ia11 . matrices such that A† = −A) and unitary matrices. e2 = √ .21) 0 e−iθ Remark.19) 2 −i 1 Thence from (4. A convenient normalisation (that results in an orthonormal basis) is to choose   1 1 −i A= √ . Mathematical Tripos: IA Algebra & Geometry (Part I) 91 c S. e1 . It is therefore useful to know what classes of matrices can be diagonalized by a change of basis.20a) −i 1 e e 2 2 This is a supervisor’s copy of the notes.3).20b) Moreover.uk. (6. e (6. Instead we note from (6. . (6.9) and (6. however.e. (6.e. e1 .16b) that he e1 i = 1 .e. You will learn more about this in in the second part of this course (and elsewhere). (6.5a). there is an equivalent for complex matrices to symmetric matrices for real matrices. i.23) Unitary matrices are to complex matrices what orthogonal matrices are to real matrices.25) Remark. i.19) it follows that.e. . if U† = U−1 . The metric G is Hermitian since from (6. e† = R R e −1 . a class that includes skew-symmetric Hermitian matrices (i.21) that R eRe† = R e†R e = I.e.Cowley@damtp.24) Example. e2 . (6.ac.18) and (6.e.J.e. A complex square matrix A is said to be Hermitian if it is equal to its own Hermitian conjugate. We know from (4. Similarly. (6. sin θ 2 (A−1 RA)21 a11 + a221 .22a) i. R e is diagonal:     e = 1 1 R i cos θ − sin θ 1 −i 2 i 1 sin θ cos θ −i 1  iθ  e 0 = . (6.