Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
Apostol,
Calculus, vol. 1 assigned to doctoral students in
years 20022003
andrea battinelli
dipartimento di scienze matematiche e informatiche “R.Magari”
dell’università di Siena
via del Capitano 15  53100 Siena
tel: +390577233769/02 fax: /01/30
email: battinelli @unisi.it
web: http//www.batman vai li
December 12, 2005
2
Contents
I Volume 1 1
1 Chapter 1 3
2 Chapter 2 5
3 Chapter 3 7
4 Chapter 4 9
5 Chapter 5 11
6 Chapter 6 13
7 Chapter 7 15
8 Chapter 8 17
9 Chapter 9 19
10 Chapter 10 21
11 Chapter 11 23
12 Vector algebra 25
12.1 Historical introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 25
12.2 The vector space of ntuples of real numbers . . . . . . . . . . . . . . 25
12.3 Geometric interpretation for n ≤ 3 . . . . . . . . . . . . . . . . . . . 25
12.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
12.4.1 n. 1 (p. 450) . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
12.4.2 n. 2 (p. 450) . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
12.4.3 n. 3 (p. 450) . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
12.4.4 n. 4 (p. 450) . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
12.4.5 n. 5 (p. 450) . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
12.4.6 n. 6 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4 CONTENTS
12.4.7 n. 7 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
12.4.8 n. 8 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
12.4.9 n. 9 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
12.4.10 n. 10 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . 30
12.4.11 n. 11 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . 30
12.4.12 n. 12 (p. 451) . . . . . . . . . . . . . . . . . . . . . . . . . . 32
12.5 The dot product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
12.6 Length or norm of a vector . . . . . . . . . . . . . . . . . . . . . . . . 33
12.7 Orthogonality of vectors . . . . . . . . . . . . . . . . . . . . . . . . . 33
12.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
12.8.1 n. 1 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
12.8.2 n. 2 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
12.8.3 n. 3 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
12.8.4 n. 5 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
12.8.5 n. 6 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
12.8.6 n. 7 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
12.8.7 n. 10 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 36
12.8.8 n. 13 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 36
12.8.9 n. 14 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 37
12.8.10 n. 15 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 39
12.8.11 n. 16 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 39
12.8.12 n. 17 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 39
12.8.13 n. 19 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 40
12.8.14 n. 20 (p. 456) . . . . . . . . . . . . . . . . . . . . . . . . . . 40
12.8.15 n. 21 (p. 457) . . . . . . . . . . . . . . . . . . . . . . . . . . 40
12.8.16 n. 22 (p. 457) . . . . . . . . . . . . . . . . . . . . . . . . . . 41
12.8.17 n. 24 (p. 457) . . . . . . . . . . . . . . . . . . . . . . . . . . 42
12.8.18 n. 25 (p. 457) . . . . . . . . . . . . . . . . . . . . . . . . . . 42
12.9 Projections. Angle between vectors in nspace . . . . . . . . . . . . . 43
12.10 The unit coordinate vectors . . . . . . . . . . . . . . . . . . . . . . . 43
12.11 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
12.11.1 n. 1 (p. 460) . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
12.11.2 n. 2 (p. 460) . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
12.11.3 n. 3 (p. 460) . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
12.11.4 n. 5 (p. 460) . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
12.11.5 n. 6 (p. 460) . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
12.11.6 n. 8 (p. 460) . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
12.11.7 n. 10 (p. 461) . . . . . . . . . . . . . . . . . . . . . . . . . . 46
12.11.8 n. 11 (p. 461) . . . . . . . . . . . . . . . . . . . . . . . . . . 47
12.11.9 n. 13 (p. 461) . . . . . . . . . . . . . . . . . . . . . . . . . . 48
12.11.10 n. 17 (p. 461) . . . . . . . . . . . . . . . . . . . . . . . . . . 48
12.12 The linear span of a ﬁnite set of vectors . . . . . . . . . . . . . . . . 50
CONTENTS 5
12.13 Linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
12.14 Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
12.15 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
12.15.1 n. 1 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
12.15.2 n. 3 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
12.15.3 n. 5 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
12.15.4 n. 6 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
12.15.5 n. 7 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
12.15.6 n. 8 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
12.15.7 n. 10 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . 52
12.15.8 n. 12 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . 53
12.15.9 n. 13 (p. 467) . . . . . . . . . . . . . . . . . . . . . . . . . . 55
12.15.10 n. 14 (p. 468) . . . . . . . . . . . . . . . . . . . . . . . . . . 56
12.15.11 n. 15 (p. 468) . . . . . . . . . . . . . . . . . . . . . . . . . . 56
12.15.12 n. 17 (p. 468) . . . . . . . . . . . . . . . . . . . . . . . . . . 56
12.15.13 n. 18 (p. 468) . . . . . . . . . . . . . . . . . . . . . . . . . . 57
12.15.14 n. 19 (p. 468) . . . . . . . . . . . . . . . . . . . . . . . . . . 57
12.15.15 n. 20 (p. 468) . . . . . . . . . . . . . . . . . . . . . . . . . . 58
12.16 The vector space V
n
(C) of ntuples of complex numbers . . . . . . . 59
12.17 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
13 Applications of vector algebra to analytic geometry 61
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.2 Lines in nspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.3 Some simple properties of straight lines . . . . . . . . . . . . . . . . . 61
13.4 Lines and vectorvalued functions . . . . . . . . . . . . . . . . . . . . 61
13.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.5.1 n. 1 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.5.2 n. 2 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
13.5.3 n. 3 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.5.4 n. 4 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.5.5 n. 5 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
13.5.6 n. 6 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
13.5.7 n. 7 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
13.5.8 n. 8 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
13.5.9 n. 9 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
13.5.10 n. 10 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . 65
13.5.11 n. 11 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . 66
13.5.12 n. 12 (p. 477) . . . . . . . . . . . . . . . . . . . . . . . . . . 67
13.6 Planes in euclidean nspaces . . . . . . . . . . . . . . . . . . . . . . . 67
13.7 Planes and vectorvalued functions . . . . . . . . . . . . . . . . . . . 67
13.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6 CONTENTS
13.8.1 n. 2 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
13.8.2 n. 3 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
13.8.3 n. 4 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
13.8.4 n. 5 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
13.8.5 n. 6 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
13.8.6 n. 7 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
13.8.7 n. 8 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
13.8.8 n. 9 (p. 482) . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
13.8.9 n. 10 (p. 483) . . . . . . . . . . . . . . . . . . . . . . . . . . 73
13.8.10 n. 11 (p. 483) . . . . . . . . . . . . . . . . . . . . . . . . . . 74
13.8.11 n. 12 (p. 483) . . . . . . . . . . . . . . . . . . . . . . . . . . 75
13.8.12 n. 13 (p. 483) . . . . . . . . . . . . . . . . . . . . . . . . . . 75
13.8.13 n. 14 (p. 483) . . . . . . . . . . . . . . . . . . . . . . . . . . 75
13.9 The cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.10 The cross product expressed as a determinant . . . . . . . . . . . . . 76
13.11 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.11.1 n. 1 (p. 487) . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.11.2 n. 2 (p. 487) . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.11.3 n. 3 (p. 487) . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.11.4 n. 4 (p. 487) . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
13.11.5 n. 5 (p. 487) . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
13.11.6 n. 6 (p. 487) . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
13.11.7 n. 7 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
13.11.8 n. 8 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
13.11.9 n. 9 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
13.11.10 n. 10 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . 80
13.11.11 n. 11 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . 80
13.11.12 n. 12 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . 82
13.11.13 n. 13 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . 82
13.11.14 n. 14 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . 83
13.11.15 n. 15 (p. 488) . . . . . . . . . . . . . . . . . . . . . . . . . . 84
13.12 The scalar triple product . . . . . . . . . . . . . . . . . . . . . . . . 85
13.13 Cramer’s rule for solving systems of three linear equations . . . . . . 85
13.14 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
13.15 Normal vectors to planes . . . . . . . . . . . . . . . . . . . . . . . . 85
13.16 Linear cartesian equations for planes . . . . . . . . . . . . . . . . . . 85
13.17 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.17.1 n. 1 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.17.2 n. 2 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.17.3 n. 3 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
13.17.4 n. 4 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
13.17.5 n. 5 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
CONTENTS 7
13.17.6 n. 6 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
13.17.7 n. 8 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
13.17.8 n. 9 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
13.17.9 n. 10 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . 89
13.17.10 n. 11 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . 90
13.17.11 n. 13 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . 90
13.17.12 n. 14 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . 90
13.17.13 n. 15 (p. 496) . . . . . . . . . . . . . . . . . . . . . . . . . . 90
13.17.14 n. 17 (p. 497) . . . . . . . . . . . . . . . . . . . . . . . . . . 91
13.17.15 n. 20 (p. 497) . . . . . . . . . . . . . . . . . . . . . . . . . . 91
13.18 The conic sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
13.19 Eccentricity of conic sections . . . . . . . . . . . . . . . . . . . . . . 91
13.20 Polar equations for conic sections . . . . . . . . . . . . . . . . . . . . 91
13.21 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
13.22 Conic sections symmetric about the origin . . . . . . . . . . . . . . . 92
13.23 Cartesian equations for the conic sections . . . . . . . . . . . . . . . 92
13.24 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
13.25 Miscellaneous exercises on conic sections . . . . . . . . . . . . . . . . 92
14 Calculus of vectorvalued functions 93
15 Linear spaces 95
15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
15.2 The deﬁnition of a linear space . . . . . . . . . . . . . . . . . . . . . 95
15.3 Examples of linear spaces . . . . . . . . . . . . . . . . . . . . . . . . 95
15.4 Elementary consequences of the axioms . . . . . . . . . . . . . . . . . 95
15.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
15.5.1 n. 1 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
15.5.2 n. 2 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
15.5.3 n. 3 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
15.5.4 n. 4 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
15.5.5 n. 5 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
15.5.6 n. 6 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
15.5.7 n. 7 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
15.5.8 n. 11 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 98
15.5.9 n. 13 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 98
15.5.10 n. 14 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 99
15.5.11 n. 16 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 99
15.5.12 n. 17 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 101
15.5.13 n. 18 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 101
15.5.14 n. 19 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 102
15.5.15 n. 22 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8 CONTENTS
15.5.16 n. 23 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 102
15.5.17 n. 24 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 102
15.5.18 n. 25 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 102
15.5.19 n. 26 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 102
15.5.20 n. 27 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 103
15.5.21 n. 28 (p. 555) . . . . . . . . . . . . . . . . . . . . . . . . . . 103
15.6 Subspaces of a linear space . . . . . . . . . . . . . . . . . . . . . . . . 103
15.7 Dependent and independent sets in a linear space . . . . . . . . . . . 103
15.8 Bases and dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
15.9 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
15.9.1 n. 1 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
15.9.2 n. 2 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
15.9.3 n. 3 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
15.9.4 n. 4 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
15.9.5 n. 5 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
15.9.6 n. 6 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
15.9.7 n. 7 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
15.9.8 n. 8 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
15.9.9 n. 9 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
15.9.10 n. 10 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 105
15.9.11 n. 11 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 105
15.9.12 n. 12 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 106
15.9.13 n. 13 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 106
15.9.14 n. 14 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 106
15.9.15 n. 15 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 107
15.9.16 n. 16 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 107
15.9.17 n. 22 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 108
15.9.18 n. 23 (p. 560) . . . . . . . . . . . . . . . . . . . . . . . . . . 108
15.10 Inner products. Euclidean spaces. Norms . . . . . . . . . . . . . . . 111
15.11 Orthogonality in a euclidean space . . . . . . . . . . . . . . . . . . . 111
15.12 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
15.12.1 n. 9 (p. 567) . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
15.12.2 n. 11 (p. 567) . . . . . . . . . . . . . . . . . . . . . . . . . . 112
15.13 Construction of orthogonal sets. The GramSchmidt process . . . . . 115
15.14 Orthogonal complements. projections . . . . . . . . . . . . . . . . . 115
15.15 Best approximation of elements in a euclidean space by elements in a
ﬁnitedimensional subspace . . . . . . . . . . . . . . . . . . . . . . . . 115
15.16 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
15.16.1 n. 1 (p. 576) . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
15.16.2 n. 2 (p. 576) . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
15.16.3 n. 3 (p. 576) . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
15.16.4 n. 4 (p. 576) . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
CONTENTS 9
16 Linear transformations and matrices 121
16.1 Linear transformations . . . . . . . . . . . . . . . . . . . . . . . . . . 121
16.2 Null space and range . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
16.3 Nullity and rank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
16.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
16.4.1 n. 1 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
16.4.2 n. 2 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
16.4.3 n. 3 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
16.4.4 n. 4 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
16.4.5 n. 5 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
16.4.6 n. 6 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
16.4.7 n. 7 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
16.4.8 n. 8 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
16.4.9 n. 9 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
16.4.10 n. 10 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . 124
16.4.11 n. 16 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . 125
16.4.12 n. 17 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . 125
16.4.13 n. 23 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . 126
16.4.14 n. 25 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . 126
16.4.15 n. 27 (p. 582) . . . . . . . . . . . . . . . . . . . . . . . . . . 127
16.5 Algebraic operations on linear transformations . . . . . . . . . . . . . 128
16.6 Inverses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
16.7 Onetoone linear transformations . . . . . . . . . . . . . . . . . . . . 128
16.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
16.8.1 n. 15 (p. 589) . . . . . . . . . . . . . . . . . . . . . . . . . . 128
16.8.2 n. 16 (p. 589) . . . . . . . . . . . . . . . . . . . . . . . . . . 128
16.8.3 n. 17 (p. 589) . . . . . . . . . . . . . . . . . . . . . . . . . . 128
16.8.4 n. 27 (p. 590) . . . . . . . . . . . . . . . . . . . . . . . . . . 129
16.9 Linear transformations with prescribed values . . . . . . . . . . . . . 129
16.10 Matrix representations of linear transformations . . . . . . . . . . . . 129
16.11 Construction of a matrix representation in diagonal form . . . . . . . 129
16.12 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
16.12.1 n. 3 (p. 596) . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
16.12.2 n. 4 (p. 596) . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
16.12.3 n. 5 (p. 596) . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
16.12.4 n. 7 (p. 597) . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
16.12.5 n. 8 (p. 597) . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
16.12.6 n. 16 (p. 597) . . . . . . . . . . . . . . . . . . . . . . . . . . 135
10 CONTENTS
Part I
Volume 1
1
Chapter 1
CHAPTER 1
4 Chapter 1
Chapter 2
CHAPTER 2
6 Chapter 2
Chapter 3
CHAPTER 3
8 Chapter 3
Chapter 4
CHAPTER 4
10 Chapter 4
Chapter 5
CHAPTER 5
12 Chapter 5
Chapter 6
CHAPTER 6
14 Chapter 6
Chapter 7
CHAPTER 7
16 Chapter 7
Chapter 8
CHAPTER 8
18 Chapter 8
Chapter 9
CHAPTER 9
20 Chapter 9
Chapter 10
CHAPTER 10
22 Chapter 10
Chapter 11
CHAPTER 11
24 Chapter 11
Chapter 12
VECTOR ALGEBRA
12.1 Historical introduction
12.2 The vector space of ntuples of real numbers
12.3 Geometric interpretation for n ≤ 3
12.4 Exercises
12.4.1 n. 1 (p. 450)
(a) a +b = (5, 0, 9).
(b) a −b = (−3, 6, 3).
(c) a +b −c = (3, −1, 4).
(d) 7a −2b −3c = (−7, 24, 21).
(e) 2a +b −3c = (0, 0, 0).
12.4.2 n. 2 (p. 450)
The seven points to be drawn are the following:
µ
7
3
, 2
¶
,
µ
5
2
,
5
2
¶
,
µ
11
4
,
13
4
¶
, (3, 4) , (4, 7) , (1, −2) , (0, −5)
The purpose of the exercise is achieved by drawing, as required, a single picture,
containing all the points (included the starting points A and B, I would say).
26 Vector algebra
It can be intuitively seen that, by letting t vary in all R, the straight line through
point A with direction given by the vector b ≡
−→
OB is obtained.
12.4.3 n. 3 (p. 450)
The seven points this time are the following:
µ
5
3
,
10
3
¶
,
µ
2,
7
2
¶
,
µ
5
2
,
15
4
¶
, (3, 4) , (5, 5) , (−1, 2) , (−3, 1)
3
2
1
0
1
2
3
4
5
3 2 1 1 2 3 4 5
It can be intuitively seen that, by letting t vary in all R, the straight line through B
with direction given by the vector a ≡
−→
OA is obtained.
Exercises 27
12.4.4 n. 4 (p. 450)
(a) The seven points to be drawn are the following:
µ
3
2
, 2
¶
,
µ
5
4
,
5
2
¶
,
µ
4
3
,
7
3
¶
, (3, −1) , (4, −3) ,
µ
1
2
, 4
¶
, (0, 5)
The whole purpose of this part of the exercise is achieved by drawing a single picture,
containing all the points (included the starting points A and B, I would say). This
is made clear, it seems to me, by the question immediately following.
4
2
0
2
4
4 2 2 4
(b) It is hard not to notice that all points belong to the same straight line; indeed,
as it is going to be clear after the second lecture, all the combinations are aﬃne.
(c) If the value of x is ﬁxed at 0 and y varies in [0, 1], the segment OB is obtained;
the same construction with the value of x ﬁxed at 1 yields the segment AD, where
−→
OD =
−→
OA+
−→
OB, and hence D is the vertex of the parallelogram with three vertices
at O, A, and B. Similarly, when x =
1
2
the segment obtained joins the midpoints of
the two sides OA and BD; and it is enough to repeat the construction a few more
times to convince oneself that the set
©
xa +yb : (x, y) ∈ [0, 1]
2
ª
is matched by the set of all points of the parallelogram OADB. The picture below
is made with the value of x ﬁxed at 0,
1
4
,
1
2
,
3
4
, 1,
5
4
,
3
2
, and 2.
28 Vector algebra
0
1
2
3
4
0.5 1 1.5 2 2.5 3
(d) All the segments in the above construction are substituted by straight lines, and
the resulting set is the (inﬁnite) stripe bounded by the lines containing the sides OB
and AD, of equation 3x −y = 0 and 3x −y = 5 respectively.
(e) The whole plane.
12.4.5 n. 5 (p. 450)
x
µ
2
1
¶
+y
µ
1
3
¶
=
µ
c
1
c
2
¶
I 2x +y = c
1
II x + 3y = c
2
3I −II 5x = 3c
1
−c
2
2II −I 5y = 2c
2
−c
1
µ
c
1
c
2
¶
=
3c
1
−c
2
5
µ
2
1
¶
+
2c
2
−c
1
5
µ
1
3
¶
12.4.6 n. 6 (p. 451)
(a)
d = x
¸
1
1
1
¸
+y
¸
0
1
1
¸
+z
¸
1
1
0
¸
=
¸
x +z
x +y +z
x +y
¸
(b)
I x +z = 0
II x +y +z = 0
III x +y = 0
I x = −z
(↑) ,→II y = 0
(↑) ,→III x = 0
(↑) ,→I z = 0
(c)
I x +z = 1
II x +y +z = 2
III x +y = 3
II −I y = 1
II −III z = −1
(↑) ,→I x = 2
Exercises 29
12.4.7 n. 7 (p. 451)
(a)
d = x
¸
1
1
1
¸
+y
¸
0
1
1
¸
+z
¸
2
1
1
¸
=
¸
x + 2z
x +y +z
x +y +z
¸
(b)
I x + 2z = 0
II x +y +z = 0
III x +y +z = 0
I x = −2z
(↑) ,→II y = z
z ←1 (−2, 1, 1)
(c)
I x + 2z = 1
II x +y +z = 2
III x +y +z = 3
III −II 0 = 1
12.4.8 n. 8 (p. 451)
(a)
d = x
¸
¸
¸
1
1
1
0
¸
+y
¸
¸
¸
0
1
1
1
¸
+z
¸
¸
¸
1
1
0
0
¸
=
¸
¸
¸
x +z
x +y +z
x +y
y
¸
(b)
I x +z = 0
II x +y +z = 0
III x +y = 0
IV y = 0
IV y = 0
IV ,→III x = 0
(↑) ,→I z = 0
II (check) 0 = 0
(c)
I x +z = 1
II x +y +z = 5
III x +y = 3
IV y = 4
IV y = 4
IV ,→III x = −1
(↑) ,→I z = 2
II (check) −1 + 4 + 2 = 5
(d)
I x +z = 1
II x +y +z = 2
III x +y = 3
IV y = 4
II −I −IV 0 = −3
30 Vector algebra
12.4.9 n. 9 (p. 451)
Let the two vectors u and v be both parallel to the vector w. According to the
deﬁnition at page 450 (just before the beginning of § 12.4), this means that there are
two real nonzero numbers α and β such that u = αw and v = βw. Then
u = αw = α
µ
v
β
¶
=
α
β
v
that is, u is parallel to v.
12.4.10 n. 10 (p. 451)
Assumptions:
I c = a +b
II ∃k ∈ R ∼ {0} , a = kd
Claim:
(∃h ∈ R ∼ {0} , c = hd) ⇔(∃l ∈ R ∼ {0} , b = ld)
I present two possible lines of reasoning (among others, probably). If you look
carefully, they diﬀer only in the phrasing.
1.
∃h ∈ R ∼ {0} , c = hd
⇔
I
∃h ∈ R ∼ {0} , a +b = hd
⇔
II
∃h ∈ R ∼ {0} , ∃k ∈ R ∼ {0} , kd +b = hd
⇔ ∃h ∈ R ∼ {0} , ∃k ∈ R ∼ {0} , b = (h −k) d
(b 6= 0) h 6= k
⇔
l≡h−k
∃l ∈ R ∼ {0} , b = ld
2. (⇒) Since (by I) we have b = c −a, if c is parallel to d, then (by II) b is the
diﬀerence of two vectors which are both parallel to d; it follows that b, which
is nonnull, is parallel to d too.
(⇐) Since (by I) we have c = a + b, if b is parallel to d, then (by II) c is
the sum of two vectors which are both parallel to d; it follows that c, which is
nonnull, is parallel to d too.
12.4.11 n. 11 (p. 451)
(b) Here is an illustration of the ﬁrst distributive law
(α +β) v = αv +βv
Exercises 31
with v = (2, 1), α = 2, β = 3. The vectors v, αv, βv, αv + βv are displayed by
means of representative oriented segments from left to right, in black, red, blue, and
redblue colour, respectively. The oriented segment representing vector (α +β) v is
above, in violet. The dotted lines are there just to make it clearer that the two
oriented segments representing αv + βv and (α +β) v are congruent, that is, the
vectors αv +βv and (α +β) v are the same.
0
2
4
6
8
2 4 6 8 10 12 14
tails are marked with a cross, heads with a diamond
An illustration of the second distributive law
α(u +v) = αu +αv
is provided by means of the vectors u = (1, 3), v = (2, 1), and the scalar α = 2. The
vectors u and αu are represented by means of blue oriented segments; the vectors
v and αv by means of red ones; u + v and α(u +v) by green ones; αu + αv is in
violet. The original vectors u, v, u + v are on the left; the “rescaled” vectors αu,
αv, α(u +v) on the right. Again, the black dotted lines are there just to emphasize
congruence.
32 Vector algebra
0
2
4
6
8
4 2 2 4 6 8
tails are marked with a cross, heads with a diamond
12.4.12 n. 12 (p. 451)
The statement to be proved is better understood if written in homogeneous form,
with all vectors represented by oriented segments:
−→
OA+
1
2
−→
AC =
1
2
−→
OB (12.1)
Since A and C are opposed vertices, the same holds for O and B; this means that
the oriented segment OB covers precisely one diagonal of the parallelogram OABC,
and AC precisely covers the other (in the rightwarddownwards orientation), hence
what needs to be proved is the following:
−→
OA+
1
2
³
−→
OC −
−→
OA
´
=
1
2
³
−→
OA+
−→
OC
´
which is now seen to be an immediate consequence of the distributive properties.
In the picture below,
−→
OA is in red,
−→
OC in blue,
−→
OB =
−→
OA +
−→
OC in green, and
−→
AC =
−→
OC −
−→
OA in violet.
The dot product 33
The geometrical theorem expressed by equation (12.1) is the following:
Theorem 1 The two diagonals of every parallelogram intersect at a point which di
vides them in two segments of equal length. In other words, the intersection point of
the two diagonals is the midpoint of both.
Indeed, the lefthand side of (12.1) is the vector represented by the oriented
segment OM, where M is the midpoint of diagonal AC (violet), whereas the righthand
side is the vector represented by the oriented segment ON, where N is the midpoint
of diagonal AB (green). More explicitly, the movement from O to M is described
as achieved by the composition of a ﬁrst movement from O to A with a second
movement from A towards C, which stops halfway (red plus half the violet); whereas
the movement from A to N is described as a single movement from A toward B,
stopping halfway (half the green). Since (12.1) asserts that
−−→
OM =
−→
ON, equality
between M and N follows, and hence a proof of (12.1) is a proof of the theorem
12.5 The dot product
12.6 Length or norm of a vector
12.7 Orthogonality of vectors
12.8 Exercises
12.8.1 n. 1 (p. 456)
(a) ha, bi = −6
(b) hb, ci = 2
(c) ha, ci = 6
(d) ha, b +ci = 0
(e) ha −b, ci = 4
12.8.2 n. 2 (p. 456)
(a) ha, bi c =(2 · 2 + 4 · 6 + (−7) · 3) (3, 4, −5) = 7 (3, 4, −5) = (21, 28, −35)
(b) ha, b +ci = 2 · (2 + 3) + 4 · (6 + 4) + (−7) (3 −5) = 64
34 Vector algebra
(c) ha +b, ci = (2 + 2) · 3 + (4 + 6) · 4 + (−7 + 3) · (−5) = 72
(d) ahb, ci = (2, 4, −7) (2 · 3 + 6 · 4 + 3 · (−5)) = (2, 4, −7) 15 = (30, 60, −105)
(e)
a
hb,ci
=
(2,4,−7)
15
=
¡
2
15
,
4
15
, −
7
15
¢
12.8.3 n. 3 (p. 456)
The statement is false. Indeed,
ha, bi = ha, ci ⇔ha, b −ci = 0 ⇔a ⊥ b −c
and the diﬀerence b −c may well be orthogonal to a without being the null vector.
The simplest example that I can conceive is in R
2
:
a = (1, 0) b = (1, 1) c = (1, 2)
See also exercise 1, question (d).
12.8.4 n. 5 (p. 456)
The required vector, of coordinates (x, y, z) must satisfy the two conditions
h(2, 1, −1) , (x, y, z)i = 0
h(1, −1, 2) , (x, y, z)i = 0
that is,
I 2x +y −z = 0
II x −y + 2z = 0
I +II 3x +z = 0
2I +II 5x +y = 0
Thus the set of all such vectors can be represented in parametric form as follows:
{(α, −5α, −3α)}
α∈R
12.8.5 n. 6 (p. 456)
hc, bi = 0
c = xa +yb
¾
⇒ hxa +yb, bi = 0
⇔ xha, bi +y hb, bi = 0
⇔ 7x + 14y = 0
Let (x, y) ≡ (−2, 1). Then c = (1, 5, −4) 6= 0 and h(1, 5, −4) , (3, 1, 2)i = 0.
Exercises 35
12.8.6 n. 7 (p. 456)
1 (solution with explicit mention of coordinates) From the last condition, for
some α to be determined,
c = (α, 2α, −2α)
Substituting in the ﬁrst condition,
d = a −c = (2 −α, −1 −2α, 2 + 2α)
Thus the second condition becomes
1 · (2 −α) + 2 · (−1 −2α) −2 (2 + 2α) = 0
that is,
−4 −9α = 0 α = −
4
9
and hence
c =
1
9
(−4, −8, 8)
d =
1
9
(22, −1, 10)
2 (same solution, with no mention of coordinates) From the last condition, for
some α to be determined,
c = αb
Substituting in the ﬁrst condition,
d = a −c = a −αb
Thus the second condition becomes
hb, a −αbi = 0
that is,
hb, ai −αhb, bi = 0 α =
hb, ai
hb, bi
=
2 · 1 + (−1) · 2 + 2 · (−2)
1
2
+ 2
2
+ (−2)
2
= −
4
9
and hence
c =
1
9
(−4, −8, 8)
d =
1
9
(22, −1, 10)
36 Vector algebra
12.8.7 n. 10 (p. 456)
(a) b = (1, −1) or b = (−1, 1).
(b) b = (−1, −1) or b = (1, 1).
(c) b = (−3, −2) or b = (3, 2).
(d) b = (b, −a) or b = (−b, a).
12.8.8 n. 13 (p. 456)
If b is the required vector, the following conditions must be satisﬁed
ha, bi = 0 kbk = kak
If a = 0, then kak = 0 and hence b = 0 as well. If a 6= 0, let the coordinates of a be
given by the couple (α, β), and the coordinates of b by the couple (x, y). The above
conditions take the form
αx +βy = 0 x
2
+y
2
= α
2
+β
2
Either α or β must be diﬀerent from 0. Suppose α is nonzero (the situation is
completely analogous if β 6= 0 is assumed). Then the ﬁrst equations gives
x = −
β
α
y (12.2)
and substitution in the second equation yields
µ
β
2
α
2
+ 1
¶
y
2
= α
2
+β
2
that is,
y
2
=
α
2
+β
2
α
2
+β
2
α
2
= α
2
and hence
y = ±α
Substituting back in (12.2) gives
x = ∓β
Thus there are only two solutions to the problem, namely, b = (−β, α) and b =
(β, −α) In particular,
(a) b = (−2, 1) or b = (2, −1).
(b) b = (2, −1) or b = (−2, 1).
(c) b = (−2, −1) or b = (2, 1).
(d) b = (−1, 2) or b = (1, −2)
Exercises 37
12.8.9 n. 14 (p. 456)
1 (right angle in C; solution with explicit mention of coordinates) Let (x, y, z)
be the coordinates of C. If the right angle is in C, the two vectors
−→
CA and
−→
CB must
be orthogonal. Thus
h[(2, −1, 1) −(x, y, z)] , [(3, −4, −4) −(x, y, z)]i = 0
that is,
h(2, −1, 1) , (3, −4, −4)i + h(x, y, z) , (x, y, z)i −h(2, −1, 1) + (3, −4, −4) , (x, y, z)i = 0
or
x
2
+y
2
+z
2
−5x + 5y + 3z + 6 = 0
The above equation, when rewritten in a more perspicuous way by “completion of
the squares”
µ
x −
5
2
¶
2
+
µ
y +
5
2
¶
2
+
µ
x +
3
2
¶
2
=
25
4
is seen to deﬁne the sphere of center P =
¡
5
2
, −
5
2
, −
3
2
¢
and radius
5
2
.
2 (right angle in C; same solution, with no mention of coordinates) Let
a ≡
−→
OA b ≡
−→
OB x ≡
−→
OC
(stressing that the point C, hence the vector
−→
OC, is unknown). Then, orthogonality
of the two vectors
−→
CA = a −x
−→
CB = b −x
is required; that is,
ha −x, b −xi = 0
or
ha, bi −ha +b, xi + hx, xi = 0
Equivalently,
kxk
2
−2
¿
a +b
2
, x
À
+
°
°
°
°
a +b
2
°
°
°
°
2
=
°
°
°
°
a +b
2
°
°
°
°
2
−ha, bi
°
°
°
°
x −
a +b
2
°
°
°
°
2
=
°
°
°
°
a −b
2
°
°
°
°
2
38 Vector algebra
The last characterization of the solution to our problem shows that the solution set is
the locus of all points having ﬁxed distance (
°
°
a−b
2
°
°
) from the midpoint of the segment
AB. Indeed, if π is any plane containing AB, and C is any point of the circle of π
having AB as diameter, it is known by elementary geometry that the triangle ACB
is rectangle in C.
3 (right angle in B; solution with explicit mention of coordinates) With
this approach, the vectors required to be orthogonal are
−→
BA and
−→
BC. Since
−→
BA =
(−1, 3, 5) and
−→
BC = (x −1, y + 4, z + 4), the following must hold
0 =
D
−→
BA,
−→
BC
E
= 1 −x + 3y + 12 + 5z + 20
that is,
x −3y −5z = 33
The solution set is the plane π through B and orthogonal to (1, −3, −5).
4 (right angle in B; same solution, with no mention of coordinates) Pro
ceeding as in the previous point, with the notation of point 2, the condition to be
required is
0 = ha −b, x −bi
that is,
ha −b, xi = ha −b, bi
Thus the solution plane π is seen to be through B and orthogonal to the segment
connecting point B to point A.
Exercises 39
5 (right angle in B) It should be clear at this stage that the solution set in this
case is the plane π
0
through A and orthogonal to AB, of equation
hb −a, xi = hb −a, ai
It is also clear that π and π
0
are parallel.
12.8.10 n. 15 (p. 456)
I c
1
−c
2
+ 2c
3
= 0
II 2c
1
+c
2
−c
3
= 0
I +II 3c
1
+c
3
= 0
I + 2II 5c
1
+c
2
= 0
c = (−1, 5, 3)
12.8.11 n. 16 (p. 456)
p = (3α, 4α)
q = (4β, −3β)
I 3α + 4β = 1
II 4α −3β = 2
4II + 3I 25α = 11
4I −3II 25β = −2
(α, β) =
1
25
(11, −2)
p =
1
25
(33, 44) q =
1
25
(−8, 6)
12.8.12 n. 17 (p. 456)
The question is identical to the one already answered in exercise 7 of this section.
Recalling solution 2 to that exercise, we have that p ≡
−→
OP must be equal to αb =
α
−→
OB, with
α =
ha, bi
hb, bi
=
10
4
Thus
p =
µ
5
2
,
5
2
,
5
2
,
5
2
¶
q =
µ
−
3
2
, −
1
2
,
1
2
,
3
2
¶
40 Vector algebra
12.8.13 n. 19 (p. 456)
It has been quickly seen in class that
ka +bk
2
= kak
2
+ kbk
2
+ 2 ha, bi
substituting −b to b, I obtain
ka −bk
2
= kak
2
+ kbk
2
−2 ha, bi
By subtraction,
ka +bk
2
−ka −bk
2
= 4 ha, bi
as required. You should notice that the above identity has been already used at the
end of point 2 in the solution to exercise 14 of the present section.
Concerning the geometrical interpretation of the special case of the above
identity
ka +bk
2
= ka −bk
2
if and only if ha, bi = 0
it is enough to notice that orthogonality of a and b is equivalent to the property
for the parallelogram OACB (in the given order around the perimeter, that is, with
vertex C opposed to the vertex in the origin O) to be a rectangle; and that in such
a rectangle ka +bk and ka −bk measure the lengths of the two diagonals.
12.8.14 n. 20 (p. 456)
ka +bk
2
+ ka −bk
2
= ha +b, a +bi + ha −b, a −bi
= ha, ai + 2 ha, bi + hb, bi + ha, ai −2 ha, bi + hb, bi
= 2 kak
2
+ 2 kbk
2
The geometric theorem expressed by the above identity can be stated as follows:
Theorem 2 In every parallelogram, the sum of the squares of the four sides equals
the sum of the squares of the diagonals.
12.8.15 n. 21 (p. 457)
Exercises 41
Let A, B, C, and D be the four vertices of the quadrilateral, starting from left
in clockwise order, and let M and N be the midpoint of the diagonals
−→
AC and
−→
DB.
In order to simplify the notation, let
u ≡
−→
AB, v ≡
−→
BC, w ≡
−→
CD, z ≡
−→
DA = −(u +v +w)
Then
c ≡
−→
AC = u +v d ≡
−→
DB = u +z = −(v +w)
−−→
MN =
−→
AN −
−−→
AM = u −
1
2
d −
1
2
c
2
−−→
MN = 2u+(v +w) −(u +v) = w +u
4
°
°
°
−−→
MN
°
°
°
2
= kwk
2
+ kuk
2
+ 2 hw, ui
kzk
2
= kuk
2
+ kvk
2
+ kwk
2
+ 2 hu, vi + 2 hu, wi + 2 hv, wi
kck
2
= kuk
2
+ kvk
2
+ 2 hu, vi
kdk
2
= kvk
2
+ kwk
2
+ 2 hv, wi
We are now in the position to prove the theorem.
¡
kuk
2
+ kvk
2
+ kwk
2
+ kzk
2
¢
−
¡
kck
2
+ kdk
2
¢
= kzk
2
−kvk
2
−2 hu, vi −2 hv, wi
= kuk
2
+ kwk
2
+ 2 hu, wi
= 4
°
°
°
−−→
MN
°
°
°
2
12.8.16 n. 22 (p. 457)
Orthogonality of xa +yb and 4ya −9xb amounts to
0 = hxa +yb, 4ya −9xbi
= −9 ha, bi x
2
+ 4 ha, bi y
2
+
¡
4 kak
2
−9 kbk
2
¢
xy
Since the above must hold for every couple (x, y), choosing x = 0 and y = 1 gives
ha, bi = 0; thus the condition becomes
¡
4 kak
2
−9 kbk
2
¢
xy = 0
and choosing now x = y = 1 gives
4 kak
2
= 9 kbk
2
Since kak is known to be equal to 6, it follows that kbk is equal to 4.
42 Vector algebra
Finally, since a and b have been shown to be orthogonal,
k2a + 3bk
2
= k2ak
2
+ k3bk
2
+ 2 h2a, 3bi
= 2
2
· 6
2
+ 3
2
· 4
2
+ 2 · 2 · 3 · 0
= 2
5
· 3
2
and hence
k2a + 3bk = 12
√
2
12.8.17 n. 24 (p. 457)
This is once again the question raised in exercises 7 and 17, in the general context of
the linear space R
n
(where n ∈ N is arbitrary). Since the coordinatefree version of
the solution procedure is completely independent from the number of coordinates of
the vectors involved, the full answer to the problem has already been seen to be
c =
hb, ai
ha, ai
a
d = b −
hb, ai
ha, ai
a
12.8.18 n. 25 (p. 457)
(a) For every x ∈ R,
ka +xbk
2
= kak
2
+x
2
kbk
2
+ 2xha, bi
= kak
2
+x
2
kbk
2
if a ⊥ b
≥ kak
2
if a ⊥ b
(b) Since the norm of any vector is nonnegative, the following biconditional is true:
(∀x ∈ R, ka +xbk ≥ kak) ⇔
¡
∀x ∈ R, ka +xbk
2
≥ kak
2
¢
Moreover, by pure computation, the following biconditional is true:
∀x ∈ R, ka +xbk
2
−kak
2
≥ 0 ⇔∀x ∈ R, x
2
kbk
2
+ 2xha, bi ≥ 0
If ha, bi is positive, the trinomial x
2
kbk
2
+ 2xha, bi is negative for all x in the open
interval
³
−
2ha,bi
kbk
2
, 0
´
; if it is negative, the trinomial is negative in the open interval
³
0, −
2ha,bi
kbk
2
´
. It follows that the trinomial can be nonnegative for every x ∈ R only
if ha, bi is zero, that is, only if it reduces to the second degree term x
2
kbk
2
. In
conclusion, I have proved that the conditional
(∀x ∈ R, ka +xbk ≥ kak) ⇒ha, bi = 0
is true.
Projections. Angle between vectors in nspace 43
12.9 Projections. Angle between vectors in nspace
12.10 The unit coordinate vectors
12.11 Exercises
12.11.1 n. 1 (p. 460)
ha, bi = 11 kbk
2
= 9
The projection of a along b is
11
9
b =
µ
11
9
,
22
9
,
22
9
¶
12.11.2 n. 2 (p. 460)
ha, bi = 10 kbk
2
= 4
The projection of a along b is
10
4
b =
µ
5
2
,
5
2
,
5
2
,
5
2
¶
12.11.3 n. 3 (p. 460)
(a)
cos
c
ai =
ha, ii
kak kik
=
6
7
cos
c
aj =
ha, ji
kak kjk
=
3
7
cos
c
ai =
ha, ki
kak kkk
= −
2
7
(b) There are just two vectors as required, the unit direction vector u of a, and its
opposite:
u =
a
kak
=
µ
6
7
,
3
7
, −
2
7
¶
−
a
kak
=
µ
−
6
7
, −
3
7
,
2
7
¶
44 Vector algebra
12.11.4 n. 5 (p. 460)
Let
A ≡ (2, −1, 1) B ≡ (1, −3, −5) C ≡ (3, −4, −4)
p ≡
−→
BC = (2, −1, 1) q ≡
−→
CA = (−1, 3, 5) r ≡
−→
AB = (−1, −2, −6)
Then
cos
b
A =
h−q, ri
k−qk krk
=
35
√
35
√
41
=
√
35
√
41
41
cos
b
B =
hp, −ri
kpk k−rk
=
6
√
6
√
41
=
√
6
√
41
41
cos
b
C =
h−p, qi
k−pk kqk
=
0
√
6
√
35
= 0
There is some funny occurrence in this exercise, which makes a wrong solution
apparently correct (or almost correct), if one looks only at numerical results. The
angles in A, B, C, as implicitly argued above, are more precisely described as B
b
AC,
C
b
BA, A
b
CB; that is, as the three angles of triangle ABC. If some confusion is made
between points and vectors, and/or angles, one may be led into operate directly with
the coordinates of points A, B, C in place of the coordinates of vectors
−→
BC,
−→
AC,
−→
AB,
respectively. This amounts to work, as a matter of fact, with the vectors
−→
OA,
−→
OB,
−→
OC, therefore computing
D
−→
OB,
−→
OC
E
°
°
°
−→
OB
°
°
°
°
°
°
−→
OC
°
°
°
= cos B
b
OC an angle of triangle OBC
D
−→
OC,
−→
OA
E
°
°
°
−→
OC
°
°
°
°
°
°
−→
OA
°
°
°
= cos C
b
OA an angle of triangle OCA
D
−→
OA,
−→
OB
E
°
°
°
−→
OA
°
°
°
°
°
°
−→
OB
°
°
°
= cos A
b
OB an angle of triangle OAB
instead of cos B
b
AC, cos C
b
BA, cos A
b
CB, respectively. Up to this point, there is
nothing funny in doing that; it’s only a mistake, and a fairly bad one, being of
conceptual type. The funny thing is in the numerical data of the exercise: it so
happens that
1. points A, B, C are coplanar with the origin O; more than that,
2.
−→
OA =
−→
BC and
−→
OB =
−→
AC; more than that,
Exercises 45
3. A
b
OB is a right angle.
Point 1 already singles out a somewhat special situation, but point 2 makes
OACB a parallelogram, and point 3 makes it even a rectangle.
−→
OA red,
−→
OB blue,
−→
OC green, AB violet
It turns out, therefore, that
°
°
°
−→
OA
°
°
° =
°
°
°
−→
BC
°
°
° = kpk
°
°
°
−→
OB
°
°
° =
°
°
°
−→
AC
°
°
° = k−qk = kqk
°
°
°
−→
OC
°
°
° =
°
°
°
−→
AB
°
°
° = krk
C
b
OA = C
b
BA B
b
OC = B
b
AC A
b
OB = A
b
CB
and such a special circumstance leads a wrong solution to yield the right numbers.
12.11.5 n. 6 (p. 460)
Since
ka +c ±bk
2
= ka +ck
2
+ kbk
2
± 2 ha +c, bi
from
ka +c +bk = ka +c −bk
it is possible to deduce
ha +c, bi = 0
This is certainly true, as a particular case, if c = −a, which immediately implies
c ac = π
c
bc = π −
c
ab =
7
8
π
46 Vector algebra
Moreover, even if a +c = 0 is not assumed, the same conclusion holds. Indeed, from
hc, bi = −ha, bi
and
kck = kak
it is easy to check that
cos
c
bc =
hb, ci
kbk kck
= −
ha, ci
kbk kak
= −cos
c
ab
and hence that
c
bc = π ±c ac =
7
8
π,
9
8
π (the second value being superﬂuous).
12.11.6 n. 8 (p. 460)
We have
kak =
√
n kb
n
k =
r
n(n + 1) (2n + 1)
6
ha, b
n
i =
n(n + 1)
2
cos
[
ab
n
=
n(n+1)
2
√
n
q
n(n+1)(2n+1)
6
=
v
u
u
t
n
2
(n+1)
2
4
n
2
(n+1)(2n+1)
6
=
r
3
2
n + 1
2n + 1
lim
n→+∞
cos
[
ab
n
=
√
3
2
lim
n→+∞
[
ab
n
=
π
3
12.11.7 n. 10 (p. 461)
(a)
ha, bi = cos ϑsin ϑ −cos ϑsin ϑ = 0
kak
2
= kbk
2
= cos
2
ϑ + sin
2
ϑ = 1
(b) The system
µ
cos ϑ sin ϑ
−sin ϑ cos ϑ
¶µ
x
y
¶
=
µ
x
y
¶
that is,
µ
cos ϑ −1 sin ϑ
−sin ϑ cos ϑ −1
¶µ
x
y
¶
=
µ
0
0
¶
has the trivial solution as its unique solution if
¯
¯
¯
¯
cos ϑ −1 sin ϑ
−sin ϑ cos ϑ −1
¯
¯
¯
¯
6= 0
Exercises 47
The computation gives
1 + cos
2
ϑ + sin
2
ϑ −2 cos ϑ 6= 0
2 (1 −cos ϑ) 6= 0
cos ϑ 6= 1
ϑ / ∈ {2kπ}
k∈Z
Thus if
ϑ ∈ {(−2kπ, (2k + 1) π)}
k∈Z
the only vector satisfying the required condition is (0, 0). On the other hand, if
ϑ ∈ {2kπ}
k∈Z
the coeﬃcient matrix of the above system is the identity matrix, and every vector in
R
2
satisﬁes the required condition.
12.11.8 n. 11 (p. 461)
Let OABC be a rhombus, which I have taken (without loss of generality) with one
vertex in the origin O and with vertex B opposed to O. Let
a ≡
−→
OA (red) b ≡
−→
OB (green) c ≡
−→
OC (blue)
Since OABC is a parallelogram, the oriented segments AB (blue, dashed) and OB
are congruent, and the same is true for CB (red, dashed) and OA. Thus
−→
AB = b
−→
CB = a
From elementary geometry (see exercise 12 of section 4) the intersection point M of
the two diagonals OB (green) and AC (violet) is the midpoint of both.
The assumption that OABC is a rhombus is expressed by the equality
kak = kck ((rhombus))
48 Vector algebra
The statement to be proved is orthogonality between the diagonals
D
−→
OB,
−→
AC
E
= 0
that is,
ha +c, c −ai = 0
or
kck
2
−kak
2
+ ha, ci −hc, ai = 0
and hence (by commutativity)
kck
2
= kak
2
The last equality is an obvious consequence of ((rhombus)). As a matter of fact, since
norms are nonnegative real numbers, the converse is true, too. Thus a parallelogram
has orthogonal diagonals if and only if it is a rhombus.
12.11.9 n. 13 (p. 461)
The equality to be proved is straightforward. The “law of cosines” is often called
Theorem 3 (Carnot) In every triangle, the square of each side is the sum of the
squares of the other two sides, minus their double product multiplied by the cosine of
the angle they form.
The equality in exam can be readily interpreted according to the theorem’s
statement, since in every parallelogram ABCD with a =
−→
AB and b =
−→
AC the
diagonal vector
−→
CB is equal to a − b, so that the triangle ABC has side vectors a,
b, and a −b. The theorem specializes to Pythagoras’ theorem when a ⊥ b.
12.11.10 n. 17 (p. 461)
(a) That the function
R
n
→R, a 7→
X
i∈n
a
i

is positive can be seen by the same arguments used for the ordinary norm; nonnega
tivity is obvious, and strict positivity still relies on the fact that a sum of concordant
numbers can only be zero if all the addends are zero. Homogeneity is clear, too:
∀a ∈ R
n
, ∀α ∈ R,
kαak =
X
i∈n
αa
i
 =
X
i∈n
α a
i
 = α
X
i∈n
a
i
 = α kak
Exercises 49
Finally, the triangle inequality is much simpler to prove for the present norm (some
times referred to as “the taxicab norm”) than for the euclidean norm:
∀a ∈ R
n
, ∀b ∈ R
n
,
ka +bk =
X
i∈n
a
i
+b
i
 ≤
X
i∈n
a
i
 + b
i
 =
X
i∈n
a
i
 +
X
i∈n
b
i

= kak + kbk
(b) The subset of R
2
to be described is
S ≡
©
(x, y) ∈ R
2
: x + y = 1
ª
= S
++
∪ S
−+
∪ S
−−
∪ S
+−
where
S
++
≡
©
(x, y) ∈ R
2
+
: x +y = 1
ª
(red)
S
−+
≡ {(x, y) ∈ R
−
×R
+
: −x +y = 1} (green)
S
−−
≡
©
(x, y) ∈ R
2
−
: −x −y = 1
ª
(violet)
S
+−
≡ {(x, y) ∈ R
+
×R
−
: x −y = 1} (blue)
Once the lines whose equations appear in the deﬁnitions of the four sets above are
drawn, it is apparent that S is a square, with sides parallel to the quadrant bisectrices
2
1
0
1
2
2 1 1 2
(c) The function
f : R
n
→R, a 7→
X
i∈n
a
i

is nonnegative, but not positive (e.g., for n = 2, f (x, −x) = 0 ∀x ∈ R). It is
homogeneous:
∀a ∈ R
n
, ∀α ∈ R,
kαak =
¯
¯
¯
¯
¯
X
i∈n
αa
i
¯
¯
¯
¯
¯
=
¯
¯
¯
¯
¯
α
X
i∈n
a
i
¯
¯
¯
¯
¯
= α
¯
¯
¯
¯
¯
X
i∈n
a
i
¯
¯
¯
¯
¯
= α kak
50 Vector algebra
Finally, f is subadditive, too (this is another way of saying that the triangle inequality
holds). Indeed,
∀a ∈ R
n
, ∀b ∈ R
n
,
ka +bk =
¯
¯
¯
¯
¯
X
i∈n
(a
i
+b
i
)
¯
¯
¯
¯
¯
=
¯
¯
¯
¯
¯
X
i∈n
a
i
+
X
i∈n
b
i
¯
¯
¯
¯
¯
≤
¯
¯
¯
¯
¯
X
i∈n
a
i
¯
¯
¯
¯
¯
+
¯
¯
¯
¯
¯
X
i∈n
b
i
¯
¯
¯
¯
¯
= kak + kbk
12.12 The linear span of a ﬁnite set of vectors
12.13 Linear independence
12.14 Bases
12.15 Exercises
12.15.1 n. 1 (p. 467)
x(i −j) +y (i +j) = (x +y, y −x)
(a) x +y = 1 and y −x = 0 yield (x, y) =
¡
1
2
,
1
2
¢
.
(b) x +y = 0 and y −x = 1 yield (x, y) =
¡
−
1
2
,
1
2
¢
.
(c) x +y = 3 and y −x = −5 yield (x, y) = (4, −1).
(d) x +y = 7 and y −x = 5 yield (x, y) = (1, 6).
12.15.2 n. 3 (p. 467)
I 2x +y = 2
II −x + 2y = −11
III x −y = 7
I +III 3x = 9
II +III y = −4
III (check) 3 + 4 = 7
The solution is (x, y) = (3, −4).
Exercises 51
12.15.3 n. 5 (p. 467)
(a) If there exists some α ∈ R ∼ {0} such that
∗
a = αb, then the linear combination
1a−αb is nontrivial and it generates the null vector; a and b are linearly dependent.
(b) The argument is best formulated by counterposition, by proving that if a and b
are linearly dependent, then they are parallel. Let a and b nontrivially generate the
null vector: αa + βb = 0 (according to the deﬁnition, at least one between α and β
is nonzero; but in the present case they are both nonzero, since a and b have been
assumed both diﬀerent from the null vector; indeed, αa + βb = 0 with α = 0 and
β 6= 0 implies b = 0, and αa + βb = 0 with β = 0 and α 6= 0 implies a = 0). Thus
αa = −βb, and hence a = −
β
α
b (or b = −
α
β
a, if you prefer).
12.15.4 n. 6 (p. 467)
Linear independency of the vectors (a, b) and (c, d) has been seen in the lectures
(proof of Steinitz’ theorem, part 1) to be equivalent to non existence of nontrivial
solutions to the system
ax +cy = 0
bx +dy = 0
The system is linear and homogeneous, and has unknowns and equations in equal
number (hence part 2 of the proof of Steinitz’ theorem does not apply). I argue by
the principle of counterposition. If a nontrivial solution (x, y) exists, both (a, c) and
(b, d) must be proportional to (y, −x) (see exercise 10, section 8), and hence to each
other. Then, from (a, c) = h(b, d) it is immediate to derive ad−bc = 0. The converse
is immediate, too
12.15.5 n. 7 (p. 467)
By the previous exercise, it is enough to require
(1 +t)
2
−(1 −t)
2
6= 0
4t 6= 0
t 6= 0
12.15.6 n. 8 (p. 467)
(a) The linear combination
1i + 1j + 1k + (−1) (i +j +k)
is nontrivial and spans the null vector.
(b) Since, for every (α, β, γ) ∈ R
3
,
αi +βj +γk = (α, β, γ)
∗
Even if parallelism is deﬁned more broadly − see the footnote in exercise 9 of section 4 − α
cannot be zero in the present case, because both a and b have been assumed diﬀerent from the null
vector.
52 Vector algebra
it is clear that
∀(α, β, γ) ∈ R
3
, (αi +βj +γk = 0) ⇒α = β = γ = 0
so that the triple (i, j, k) is linearly independent.
(c) Similarly, for every (α, β, γ) ∈ R
3
,
αi +βj +γ (i +j +k) = (α +γ, β +γ, γ)
and hence from
αi +βj +γ (i +j +k) = 0
it follows
α +γ = 0 β +γ = 0 γ = 0
that is,
α = β = γ = 0
showing that the triple (i, j, i +j +k) is linearly independent.
(d) The last argument can be repeated almost verbatim for triples (i, i +j +k, k)
and (i +j +k, j, k), taking into account that
αi +β (i +j +k) +γk = (α +β, β, β + γ)
α(i +j +k) +βj +γk = (α, α +β, α +γ)
12.15.7 n. 10 (p. 467)
(a) Again from the proof of Steinitz’ theorem, part 1, consider the system
I x +y +z = 0
II y +z = 0
III 3z = 0
It is immediate to derive that its unique solution is the trivial one, and hence that
the given triple is linearly independent.
(b) We need to consider the following two systems:
I x +y +z = 0
II y +z = 1
III 3z = 0
I x +y +z = 0
II y +z = 0
III 3z = 1
It is again immediate that the unique solutions to the systems are (−1, 1, 0) and
¡
0, −
1
3
,
1
3
¢
respectively.
Exercises 53
(c) The system to study is the following:
I x +y +z = 2
II y +z = −3
III 3z = 5
III z =
5
3
(↑) ,→II y = −
14
3
(↑) ,→I x = 5
(d) For an arbitrary triple (a, b, c), the system
I x +y +z = a
II y +z = b
III 3z = c
has the (unique) solution
¡
a −b, b −
c
3
,
c
3
¢
. Thus the given triple spans R
3
, and it is
linearly independent (as seen at a).
12.15.8 n. 12 (p. 467)
(a) Let xa +yb +zc = 0, that is,
x +z = 0
x +y +z = 0
x +y = 0
y = 0
then y, x, and z are in turn seen to be equal to 0, from the fourth, third, and ﬁrst
equation in that order. Thus (a, b, c) is a linearly independent triple.
(b) Any nontrivial linear combination d of the given vectors makes (a, b, c, d) a
linearly dependent quadruple. For example, d ≡ a +b +c = (2, 3, 2, 1)
(c) Let e ≡ (−1, −1, 1, 3), and suppose xa +yb +zc +te = 0, that is,
x +z −t = 0
x +y +z −t = 0
x +y +t = 0
y + 3t = 0
Then, subtracting the ﬁrst equation from the second, gives y = 0, and hence t, x,
and z are in turn seen to be equal to 0, from the fourth, third, and ﬁrst equation in
that order. Thus (a, b, c, e) is linearly independent.
(d) The coordinates of x with respect to {a, b, c, e}, which has just been seen to be
a basis of R
4
, are given as the solution (s, t, u, v) to the following system:
s +u −v = 1
s +t +u −v = 2
s +t +v = 3
t + 3v = 4
54 Vector algebra
It is more direct and orderly to work just with the table formed by the system
(extended) matrix
¡
A x
¢
, and to perform elementary row operations and column
exchanges.
¸
¸
¸
¸
¸
¸
¸
s t u v x
1 0 1 −1 1
1 1 1 −1 2
1 1 0 1 3
0 1 0 3 4
¸
a
1
↔a
3
,
2
a
0
←
2
a
0
−
1
a
0
¸
¸
¸
¸
¸
¸
¸
u t s v x
1 0 1 −1 1
0 1 0 0 1
0 1 1 1 3
0 1 0 3 4
¸
a
2
↔a
3
,
2
a
0
↔
3
a
0
¸
¸
¸
¸
¸
¸
¸
u s t v x
1 1 0 −1 1
0 1 1 1 3
0 0 1 0 1
0 0 1 3 4
¸
4
a
0
←
4
a
0
−
3
a
0
¸
¸
¸
¸
¸
¸
¸
u s t v x
1 1 0 −1 1
0 1 1 1 3
0 0 1 0 1
0 0 0 3 3
¸
Thus the system has been given the following form:
u +s −v = 1
s +t +v = 3
t = 1
3v = 3
which is easily solved from the bottom to the top: (s, t, u, v) = (1, 1, 1, 1).
Exercises 55
12.15.9 n. 13 (p. 467)
(a)
α
¡√
3, 1, 0
¢
+β
¡
1,
√
3, 1
¢
+γ
¡
0, 1,
√
3
¢
= (0, 0, 0)
⇔
I
√
3α +β = 0
II α +
√
3β +γ = 0
III β +
√
3γ = 0
⇔
I
√
3α +β = 0
√
3II −III −I 2β = 0
III β +
√
3γ = 0
⇔ (α, β, γ) = (0, 0, 0)
and the three given vectors are linearly independent.
(b)
α
¡√
2, 1, 0
¢
+β
¡
1,
√
2, 1
¢
+γ
¡
0, 1,
√
2
¢
= (0, 0, 0)
⇔
I
√
2α +β = 0
II α +
√
2β +γ = 0
III β +
√
2γ = 0
This time the sum of equations I and III (multiplied by
√
2) is the same as twice
equation II, and a linear dependence in the equation system suggests that it may
well have nontrivial solutions. Indeed,
¡√
2, −2,
√
2
¢
is such a solution. Thus the
three given triple of vectors is linearly dependent.
(c)
α(t, 1, 0) +β (1, t, 1) +γ (0, 1, t) = (0, 0, 0)
⇔
I tα +β = 0
II α +tβ +γ = 0
III β +tγ = 0
It is clear that t = 0 makes the triple linearly dependent (the ﬁrst and third vector
coincide in this case). Let us suppose, then, t 6= 0. From I and III, as already
noticed, I deduce that α = γ. Equations II and III then become
tβ + 2γ = 0
β +tγ = 0
a 2 by 2 homogeneous system with determinant of coeﬃcient matrix equal to t
2
−2.
Such a system has nontrivial solutions for t ∈
©√
2, −
√
2
ª
. In conclusion, the given
triple is linearly dependent for t ∈
©
0,
√
2, −
√
2
ª
.
56 Vector algebra
12.15.10 n. 14 (p. 468)
Call as usual u, v, w, and z the four vectors given in each case, in the order.
(a) It is clear that v = u + w, so that v can be dropped. Moreover, every linear
combination of u and w has the form (x, y, x, y), and cannot be equal to z. Thus
(u, w, z) is a maximal linearly independent triple.
(b) Notice that
1
2
(u +z) = e
(1)
1
2
(u −v) = e
(2)
1
2
(v −w) = e
(3)
1
2
(w−z) = e
(4)
Since (u, v, w, z) spans the four canonical vectors, it is a basis of R
4
. Thus (u, v, w, z)
is maximal linearly independent.
(c) Similarly,
u −v = e
(1)
v −w = e
(2)
w−z = e
(3)
z = e
(4)
and (u, v, w, z) is maximal linearly independent.
12.15.11 n. 15 (p. 468)
(a) Since the triple (a, b, c) is linearly independent,
α(a +b) +β (b +c) +γ (a +c) = 0
⇔ (α +γ) a + (α +β) b + (β +γ) c = 0
⇔
I α +γ = 0
II α +β = 0
III β +γ = 0
⇔
I +II −III 2α = 0
−I +II +III 2β = 0
I −II +III 2γ = 0
and the triple (a +b, b +c, a +c) is linearly independent, too
(b) On the contrary, choosing as nontrivial coeﬃcient triple (α, β, γ) ≡ (1, −1, 1),
(a −b) −(b +c) + (a +c) = 0
it is seen that the triple (a −b, b +c, a +c) is linearly dependent.
12.15.12 n. 17 (p. 468)
Let a ≡ (0, 1, 1) and b ≡ (1, 1, 1); I look for two possible alternative choices of a
vector c ≡ (x, y, z) such that the triple (a, b, c) is linearly independent. Since
αa +βb +γc = (β +γx, α +β +γy, α +β +γz)
Exercises 57
my choice of x, y, and z must be such to make the following conditional statement
true for each (α, β, γ) ∈ R
3
:
I β +γx = 0
II α +β +γy = 0
III α +β +γz = 0
⇒
α = 0
β = 0
γ = 0
Subtracting equation III from equation II, I obtain
γ (y −z) = 0
Thus any choice of c with y 6= z makes γ = 0 a consequence of IIIII; in such a case,
I yields β = 0 (independently of the value assigned to x), and then either II or III
yields α = 0. Conversely, if y = z, equations II and III are the same, and system
IIII has inﬁnitely many nontrivial solutions
α = −γy −β
β = −γx
γ free
provided either x or y (hence z) is diﬀerent from zero.
As an example, possible choices for c are (0, 0, 1), (0, 1, 0), (1, 0, 1), (1, 1, 0).
12.15.13 n. 18 (p. 468)
The ﬁrst example of basis containing the two given vectors is in point (c) of exercise
14 in this section, since the vectors u and v there coincide with the present ones.
Keeping the same notation, a second example is (u, v, w +z, w−z).
12.15.14 n. 19 (p. 468)
(a) It is enough to prove that each element of T belongs to lin S, since each element
of lin T is a linear combination in T, and S, as any subspace of a vector space, is
closed with respect to formation of linear combinations. Let u, v, and w be the three
elements of S (in the given order), and let a and b be the two elements of T (still in
the given order); it is then readily checked that
a = u −w b = 2w
(b) The converse inclusion holds as well. Indeed,
v = a −b w =
1
2
b u = v +w = a −
1
2
b
Thus
lin S = lin T
58 Vector algebra
Similarly, if c and d are the two elements of U,
c = u +v d = u + 2v
which proves that lin U ⊆ lin S. Inverting the above formulas,
u = 2c −d v = d −c
It remains to be established whether or not w is an element of lin U. Notice that
αc +βd = (α +β, 2α + 3β, 3α + 5β)
It follows that w is an element of lin U if and only if there exists (α, β) ∈ R
2
such
that
I α +β = 1
II 2α + 3β = 0
III 3α + 5β = −1
The above system has the unique solution (3, −2), which proves that lin S ⊆ linU.
In conclusion,
linS = linT = lin U
12.15.15 n. 20 (p. 468)
(a) The claim has already been proved in the last exercise, since from A ⊆ B I can
infer A ⊆ lin B, and hence lin A ⊆ lin B.
(b) By the last result, from A ∩ B ⊆ A and A ∩ B ⊆ B I infer
lin A∩ B ⊆ lin A linA ∩ B ⊆ lin B
which yields
lin A∩ B ⊆ lin A∩ linB
(c) It is enough to deﬁne
A ≡ {a, b} B ≡ {a +b}
where the couple (a, b) is linearly independent. Indeed,
B ⊆ lin A
and hence, by part (a) of last exercise,
lin B ⊆ lin lin A = lin A
lin A∩ lin B = lin B
On the other hand,
A∩ B = ∅ lin A∩ B = {0}
The vector space V
n
(C) of ntuples of complex numbers 59
12.16 The vector space V
n
(C) of ntuples of complex numbers
12.17 Exercises
60 Vector algebra
Chapter 13
APPLICATIONS OF VECTOR ALGEBRA TO ANALYTIC
GEOMETRY
13.1 Introduction
13.2 Lines in nspace
13.3 Some simple properties of straight lines
13.4 Lines and vectorvalued functions
13.5 Exercises
13.5.1 n. 1 (p. 477)
A direction vector for the line L is
−→
PQ = (4, 0), showing that L is horizontal. Thus
a point belongs to L if and only if its second coordinate is equal to 1. Among the
given points, (b), (d), and (e) belong to L.
13.5.2 n. 2 (p. 477)
A direction vector for the line L is v ≡
1
2
−→
PQ = (−2, 1). The parametric equations
for L are
x = 2 −2t
y = −1 +t
If t = 1 I get point (a) (the origin). Points (b), (d) and (e) have the second coordinate
equal to 1, which requires t = 2. This gives x = −2, showing that of the three points
only (e) belongs to L. Finally, point (c) does not belong to L, because y = 2 requires
t = 3, which yields x = −4 6= 1.
62 Applications of vector algebra to analytic geometry
13.5.3 n. 3 (p. 477)
The parametric equations for L are
x = −3 +h
y = 1 −2h
z = 1 + 3h
The following points belong to L:
(c) (h = 1) (d) (h = −1) (e) (h = 5)
13.5.4 n. 4 (p. 477)
The parametric equations for L are
x = −3 + 4k
y = 1 +k
z = 1 + 6h
The following points belong to L:
(b) (h = −1) (e)
µ
h =
1
2
¶
(f)
µ
h =
1
3
¶
A direction vector for the line L is
−→
PQ = (4, 0), showing that L is horizontal.
Thus a point belongs to L if and only if its second coordinate is equal to 1. Among
the given points, (b), (d), and (e) belong to L.
13.5.5 n. 5 (p. 477)
I solve each case in a diﬀerent a way.
(a)
−→
PQ = (2, 0, −2)
−→
QR = (−1, −2, 2)
The two vectors are not parallel, hence the three points do not belong to the same
line.
(b) Testing aﬃne dependence,
I 2h + 2k + 3l = 0
II −2h + 3k +l = 0
III −6h + 4k +l = 0
I −2II +III 2l = 0
I +II 5k + 4l = 0
2I −III 10h + 5l = 0
the only combination which is equal to the null vector is the trivial one. The three
points do not belong to the same line.
Exercises 63
(c) The line through P and R has equations
x = 2 + 3k
y = 1 −2k
z = 1
Trying to solve for k with the coordinates of Q, I get k = −
4
3
from the ﬁrst equation
and k = −1 from the second; Q does not belong to L(P, R).
13.5.6 n. 6 (p. 477)
The question is easy, but it must be answered by following some orderly path, in
order to achieve some economy of thought and of computations (there are
¡
8
2
¢
= 28
diﬀerent oriented segments joining two of the eight given points. First, all the eight
points have their third coordinate equal to 1, and hence they belong to the plane
π of equation z = 1. We can concentrate only on the ﬁrst two coordinates, and,
as a matter of fact, we can have a very good hint on the situation by drawing a
twodimensional picture, to be considered as a picture of the π
15
10
5
0
5
10
15
15 10 5 5 10 15
Since the ﬁrst two components of
−→
AB are (4, −2), I check that among the
twodimensional projections of the oriented segments connecting A with points D to
H
p
xy
−→
AD = (−4, 2) p
xy
−→
AE = (−1, 1) p
xy
−→
AF = (−6, 3)
p
xy
−→
AG = (−15, 8) p
xy
−→
AH = (12, −7)
64 Applications of vector algebra to analytic geometry
only the ﬁrst and third are parallel to p
xy
−→
AB. Thus all elements of the set P
1
≡
{A, B, C, D, F} belong to the same (black) line L
ABC
, and hence no two of them can
both belong to a diﬀerent line. Therefore, it only remains to be checked whether or
not the lines through the couples of points (E, G) (red), (G, H) (blue), (H, E) (green)
coincide, and whether or not any elements of P
1
belong to them. Direction vectors
for these three lines are
1
2
p
xy
−→
EG = (−7, 4)
1
3
p
xy
−→
GH = (9, −5) p
xy
−−→
HE = (−13, 7)
and as normal vectors for them I may take
n
EG
≡ (4, 7) n
GH
≡ (5, 9) n
HE
≡ (7, 13)
By requiring point E to belong to the ﬁrst and last, and point H to the second, I end
up with their equations as follows
L
EG
: 4x + 7y = 11 L
GH
: 5x + 9y = 16 L
HE
: 7x + 13y = 20
All these three lines are deﬁnitely not parallel to L
ABC
, hence they intersect it exactly
at one point. It is seen by direct inspection that, within the set P
1
, C belongs to
L
EG
, F belongs to L
GH
, and neither A nor B, nor D belong to L
HE
.
Thus there are three (maximal) sets of at least three collinear points, namely
P
1
≡ {A, B, C, D, F} P
2
≡ {C, E, G} P
3
≡ {F, G, H}
13.5.7 n. 7 (p. 477)
The coordinates of the intersection point are determined by the following equation
system
I 1 +h = 2 + 3k
II 1 + 2h = 1 + 8k
III 1 + 3h = 13k
Subtracting the third equation from the sum of the ﬁrst two, I get k = 1; substituting
this value in any of the three equations, I get h = 4. The three equations are
consistent, and the two lines intersect at the point of coordinates (5, 9, 13).
13.5.8 n. 8 (p. 477)
(a) The coordinates of the intersection point of L(P; a) and L(Q; b) are determined
by the vector equation
P +ha = Q+kb (13.1)
which gives
P −Q = kb −ha
that is,
−→
PQ ∈ span {a, b}
Exercises 65
13.5.9 n. 9 (p. 477)
X (t) = (1 +t, 2 −2t, 3 + 2t)
(a)
d (t) ≡ kQ−X (t)k
2
= (2 −t)
2
+ (1 + 2t)
2
+ (−2 −2t)
2
= 9t
2
+ 8t + 9
(b) The graph of the function t 7→ d (t) ≡ 9t
2
+ 8t + 9 is a parabola with the
point (t
0
, d (t
0
)) =
¡
−
4
9
,
65
9
¢
as vertex. The minimum squared distance is
65
9
, and the
minimum distance is
√
65
3
.
(c)
X (t
0
) =
µ
5
9
,
26
9
,
19
9
¶
Q−X (t
0
) =
µ
22
9
,
1
9
, −
10
9
¶
hQ−X (t
0
) , Ai =
22 −2 + 20
9
= 0
that is, the point on L of minimum distance from Q is the orthogonal projection of
Q on L.
13.5.10 n. 10 (p. 477)
(a) Let A ≡ (α, β, γ), P ≡ (λ, µ, ν) and Q ≡ (%, σ, τ). The points of the line L
through P with direction vector
−→
OA are represented in parametric form (the generic
point of L is denoted X (t)). Then
X (t) ≡ P +At = (λ +αt, µ +βt, ν +γt)
f (t) ≡ kQ−X (t)k
2
= kQ−P −Atk
2
= kQk
2
+ kPk
2
+ kAk
2
t −2 hQ, Pi + 2 hP −Q, Ai t
= at
2
+bt +c
where
a ≡ kAk
2
= α
2
+β
2
+γ
2
b
2
≡ hP −Q, Ai = α(λ −%) +β (µ −σ) +γ (ν −τ)
c ≡ kQk
2
+ kPk
2
−2 hQ, Pi = kP −Qk
2
= (λ −%)
2
+ (µ −σ)
2
+ (ν −τ)
2
The above quadratic polynomial has a second degree term with a positive coeﬃcient;
its graph is a parabola with vertical axis and vertex in the point of coordinates
µ
−
b
2a
, −
∆
4a
¶
=
Ã
hQ−P, Ai
kAk
2
,
kAk
2
kP −Qk
2
−hP −Q, Ai
2
kAk
2
!
66 Applications of vector algebra to analytic geometry
Thus the minimum value of this polynomial is achieved at
t
0
≡
hQ−P, Ai
kAk
2
=
α(% −λ) +β (σ −β) +γ (τ −ν)
α
2
+β
2
+γ
2
and it is equal to
kAk
2
kP −Qk
2
−kAk
2
kP −Qk
2
cos
2
ϑ
kAk
2
= kP −Qk
2
sin
2
ϑ
where
ϑ ≡
\
−→
OA,
−→
QP
is the angle formed by direction vector of the line and the vector carrying the point
Q not on L to the point P on L.
(b)
Q−X (t
0
) = Q−(P +At
0
) = (Q−P) −
hQ−P, Ai
kAk
2
A
hQ−X (t
0
) , Ai = hQ−P, Ai −
hQ−P, Ai
kAk
2
hA, Ai = 0
13.5.11 n. 11 (p. 477)
The vector equation for the coordinates of an intersection point of L(P; a) and
L(Q; a) is
P +ha = Q+ka (13.2)
It gives
P −Q = (h −k) a
and there are two cases: either
−→
PQ / ∈ span {a}
and equation (13.2) has no solution (the two lines are parallel), or
−→
PQ ∈ span {a}
that is,
∃c ∈ R, Q−P = ca
and equation (13.2) is satisﬁed by all couples (h, k) such that k −h = c (the two lines
intersect in inﬁnitely many points, i.e., they coincide). The two expressions P + ha
and Q + ka are seen to provide alternative parametrizations of the same line. For
each given point of L, the transformation h 7→k (h) ≡ h +c identiﬁes the parameter
change which allows to shift from the ﬁrst parametrization to the second.
Planes in euclidean nspaces 67
13.5.12 n. 12 (p. 477)
13.6 Planes in euclidean nspaces
13.7 Planes and vectorvalued functions
13.8 Exercises
13.8.1 n. 2 (p. 482)
We have:
−→
PQ = (2, 2, 3)
−→
PR = (2, −2, −1)
so that the parametric equations of the plane are
x = 1 + 2s + 2t (13.3)
y = 1 + 2s −2t
z = −1 + 3s −t
(a) Equating (x, y, z) to
¡
2, 2,
1
2
¢
in (13.3) we get
2s + 2t = 1
2s −2t = 1
3s −t =
3
2
yielding s =
1
2
and t = 0, so that the point with coordinates
¡
2, 2,
1
2
¢
belongs to the
plane.
(b) Similarly, equating (x, y, z) to
¡
4, 0,
3
2
¢
in (13.3) we get
2s + 2t = 3
2s −2t = −1
3s −t =
1
2
yielding s =
1
2
and t = 1, so that the point with coordinates
¡
4, 0,
3
2
¢
belongs to the
plane.
(c) Again, proceeding in the same way with the triple (−3, 1, −1), we get
2s + 2t = −4
2s −2t = 0
3s −t = −2
68 Applications of vector algebra to analytic geometry
yielding s = t = −1, so that the point with coordinates (−3, 1, −1) belongs to the
plane.
(d) The point of coordinates (3, 1, 5) does not belong to the plane, because the system
2s + 2t = 2
2s −2t = 0
3s −t = 4
is inconsistent (from the ﬁrst two equations we get s = t = 1, contradicting the third
equation).
(e) Finally, the point of coordinates (0, 0, 2) does not belong to the plane, because
the system
2s + 2t = −1
2s −2t = −1
3s −t = 1
is inconsistent (the ﬁrst two equations yield s = −
1
2
and t = 0, contradicting the
third equation).
13.8.2 n. 3 (p. 482)
(a)
x = 1 +t
y = 2 +s +t
z = 1 + 4t
(b)
u = (1, 2, 1) −(0, 1, 0) = (1, 1, 1)
v = (1, 1, 4) −(0, 1, 0) = (1, 0, 4)
x = s +t
y = 1 +s
z = s + 4t
Exercises 69
13.8.3 n. 4 (p. 482)
(a) Solving for s and t with the coordinates of the ﬁrst point, I get
I s −2t + 1 = 0
II s + 4t + 2 = 0
III 2s +t = 0
I +II −III t + 3 = 0
2III −I −II 2s −3 = 0
check on I
3
2
−6 + 1 6= 0
and (0, 0, 0) does not belong to the plane M. The second point is directly seen to
belong to M from the parametric equations
(1, 2, 0) +s (1, 1, 2) +t (−2, 4, 1) (13.4)
Finally, with the third point I get
I s −2t + 1 = 2
II s + 4t + 2 = −3
III 2s +t = −3
I +II −III t + 3 = 2
2III −I −II 2s −3 = −5
check on I −1 + 2 + 1 = 2
check on II −1 −4 + 2 = −3
check on III −2 −1 = −3
and (2, −3, −3) belongs to M.
(b) The answer has already been given by writing equation (13.4):
P ≡ (1, 2, 0) a = (1, 1, 2) b = (−2, 4, 1)
13.8.4 n. 5 (p. 482)
This exercise is a replica of material already presented in class and inserted in the
notes.
(a) If p +q +r = 1, then
pP + qQ +rR = P −(1 −p) P +qQ+rR
= P + (q +r) P +qQ +rR
= P +q (Q−P) +r (R −P)
and the coordinates of pP + qQ + rR satisfy the parametric equations of the plane
M through P, Q, and R..
(b) If S is a point of M, there exist real numbers q and r such that
S = P +q (Q−P) +r (R −P)
Then, deﬁning p ≡ 1 −(q +r),
S = (1 −q −r) P +qQ+rR
= pP +qQ +rR
70 Applications of vector algebra to analytic geometry
13.8.5 n. 6 (p. 482)
I use three diﬀerent methods for the three cases.
(a) The parametric equations for the ﬁrst plane π
1
are
I x = 2 + 3h −k
II y = 3 + 2h −2k
III z = 1 +h −3k
Eliminating the parameter h (I −II −III) I get
4k = x −y −z + 2
Eliminating the parameter k (I +II −III) I get
4h = x +y −z −4
Substituting in III (or, better, in 4 · III),
4z = 4 +x +y −z −4 −3x + 3y + 3z −6
I obtain a cartesian equation for π
1
x −2y +z + 3 = 0
(b) I consider the four coeﬃcients (a, b, c, d) of the cartesian equation of π
2
as un
known, and I require that the three given points belong to π
2
; I obtain the system
2a + 3b +c +d = 0
−2a −b −3c +d = 0
4a + 3b −c +d = 0
which I solve by elimination
2 3 1 1
−2 −1 −3 1
4 3 −1 1
¸
¸
µ
2
a
0
←
1
2
(
2
a
0
+
1
a
0
)
3
a
0
←
3
a
0
−2
1
a
0
¶
2 3 1 1
0 1 −1 1
0 −3 −3 −1
¸
¸
¡
3
a
0
←
1
2
(
3
a
0
+ 3
1
a
0
)
¢
2 3 1 1
0 1 −1 1
0 0 −3 1
¸
¸
Exercises 71
From the last equation (which reads −3c +d = 0) I assign values 1 and 3 to c and d,
respectively; from the second (which reads b −c +d = 0) I get b = −2, and from the
ﬁrst (which is unchanged) I get a = 1. The cartesian equation of π
2
is
x −2y +z + 3 = 0
(notice that π
1
and π
2
coincide)
(c) This method requires the vector product (called cross product by Apostol ), which
is presented in the subsequent section; however, I have thought it better to show its
application here already. The plane π
3
and the given plane being parallel, they have
the same normal direction. Such a normal is
n =(2, 0, −2) × (1, 1, 1) = (2, −4, 2)
Thus π
3
coincides with π
2
, since it has the same normal and has a common point
with it. At any rate (just to show how the method applies in general), a cartesian
equation for π
3
is
x −2y +z +d = 0
and the coeﬃcient d is determined by the requirement that π
3
contains (2, 3, 1)
2 −6 + 1 +d = 0
yielding
x −2y +z + 3 = 0
13.8.6 n. 7 (p. 482)
(a) Only the ﬁrst two points belong to the given plane.
(b) I assign the values −1 and 1 to y and z in the cartesian equation of the plane M,
in order to obtain the third coordinate of a point of M, and I obtain P = (1, −1, 1) (I
may have as well taken one of the ﬁrst two points of part a). Any two vectors which
are orthogonal to the normal n = (3, −5, 1) and form a linearly independent couple
can be chosen as direction vectors for M. Thus I assign values arbitrarily to d
2
and
d
3
in the orthogonality condition
3d
1
−5d
2
+d
3
= 0
say, (0, 3) and (3, 0), and I get
d
I
= (−1, 0, 3) d
II
= (5, 3, 0)
The parametric equations for M are
x = 1 −s + 5t
y = −1 + 3t
z = 1 + 3s
72 Applications of vector algebra to analytic geometry
13.8.7 n. 8 (p. 482)
The question is formulated in a slightly insidious way (perhaps Tom did it on pur
pose...), because the parametric equations of the two planes M and M
0
are written
using the same names for the two parameters. This may easily lead to an incorrect
attempt two solve the problem by setting up a system of three equations in the two
unknown s and t, which is likely to have no solutions, thereby suggesting the wrong
answer that M and M
0
are parallel. On the contrary, the equation system for the
coordinates of points in M ∩ M
0
has four unknowns:
I 1 + 2s −t = 2 +h + 3k
II 1 −s = 3 + 2h + 2k
III 1 + 3s + 2t = 1 + 3h +k
(13.5)
I take all the variables on the lefthand side, and the constant terms on the righthand
one, and I proceed by elimination:
s t h k  const.
2 −1 −1 −3  1
−1 0 −2 −2  2
3 2 −3 −1  0
¸
1
r
0
←
1
r
0
+ 2
2
r
0
3
r
0
←
3
r
0
+ 3
2
r
0
1
r
0
↔
2
r
0
¸
s t h k  const.
−1 0 −2 −2  2
0 −1 −5 −7  5
0 2 −9 −7  6
µ
3
r
0
←
3
r
0
+ 2
2
r
0
1
r
0
↔
2
r
0
¶
s t h k  const.
−1 0 −2 −2  2
0 −1 −5 −7  5
0 0 −19 −21  16
A handy solution to the last equation (which reads −19h − 21k = 16) is obtained
by assigning values 8 and −8 to h and k, respectively. This is already enough to
get the ﬁrst point from the parametric equation of M
0
, which is apparent (though
incomplete) from the righthand side of (13.5). Thus Q = (−14, 3, 17). However, just
to check on the computations, I proceed to ﬁnish up the elimination, which leads
to t = 11 from the second row, and to s = −2 from the ﬁrst. Substituting in the
lefthand side of (13.5), which is a trace of the parametric equations of M, I do get
indeed (−14, 3, 17). Another handy solution to the equation corresponding to the last
row of the ﬁnal elimination table is (h, k) =
¡
−
2
5
, −
2
5
¢
. This gives R =
¡
2
5
,
7
5
, −
3
5
¢
,
and the check by means of (s, t) =
¡
−
2
5
, −
1
5
¢
is all right.
13.8.8 n. 9 (p. 482)
(a) A normal vector for M is
n = (1, 2, 3) × (3, 2, 1) = (−4, 8, −4)
Exercises 73
whereas the coeﬃcient vector in the equation of M
0
is (1, −2, 1). Since the latter is
parallel to the former, and the coordinates of the point P ∈ M do not satisfy the
equation of M
0
, the two planes are parallel.
(b) A cartesian equation for M is
x −2y +z +d = 0
From substitution of the coordinates of P, the coeﬃcient d is seen to be equal to 3.
The coordinates of the points of the intersection line L ≡ M∩M
00
satisfy the system
½
x −2y +z + 3 = 0
x + 2y +z = 0
By sum and subtraction, the line L can be represented by the simpler system
½
2x + 2z = −3
4y = 3
as the intersection of two diﬀerent planes π and π
0
, with π parallel to the yaxis, and
π
0
parallel to the xzplane. Coordinate of points on L are now easy to produce, e.g.,
Q =
¡
−
3
2
,
3
4
, 0
¢
and R =
¡
0,
3
4
, −
3
2
¢
13.8.9 n. 10 (p. 483)
The parametric equations for the line L are
x = 1 + 2r
y = 1 −r
z = 1 + 3r
and the parametric equations for the plane M are
x = 1 + 2s
y = 1 +s +t
z = −2 + 3s +t
The coordinates of a point of intersection between L and M must satisfy the system
of equations
2r −2s = 0
r +s +t = 0
3r −3s −t = −3
The ﬁrst equation yields r = s, from which the third and second equation give t = 3,
r = s = −
3
2
. Thus L ∩ M consists of a single point, namely, the point of coordinates
¡
−2,
5
2
. −
7
2
¢
.
74 Applications of vector algebra to analytic geometry
13.8.10 n. 11 (p. 483)
The parametric equations for L are
x = 1 + 2t
y = 1 −t
z = 1 + 3t
and a direction vector for L is v =(2, −1, 3)
(a) L is parallel to the given plane (which I am going to denote π
a
) if v is a linear
combination of (2, 1, 3) and
¡
3
4
, 1, 1
¢
. Thus I consider the system
2x +
3
4
y = 2
x +y = −1
3x +y = 3
where subtraction of the second equation from the third gives 2x = 4, hence x = 2;
this yields y = −3 in the last two equations, contradicting the ﬁrst.
Alternatively, computing the determinant of the matrix having v, a, b as
columns
¯
¯
¯
¯
¯
¯
2 2
3
4
−1 1 1
3 3 1
¯
¯
¯
¯
¯
¯
=
¯
¯
¯
¯
¯
¯
4 0 −
5
4
−1 1 1
6 0 −2
¯
¯
¯
¯
¯
¯
=
¯
¯
¯
¯
4 −
5
4
6 −2
¯
¯
¯
¯
= −
1
2
shows that {v, a, b} is a linearly independent set.
With either reasoning, it is seen that L is not parallel to π
a
.
(b) By computing two independent directions for the plane π
b
, e.g.,
−→
PQ = (3, 5, 2) −
(1, 1, −2) and
−→
PR = (2, 4, −1) − (1, 1, −2), I am reduced to the previous case. The
system to be studied now is
2x +y = 2
4x + 3y = −1
4x +y = 3
and the three equations are again inconsistent, because subtraction of the third equa
tion from the second gives y = −2, whereas subtraction of the ﬁrst from the third
yields x =
1
2
, these two values contradicting all equations. L is not parallel to π
b
.
(c) Since a normal vector to π
c
has for components the coeﬃcients of the unknowns
in the equation of π
c
, it suﬃces to check orthogonality between v and (1, 2, 3):
h(2, −1, 3) , (1, 2, 3)i = 9 6= 0
L is not parallel to π
c
.
Exercises 75
13.8.11 n. 12 (p. 483)
Let R be any point of the given plane π, other than P or Q. A point S then belongs
to M if and only if there exists real numbers q and r such that
S = P +q (Q−P) +r (R −P) (13.6)
If S belongs to the line through P and Q, there exists a real number p such that
S = P +p (Q−P)
Then, by deﬁning
q ≡ p r ≡ 0
it is immediately seen that condition (13.6) is satisﬁed.
13.8.12 n. 13 (p. 483)
Since every point of L belongs to M, M contains the points having coordinates equal
to (1, 2, 3), (1, 2, 3) +t (1, 1, 1) for each t ∈ R, and (2, 3, 5). Choosing, e.g., t = 1, the
parametric equations for M are
x = 1 +r +s
y = 2 +r +s
z = 3 +r + 2s
It is possible to eliminate the two parameters at once, by subtracting the ﬁrst equaiton
from the second, obtaining
x −y + 1 = 0
13.8.13 n. 14 (p. 483)
Let d be a direction vector for the line L, and let Q ≡ (x
Q
, y
Q
, z
Q
) be a point of L.
A plane containing L and the point P ≡ (x
P
, y
P
, z
P
) is the plane π through Q with
direction vectors d and
−→
QP
x = x
Q
+hd
1
+k (x
P
−x
Q
)
y = y
Q
+hd
2
+k (y
P
−y
Q
)
z = z
Q
+hd
3
+k (z
P
−z
Q
)
L belongs to π because the parametric equation of π reduces to that of L when k is
assigned value 0, and P belongs to π because the coordinates of P are obtained from
the parametric equation of π when h is assigned value 0 and k is assigned value 1.
Since any two distinct points on L and P determine a unique plane, π is the only
plane containing L and P.
76 Applications of vector algebra to analytic geometry
13.9 The cross product
13.10 The cross product expressed as a determinant
13.11 Exercises
13.11.1 n. 1 (p. 487)
(a) A×B = −2i + 3j −k.
(b) B ×C = 4i −5j + 3k.
(c) C ×A = 4i −4j + 2k.
(d) A× (C ×A) = 8i + 10j + 4k.
(e) (A×B) ×C = 8i + 3j −7k.
(f) A× (B ×C) = 10i + 11j + 5k.
(g) (A×C) ×B = −2i −8j −12k.
(h) (A+B) × (A−C) = 2i −2j.
(i) (A×B) × (A×C) = −2i + 4k.
13.11.2 n. 2 (p. 487)
(a)
A×B
kA×Bk
= −
4
√
26
i +
3
√
26
j +
1
√
26
k or −
A×B
kA×Bk
=
4
√
26
i −
3
√
26
j −
1
√
26
k.
(b)
A×B
kA×Bk
= −
41
√
2054
i −
18
√
2054
j +
7
√
2054
k or −
A×B
kA×Bk
=
41
√
2054
i +
18
√
2054
j −
7
√
2054
k.
(c)
A×B
kA×Bk
= −
1
√
6
i −
2
√
6
j −
1
√
6
k or −
A×B
kA×Bk
=
1
√
6
i +
2
√
6
j +
1
√
6
k.
13.11.3 n. 3 (p. 487)
(a)
−→
AB ×
−→
AC = (2, −2, −3) × (3, 2, −2) = (10, −5, 10)
area ABC =
1
2
°
°
°
−→
AB ×
−→
AC
°
°
° =
15
2
(b)
−→
AB ×
−→
AC = (3, −6, 3) × (3, −1, 0) = (3, 9, 15)
area ABC =
1
2
°
°
°
−→
AB ×
−→
AC
°
°
° =
3
√
35
2
Exercises 77
(c)
−→
AB ×
−→
AC = (0, 1, 1) × (1, 0, 1) = (1, 1, −1)
area ABC =
1
2
°
°
°
−→
AB ×
−→
AC
°
°
° =
√
3
2
13.11.4 n. 4 (p. 487)
−→
CA×
−→
AB = (−i + 2j −3k) × (2j +k) = 8i −j −2k
13.11.5 n. 5 (p. 487)
Let a ≡ (l, m, n) and b ≡ (p, q, r) for the sake of notational simplicity. Then
ka ×bk
2
= (mr −nq)
2
+ (np −lr)
2
+ (lq −mp)
2
= m
2
r
2
+n
2
q
2
−2mnrq +n
2
p
2
+l
2
r
2
−2lnpr +l
2
q
2
+m
2
p
2
−2lmpq
kak
2
kbk
2
=
¡
l
2
+m
2
+n
2
¢ ¡
p
2
+q
2
+r
2
¢
= l
2
p
2
+l
2
q
2
+l
2
r
2
+m
2
p
2
+m
2
q
2
+m
2
r
2
+n
2
p
2
+n
2
q
2
+n
2
r
2
(ka ×bk = kak kbk) ⇔
¡
ka ×bk
2
= kak
2
kbk
2
¢
⇔ (lp +mq +nr)
2
= 0
⇔ ha, bi = 0
13.11.6 n. 6 (p. 487)
(a)
ha, b +ci = ha, b ×ai = 0
because b ×a is orthogonal to a.
(b)
hb, ci = hb, (b ×a) −bi = hb, (b ×a)i + hb, −bi = −kbk
2
because b ×a is orthogonal to b. Thus
hb, ci < 0 cos
c
bc = −
kbk
kck
< 0
c
bc ∈
i
π
2
, π
i
. Moreover,
c
bc = −π is impossible, because
kck
2
= kb ×ak
2
+ kbk
2
−2 kb ×ak kbk cos
\
(b ×a) b
= kb ×ak
2
+ kbk
2
> kbk
2
since (a, b) is linearly independent and kb ×ak > 0.
(c) By the formula above, kck
2
= 2
2
+ 1
2
, and kck =
√
5.
78 Applications of vector algebra to analytic geometry
13.11.7 n. 7 (p. 488)
(a) Since kak = kbk = 1 and ha, bi = 0, by Lagrange’s identity (theorem 13.12.f)
ka ×bk
2
= kak
2
kbk
2
−ha, bi
2
= 1
so that a × b is a unit vector as well. The three vectors a, b, a × b are mutually
orthogonal either by assumption or by the properties of the vector product (theorem
13.12.de).
(b) By Lagrange’s identity again,
kck
2
= ka ×bk
2
kak
2
−ha ×b, ai
2
= 1
(c) And again,
k(a ×b) ×bk
2
= ka ×bk
2
kbk
2
−ha ×b, bi
2
= 1
k(a ×b) ×ak
2
= kck
2
= 1
Since the direction which is orthogonal to (a ×b) and to b is spanned by a, (a ×b)×b
is either equal to a or to −a. Similarly, (a ×b) × a is either equal to b or to −b.
Both the righthand rule and the lefthand rule yield now
(a ×b) ×a = b (a ×b) ×b = −a
(d)
(a ×b) ×a =
¸
(a
3
b
1
−a
1
b
3
) a
3
−(a
1
b
2
−a
2
b
1
) a
2
(a
1
b
2
−a
2
b
1
) a
1
−(a
2
b
3
−a
3
b
2
) a
3
(a
2
b
3
−a
3
b
2
) a
2
−(a
3
b
1
−a
1
b
3
) a
1
¸
=
¸
(a
2
2
+a
2
3
) b
1
−a
1
(a
3
b
3
+a
2
b
2
)
(a
2
1
+a
2
3
) b
2
−a
2
(a
1
b
1
+a
3
b
3
)
(a
2
1
+a
2
2
) b
3
−a
3
(a
1
b
1
+a
2
b
2
)
¸
Since a and b are orthogonal,
(a ×b) ×a =
¸
(a
2
2
+a
2
3
) b
1
+a
1
(a
1
b
1
)
(a
2
1
+a
2
3
) b
2
+a
2
(a
2
b
2
)
(a
2
1
+a
2
2
) b
3
+a
3
(a
3
b
3
)
¸
Since a is a unit vector,
(a ×b) ×a =
¸
b
1
b
2
b
3
¸
The proof that (a ×b) ×b = −a is identical.
Exercises 79
13.11.8 n. 8 (p. 488)
(a) From a ×b = 0, either there exists some h ∈ R such that a = hb, or there exists
some k ∈ R such that b = ka. Then either ha, bi = hkbk
2
or ha, bi = k kak
2
, that
is, either hkbk = 0 or k kak = 0. In the ﬁrst case, either h = 0 (which means a = 0),
or kbk = 0 (which is equivalent to b = 0). In the second case, either k = 0 (which
means b = 0), or kak = 0 (which is equivalent to a = 0). Geometrically, suppose
that a 6= 0. Then both the projection
ha,bi
kak
2
a of b along a and the projecting vector
(of length
kb×ak
kak
) are null, which can only happen if b = 0.
(b) From the previous point, the hypotheses a 6= 0, a×(b −c) = 0, and ha, b −ci =
0 imply that b −c = 0.
13.11.9 n. 9 (p. 488)
(a) Let ϑ ≡
c
ab, and observe that a and c are orthogonal. Froma×b = kak kbk sin ϑ,
in order to satisfy the condition
a ×b = c
the vector b must be orthogonal to c, and its norm must depend on ϑ according to
the relation
kbk =
kck
kak sin ϑ
(13.7)
Thus
kck
kak
≤ kbk < +∞. In particular, b can be taken orthogonal to a as well, in
which case it has to be equal to
a×c
kak
2
or to
c×a
kak
2
. Thus two solutions to the problem
are
±
µ
7
9
, −
8
9
, −
11
9
¶
(b) Let (p, q, r) be the coordinates of b; the conditions a ×b = c and ha, bi = 1 are:
I −2q −r = 3
II 2p −2r = 4
III p + 2q = −1
IV 2p −q + 2r = 1
Standard manipulations yield
2II + 2IV +III 9p = 9
(↑) ,→II 2 −2r = 4
(↑) ,→I −2q + 1 = 3
check on IV 2 + 1 −2 = 1
check on III 1 −2 = 1
that is, the unique solution is
b =(1, −1, −1)
The solution to this exercise given by Apostol at page 645 is wrong.
80 Applications of vector algebra to analytic geometry
13.11.10 n. 10 (p. 488)
Replacing b with c in the ﬁrst result of exercise7,
(a ×c) ×a = c
Therefore, by skewsimmetry of the vector product, it is seen that the c × a is a
solution to the equation
a ×x = c (13.8)
However, c ×a is orthogonal to a, and hence it does not meet the additional require
ment
ha, xi = 1 (13.9)
We are bound to look for other solutions to (13.8). If x and z both solve (13.8),
a × (x −z) = a ×x −a ×z = c −c = 0
which implies that x − z is parallel to a. Thus the set of all solutions to equation
(13.8) is
©
x ∈ R
3
: ∃α ∈ R, x = a ×c +αa
ª
From condition (13.9),
ha, a ×c +αai = 1
that is,
α = −
1
kak
2
Thus the unique solution to the problem is
b ≡ a ×c−
a
kak
2
13.11.11 n. 11 (p. 488)
(a) Let
u ≡
−→
AB = (−2, 1, 0)
v ≡
−→
BC = (3, −2, 1)
w ≡
−→
CA = (−1, 1, −1)
Each side of the triangle ABC can be one of the two diagonals of the parallelogram
to be determined.
Exercises 81
If one of the diagonals is BC, the other is AD, where
−→
AD =
−→
AB +
−→
AC = u −w. In
this case
D = A+u −w = “B +C −A” = (0, 0, 2)
If one of the diagonals is CA, the other is BE, where
−→
BE =
−→
BC +
−→
BA = v −u. In
this case
E = B +v −u = “A+C −B” = (4, −2, 2)
If one of the diagonals is AB, the other is CF, where
−→
CF =
−→
CA +
−→
CB = −v + w.
In this case
F = A−v +w = “A +B −C” = (−2, 2, 0)
(b)
u ×w =
µ¯
¯
¯
¯
1 0
1 −1
¯
¯
¯
¯
, −
¯
¯
¯
¯
−2 0
−1 −1
¯
¯
¯
¯
,
¯
¯
¯
¯
−2 1
−1 1
¯
¯
¯
¯
¶
= (−1, −2, −1)
ku ×wk =
√
6 = area (ABC) =
√
6
2
82 Applications of vector algebra to analytic geometry
13.11.12 n. 12 (p. 488)
b +c = 2 (a ×b) −2b
ha, b +ci = −2 ha, bi = −4
cos
c
ab =
ha, bi
kak kbk
=
1
2
ka ×bk
2
= kak
2
kbk
2
sin
2
c
ab = 12
kck
2
= 4 ka ×bk
2
−12 ha ×b, bi + 9 kbk
2
= 192
kck = 8
√
3
hb, ci = 2 hb, a ×bi −3 kbk
2
= 48
cos
c
bc =
hb, ci
kbk kck
=
√
3
2
13.11.13 n. 13 (p. 488)
(a) It is true that, if the couple (a, b) is linearly independent, then the triple
(a +b, a −b, a ×b)
is linearly independent, too. Indeed, in the ﬁrst place a and b are nonnull, since every
ntuple having the null vector among its components is linearly dependent. Secondly,
(a ×b) is nonnull, too, because (a, b) is linearly independent. Third, (a ×b) is
orthogonal to both a + b and a −b (as well as to any vector in lin {a, b}), because
it is orthogonal to a and b; in particular, (a ×b) / ∈ lin {a.b}, since the only vector
which is orthogonal to itself is the null vector. Fourth, suppose that
x(a +b) +y (a −b) +z (a ×b) = 0 (13.10)
Then z cannot be diﬀerent from zero, since otherwise
(a ×b) = −
x
z
(a +b) −
y
z
(a −b)
= −
x +y
z
a +
y −x
z
b
would belong to lin{a, b}. Thus z = 0, and (13.10) becomes
x(a +b) +y (a −b) = 0
which is equivalent to
(x +y) a+(x −y) b = 0
By linear independence of (a, b),
x +y = 0 and x −y = 0
Exercises 83
and hence x = y = 0.
(b) It is true that, if the couple (a, b) is linearly independent, then the triple
(a +b, a +a ×b, b +a ×b)
is linearly independent, too. Indeed, let (x, y, z) be such that
x (a +b) +y (a +a ×b) +z (b +a ×b) = 0
which is equivalent to
(x +y) a + (x +z) b + (y +z) a ×b = 0
Since, arguing as in the fourth part of the previous point, y + z must be null, and
this in turn implies that both x +y and x +z are also null. The system
x + y = 0
x +z = 0
y +z = 0
has the trivial solution as the only solution.
(c) It is true that, if the couple (a, b) is linearly independent, then the triple
(a, b, (a + b) × (a −b))
is linearly independent, too. Indeed,
(a +b) × (a −b) = a ×a −a ×b +b ×a −b ×b
= −2a ×b
and the triple
(a, b, a ×b)
is linearly independent when the couple (a, b) is so.
13.11.14 n. 14 (p. 488)
(a) The cross product
−→
AB ×
−→
AC equals the null vector if and only if the couple
³
−→
AB,
−→
AC
´
is linearly dependent. In such a case, if
−→
AC is null, then A and C coin
cide and it is clear that the line through them and B contains all the three points.
Otherwise, if
−→
AC is nonnull, then in the nontrivial null combination
x
−→
AB +y
−→
AC = 0
x must be nonnull, yielding
−→
AB = −
y
x
−→
AC
84 Applications of vector algebra to analytic geometry
that is,
B = A−
y
x
−→
AC
which means that B belongs to the line through A and C.
(b) By the previous point, the set
n
P :
−→
AP ×
−→
BP = 0
o
is the set of all points P such that A, B, and P belong to the same line, that is, the
set of all points P belonging to the line through A and B.
13.11.15 n. 15 (p. 488)
(a) From the assumption (p ×b) +p = a,
hb, p ×bi + hb, pi = hb, ai
hb, pi = 0
that is, p is orthogonal to b. This gives
kp ×bk = kpk kbk = kpk (13.11)
and hence
1 = kak
2
= kp ×bk
2
+ kpk
2
= 2 kpk
2
that is, kpk =
√
2
2
.
(b) Since p, b, and p×b are pairwise orthogonal, (p, b, p ×b) is linearly independent,
and {p, b, p ×b} is a basis of R
3
.
(c) Since p is orthogonal both to p×b (deﬁnition of vector product) and to b (point
a), there exists some h ∈ R such that
(p ×b) ×b = hp
Thus, taking into account (13.11),
h kpk = kp ×bk kbk
¯
¯
¯sin
π
2
¯
¯
¯ = kpk
and h ∈ {1, −1}. Since the triples (p, b, p ×b) and (p ×b, p, b) deﬁne the same
orientation, and (p ×b, b, p) deﬁnes the opposite one, h = −1.
(d) Still from the assumption (p ×b) +p = a,
hp, p ×bi + hp, pi = hp, ai
hp, ai = kpk
2
=
1
2
p × (p ×b) +p ×p = p ×a
kp ×ak = kpk
2
kbk =
1
2
The scalar triple product 85
Thus
p =
1
2
a +
1
2
q (13.12)
where q = 2p−a is a vector in the plane π generated by p and a, which is orthogonal
to a. Since a normal vector to π is b, there exists some k ∈ R such that
q = k (b ×a)
Now
kqk
2
= 4 kpk
2
+ kak
2
−4 kpk kak cos c pa
= 2 + 1 −4
√
2
2
√
2
2
= 1
kqk = 1 k = 1
If on the plane orthogonal to b the mapping u 7→b×u rotates counterclockwise, the
vectors a = p + b × p, b × p, and b × a form angles of
π
4
,
π
2
, and
3π
4
, respectively,
with p. On the other hand, the decomposition of p obtained in (13.12) requires that
the angle formed with p by q is −
π
4
, so that q is discordant with b × a. It follows
that k = −1. I have ﬁnally obtained
p =
1
2
a −
1
2
(b ×a)
13.12 The scalar triple product
13.13 Cramer’s rule for solving systems of three linear equations
13.14 Exercises
13.15 Normal vectors to planes
13.16 Linear cartesian equations for planes
86 Applications of vector algebra to analytic geometry
13.17 Exercises
13.17.1 n. 1 (p. 496)
(a) Out of the inifnitely many vectors satisfying the requirement, a distinguished one
is
n ≡ (2i + 3j −4k) × (j +k) = 7i −2j + 2k
(b)
hn, (x, y, z)i = 0 or 7x −2y + 2z = 0
(c)
hn, (x, y, z)i = hn, (1, 2, 3)i or 7x −2y + 2z = 9
13.17.2 n. 2 (p. 496)
(a)
n
knk
=
µ
1
3
,
2
3
, −
2
3
¶
(b) The three intersection points are:
Xaxis : (−7, 0, 0) Y axis :
µ
0, −
7
2
, 0
¶
Zaxis :
µ
0, 0,
7
2
¶
.
(c) The distance from the origin is
7
3
.
(d) Intersecting with π the line through the origin which is directed by the normal
to π
x = h
y = 2h
z = −2h
x + 2y −2z + 7 = 0
yields
h + 4h + 4h + 7 = 0
h = −
7
9
(x, y, z) =
µ
−
7
9
, −
14
9
,
14
9
¶
Exercises 87
13.17.3 n. 3 (p. 496)
A cartesian equation for the plane which passes through the point P ≡ (1, 2, −3) and
is parallel to the plane of equation
3x −y + 2z = 4
is
3 (x −1) −(y −2) + 2 (z + 3) = 0
or
3x −y + 2z + 5 = 0
The distance between the two planes is
4 −(−5)
√
9 + 1 + 4
=
9
√
14
14
13.17.4 n. 4 (p. 496)
π
1
: x + 2y −2z = 5
π
2
: 3x −6y + 3z = 2
π
3
: 2x +y + 2z = −1
π
4
: x −2y +z = 7
(a) π
2
and π
4
are parallel because n
2
= 3n
4
; π
1
and π
3
are orthogonal because
hn
1
, n
3
i = 0.
(b) The straight line through the origin having direction vector n
4
has equations
x = t
y = −2t
z = t
and intersects π
2
and π
4
at points C and D, respectively, which can be determined
by the equations
3t
C
−6 (−2t
C
) + 3t
C
= 2
t
D
−2 (−2t
D
) +t
D
= 7
Thus t
C
=
1
9
, t
D
=
7
6
,
−→
CD =
¡
7
6
−
1
9
¢
n
4
, and
°
°
°
−→
CD
°
°
° =
19
18
kn
4
k =
19
18
√
6
Alternatively, rewriting the equation of π
4
with normal vector n
2
3x −6y + 3z = 21
dist (π, π
0
) =
d
2
−d
4

kn
2
k
=
19
√
54
88 Applications of vector algebra to analytic geometry
13.17.5 n. 5 (p. 496)
(a) A normal vector for the plane π through the points P ≡ (1, 1, −1), Q ≡ (3, 3, 2),
R ≡ (3, −1, −2) is
n ≡
−→
PQ×
−→
QR = (2, 2, 3) × (0, −4, −4) = (4, 8, −8)
(b) A cartesian equation for π is
x + 2y −2z = 5
(c) The distance of π from the origin is
5
3
.
13.17.6 n. 6 (p. 496)
Proceeding as in points (a) and (b) of the previous exercise, a normal vector for the
plane through the points P ≡ (1, 2, 3), Q ≡ (2, 3, 4), R ≡ (−1, 7, −2) is
n ≡
−→
PQ×
−→
RQ = (1, 1, 1) × (3, −4, 6) = (10, −3, −7)
and a cartesian equation for it is
10x −3y −7z + 17 = 0
13.17.7 n. 8 (p. 496)
A normal vector to the plane is given by any direction vector for the given line
n ≡ (2, 4, 12) −(1, 2, 3) = (1, 2, 9)
A cartesian equation for the plane is
(x −2) + 2 (y −3) + 9 (z + 7) = 0
or
x + 2y + 9z = 55
13.17.8 n. 9 (p. 496)
A direction vector for the line is just the normal vector of the plane. Thus the
parametric equations are
x = 2 + 4h y = 1 −3h z = −3 +h
Exercises 89
13.17.9 n. 10 (p. 496)
(a) The position of the point at time t can be written as follows:
x = 1 −t y = 2 −3t z = −1 + 2t
which are just the parametric equations of a line.
(b) The direction vector of the line L is
d = (−1, −3, 2)
(c) The hitting instant is determined by substituting the coordinates of the moving
point into the equation of the given plane π
2 (1 −t) + 3 (2 −3t) + 2 (−1 + 2t) + 1 = 0
−7t + 7 = 0
yielding t = 1. Hence the hitting point is (0, −1, 1).
(d) At time t = 3, the moving point has coordinates (−2, −7, 5). Substituting,
2 · (−2) + 3 · (−7) + 2 · 5 +d = 0
d = 15
A cartesian equation for the plane π
0
which is parallel to π and contains (−2, −7, 5)
is
2x + 3y + 2z + 15 = 0
(e) At time t = 2, the moving point has coordinates (−1, −4, 3). A normal vector for
the plane π
00
which is orthogonal to L and contains (−1, −4, 3) is −d. Thus
−1 · −1 −3 · −4 + 2 · 3 +d = 0
d = −19
A cartesian equation for π
00
is
(x + 1) + 3 (y + 4) −2 (z −3) = 0
or
x + 3y −2z + 19 = 0
90 Applications of vector algebra to analytic geometry
13.17.10 n. 11 (p. 496)
From
c
ni =
π
3
c
nj =
π
4
c
nk =
π
3
we get
n
knk
=
1
2
³
1,
√
2, 1
´
A cartesian equation for the plane in consideration is
(x −1) +
√
2 (y −1) + (z −1) = 0
or
x +
√
2y +z = 2 +
√
2
13.17.11 n. 13 (p. 496)
First I ﬁnd a vector of arbitrary norm which satisﬁes the given conditions:
l + 2m−3n = 0
l −m+ 5n = 0
Assigning value 1 to n, the other values are easily obtained: (l, m) =
¡
−
7
3
,
8
3
¢
. Then
the required vector is
³
−
7
√
122
,
8
√
122
,
3
√
122
´
13.17.12 n. 14 (p. 496)
A normal vector for the plane π which is parallel to both vectors i +j and j +k is
n ≡ (i +j) × (j +k) = i −j +k
Since the intercept of π with the Xaxis is (2, 0, 0), a cartesian equation for π is
x −y +z = 2
13.17.13 n. 15 (p. 496)
I work directly on the coeﬃcient matrix for the equation system:
3 1 1  5
3 1 5  7
1 −1 3  3
With obvious manipulations
0 4 −8  −4
0 0 4  2
1 −1 3  3
I obtain a unique solution (x, y, z) =
¡
3
2
, 0,
1
2
¢
.
The conic sections 91
13.17.14 n. 17 (p. 497)
If the line ` under consideration is parallel to the two given planes, which have as
normal vectors n
1
≡ (1, 2, 3) and n
2
≡ (2, 3, 4), a direction vector for ` is
d ≡ (2, 3, 4) × (1, 2, 3) = (1, −2, 1)
Since ` goes through the point P ≡ (1, 2, 3), parametric eqautions for ` are the
following:
x = 1 +t y = 2 −2t z = 3 +t
13.17.15 n. 20 (p. 497)
A cartesian equation for the plane under consideration is
2x −y + 2z +d = 0
The condition of equal distance from the point P ≡ (3, 2, −1) yields
6 −2 −2 +d
3
=
6 −2 −2 + 4
3
that is,
2 +d = 6
The two solutions of the above equation are d
1
= 4 (corresponding to the plane
already given) and d
2
= −8. Thus the required equation is
2x −y + 2z = 8
13.18 The conic sections
13.19 Eccentricity of conic sections
13.20 Polar equations for conic sections
13.21 Exercises
92 Applications of vector algebra to analytic geometry
13.22 Conic sections symmetric about the origin
13.23 Cartesian equations for the conic sections
13.24 Exercises
13.25 Miscellaneous exercises on conic sections
Chapter 14
CALCULUS OF VECTORVALUED FUNCTIONS
94 Calculus of vectorvalued functions
Chapter 15
LINEAR SPACES
15.1 Introduction
15.2 The deﬁnition of a linear space
15.3 Examples of linear spaces
15.4 Elementary consequences of the axioms
15.5 Exercises
15.5.1 n. 1 (p. 555)
The set of all real rational functions is a real linear space. Indeed, let P, Q, R, and
S be any four real polynomials, and let
f : x 7→
P (x)
Q(x)
g : x 7→
R(x)
S (x)
Then for every two real numbers α and β
αf +βg : x 7→
αP (x) S (x) +βQ(x) R(x)
Q(x) S (x)
is a well deﬁned real rational function. This shows that the set in question is closed
with respect to the two linear space operations of function sum and function multipli
cation by a scalar. From this the other two existence axioms (of the zero element and
of negatives) also follow as particular cases, for (α, β) = (0, 0) and for (α, β) = (0, −1)
respectively. Of course it can also be seen directly that the identically null function
x 7→ 0 is a rational function, as the quotient of the constant polynomials P : x 7→ 0
96 Linear spaces
and Q : x 7→1. The remaining linear space axioms are immediately seen to hold, tak
ing into account the general properties of the operations of function sum and function
multiplication by a scalar
15.5.2 n. 2 (p. 555)
The set of all real rational functions having numerator of degree not exceeding the
degree of the denominator is a real linear space. Indeed, taking into account exercise
1, and using the same notation, it only needs to be proved that if deg P ≤ deg Q and
deg R ≤ deg S, then
deg [αPS +βQR] ≤ deg QS
This is clear, since for every two real numbers α and β
deg [αPS +βQR] ≤ max {deg PS, deg QR}
deg PS = deg P deg S ≤ deg Qdeg S
deg QR = deg Qdeg R ≤ deg Qdeg S
so that the closure axioms hold. It may be also noticed that the degree of both
polynomials P and Q occurring in the representation of the identically null function
as a rational function is zero.
15.5.3 n. 3 (p. 555)
The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0
and 1, and which have the same value at 0 and 1, is a real linear space. Indeed, for
every two real numbers α and β,
(αf +βg) (0) = αf (0) +βg (0) = αf (1) +βg (1) = (αf +βg) (1)
so that the closure axioms hold. Again, the two existence axioms follow as particular
cases; and it is anyway clear that the identically null function x 7→0 achieves the same
value at 0 and 1. Similarly, that the other linear space axioms hold is a straightforward
consequence of the general properties of the operations of function sum and function
multiplication by a scalar.
15.5.4 n. 4 (p. 555)
The set of all real valued functions which are deﬁned on a ﬁxed domain containing
0 and 1, and which achieve at 0 the double value they achieve at 1 is a real linear
space. Indeed, for every two real numbers α and β,
(αf +βg) (0) = αf (0) +βg (0) = α2f (1) +β2g (1) = 2 (αf +βg) (1)
so that the closure axioms hold. A ﬁnal remark concerning the other axioms, of a
type which has become usual at this point, allows to conclude.
Exercises 97
15.5.5 n. 5 (p. 555)
The set of all real valued functions which are deﬁned on a ﬁxed domain containing
0 and 1, and which have value at 1 which exceeds the value at 0 by 1, is not a real
linear space. Indeed, let α and β be any two real numbers such that α+β 6= 1. Then
(αf +βg) (1) = αf (1) +βg (1) = α[1 +f (0)] +β [1 +g (0)]
= α +β +αf (0) +βg (0) = α +β + (αf +βg) (0)
6= 1 + (αf +βg) (0)
and the two closure axioms fail to hold (the above shows, however, that the set in
question is an aﬃne subspace of the real linear space). The other failing axioms are:
existence of the zero element (the identically null function has the same value at 0
and at 1), and existence of negatives
f (1) = 1 +f (0) ⇒(−f) (1) = −1 + (−f) (0) 6= 1 + (−f) (0)
15.5.6 n. 6 (p. 555)
The set of all real valued step functions which are deﬁned on [0, 1] is a real linear
space. Indeed, let f and g be any two such functions, so that for some nonnegative
integers m and n, some increasing (m+ 1)tuple (σ
r
)
r∈{0}∪m
of elements of [0, 1] with
σ
0
= 0 and σ
m
= 1, some increasing (n + 1)tuple (τ
s
)
s∈{0}∪n
of elements of [0, 1]
with τ
0
= 0 and τ
m
= 1, some (m+ 1)tuple (γ
r
)
r∈{0}∪m
of real numbers, and some
ntuple (δ
s
)
s∈{0}∪n
of real numbers
∗
,
f ≡ γ
0
χ
{0}
+
X
r∈m
γ
r
χ
Ir
g ≡ δ
0
χ
{0}
+
X
s∈n
δ
s
χ
Js
where for each subset C of [0, 1], χ
C
is the characteristic function of C
χ
C
≡ x 7→
¿
1 if x ∈ C
0 if x / ∈ C
and (I
r
)
r∈m
, (J
s
)
s∈n
are the partitions of (0, 1] associated to (σ
r
)
r∈{0}∪m
, (τ
s
)
s∈{0}∪n
I
r
≡ (σ
r−1
, σ
r
] (r ∈ m)
J
s
≡ (τ
s−1
, τ
s
] (s ∈ n)
Then, for any two real numbers α and β,
αf + βg = (αγ
0
+βδ
0
) χ
{0}
+
X
(r,s)∈m×n
(αγ
r
+βδ
s
) χ
K
rs
∗
I am using here the following slight abuse of notation for degenerate intervals: (σ, σ] ≡ {σ};
This is necessary in order to allow for “point steps”.
98 Linear spaces
where (K
rs
)
(r,s)∈m×n
is the “meet” partition of (I
r
)
r∈m
and (J
s
)
s∈n
K
rs
≡ I
r
∩ J
s
( (r, s) ∈ m×n)
which shows that αf +βg is a step function too, with no more
†
than m+n steps on
[0, 1].
Thus the closure axioms hold, and a ﬁnal, usual remark concenrning the other
linear space axioms applies. It may also be noticed independently that the identically
null function [0, 1] 7→R, x 7→0 is indeed a step function, with just one step (m = 1,
γ
0
= γ
1
= 0), and that the opposite of a step function is a step function too, with
the same number of steps.
15.5.7 n. 7 (p. 555)
The set of all real valued functions which are deﬁned on R and convergent to 0 at
+∞ is a real linear space. Indeed, by classical theorems in the theory of limits, for
every two such functions f and g, and for any two real numbers α and β,
lim
x→+∞
(αf +βg) (x) = α lim
x→+∞
f (x) + β lim
x→+∞
g (x)
= 0
so that the closure axioms hold. Final remark concerning the other linear space
axioms. It may be noticed independently that the identically null function [0, 1] 7→
R, x 7→0 indeed converges to 0 at +∞(and, for that matter, at any x
0
∈ R∪{−∞}).
15.5.8 n. 11 (p. 555)
The set of all real valued and increasing functions of a real variable is not a real linear
space. The ﬁrst closure axiom holds, because the sum of two increasing functions is an
increasing function too. The second closure axiom does not hold, however, because
the function αf is decreasing if f is increasing and α is a negative real number.
The axiom of existence of the zero element holds or fails, depending on whether
monotonicity is meant in the weak or in the strict sense, since the identically null
function is weakly increasing (and, for that matter, weakly decreasing too, as every
constant function) but not strictly increasing. Thus, in the former case, the set of all
real valued and (weakly) increasing functions of a real variable is a convex cone. The
axiom of existence of negatives fails, too. All the other linear space axioms hold, as
in the previous examples.
15.5.9 n. 13 (p. 555)
The set of all real valued functions which are deﬁned and integrable on [0, 1], with
the integral over [0, 1] equal to zero, is a real linear space. Indeed, for any two such
functions f and g, and any two real numbers α and β,
Z
1
0
(αf +βg) (x) dx = α
Z
1
0
f (x) dx +β
Z
1
0
g (x) dx = 0
†
Some potential steps may “collapse” if, for some (r, s) ∈ m×n, σ
r−1
= τ
s−1
and ασ
r
+βτ
s
=
ασ
r−1
+βτ
s−1
.
Exercises 99
15.5.10 n. 14 (p. 555)
The set of all real valued functions which are deﬁned and integrable on [0, 1], with
nonnegative integral over [0, 1], is a convex cone, but not a linear space. For any two
such functions f and g, and any two nonnegative real numbers α and β,
Z
1
0
(αf +βg) (x) dx = α
Z
1
0
f (x) dx +β
Z
1
0
g (x) dx ≥ 0
It is clear that if α < 0 and
R
1
0
(αf) (x) dx > 0, then
R
1
0
(αf) (x) dx < 0, so that
axioms 2 and 6 fail to hold.
15.5.11 n. 16 (p. 555)
First solution. The set of all real Taylor polynomials of degree less than or equal to
n (including the zero polynomial) is a real linear space, since it coincides with the set
of all real polynomials of degree less than or equal to n, which is already known to be
a real linear space. The discussion of the above statement is a bit complicated by the
fact that nothing is said concerning the point where our Taylor polynomials are to be
centered. If the center is taken to be 0, the issue is really easy: every polynomial is
the Taylor polynomial centered at 0 of itself (considered, as it is, as a real function of
a real variable). This is immediate, if one thinks at the motivation for the deﬁnition
of Taylor polynomials: best ndegree polynomial approximation of a given function.
At any rate, if one takes the standard formula as a deﬁnition
Taylor
n
(f) at 0 ≡ x 7→
n
X
j=0
f
(j)
(0)
j!
x
j
here are the computations. Let
P : x 7→
n
X
i=0
p
i
x
n−i
Then
P
0
: x 7→
P
n−1
i=0
(n −i) p
i
x
n−1−i
P
0
(0) = p
n−1
P
00
: x 7→
P
n−2
i=0
(n −i) (n −(i + 1)) p
i
x
n−2−i
P
00
(0) = 2p
n−2
.
.
.
.
.
.
.
.
.
P
(j)
: x 7→
P
n−j
i=0
(n−i)!
(n−i−j)!
p
i
x
n−j−i
P
(j)
(0) = j!p
n−j
.
.
.
.
.
.
.
.
.
P
(n−1)
: x 7→n!p
0
x + (n −1)!p
1
P
(n−1)
(0) = (n −1)!p
1
P
(n)
: x 7→n!p
0
P
(n)
(0) = n!p
0
100 Linear spaces
and hence
n
X
j=0
P
(j)
(0)
j!
x
j
=
n
X
j=0
j!p
n−j
j!
x
j
=
n
X
i=0
p
i
x
n−i
= P (x)
On the other hand, if the Taylor polynomials are meant to be centered at some x
0
6= 0
Taylor
n
(f) at x
0
≡ x 7→
n
X
j=0
f
(j)
(x
0
)
j!
(x −x
0
)
j
it must be shown by more lengthy arguments that for each polynomial P the following
holds:
n
X
j=0
P
(j)
(x
0
)
j!
(x −x
0
)
j
=
n
X
i=0
p
i
x
n−i
Second solution. The set of all real Taylor polynomials (centered at x
0
∈ R)
of degree less than or equal to n (including the zero polynomial) is a real linear space.
Indeed, let P and Q be any two such polynomials; that is, let f and g be two real
functions of a real variable which are m and n times diﬀerentiable in x
0
, and let
P ≡Taylor
m
(f) at x
0
, Q ≡Taylor
n
(g) at x
0
, that is,
P : x 7→
m
X
i=0
f
(i)
(x
0
)
i!
(x −x
0
)
i
Q : x 7→
n
X
j=0
g
(j)
(x
0
)
j!
(x −x
0
)
j
Suppose ﬁrst that m = n. Then for any two real numbers α andβ
αP +βQ =
m
X
k=0
½
α
·
f
(k)
(x
0
)
k!
¸
+β
·
g
(k)
(x
0
)
k!
¸¾
(x −x
0
)
k
=
m
X
k=0
(αf +βg)
(k)
(x
0
)
k!
(x −x
0
)
k
= Taylor
m
(αf +βg) at x
0
Second, suppose (without loss of generality) that m > n. In this case, however,
αP +βQ =
n
X
k=0
αf
(k)
(x
0
) +βg
(k)
(x
0
)
k!
(x −x
0
)
k
+
m
X
k=n+1
αf
(k)
(x
0
)
k!
(x −x
0
)
k
a polynomial which can be legitimately considered the Taylor polynomial of degree
n at x
0
of the function αf + βg only if all the derivatives of g of order from n + 1
to m are null at x
0
. This is certainly true if g itself a polynomial of degree n. In
fact, this is true only in such a case, as it can be seen by repeated integration. It is
hence necessary, in addition, to state and prove the result asserting that each Taylor
polynomial of any degree and centered at any x
0
∈ R can be seen as the Taylor
polynomial of itself.
Exercises 101
15.5.12 n. 17 (p. 555)
The set S of all solutions of a linear secondorder homogeneous diﬀerential equation
∀x ∈ (a, b) , y
00
(x) +P (x) y
0
(x) +Q(x) y (x) = 0
where P and Q are given everywhere continuous real functions of a real variable,
and (a, b) is some open interval to be determined together with the solution, is a real
linear space. First of all, it must be noticed that S is nonempty, and that its elements
are indeed real functions which are everywhere deﬁned (that is, with (a, b) = R), due
to the main existence and solution continuation theorems in the theory of diﬀerential
equations. Second, the operator
L : D
2
→R
R
, y 7→y
00
+Py
0
+Qy
where R
R
is the set of all the real functions of a real variable, and D
2
is the subset
of R
R
of all the functions having a second derivative which is everywhere deﬁned, is
linear:
∀y ∈ D
2
, ∀z ∈ D
2
, ∀α ∈ R, ∀β ∈ R,
L(αy +βz) = (αy +βz)
00
+P (αy +βz)
0
+Q(αy +βz)
= α(y
00
+Py
0
+Qy) +β (z
00
+Pz
0
+Qz)
= αL(y) +βL(z)
Third, D
2
is a real linear space, since
∀y ∈ D
2
, ∀z ∈ D
2
, ∀α ∈ R, ∀β ∈ R,
αy +βz ∈ D
2
by standard theorems on the derivative of a linear combination of diﬀerentiable
functions, and the usual remark concerning the other linear space axioms. Finally,
S =ker L is a linear subspace of D
2
, by the following standard argument:
L(y) = 0 ∧ L(z) = 0 ⇒L(αy +βz) = αL(y) +βL(z) = 0
and the usual remark.
15.5.13 n. 18 (p. 555)
The set of all bounded real sequences is a real linear space. Indeed, for every two real
sequences x ≡ (x
n
)
n∈N
and y ≡ (y
n
)
n∈N
and every two real numbers α and β, if x
and y are bounded, so that for some positive numbers ε and η and for every n ∈ N
the following holds:
x
n
 < ε y
n
 < η
then the sequence αx +βy is also bounded, since for every n ∈ N
αx
n
+βy
n
 ≤ α x
n
 + β y
n
 < α ε + β η
Thus the closure axioms hold, and the usual remark concerning the other linear space
axioms applies.
102 Linear spaces
15.5.14 n. 19 (p. 555)
The set of all convergent real sequences is a real linear space. Indeed, for every two
real sequences x ≡ (x
n
)
n∈N
and y ≡ (y
n
)
n∈N
and every two real numbers α and β, if
x and y are convergent to x and y respectively, then the sequence αx+βy converges
to αx +βy. Thus the closure axioms hold. Usual remark concerning the other linear
space axioms.
15.5.15 n. 22 (p. 555)
The set U of all elements of R
3
with their third component equal to 0 is a real linear
space. Indeed, by linear combination the third component remains equal to 0; U is
the kernel of the linear function R
3
→R, (x, y, z) 7→z.
15.5.16 n. 23 (p. 555)
The set W of all elements of R
3
with their second or third component equal to 0 is
not a linear subspace of R
3
. For example, (0, 1, 0) and (0, 0, 1) are in the set, but
their sum (0, 1, 1) is not. The second closure axiom and the existence of negatives
axiom fail too. The other axioms hold. W is not even an aﬃne subspace, nor it is
convex; however, it is a cone.
15.5.17 n. 24 (p. 555)
The set π of all elements of R
3
with their second component which is equal to the
third multiplied by 5 is a real linear space, being the kernel of the linear function
R
3
→R, (x, y, z) 7→5x −y.
15.5.18 n. 25 (p. 555)
The set ` of all elements (x, y, z) of R
3
such that 3x + 4y = 1 and z = 0 (the line
through the point P ≡ (−1, 1, 0) with direction vector v ≡ (4, −3, 0)) is an aﬃne
subspace of R
3
, hence a convex set, but not a linear subspace of R
3
, hence not a
linear space itself. Indeed, for any two triples (x, y, z) and (u, v, w) of R
3
, and any
two real numbers α and β
3x + 4y = 1
z = 0
3u + 4v = 1
w = 0
⇒
½
3 (αx +βu) + 4 (αy +βv) = α +β
z + w = 0
Thus both closure axioms fail for `. Both existence axioms also fail, since neither
the null triple, nor the opposite of any triple in `, belong to `. The associative and
distributive laws have a defective status too, since they make sense in ` only under
quite restrictive assumptions on the elements of R
3
or the real numbers appearing in
them.
15.5.19 n. 26 (p. 555)
The set r of all elements (x, y, z) of R
3
which are scalar multiples of (1, 2, 3) (the line
through the origin having (1, 2, 3) as direction vector) is a real linear space. For any
Subspaces of a linear space 103
two elements (h, 2h, 3h) and (k, 2k, 3k) of r, and for any two real numbers α and β,
the linear combination
α(h, 2h, 3h) +β (k, 2k, 3k) = (αh +βk) (1, 2, 3)
belongs to r. r is the kernel of the linear function R
3
→R
2
, (x, y, z) 7→(2x −y, 3x −z).
15.5.20 n. 27 (p. 555)
The set of solutions of the linear homogenous system of equations
A(x, y, z)
0
= 0
is the kernel of the linear function R
3
→ R, (x, y, z) 7→ A(x, y, z)
0
and hence a real
linear space.
15.5.21 n. 28 (p. 555)
The subset of R
n
of all the linear combinations of two given vectors a and b is a
vector subspace of R
n
, namely span {a, b}. It is immediate to check that every linear
combination of linear combinations of a and b is again a linear combination of a and
b.
15.6 Subspaces of a linear space
15.7 Dependent and independent sets in a linear space
15.8 Bases and dimension
15.9 Exercises
15.9.1 n. 1 (p. 560)
The set S
1
of all elements of R
3
with their ﬁrst coordinate equal to 0 is a linear
subspace of R
3
(see exercise 5.22 above). S
1
is the coordinate Y Zplane, its dimension
is 2. A standard basis for it is {(0, 1, 0) , (0, 0, 1)}. It is clear that the two vectors are
linearly independent, and that for every real numbers y and z
(0, y, z) = y (0, 1, 0) +z (0, 0, 1)
104 Linear spaces
15.9.2 n. 2 (p. 560)
The set S
2
of all elements of R
3
with the ﬁrst and second coordinate summing up to
0 is a linear subspace of R
3
(S
2
is the plane containing the Zaxis and the bisectrix of
the evennumbered quadrants of the XY plane). A basis for it is {(1, −1, 0) , (0, 0, 1)};
its dimension is 2. It is clear that the two vectors are linearly independent, and that
for every three real numbers x, y and z such that x +y = 0
(x, y, z) = x(1, −1, 0) +z (0, 0, 1)
15.9.3 n. 3 (p. 560)
The set S
3
of all elements of R
3
with the coordinates summing up to 0 is a linear
subspace of R
3
(S
3
is the plane through the origin and normal vector n ≡ (1, 1, 1)).
A basis for it is {(1, −1, 0) , (0, −1, 1)}; its dimension is 2. It is clear that the two
vectors are linearly independent, and that for every three real numbers x, y and z
such that x +y +z = 0
(x, y, z) = x(1, −1, 0) +z (0, −1, 1)
15.9.4 n. 4 (p. 560)
The set S
4
of all elements of R
3
with the ﬁrst two coordinates equal is a linear subspace
of R
3
(S
4
is the plane containing the Zaxis and the bisectrix of the oddnumbered
quadrants of the XY plane). A basis for it is {(1, 1, 0) , (0, 0, 1)}; its dimension is 2.
It is clear that the two vectors are linearly independent, and that for every three real
numbers x, y and z such that x = y
(x, y, z) = x(1, 1, 0) +z (0, 0, 1)
15.9.5 n. 5 (p. 560)
The set S
5
of all elements of R
3
with all the coordinates equal is a linear subspace of
R
3
(S
5
is the line through the origin and direction vector d ≡ (1, 1, 1)). A basis for it
is {d}; its dimension is 1.
15.9.6 n. 6 (p. 560)
The set S
6
of all elements of R
3
with the ﬁrst coordinate equal either to the second
or to the third is not a linear subspace of R
3
(S
6
is the union of the plane S
4
of
exercise 4 and the plane containing the Y axis and the bisectrix of the oddnumbered
quadrants of the XZplane). For example, (1, 1, 0) and (1, 0, 1) both belong to S
6
,
but their sum (2, 1, 1) does not.
15.9.7 n. 7 (p. 560)
The set S
7
of all elements of R
3
with the ﬁrst and second coordinates having identical
square is not a linear subspace of R
3
(S
7
is the union of the two planes containing
the Zaxis and one of the bisectrices of the odd and evennumbered quadrants of the
XY plane). For example, (1, −1, 0) and (1, 1, 0) both belong to S
7
, but their sum
Exercises 105
(2, 0, 0) or their semisum (1, 0, 0) do not. S
7
is not an aﬃne subspace of R
3
, and it
is not even convex. However, S
7
is a cone, and it is even closed with respect to the
external product (multiplication by arbitrary real numbers).
15.9.8 n. 8 (p. 560)
The set S
8
of all elements of R
3
with the ﬁrst and second coordinates summing up to
1 is not a linear subspace of R
3
(S
8
is the vertical plane containing the line through
the points P ≡ (1, 0, 0) and Q ≡ (0, 1, 0)) For example, for every α 6= 1, α(1, 0, 0) ,
and α(0, 1, 0) do not belong to S
8
. S
8
is an aﬃne subspace of R
3
, since
∀(x, y, z) ∈ R
3
, ∀(u, v, w) ∈ R
3
, ∀α ∈ R
(x +y = 1) ∧ (u +v = 1) ⇒[(1 −α) x +αu] + [(1 −α) y +αv] = 1
15.9.9 n. 9 (p. 560)
See exercise 26, p.555.
15.9.10 n. 10 (p. 560)
The set S
10
of all elements (x, y, z) of R
3
such that
x +y +z = 0
x −y −z = 0
is a line containing the origin, a subspace of R
3
of dimension 1. A base for it is
{(0, 1, −1)}. If (x, y, z) is any element of S
10
, then
(x, y, z) = y (0, 1, −1)
15.9.11 n. 11 (p. 560)
The set S
11
of all polynomials of degree not exceeding n, and taking value 0 at 0 is a
linear subspace of the set of all polynomials of degree not exceeding n, and hence a
linear space. If P is any polynomial of degree not exceeding n,
P : t 7→p
0
+
X
h∈n
p
h
t
h
(15.1)
then P belongs to S
11
if and only if p
0
= 0. It is clear that any linear combination of
polynomials in S
11
takes value 0 at 0. A basis for S
11
is
B ≡
©
t 7→t
h
ª
h∈n
Indeed, by the general principle of polynomial identity (which is more or less explicitly
proved in example 6, p.558), a linear combination of B is the null polynomial, if and
only if all its coeﬃcients are null. Moreover, every polynomial belonging to S
11
is a
linear combination of B. It follows that dimS
11
= n.
106 Linear spaces
15.9.12 n. 12 (p. 560)
The set S
12
of all polynomials of degree not exceeding n, with ﬁrst derivative taking
value 0 at 0, is a linear subspace of the set of all polynomials of degree not exceeding
n. Indeed, for any two suach polynomials P and Q, and any two real numbers α and
β,
(αP +βQ)
0
(0) = αP
0
(0) +βQ
0
(0) = α · 0 +β · 0 = 0
A polynomial P as in (15.1) belongs to S
12
if and only if p
1
= 0. A basis for S is
B ≡
n
¡
t 7→t
h+1
¢
h∈n−1
, (t 7→1)
o
That B is a basis can be seen as in the previous exercise. Thus dimS
12
= n.
15.9.13 n. 13 (p. 560)
The set S
13
of all polynomials of degree not exceeding n, with second derivative
taking value 0 at 0, is a linear subspace of the set of all polynomials of degree not
exceeding n. This is proved exactly as in the previous exercise (just replace
0
with
00
).
A polynomial P as in (15.1) belongs to S
13
if and only if p
2
= 0. A basis for S is
B ≡
n
¡
t 7→t
h+2
¢
h∈n−2
, (t 7→1) , (t 7→t)
o
That B is a basis can be seen as in the exercise 11. Thus dimS
13
= n.
15.9.14 n. 14 (p. 560)
The set S
14
of all polynomials of degree not exceeding n, with ﬁrst derivative taking
value at 0 which is the opposite of the polynomial value at 0, is a linear subspace of
the set of all polynomials of degree not exceeding n. A polynomial P as in (15.1)
belongs to S
14
if and only if p
0
+ p
1
= 0. If P and Q are any two polynomials in S,
and α and β are any two real numbers,
(αP +βQ) (0) + (αP +βQ)
0
(1) = αP (0) +βQ(0) +αP
0
(0) +βQ
0
(0)
= α[P (0) +P
0
(0)] +β [Q(0) +Q
0
(0)]
= α · 0 +β · 0
A basis for S
14
is
B ≡
n
t 7→1 −t,
¡
t 7→t
h+1
¢
h∈n−1
o
Indeed, if α is the ntuple of coeﬃcients of a linear combination R of B, then
R : t 7→α
1
−α
1
t +
X
h∈n−1
α
h+1
t
h+1
Exercises 107
By the general principle of polynomial identity, the ntuple α of coeﬃcients of any
linear combination of B spanning the null vector must satisfy the conditions
α
1
= 0
α
h+1
= 0 (h ∈ n −1)
that is, the combination must be the trivial one. If P is any polynomial such that
p
0
+p
1
= 0, P belongs to span B, since by the position
α
1
= p
0
= −p
1
α
h+1
= p
h+1
(h ∈ n −1)
the linear combination of B with α as ntuple of coeﬃcients is equal to P. It follows
that B is a basis of S
14
, and hence that dimS = n.
15.9.15 n. 15 (p. 560)
The set S
15
of all polynomials of degree not exceeding n, and taking the same value
at 0 and at 1 is a linear subspace of the set of all polynomials of degree not exceeding
n, and hence a linear space. This can be seen exactly as in exercise 3, p.555. A
polynomial P belongs to S
15
if and only if
p
0
= p
0
+
X
h∈n
p
h
that is, if and only if
X
h∈n
p
h
= 0
A basis for S
15
is
B ≡
n
(t 7→1) ,
¡
t 7→(1 −t) t
h
¢
h∈n−1
o
and the dimension of S
15
is n.
15.9.16 n. 16 (p. 560)
The set S
16
of all polynomials of degree not exceeding n, and taking the same value
at 0 and at 2 is a linear subspace of the set of all polynomials of degree not exceeding
n, and hence a linear space. This can be seen exactly as in exercise 3, p.555. A
polynomial P belongs to S
16
if and only if
p
0
= p
0
+
X
h∈n
2
h
p
h
that is, if and only if
X
h∈n
2
h
p
h
= 0
108 Linear spaces
A basis for S
16
is
B ≡
n
(t 7→1) ,
¡
t 7→(2 −t) t
h
¢
h∈n−1
o
and the dimension of S
16
is n.
15.9.17 n. 22 (p. 560)
(b) It has alredy been shown in the notes that
lin S ≡
\
W subspace of V
S⊆W
W
=
(
v : ∃n ∈ N, ∃u ≡ (u
i
)
i∈n
∈ V
n
, ∃a ≡ (α
i
)
i∈n
∈ F
n
, v =
X
i∈n
α
i
u
i
)
hence if T is a subspace, and it contains S, then T contains lin S.
(c) If S = lin S, then of course S is a subspace of V because lin S is so. Conversely, if
S is a subspace of V , then S is one among the subspaces appearing in the deﬁnition
of lin S, and hence lin S ⊆ S; since S ⊆ lin S always, it follows that S = linS.
(e) If S and T are subspaces of V , then by point c above lin S = S and lin T = T,
and hence
S ∩ T = lin S ∩ lin T
Since the intersection of any family of subspaces is a subspace, S ∩ T is a subspace
of V .
(g) Let V ≡ R
2
, S ≡ {(1, 0) , (0, 1)}, T ≡ {(1, 0) , (1, 1)}. Then
S ∩ T = {(1, 0)}
L(S) = L(T) = R
2
L(S ∩ T) = {(x, y) : y = 0}
15.9.18 n. 23 (p. 560)
(a) Let
f : R →R, x 7→1
g : R →R, x 7→e
ax
h : R →R, x 7→e
bx
0 : R →R, x 7→0
and let (u, v, w) ∈ R
3
be such that
uf +vg +wh = 0
Exercises 109
In particular, for x = 0, x = 1, and x = 2, we have
u +v +w = 0
u +e
a
v +e
b
w = 0 (15.2)
u +e
2a
v +e
2b
w = 0
The determinant of the coeﬃcient matrix of the above system is
¯
¯
¯
¯
¯
¯
1 1 1
1 e
a
e
b
1 e
2a
e
2b
¯
¯
¯
¯
¯
¯
=
¯
¯
¯
¯
¯
¯
1 1 1
0 e
a
−1 e
b
−1
0 e
2a
−1 e
2b
−1
¯
¯
¯
¯
¯
¯
= (e
a
−1)
¡
e
b
−1
¢ £
e
b
+ 1 −(e
a
+ 1)
¤
= (e
a
−1)
¡
e
b
−1
¢ ¡
e
b
−e
a
¢
Since by assumption a 6= b, at most one of the two numbers can be equal to zero. If
none of them is null, system (15.2) has only the trivial solution, the triple (f, g, h)
is linearly independent, and dimlin {f, g, h} = 3. If a = 0, then g = f; similarly, if
b = 0, then h = f. Thus if either a or b is null then (f, g, h) is linearly dependent (it
suﬃces to take (u, v, w) ≡ (1, −1, 0) or (u, v, w) ≡ (1, 0, −1), respectively).
It is convenient to state and prove the following (very) simple
Lemma 4 Let X be a set containing at least two distinct elements, and let f : X →R
be a nonnull constant function. For any function g : X → R, the couple (f, g) is
linearly dependent if and only if g is constant, too.
Proof. If g is constant, let Imf ≡ {y
f
} and Img ≡ {y
g
}. Then the function
y
g
f −y
f
g is the null function. Conversely, if (f, g) is linearly dependent, let (u, v) ∈
R
2
∼ (0, 0) be such that uf +vg is the null function. Thus
∀x ∈ X, uy
f
+vg (x) = 0
Since f is nonnull, y
f
6= 0. This yields
v = 0 ⇒u = 0
and hence v 6= 0. Then
∀x ∈ X, g (x) = −
u
v
y
f
and g is constant, too.
It immediately follows from the above lemma (and from the assumption a 6= b)
that if either a = 0 or b = 0, then dimlin {f, g, h} = 2.
(b) The two functions
f : x 7→e
ax
g : x 7→xe
ax
110 Linear spaces
are linearly independent, since
[∀x ∈ R, αe
x
+βxe
ax
= 0] ⇔ [∀x ∈ R, (α +βx) e
ax
= 0]
⇔ [∀x ∈ R, (α +βx) = 0]
⇔ α = β = 0
Notice that the argument holds even in the case a = 0, Thus dimlin {f, g} = 2.
(c) Arguing as in point a, let (u, v, w) ∈ R
3
be such that
uf +vg +wh = 0
where
f : x 7→1 g : x 7→e
ax
h : x 7→xe
ax
In particular, for x = 0, x = 1, and x = −1, we have
u +v +w = 0
u +e
a
v +e
a
w = 0 (15.3)
u +e
−a
v −e
−a
w = 0
a homogeneous system of equations whose coeﬃcient matrix has determinant
¯
¯
¯
¯
¯
¯
1 1 1
1 e
a
e
a
1 e
−a
−e
−a
¯
¯
¯
¯
¯
¯
=
¯
¯
¯
¯
¯
¯
1 1 1
0 e
a
−1 e
a
−1
0 e
−a
−1 −e
−a
−1
¯
¯
¯
¯
¯
¯
= −(e
a
−1)
¡
e
−a
+ 1 +e
−a
−1
¢
= 2e
−a
(1 −e
a
)
Thus if a 6= 0 the above determinant is diﬀerent from zero, yielding (u, v, w) =
(0, 0, 0). The triple (f, g, h) is linearly independent, and dimlin {f, g, h} = 3. On the
other hand, if a = 0 then f = g and h = Id
R
, so that dimlin {f, g, h} = 2.
(f) The two functions
f : x 7→sinx g : x 7→cos x
are linearly independent. Indeed, for every two real numbers α and β which are not
both null,
∀x ∈ R, αsinx +β cos x = 0
m
∀x ∈ R,
α
√
α
2
+β
2
sinx +
β
√
α
2
+β
2
cos x = 0
m
∀x ∈ R, sin (γ +x) = 0
Inner products. Euclidean spaces. Norms 111
where
γ ≡ signβ arccos
α
p
α
2
+β
2
since the sin function is not identically null, the last condition cannot hold. Thus
dimlin{x 7→sin x, x 7→cos x} = 2.
(h) From the trigonometric addition formulas
∀x ∈ R cos 2x = cos
2
x −sin
2
x = 1 −2 sin
2
x
Let then
f : x 7→1 g : x 7→cos x g : x 7→sin
2
x
The triple (f, g, h) is linearly dependent, since
f −g + 2h = 0
By the lemma discussed at point a, dimlin {f, g, h} = 2.
15.10 Inner products. Euclidean spaces. Norms
15.11 Orthogonality in a euclidean space
112 Linear spaces
15.12 Exercises
15.12.1 n. 9 (p. 567)
hu
1
, u
2
i =
Z
1
−1
t dt =
t
2
2
¯
¯
¯
¯
1
−1
=
1
2
−
1
2
= 0
hu
2
, u
3
i =
Z
1
−1
t +t
2
dt = 0 +
t
3
3
¯
¯
¯
¯
1
−1
=
1
3
−
µ
−
1
3
¶
=
2
3
hu
3
, u
1
i =
Z
1
−1
1 +t dt = t
1
−1
+ 0 = 1 −(−1) = 2
ku
1
k
2
=
Z
1
−1
1 dt = 2 ku
1
k =
√
2
ku
2
k
2
=
Z
1
−1
t
2
dt =
2
3
ku
2
k =
r
2
3
ku
3
k
2
=
Z
1
−1
1 + 2t +t
2
dt = 2 +
2
3
ku
3
k =
r
8
3
cos [ u
2
u
3
=
2
3
q
2
3
q
8
3
=
1
2
cos [ u
1
u
3
=
2
√
2
q
8
3
=
√
3
2
[ u
2
u
3
=
π
3
[ u
1
u
3
=
π
6
15.12.2 n. 11 (p. 567)
(a) Let for each n ∈ N
I
n
≡
Z
+∞
0
e
−t
t
n
dt
Then
I
0
=
Z
+∞
0
e
−t
dt = lim
x→+∞
Z
x
0
e
−t
dt
= lim
x→+∞
−e
−t
¯
¯
x
0
= lim
x→+∞
−e
−x
+ 1
= 1
Exercises 113
and, for each n ∈ N,
I
n+1
=
Z
+∞
0
e
−t
t
n+1
dt = lim
x→+∞
Z
x
0
e
−t
t
n+1
dt
= lim
x→+∞
½
−e
−t
t
n+1
¯
¯
x
0
+ (n + 1)
Z
x
0
e
−t
t
n
dt
¾
= lim
x→+∞
©
−e
−x
x
n+1
−0
ª
+ (n + 1) lim
x→+∞
½Z
x
0
e
−t
t
n
dt
¾
= 0 + (n + 1)
Z
+∞
0
e
−t
t
n
dt
= (n + 1) I
n
The integral involved in the deﬁnition of I
n
is always convergent, since for each n ∈ N
lim
x→+∞
e
−x
x
n
= 0
It follows that
∀n ∈ N, I
n
= n!
Let now
f : t 7→α
0
+
X
i∈m
α
i
t
i
g : t 7→β
0
+
X
i∈n
β
j
t
j
be two real polynomials, of degree m and n respectively. Then the product fg is a
real polynomial of degree m+n, containing for each k ∈ m+n a monomial of degree
k of the form α
i
t
i
β
j
t
j
= α
i
β
j
t
i+j
whenever i +j = k. Thus
fg = α
0
β
0
+
X
k∈m+n
¸
¸
¸
X
i∈m, j∈n
i+j=k
α
i
β
j
¸
t
k
114 Linear spaces
and the scalar product of f and g is results in the sum of m + n + 1 converging
integrals
hf, gi =
Z
+∞
0
e
−t
α
0
β
0
+
X
k∈m+n
¸
¸
¸
X
i∈m, j∈n
i+j=k
α
i
β
j
¸
t
k
¸
¸
¸
¸
dt
= α
0
β
0
Z
+∞
0
e
−t
dt +
X
k∈m+n
¸
¸
¸
X
i∈m, j∈n
i+j=k
α
i
β
j
¸
Z
+∞
0
e
−t
t
k
dt
= α
0
β
0
I
0
+
X
k∈m+n
¸
¸
¸
X
i∈m, j∈n
i+j=k
α
i
β
j
¸
I
k
= α
0
β
0
+
X
k∈m+n
¸
¸
¸
X
i∈m, j∈n
i+j=k
α
i
β
j
¸
k!
(b) If for each n ∈ N
x
n
: t 7→t
n
then for each m ∈ N and each n ∈ N
hx
m
, x
n
i =
Z
+∞
0
e
−t
t
m
t
n
dt
= I
m+n
= (m+n)!
(c) If
f : t 7→(t + 1)
2
g : t 7→t
2
+ 1
then
fg : t 7→t
4
+ 2t
3
+ 2t
2
+ 2t + 1
hf, gi =
Z
+∞
0
e
−t
¡
t
4
+ 2t
3
+ 2t
2
+ 2t + 1
¢
dt
= I
4
+ 2I
3
+ 2I
2
+ 2I
1
+I
0
= 4! + 2 · 3! + 2 · 2! + 2 · 1! + 1
= 43
Construction of orthogonal sets. The GramSchmidt process 115
(d) If
f : t 7→t + 1 g : t 7→αt +β
then
fg : t 7→αt
2
+ (α +β) t +β
hf, gi = αI
2
+ (α +β) I
1
+βI
0
= 3α + 2β
and
hf, gi = 0 ⇔(α, β) = (2γ, −3γ) (γ ∈ R)
that is, the set of all polynomials of degree less than or equal to 1 which are orthogonal
to f is
P
1
∩ perp {f} = {g
γ
: t 7→2γt −3γ}
γ∈R
15.13 Construction of orthogonal sets. The GramSchmidt process
15.14 Orthogonal complements. projections
15.15 Best approximation of elements in a euclidean space by elements
in a ﬁnitedimensional subspace
15.16 Exercises
15.16.1 n. 1 (p. 576)
(a) By direct inspection of the coordinates of the three given vectors. it is seen that
they all belong to the plane π of equation x−z = 0. Since no two of them are parallel,
lin {x
1
, x
2
, x
3
} = π and dimlin{x
1
, x
2
, x
3
} = 2. A unit vector in π is given by
v ≡
1
3
(2, 1, 2)
Every vector belonging to the line π ∩ perp {v} must have the form
(x, −4x, x)
116 Linear spaces
with norm 3
√
2 x. Thus a second unit vector which together with v spans π, and
which is orthogonal to v, is
w ≡
√
2
6
(1, −4, 1)
The required orthonormal basis for lin {x
1
, x
2
, x
3
} is {v, w}.
(b) The answer is identical to the one just given for case a.
15.16.2 n. 2 (p. 576)
(a) First solution. It is easily checked that
x
2
/ ∈ lin {x
1
}
x
3
/ ∈ lin {x
1
, x
2
}
0 = x
1
−x
2
+x
3
−x
4
Thus dimW ≡ dimlin {x
1
, x
2
, x
3
, x
4
} = 3, and three vectors are required for any
basis of W. The following vectors form an orthonormal one:
y
1
≡
x
1
kx
1
k
=
√
2
2
(1, 1, 0, 0)
y
2
≡
x
2
−hx
2
, y
1
i y
1
kx
2
−hx
2
, y
1
i y
1
k
=
(0, 1, 1, 0) −
√
2
2
√
2
2
(1, 1, 0, 0)
°
°
°(0, 1, 1, 0) −
√
2
2
√
2
2
(1, 1, 0, 0)
°
°
°
=
¡
−
1
2
,
1
2
, 1, 0
¢
q
1
4
+
1
4
+ 1
=
√
6
6
(−1, 1, 2, 0)
y
3
≡
x
3
−hx
3
, y
1
i y
1
−−hx
3
, y
2
i y
2
kx
3
−hx
3
, y
1
i y
1
−−hx
3
, y
2
i y
2
k
=
(0, 0, 1, 1) −0 −2
√
6
6
√
6
6
(−1, 1, 2, 0)
°
°
°(0, 0, 1, 1) −0 −2
√
6
6
√
6
6
(−1, 1, 2, 0)
°
°
°
=
¡
1
3
, −
1
3
,
1
3
, 1
¢
q
1
9
+
1
9
+
1
9
+ 1
=
√
3
6
(1, −1, 1, 3)
Second solution. More easily, it suﬃces to realize that
lin {x
1
, x
2
, x
3
, x
4
} = H
(1,−1,1,−1),0
≡
©
(x, y, z, t) ∈ R
4
: x −y +z −t = 0
ª
is the hyperplane of R
4
through the origin, with (1, −1, 1, −1) as normal vector, to
directly exhibit two mutually orthogonal unit vectors in lin {x
1
, x
2
, x
3
, x
4
}
u ≡
√
2
2
(1, 1, 0, 0) v ≡
√
2
2
(0, 0, 1, 1)
Exercises 117
It remains to ﬁnd a unit vector in H
(1,−1,1,−1),0
∩perp{u, v}. The following equations
characterize H
(1,−1,1,−1),0
∩ perp{u, v}:
x −y +z −t = 0
x +y = 0
z +t = 0
yielding (by addition) 2 (x +z) = 0, and hence
w ≡
1
2
(1, −1, −1, 1) or w ≡ −
1
2
(1, −1, −1, 1)
Thus an orthonormal basis for lin {x
1
, x
2
, x
3
, x
4
} is {u, v, w}.
(b) It is easily checked that
x
2
/ ∈ lin {x
1
}
0 = 2x
1
−x
2
−x
3
Thus dimW ≡ dimlin {x
1
, x
2
, x
3
} = 2, and two vectors are required for any basis of
W. The following vectors form an orthonormal one:
y
1
≡
x
1
kx
1
k
=
√
3
3
(1, 1, 0, 1)
y
2
≡
x
2
−hx
2
, y
1
i y
1
kx
2
−hx
2
, y
1
i y
1
k
=
(1, 0, 2, 1) −2
√
3
3
√
3
3
(1, 1, 0, 1)
°
°
°(0, 1, 1, 0) −2
√
3
3
√
3
3
(1, 1, 0, 1)
°
°
°
=
¡
1
3
, −
2
3
, 2,
1
3
¢
q
1
9
+
4
9
+ 4 +
1
9
=
√
42
42
(1, −2, 6, 1)
15.16.3 n. 3 (p. 576)
hy
0
, y
0
i =
Z
π
0
1
√
π
1
√
π
dt = π
1
π
= 1
hy
n
, y
n
i =
Z
π
0
r
2
π
cos nt
r
2
π
cos nt dt =
2
π
Z
π
0
cos
2
nt dt
Let
I
n
≡
Z
π
0
cos
2
ntdt =
Z
π
0
cos nt cos nt dt
integrating by parts
I
n
= cos nt
sinnt
n
¯
¯
¯
¯
π
0
−
Z
π
0
−nsin nt
sin nt
n
dt
=
Z
π
0
sin
2
nt dt =
Z
π
0
1 −cos
2
nt dt
= π −I
n
118 Linear spaces
Thus
I
n
=
π
2
hy
n
, y
n
i = 1
and every function y
n
has norm equal to one.
To check mutual orthogonality, let us compute
hy
0
, y
n
i =
Z
π
0
1
√
π
cos nt dt =
1
√
π
sin nt
n
¯
¯
¯
¯
π
0
= 0
hy
m
, y
n
i =
Z
π
0
cos mt cos nt dt
= cos mt
sinnt
n
¯
¯
¯
¯
π
0
−
Z
π
0
−msin mt
sin nt
n
dt
=
m
n
Z
π
0
sin mt sinnt dt
=
m
n
sin mt
µ
−
cos nt
n
¶¯
¯
¯
¯
π
0
−
m
n
Z
π
0
mcos mt
µ
−
cos nt
n
¶
dt
=
³
m
n
´
2
hy
m
, y
n
i
Since m and n are distinct positive integers,
¡
m
n
¢
2
is diﬀerent from 1, and the equation
hy
m
, y
n
i =
³
m
n
´
2
hy
m
, y
n
i
can hold only if hy
m
, y
n
i = 0.
That the set {y
n
}
+∞
n=0
generates the same space generated by the set {x
n
}
+∞
n=0
is trivial, since, for each n, y
n
is a multiple of x
n
.
15.16.4 n. 4 (p. 576)
We have
hy
0
, y
0
i =
Z
1
0
1dt = 1
hy
1
, y
1
i =
Z
1
0
3
¡
4t
2
−4t + 1
¢
dt = 3
Ã
4
3
t
3
−2t
2
+t
¯
¯
¯
¯
1
0
!
= 1
hy
2
, y
2
i =
Z
1
0
5
¡
36t
4
−72t
3
+ 48t
2
−12t + 1
¢
dt
= 5
Ã
36
5
t
5
−18t
4
+ 16t
3
−6t
2
+t
¯
¯
¯
¯
1
0
!
= 1
Exercises 119
which proves that the three given functions are unit vectors with respect to the given
inner product.
Moreover,
hy
0
, y
1
i =
Z
1
0
√
3 (2t −1) dt =
√
3
³
t
2
−t
¯
¯
1
0
´
= 0
hy
0
, y
2
i =
Z
1
0
√
5
¡
6t
2
−6t + 1
¢
dt =
√
5
³
2t
3
−3t
2
+t
¯
¯
1
0
´
= 0
hy
1
, y
2
i =
Z
1
0
√
15
¡
12t
3
−18t
2
+ 8t −1
¢
dt
=
√
15
³
3t
4
−6t
3
+ 4t
2
−t
¯
¯
1
0
´
= 0
which proves that the three given functions are mutually orthogonal. Thus {y
1
, y
2
, y
3
}
is an orthonormal set, and hence linearly independent.
Finally,
x
0
= y
0
x
1
=
1
2
y
0
+
√
3
2
y
1
x
2
=
Ã
1 −
√
5
30
!
y
1
+
√
3
3
y
1
+
√
5
30
y
2
which proves that
lin {y
1
, y
2
, y
3
} = lin {x
1
, x
2
, x
3
}
120 Linear spaces
Chapter 16
LINEAR TRANSFORMATIONS AND MATRICES
16.1 Linear transformations
16.2 Null space and range
16.3 Nullity and rank
16.4 Exercises
16.4.1 n. 1 (p. 582)
T is linear, since for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R, for
every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αy +βv, αx +βu)
= α(y, x) +β (v, u)
= αT [(x, y)] +βT [(u, v)]
The null space of T is the trivial subspace {(0, 0)}, the range of T is R
2
, and hence
rank T = 2 nullity T = 0
T is the orthogonal symmetry with respect to the bisectrix r of the ﬁrst and third
quadrant, since, for every (x, y) ∈ R
2
, T (x, y) − (x, y) is orthogonal to r, and the
midpoint of [(x, y) , T (x, y)], which is (x +y, x +y), lies on r.
122 Linear transformations and matrices
16.4.2 n. 2 (p. 582)
T is linear, since for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R, and for
every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αy +βv, αx +βu)
= α(y, x) +β (v, u)
= αT [(x, y)] +βT [(u, v)]
The null space of T is the trivial subspace {(0, 0)}, the range of T is R
2
, and hence
rank T = 2 nullity T = 0
T is the orthogonal symmetry with respect to the Xaxis.
16.4.3 n. 3 (p. 582)
T is linear , since for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R, and
for every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αx +βu, 0)
= α(x, 0) +β (u, 0)
= αT [(x, y)] +βT [(u, v)]
The null space of T is the Y axis, the range of T is the Xaxis, and hence
rank T = 1 nullity T = 1
T is the orthogonal projection on the Xaxis.
16.4.4 n. 4 (p. 582)
T is linear, since for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R, and for
every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αx +βu, αx +βu)
= α(x, x) +β (u, u)
= αT [(x, y)] +βT [(u, v)]
The null space of T is the Y axis, the range of T is the bisectrix of the I and III
quadrant, and hence
rank T = nullity T = 1
Exercises 123
16.4.5 n. 5 (p. 582)
T is not a linear function, since, e.g., for x and u both diﬀerent from 0,
T (x +u, 0) =
¡
x
2
+ 2xy +u
2
, 0
¢
T (x, 0) +T (u, 0) =
¡
x
2
+u
2
, 0
¢
16.4.6 n. 6 (p. 582)
T is not linear; indeed, for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R,
for every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
=
¡
e
αx+βu
, e
αy+βv
¢
=
h
(e
x
)
α
· (e
u
)
β
, (e
y
)
α
· (e
v
)
β
i
αT [(x, y)] +βT [(u, v)] = α(e
x
, e
y
) +β (e
u
, e
v
)
= (αe
x
+βe
u
, αe
y
+βe
v
)
so that, e.g., when x = y = 0 and u = v = α = β = 1,
T [(0, 0) + (1, 1)] = (e, e)
T [(0, 0)] +T [(1, 1)] = (1 +e, 1 +e)
16.4.7 n. 7 (p. 582)
T is not an aﬃne function, but it is not linear. Indeed, for every (x, y) ∈ R
2
, for
every (u, v) ∈ R
2
, for every α ∈ R, and for every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αx +βu, 1)
αT [(x, y)] +βT [(u, v)] = α(x, 1) +β (u, 1)
= (αx +βu, α +β)
16.4.8 n. 8 (p. 582)
T is aﬃne, but not linear; indeed, for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for
every α ∈ R, and for every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αx +βu + 1, αy +βv + 1)
αT [(x, y)] +βT [(u, v)] = α(x + 1, y + 1) +β (u + 1, v + 1)
= (αx +βu +α +β, αy +βv +α +β)
124 Linear transformations and matrices
16.4.9 n. 9 (p. 582)
T is linear, since for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R, and for
every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (αx +βu −αy −βv, αx +βu +αy +βv)
= α(x −y, x +y) +β (u −v, u +v)
= αT [(x, y)] +βT [(u, v)]
The null space of T is the trivial subspace {(0, 0)}, the range of T is R
2
, and hence
rank T = 2 nullity T = 0
The matrix representing T is
A ≡
·
1 −1
1 1
¸
=
√
2
"
√
2
2
−
√
2
2 √
2
2
√
2
2
#
=
· √
2 0
0
√
2
¸ ·
cos
π
4
−sin
π
4
sin
π
4
cos
π
4
¸
Thus T is the composition of a counterclockwise rotation by an angle of
π
4
with a
homothety of modulus
√
2.
16.4.10 n. 10 (p. 582)
T is linear, since for every (x, y) ∈ R
2
, for every (u, v) ∈ R
2
, for every α ∈ R, and for
every β ∈ R,
T [α(x, y) +β (u, v)] = T [(αx +βu, αy +βv)]
= (2 (αx +βu) −(αy +βv) , (αx +βu) + (αy +βv))
= α(2x −y, x +y) +β (2u −v, u +v)
= αT [(x, y)] +βT [(u, v)]
The null space of T is the trivial subspace {(0, 0)}, the range of T is R
2
, and hence
rank T = 2 nullity T = 0
The matrix representing T is
A ≡
·
2 −1
1 1
¸
the characteristic polynomial of A is
λ
2
−3λ + 3
Exercises 125
and the eigenvalues of A are
λ
1
≡
3 +
√
3i
2
λ
2
≡
3 −
√
3i
2
An eigenvector associated to λ
1
is
z ≡
µ
2
1 −
√
3i
¶
The matrix representing T with respect to the basis
{Imz, Re z} =
½µ
0
−
√
3
¶
,
µ
2
1
¶¾
is
B ≡
"
3
2
−
√
3
2 √
3
2
3
2
#
=
√
3
"
√
3
2
−
1
2
1
2
√
3
2
#
=
· √
3 0
0
√
3
¸ ·
cos
π
6
−sin
π
6
sin
π
6
cos
π
6
¸
Thus T is the composition of a counterclockwise rotation by an angle of
π
6
with a
homothety of modulus
√
3.
16.4.11 n. 16 (p. 582)
T is linear, since for every (x, y, z) ∈ R
3
, for every (u, v, w) ∈ R
3
, for every α ∈ R,
for every β ∈ R,
T [α(x, y, z) +β (u, v, w)] = T [(αx +βu, αy + βv, αz +βw)]
= (αz +βw, αy +βv, αx +βu)
= α(z, y, x) +β (w, v, u)
= αT [(x, y, z)] +βT [(u, v, w)]
The null space of T is the trivial subspace {(0, 0, 0)}, the range of T is R
3
, and hence
rank T = 3 nark T = 0
16.4.12 n. 17 (p. 582)
T is linear (as every projection on a coordinate hyperplane), since for every (x, y, z) ∈
R
3
, for every (u, v, w) ∈ R
3
, for every α ∈ R, for every β ∈ R,
T [α(x, y, z) +β (u, v, w)] = T [(αx +βu, αy + βv, αz +βw)]
= (αx +βu, αy +βv, 0)
= α(x, y, 0) +β (u, v, 0)
= αT [(x, y, z)] +βT [(u, v, w)]
The null space of T is the Zaxis, the range of T is the XY plane, and hence
rank T = 2 nark T = 1
126 Linear transformations and matrices
16.4.13 n. 23 (p. 582)
T is linear, since for every (x, y, z) ∈ R
3
, for every (u, v, w) ∈ R
3
, for every α ∈ R,
for every β ∈ R,
T [α(x, y, z) +β (u, v, w)] = T [(αx +βu, αy +βv, αz +βw)]
= (αx +βu +αz +βw, 0, αx +βu +αy +βv)
= [α(x +z) , 0, α(x +y)] + [β (u +w) , 0, β (u +v)]
= α(x, y, z) +β (u, v, w)
= αT [(x, y, z)] +βT [(u, v, w)]
The null space of T is the axis of central symmetry of the (+, −, −)orthant and of
the (−, +, +)orthant, of parametric equations
x = t y = −t z = −t
the range of T is the XZplane, and hence
rank T = 2 nark T = 1
16.4.14 n. 25 (p. 582)
Let D
1
(−1, 1) or, more shortly, D
1
be the linear space of all real functions of a real
variable which are deﬁned and everywhere diﬀerentiable on (−1, 1). If
T : D
1
→R
(−1,1)
, f 7→(x 7→xf
0
(x))
then T is a linear operator. Indeed,
T (f +g) =
¡
x 7→x[f +g]
0
(x)
¢
= (x 7→x[f
0
(x) +g
0
(x)])
= (x 7→xf
0
(x) +xg
0
(x))
= (x 7→xf
0
(x)) + (x 7→+xg
0
(x))
= T (f) +T (g)
Moreover,
ker T = {f ∈ D
1
: ∀x ∈ (−1, 1) , xf
0
(x) = 0}
= {f ∈ D
1
: ∀x ∈ (−1, 0) ∪ (0, 1) , f
0
(x) = 0}
By an important though sometimes overlooked theorem in calculus, every function f
which is diﬀerentiable in (−1, 0) ∪ (0, 1) and continuous in 0, is diﬀerentiable in 0 as
well, provided lim
x→0
f
0
(x) exists and is ﬁnite, in which case
f
0
(0) = lim
x→0
f
0
(x)
Exercises 127
Thus
ker T = {f ∈ D
1
: f
0
= 0}
(here 0 is the identically null function deﬁned on (−1, 1)). If f belongs to ker T, by
the classical theorem due to Lagrange,
∀x ∈ (−1, 1) , ∃ϑ
x
∈ (0, x)
f (x) = f (0) +xf
0
(ϑ
x
)
and hence f is constant on (−1, 1). It follows that a basis of ker T is given by the
constant function (x 7→1), and nullity T = 1. Since T (D
1
) contains, e.g., all the
restrictions to (−1, 1) of the polynomials with vanishing zerodegree monomial, and
it is already known that the linear space of all polynomials has an inﬁnite basis, the
dimension of T (D
1
) is inﬁnite.
16.4.15 n. 27 (p. 582)
Let D
2
be the linear space of all real functions of a real variable which are deﬁned
and everywhere diﬀerentiable on R. If
L : D
2
→R
R
, y 7→y
00
+Py
0
+Q
where P and Q are real functions of a real variable which are continuous on R, it has
been shown in the solution to exercise 17, p.555, that L is a linear operator. Thus
ker L is the set of all solutions to the linear diﬀerential equation of the second order
∀x ∈ R, y
00
(x) +P (x) y
0
(x) +Q(x) y (x) = 0 (16.1)
By the uniqueness theorem for Cauchy’s problems in the theory of diﬀerential equa
tions, for each (y
0
, y
0
0
) ∈ R
2
there exists a unique solution to equation (16.1) such
that y (0) = y
0
and y
0
(0) = y
0
0
. Hence the function
ϕ : ker L →R
2
, y 7→
µ
y (0)
y
0
(0)
¶
is injective and surjective. Moreover, let u be the solution to equation (16.1) such
that (u (0) , u
0
(0)) = (1, 0), and let v be the solution to equation (16.1) such that
(v (0) , v
0
(0)) = (0, 1). Then for each (y
0
, y
0
0
) ∈ R
2
, since ker L is a linear subspace of
D
2
, the function
y : x 7→y
0
u(x) +y
0
0
v (x)
is a solution to equation (16.1), and by direct inspection it is seen that y (0) = y
0
and
y
0
(0) = y
0
0
. In other words, ϕ
−1
((y
0
, y
0
0
)), and
ker L = span {u, v}
Finally, u and v are linearly independent; indeed, since u(0) = 1 and v (0) = 0,
αu(x) + βv (x) = 0 for each x ∈ R can only hold (by evaluating at x = 0) if α = 0;
in such a case, from βv (x) = 0 for each x ∈ R and v
0
(0) = 1 it is easily deduced that
β = 0 as well. Thus nullity L = 2.
128 Linear transformations and matrices
16.5 Algebraic operations on linear transformations
16.6 Inverses
16.7 Onetoone linear transformations
16.8 Exercises
16.8.1 n. 15 (p. 589)
T is injective (or one to one), since
T (x, y, z) = T (x
0
, y
0
, z
0
) ⇔
x = x
0
2y = 2y
0
3z = 3z
0
⇔(x, y, z) = (x
0
, y
0
, z
0
)
T
−1
(u, v, w) =
³
u,
v
2
,
w
3
´
16.8.2 n. 16 (p. 589)
T is injective, since
T (x, y, z) = T (x
0
, y
0
, z
0
) ⇔
x = x
0
y = y
0
x +y +z = x
0
+y
0
+z
0
⇔(x, y, z) = (x
0
, y
0
, z
0
)
T
−1
(u, v, w) = (u, v, w −u −v)
16.8.3 n. 17 (p. 589)
T is injective, since
T (x, y, z) = T (x
0
, y
0
, z
0
) ⇔
x + 1 = x
0
+ 1
y + 1 = y
0
+ 1
z −1 = z
0
−1
⇔(x, y, z) = (x
0
, y
0
, z
0
)
T
−1
(u, v, w) = (u −1, v −1, w + 1)
Linear transformations with prescribed values 129
16.8.4 n. 27 (p. 590)
Let
p = x 7→p
0
+p
1
x +p
2
x
2
+ · · · +p
n−1
x
n−1
+p
n
x
n
DT (p) = D[T (p)] = D
·
x 7→
Z
x
0
p(t) dt
¸
= x 7→p(x)
the last equality being a consequence of the fundamental theorem of integral calculus.
TD(p) = T [D(p)] = T
£
x 7→p
1
+ 2p
2
x + · · · (n −1) p
n−1
x
n−2
+np
n
x
n−1
¤
= x 7→
Z
x
0
p
1
+ 2p
2
t + · · · (n −1) p
n−1
t
n−2
+np
n
t
n−1
dt
= x 7→ p
1
t +p
2
t
2
+ · · · p
n−1
t
n−1
+p
n
t
n
¯
¯
x
0
= x 7→p
1
x +p
2
x
2
+ · · · p
n−1
x
n−1
+p
n
x
n
= p −p
0
Thus TD acts as the identity map only on the subspace W of V containing all
polynomials having the zero degree monomial (p
0
) equal to zero. ImTD is equal to
W, and ker TD is the set of all constant polynomials, i.e., zerodegree polynomials.
16.9 Linear transformations with prescribed values
16.10 Matrix representations of linear transformations
16.11 Construction of a matrix representation in diagonal form
16.12 Exercises
16.12.1 n. 3 (p. 596)
(a) Since T (i) = i +j and T (j) = 2i −j,
T (3i −4j) = 3T (i) −4T (j) = 3 (i +j) −4 (2i −j)
= −5i + 7j
T
2
(3i −4j) = T (−5i + 7j) = −5T (i) + 7T (j)
= −5 (i +j) + 7 (2i −j) = 9i −12j
130 Linear transformations and matrices
(b) The matrix of T with respect to the basis {i, j} is
A ≡
£
T (i) T (j)
¤
=
·
1 2
1 −1
¸
and the matrix of T
2
with respect to the same basis is
A
2
≡
·
3 0
0 3
¸
(c) First solution (matrix for T). If e
1
= i −j and e
2
= 3i +j, the matrix
P ≡
£
e
1
e
2
¤
=
·
1 3
−1 1
¸
transforms the (canonical) coordinates (1, 0) and (0, 1) of e
1
and e
2
with respect to
the basis {e
1
, e
2
} into their coordinates (1, −1) and (3, 1) with respect to the basis
{i, j}; hence P is the matrix of coordinate change from basis {e
1
, e
2
} to basis {i, j},
and P
−1
is the matrix of coordinate change from basis {i, j} to basis {e
1
, e
2
}. The
operation of the matrix B representing T with respect to the basis {e
1
, e
2
} can be
described as the combined eﬀect of the following three actions: 1) coordinate change
from coordinates w.r. to basis {e
1
, e
2
} into coordinates w.r. to the basis {i, j} (that
is, multiplication by matrix P); 2) transformation according to T as expressed by
the matrix representing T w.r. to the basis {i, j} (that is, multiplication by A); 3)
coordinate change from coordinates w.r. to the basis {i, j} into coordinates w.r. to
basis {e
1
, e
2
} (that is, multiplication by P
−1
). Thus
B = P
−1
AP =
1
4
·
1 −3
1 1
¸ ·
1 2
1 −1
¸ ·
1 3
−1 1
¸
=
1
4
·
−2 5
2 1
¸ ·
1 3
−1 1
¸
=
1
4
·
−7 −1
1 7
¸
Second solution (matrix for T). The matrix B representing T with respect to the
basis {e
1
, e
2
} is
B ≡
£
T (e
1
) T (e
2
)
¤
provided T (e
1
) and T (e
2
) are meant as coordinate vectors (α, β) and (γ, δ) with
respect to the basis {e
1
, e
2
}. Since, on the other hand, with respect to the basis {i, j}
we have
T (e
1
) = T (i −j) = T (i) −T (j) = (i +j) −(2i −j)
= −i + 2j
T (e
2
) = T (3i +j) = 3T (i) +T (j) = 3 (i +j) + (2i −j)
= 5i + 2j
Exercises 131
it suﬃces to ﬁnd the {e
1
, e
2
}coordinates (α, β) and (γ, δ) which correspond to the
{i, j}coordinates (−1, 2) and (5, 2). Thus we want to solve the two equation systems
(in the unknowns (α, β) and (γ, δ), respectively)
αe
1
+βe
2
= −i + 2j
γe
1
+δe
2
= 5i + 2j
that is,
½
α + 3β = −1
−α +β = 2
½
γ + 3δ = 5
−γ +δ = 2
The (unique) solutions are (α, β) =
1
4
(−7, 1) and (γ, δ) =
1
4
(−1, 7), so that
B =
1
7
·
−7 −1
1 7
¸
Continuation (matrix for T
2
). Since T
2
is represented by a scalar diagonal matrix
with respect to the initially given basis, it is represented by the same matrix with
respect to every basis (indeed, since scalar diagonal matrices commute with every
matrix of the same order, P
−1
DP = D for every scalar diagonal matrix D).
16.12.2 n. 4 (p. 596)
T : (x, y) (reﬂection w.r. to the Y axis) 7→ (−x, y)
(length doubling) 7→ (−2x, 2y)
Thus T may be represented by the matrix
A
T
≡
µ
−2 0
0 2
¶
and hence T
2
by the matrix
A
2
T
≡
µ
4 0
0 4
¶
16.12.3 n. 5 (p. 596)
a)
T (i + 2j + 3k) = T (k) +T (j +k) +T (i +j +k)
= (2i + 3j + 5k) +i + (j −k)
= 3i + 4j + 4k
132 Linear transformations and matrices
The three image vectors T (k), T (j +k), T (i +j +k) form a linearly independent
triple. Indeed, the linear combination
xT (k) +yT (j +k) +zT (i +j +k) = x(2i + 3j + 5k) +yi +z (j −k)
= (2x +y) i + (3x +z) j + (5x −z) k
spans the null vector if and only if
2x +y = 0
3x +z = 0
5x −z = 0
which yields (II + III) x = 0, and hence (by substitution in I and II) y = z = 0.
Thus the range space of T is R
3
, and rank T is 3. It follows that the null space of T
is the trivial subspace {(0, 0, 0)}, and nark T is 0.
b) The matrix of T with respect to the basis {i, j, k} is obtained by aligning as
columns the coordinates w.r. to {i, j, k} of the image vectors T (i), T (j), T (k). The
last one is given in the problem statement, but the ﬁrst two need to be computed.
T (j) = T (j +k −k) = T (j +k) −T (k)
= −i −3j −5k
T (i) = T (i +j +k −j −k) = T (i +j +k) −T (j +k)
= −i +j −k
A
T
≡
¸
−1 −1 2
1 −3 3
−1 −5 3
¸
16.12.4 n. 7 (p. 597)
(a)
T (4i −j + k) = 4T (i) −T (j) +T (k) = (0, −2)
Since {T (j) , T (k)} is a linearly independent set,
ker T = {0} rank T = 2
(b)
A
T
=
µ
0 1 1
0 1 −1
¶
Exercises 133
(c)
Let C = (w
1
, w
2
)
T (i) = 0w
1
+ 0w
2
T (j) = 1w
1
+ 0w
2
T (k) = −
1
2
w
1
+
3
2
w
2
(A
T
)
C
=
µ
0 1 −
1
2
0 0
3
2
¶
(d)
Let B ≡ {j, k, i} and C = {w
1
, w
2
} = {(1, 1) , (1, −1)}. Then
(A
T
)
B,C
=
µ
1 0 0
0 1 0
¶
16.12.5 n. 8 (p. 597)
(a) I shall distinguish the canonical unit vectors of R
2
and R
3
by marking the former
with an underbar. Since T (i) = i +k and T
¡
j
¢
= −i +k,
T
¡
2i −3j
¢
= 2T (i) −3T
¡
j
¢
= 2 (i +k) −3 (i −k)
= −i + 5k
For any two real numbers x and y,
T
¡
xi +yj
¢
= xT (i) +yT
¡
j
¢
= x(i +k) +y (i −k)
= (x + y) i + (x −y) k
It follows that
ImT = lin {i, k}
rank T = 2
nullity T = 2 −2 = 0
(b) The matrix of T with respect to the bases
©
i, j
ª
and {i, j, k} is
A ≡
£
T (i) T
¡
j
¢ ¤
=
1 1
0 0
1 −1
¸
¸
(c) First solution. If w
1
= i +j and w
2
= i + 2j, the matrix
P ≡
£
w
1
w
2
¤
=
·
1 1
1 2
¸
134 Linear transformations and matrices
transforms the (canonical) coordinates (1, 0) and (0, 1) of w
1
and w
2
with respect to
the basis {w
1
, w
2
} into their coordinates (1, 1) and (1, 2) with respect to the basis
©
i, j
ª
; hence P is the matrix of coordinate change from basis {w
1
, w
2
} to basis
©
i, j
ª
,
and P
−1
is the matrix of coordinate change from basis
©
i, j
ª
to basis {w
1
, w
2
}. The
operation of the matrix B of T with respect to the bases {w
1
, w
2
} and {i, j, k} can be
described as the combined eﬀect of the following two actions: 1) coordinate change
from coordinates w.r. to basis {w
1
, w
2
} into coordinates w.r. to the basis
©
i, j
ª
(that
is, multiplication by matrix P); 2) transformation according to T as expressed by the
matrix representing T w.r. to the bases
©
i, j
ª
and {i, j, k} (that is, multiplication by
A). Thus
B = AP =
1 1
0 0
1 −1
¸
¸
·
1 1
1 2
¸
=
2 3
0 0
0 −1
¸
¸
Second solution (matrix for T). The matrix B representing T with respect to the
bases {w
1
, w
2
} and {i, j, k} is
B ≡
£
T (w
1
) T (w
2
)
¤
where
T (w
1
) = T
¡
i +j
¢
= T (i) +T
¡
j
¢
= (i +k) + (i −k)
= i
T (w
2
) = T
¡
i + 2j
¢
= T (i) + 2T
¡
j
¢
= (i +k) + 2 (i −k)
= 3i −k
Thus
B =
1 3
0 0
0 −1
¸
¸
(c) The matrix C representing T w.r. to bases {u
1
, u
2
} and {v
1
, v
2
, v
3
} is diagonal,
that is,
C =
γ
11
0
0 γ
22
0 0
¸
¸
if and only if the following holds
T (u
1
) = γ
11
v
1
T (u
2
) = γ
22
v
2
Exercises 135
There are indeed very many ways to achieve that. In the present situation the simplest
way seems to me to be given by deﬁning
u
1
≡ i v
1
≡ T (u
1
) = i +k
u
2
≡ j v
2
≡ T (u
2
) = i −k
v
3
≡any vector such that {v
1
, v
2
, v
3
} is lin. independent
thereby obtaining
C =
1 0
0 1
0 0
¸
¸
16.12.6 n. 16 (p. 597)
We have
D(sin) = cos D
2
(sin) = D(cos) = −sin
D(cos) = −sin D
2
(cos) = D(−sin) = −cos
D(id · sin) = sin +id · cos D
2
(id· sin) = D(sin +id · cos) = 2 cos −id · sin
D(id · cos) = cos −id· sin D
2
(id· cos) = D(cos −id · sin) = −2 sin −id · cos
and hence the matrix representing the diﬀerentiation operator D and its square D
2
acting on lin {sin, cos, id · sin, id · cos} with respect to the basis {sin, cos, id · sin, id · cos}
is
A =
0 −1 1 0
1 0 0 1
0 0 0 −1
0 0 1 0
¸
¸
¸
¸
A
2
=
−1 0 0 −2
0 −1 2 0
0 0 −1 0
0 0 0 −1
¸
¸
¸
¸
2
Contents
I Volume 1 1
3 5 7 9 11 13 15 17 19 21 23 25 25 25 25 25 25 25 26 27 28 28
1 Chapter 1 2 Chapter 2 3 Chapter 3 4 Chapter 4 5 Chapter 5 6 Chapter 6 7 Chapter 7 8 Chapter 8 9 Chapter 9 10 Chapter 10 11 Chapter 11 12 Vector algebra 12.1 Historical introduction . . . . . . . . . . . . 12.2 The vector space of ntuples of real numbers 12.3 Geometric interpretation for n ≤ 3 . . . . . 12.4 Exercises . . . . . . . . . . . . . . . . . . . . 12.4.1 n. 1 (p. 450) . . . . . . . . . . . . . 12.4.2 n. 2 (p. 450) . . . . . . . . . . . . . 12.4.3 n. 3 (p. 450) . . . . . . . . . . . . . 12.4.4 n. 4 (p. 450) . . . . . . . . . . . . . 12.4.5 n. 5 (p. 450) . . . . . . . . . . . . . 12.4.6 n. 6 (p. 451) . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . 7 (p. . . . . . 9 (p. .8. . . . . . . . . . . . 12.4 CONTENTS 12. . . . . . . . . . . . . . . . . . . . . . . . . . . 12. . . . . . . . . . . . . . . . . .8. . 12. . . . . . . . . . . . . .5 n. . .12 n. 12. 451) . . . 12. . .8. . . . . . . . . . 12. . . . . . 460) . . . . . . . .11. . . . . . . . . . . . .4. . . . . . . . . . .4. . . . . . . . 12. . . 456) . . . . . . 12. . 456) . . . . . . . . . . .9 n. . . . . . . . . . . . . . . . . . . . . . . . 12. .8. . . . . . . .11 n. .7 n. . . . . . . . . . . . . . . . . . . . . . . . . . . 12. . . . . . . . . . 451) . . . . . .11 n. . . . .8. . . .6 n. . . 11 (p.11. . . . . . . . . . 451) . .13 n. . 12. 451) . . . . . . . . . . . . . . . . . . . . 8 (p. 10 (p. . . . . . . . . . . . . .5 The dot product . . . . . . 14 (p.8. 12. .8. . . .7 n. . . . . 456) . . . . .4 n. . . . . . . . . . . .11. . . . . . . . 12. . . . . . . . . . . 12. . . . . . . 460) . . . . . 12. . . . .5 n. . . . . . . . . . . . . 457) . . . . . . .1 n. . . . . . . .9 Projections. . . . 22 (p. 456) . 456) . . .11. . 3 (p. . . .9 n. 15 (p. . . . . . . . . . . . 456) . 460) . . . . . . . . . . . . 12. . . . . . . 12. . . . . 12. . . . 8 (p. 456) . 457) . . . . . . 20 (p. . . . . . . . . . . . . . . . . .11.8. . . .7 n. . . .2 n. . . . . . . . . . . . . . . . 16 (p. . . . . . .8. .18 n. 456) . .17 n. . . . . 451) . . . . . . . . . . . . . . . . . 461) . . . . . . . 460) . . .11. . . . . .4 n. . . . . . 12. 24 (p. . . . . 12. . . . . . . . . . . . . . . . . 10 (p. . . 6 (p. . . . . . .8. . . . . . . . .10 The unit coordinate vectors . . . . . . . . . . .11. . .8. . . 456) . . . . . . .4. . 25 (p. . . . . . . . . . 12. 12. . . . . . . . . .8 n. . . . .10 n. . . . . . . . . . 13 (p.8 n. . . . . . . . . . . . . . .3 n. . . . . 12. . . . . . 12. . . . . . . . . . . . . . . .4. . 5 (p. 12. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. . 11 (p. .8. . . . . 10 (p. 457) . .11. . . . . .8.9 n. 5 (p. . . . 2 (p. . . . . . . . . . . . . . . . . .7 Orthogonality of vectors . . . 12. . . . . . . . . . . 3 (p. . . . . . . . .6 Length or norm of a vector . . . . . . 456) . . 460) . . . . . . . . . . . . 1 (p. .10 n. . . . . . . . . . . . . . . . . .11 Exercises . . . . . . . . . . . . . 1 (p. . . . . . . . . . 460) . 12. . . .8. .12 n. . . . . . . . . . .4. . . . . . . 461) . . . .12 The linear span of a ﬁnite set of vectors . 17 (p. . . . . . . . . . . . 456) . . .8. . . 19 (p. . . . . 12. . .8. .10 n. .11. . .2 n. . . . . . . .8. . . . . . . . .6 n. . . . 12. . 2 (p. . . . . . . . . . . . 12. . 456) . . . 7 (p. . . . 12. . 456) .8 n. . . . . . 21 (p. . . . . . . . . . . . .11. . . . . . . . . . . . . . . . .8. . . . . . . 12. . . .14 n. 461) . . . . 17 (p. . . . . . . .8 Exercises . . . . . 457) .15 n. . . . . . . . . . . . . . . . . . . . . .1 n. . . . . 12. 29 29 30 30 30 32 33 33 33 33 33 33 34 34 34 35 36 36 37 39 39 39 40 40 40 41 42 42 43 43 43 43 43 43 44 45 46 46 47 48 48 50 . . . . . . . 12. . . . . . . . 13 (p. . .4. 12. . . 456) . . . . . . . . . . . . 461) . . . . . . . . 12. . . . . . . . . . 12. . 12 (p. . . . . . . . . . . . . . 12. . . . . . . 6 (p. . . . . . . . . . . . . . 12. . . . . . Angle between vectors in nspace 12. .16 n. . . . . . 451) . . . . . . . . . . . . . . . . . . . . . . . . . . .3 n. . . . . .
. . . . . . . . . . . . 8 (p. . . .12 n. . . . . .17 Exercises . . . . . . . . . . . . . . . .14 Bases . . . . . 15 (p. . . . .15. . . . .15 n. . . . .8 n. . . . . . . 18 (p. 13. . . . . .1 Introduction . . . . . . . . . 19 (p. . . . . . . . . . . .15 Exercises . . . . 12. . . . . . . . . .3 n. . . . . . . . . . . . . . . . . . . . 13. . . . .15. . . . . . . . . . . . . . . . . . . . . 468) . . . . . . . . 477) . 12. 467) . . . . . . 12 (p. . . . . . . . . . .15. . . . 6 (p. . . . 468) . .5 Exercises . . . . . 477) . . . . . . . . . . . . . . . . . . . . 468) . . . . . .5. . . . . . . . . . . . . . 12. . . . . . . .5. . . . . . . . . . . . . . . 4 (p. .13 Linear independence . .9 n. . . . . 477) . . . . . . . . . . . . . . . . . 13. . . . . . . .15. . . . 13. . . 2 (p. . . . . . .6 n. . . . . . . . . . . . . . . . . . . . 13. . . . . .CONTENTS 5 12. . . . . . 467) . . . . . . 10 (p. . . . . . . . . . . . . . . . . . 477) . . . . . . . . . . . . . . . . . . . 12. . . . .3 Some simple properties of straight lines . . .6 Planes in euclidean nspaces . . . . . . . . . . . . .5. . . . .5. . . 13. . . . .5. . . 467) . . . . . . . . . 10 (p. . 13. . . . . . 14 (p. . . . . . . . . . . . . 12. 467) . . 8 (p. . . . 13. 12. . . . . . . . .5. . . . . . . . . .5. . .4 Lines and vectorvalued functions . . . 467) . . 17 (p. 6 (p. 12. . . . . . 12. . . . 12. . . . . 12. . . . 477) . . . . . . . . . . . . . . . . . . . 468) . . . . . . . . . . .13 n. . .15. . . . . . 50 50 50 50 50 51 51 51 51 52 53 55 56 56 56 57 57 58 59 59 61 61 61 61 61 61 61 61 62 62 62 63 64 64 65 65 66 67 67 67 67 13 Applications of vector algebra to analytic geometry 13. . 7 (p. . . . . . . . . . 13. . . . . . . . . .14 n. . .10 n. . . . . . . . . . . . . . . . . . . .1 n. . . . . . . . .2 n. . . . . . . 477) . . . . . . . . 12. . . . . . . . . . . . . . 13. . . 12 (p. . . .7 n. . . . . . . . . 468) .11 n. . . 7 (p. . 13. . . . . . . . . .4 n. 477) . . . . .15. . 477) . . . . . . . . . . . .15. . . . . . . . . . . . . . . . . . 9 (p. . . . 477) . 12. . . . . . . . . . . . 467) . . . . . . . . . . . . . . . . . . . . . 467) . . . .2 Lines in nspace . . . . . . . . . . . . . . . . . . 3 (p. . . . . . . 12. . . . . . . . . .5. . . . . . . . . . . . . . . 12. . . . . . . . . . .2 n. . . . . . . . 1 (p.15. .5 n. . . . . . . . . . . 12. . . . . 13. . . . . . . . . . .5. . . . . . . 13. . . . . . . . . . 20 (p. . . . . . . . . . 13. . . . . . . . . . . 13. . . . . 468) .5.7 Planes and vectorvalued functions . . . . . . . . .15. . . . . . . .5 n. . . .8 Exercises . . . . . . . . . . . . . 12. .12 n. . . . . . . . . . . . .5. . . . . . 13. . . 13.15. . . . 477) . . . . . .9 n. . . 13 (p. . . . .10 n. . 12. . 467) . . .7 n. . . . . . . 13. . . . 477) . 13. . . . . . . 3 (p. . . 5 (p. . 5 (p. . . . . . .6 n. . . . . . 467) . of complex numbers . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 n. . . .4 n. . . . .15. . . . .15. . . .15. . . . . . . . .1 n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8 n. . . . . . . . . . . . . . . . . .3 n. . 1 (p. . . . . . . . . . . . . . . . . . 477) . . . . 12. . . . . .15. . . . . .15. . . . . 11 (p. .16 The vector space Vn (C) of ntuples 12.
. . . . . . . 9 (p. . . . . . . . . . . 13 (p. . . . 13. . 5 (p. . . . . . . . . . . . .17. . . . . . . . . .9 The cross product . . . . 13. . . . . . . .12 n. . . . .5 n. . . . . . . . . . . 13. . . . . . . 4 (p.8. . . . . . 487) . . .11 n. .11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 (p. . . . 13. . . . . . . 13. . . . . . . . . .11. . . . 5 (p. . . . . 13.11 n. . . . . . . . . . . . . . . . 483) . .5 n. . . . . . . . . . 9 (p. . . 13. . . . . . . . . . 487) . . . . . . . . . . . . . . . .11. . . . . .10 n. . .11 Exercises . 488) . 13. . . . . . . . . . . . . . . . . . 8 (p. . 3 (p. . 13. . . . . . . . . . . 13. . . .8. . . . . . . .8. . . . . . . . . . . . . . . . . . . 496) . . . . . . . . .1 n. . . . . . . . . .6 n. . . . 5 (p.8. . 487) . . 482) . . . . . . 13. . . . .11. . . 487) . . .13 Cramer’s rule for solving systems of three linear 13. . . . . . . . . 483) . . . . . . . . . . . . . . . . .9 n. . . . . . . . . . . . . . . . . . . . 13. . . 3 (p. . . 488) . . . . . . . . . . . . . . . . . . . . .6 n. 12 (p. . . . . . . . . . 10 (p. . . .11. . . . . . . . . . . . . . . . . . . . . 4 (p. . . .17 Exercises . . . . . . . . . . . . .13 n. . . . . . . . .8. .11. . 13. 3 (p. . 13. . 488) . . 1 (p. . . . . . . . . . 482) . . . . . . . . . . . . . . . . . . . 483) . .12 n. 13. . . . . . .8. . . . . . . 488) . . . . . . . . . . . . . . 13. . 13. . . . . . . . . .4 n. 13. . . . . . 2 (p. . . . . . . . . . . . . .8 n. 482) . . . . . .13 n. . .2 n. . . . . . . . . . . 13. . .10 The cross product expressed as a determinant . . . .11.12 The scalar triple product .17. 13. . . . . . . . . . . . . . . 4 (p. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 n.5 n. . . 488) . . . . . 483) .14 n. . . . . . . . . 13. . . .11. . . . . . . . . . . . . . . .4 n. . . . . . . 67 68 69 69 70 71 72 72 73 74 75 75 75 76 76 76 76 76 76 77 77 77 78 79 79 80 80 82 82 83 84 85 85 85 85 85 86 86 86 87 87 88 . . . . . . . . 483) . . . . . . . . . . . . . . . . . . . .8. . . .17. . . 487) . . . . . 482) . . . . . . . . . . 10 (p. . . . 488) . . . . .11. .4 n. . . . . . . . . . . . . . . 11 (p. . . 11 (p. . . . . . . 13. . . .2 n. . . 6 (p. . . .7 n. .11. . . . . . . . . . . . . . . . . . . 13. . . . . 13. . . . . . . 2 (p.1 n.7 n. . . .2 n. . .11.9 n. . 14 (p. . . . . . . . . . . . . . 1 (p.8 n. 13. . .8. . . . . . . . . . . . 488) .11. . . . . . equations . . . . . . . . . . . . . . .11. . . . 487) . . . . 482) . .17. . . . . . . . . . . . . . . . 13. . . .10 n. . . . 13. . . . . . 7 (p. 13. . . .1 n. .3 n. . . . . . . . . . . . . . .8. .8. . 496) . . . . . . 13. . . . . . . . .15 n. . . . . . . . .11. . . . . . . 496) . . 13. . . . . . . . . . . 488) . . . . . . 482) . . . . . . . . .11. . . . . . . . . . . . . . . . . . . . . . 2 (p. .8. . 13. . . . . . . . . .8. . . 6 (p. 496) . .3 n. 14 (p. .15 Normal vectors to planes . . . . . 13. . . . . . . . .17. . . . 13. . . . . 12 (p. . . . 13. . . . 13. . . . 15 (p. . . . . . . . .6 CONTENTS 13. 482) . . . 13. . . 13. 7 (p. . . . . . . . . . . . . . . . . .16 Linear cartesian equations for planes . . . . 482) . 13. . . . . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . .14 Exercises . . . . 13 (p. . . . 13. . . . . . . . . . 488) . . . . . 496) . . 13. .
. . . . . . . 13. . . . . . . . . . . . . . . . . . . . . . . . . . . 555) . . . . . . . . . . . . . . . . .10 n. . . . 13. . . . . . . . . .9 n. . . . .2 n. . .5. . 555) . . . . 14 (p. . . . . . 15. . . .5. . . . . . . . . . . 13. . . .14 n. . . . . . . 15. . . . . . . . . . . . . . . 2 (p. 496) . 15. . . . . . . . . . . .11 n. . . . .5.13 n. . . . . . . . . . . . . . . . . . . . . . . . . . .12 n. .5. . . . . . . . . . . . . . 497) .5. . . .3 Examples of linear spaces . . . . . . . 555) . . . . . . . . .5. . . . . . . . . . . . . . . 496) . . . . . . . . 496) . . . . 555) . . . . . . . . . . . . . . . . . 555) . . .17. . . . . . . . . . . . . . . 496) . . .17. . . . . . . . . . . . . . . . . . . . . . . 496) . . . . . . . . . . . . . . . . . . 14 (p. . . . . .17. . . 6 (p.9 n. . . . . . . . . . . . . . . . . . . . . . . . .8 n. . . 11 (p. . 11 (p. .18 The conic sections . . . . . .5. . . . . . . . 15 (p. . .17. . . . . . . .17. . . . . . . . . . .15 n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17. . . . . . . . . . .6 n. . . . . .20 Polar equations for conic sections . . . . 13. . . . .17. . . . . . .7 n. . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . 15.22 Conic sections symmetric about the origin 13.5 n. . . . . . .7 n.8 n.5. . .5. . 555) . . . 13. . . . . 13. . .3 n. . . . 13. . . . 496) . . . . . 13. . . . . . .CONTENTS 7 13. . . .25 Miscellaneous exercises on conic sections . . 88 88 88 89 90 90 90 90 91 91 91 91 91 91 92 92 92 92 93 95 95 95 95 95 95 95 96 96 96 97 97 98 98 98 99 99 101 101 102 102 . . . . . . . 15.4 n. . . . . . 15. . . . 22 (p. . . . . . . . . . . . . . . . . . . . . . . . . . 555) . . 13 (p. . . . . .5. 15. . . . . . . . . . . . . . . . . . . . . .5. .1 Introduction . . . 14 Calculus of vectorvalued functions 15 Linear spaces 15. . . 8 (p. . . .19 Eccentricity of conic sections . 13. 17 (p. . 497) . . .5 Exercises . . . .6 n. 15. 555) . . . . . . . . . . . . . . . . . . 18 (p. . 13. . . . . . . . . . . . . . . . . . . . . 6 (p. . . 555) . 15. . 555) . . . . . . . . . . . . 20 (p. . . . . . . . . . . . . . . . . .15 n. . . . . 15. . . . . 13 (p. 15. 4 (p. . . . . . .21 Exercises . . . . . 15.11 n. . . . . . .10 n.17. . . . . . . . 496) . . . . . . . . . . . . . . . . . . . . . . . . 555) . . . . . . 555) . . . . . . . . .1 n. . . . .17. . . . . . . . . . . . . . . . . . .5. . . . . . . . 13. . . .4 Elementary consequences of the axioms 15. . 9 (p. . . . . . . . . . . . . . . . . . . . . 10 (p. 3 (p. . . . . . . . 7 (p. . . . . 19 (p. 555) .5. . 496) . . 15.12 n. . . 1 (p. 13. . .2 The deﬁnition of a linear space . . . . . . . . . . . . . . . . . . . . . . . . .13 n. . 13. . . . . . . . . . . 15. . . . . . .24 Exercises . . . . . . . . . 15. . . . . . . . . . . 15. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13. . . . . .14 n. . 5 (p. . . . . . . . . . . . . . . . . . . . . . 15. . . . 15.5. . . 16 (p. . . . . . . . . . . .23 Cartesian equations for the conic sections 13. . . . . . . . . . . . 13.17. . . . . . . . . . . 17 (p. . . 555) . . . . . . . 555) . . . . . . . . . . . . . .
. . . . . . . . . . . . 560) . . . . . . . . . . . . . . . . . . . 560) . 6 (p. . . . . 15. . . . . . 15. . 15. . .10 n. . . . . . 560) . . .5. . . . . 26 (p. . . . 15. . . . . The GramSchmidt process . . . . . . . . . . 560) . .1 n. 576) . . . . . . . .12. . . 15. . . . . . . . 560) . . . . . . . . . . . . . 15. . . . . . . . . . . . . . . . . . . . . . 576) . . . . 15. . . . 560) . . . . . .9 Exercises . . . . . . . 12 (p. . . . . . 15. . . . . . Norms . 15. . . . . . . . . . . . . . . 15. . . .17 n. . . . . .9 n. . . . 15. . . . . . .9. . . 4 (p. . . . . . . . . . .8 CONTENTS 15. . . . . . . . . . . . . . . . . 15.15 Best approximation of elements in a euclidean space by elements in a ﬁnitedimensional subspace . . . . . . . . . . . 1 (p. . . . . . . . . . .7 n.6 Subspaces of a linear space . . . 560) . . . . .1 n.9. . . . . . . . .5. . . . 15. . 27 (p. . . . . . . 560) . . . . . . . . . .2 n. 11 (p. . . 13 (p. . 15. . . 15. . . . . . . . . 15. . . . . . . . . . . . . . . 576) . . . . 15. . . .8 n. . . . 5 (p. . . . . .9. .18 n. . .17 n. . . . . . . . . . . . . . . . . . . . 555) .9. . . . . . . . 4 (p. . .9.9. . . . . . 15. . . . . . . . . . . 555) . . . . . . 15. . . . . . projections . . 15.4 n. . . . . . . . . 23 (p. . . . . . 15. . . . . 560) . . . . . . . .13 Construction of orthogonal sets. .16. Euclidean spaces. . 15. . . . 9 (p. . . . . 10 (p. . . . .16 Exercises . . . . . . . . . . . . . . . . . . 15. . . . . 15. . . . . . . 1 (p. . . . . . . . . 15. . . . . . . . . . . . . . . . . . . .11 n. . . . . . . . . . . . .5. . . . . . . . .10 Inner products. . .9. . 15.5. . . . .2 n. 23 (p. . . . . . . 14 (p. . .3 n. . . .9. 2 (p. . . . . . . . . . . . . . . . .14 Orthogonal complements. .7 Dependent and independent sets in a linear space . . . . . . . . .19 n.21 n. 24 (p. . . . . . . . . . .9. . . . . . . . . . . . . . . . .18 n. . . 15. 15. . 15. . . . . . . . . . . . 560) . . . 576) . . . . . . . . . . . . . . . . . . . . . 15. . . 15. . . . . . . . .1 n. . . . . . . . . . . . . . . . . . . 555) . . . . . . . . . . . . . . . . . . 16 (p. . . . . . . . . . . . . . . . . . . . . . . . . 15. . . . . . . . . .12 n. . . . . 560) . . . . . . . . 7 (p. . .6 n. . . . . . . . . . . . 9 (p. . . . . . . 15.9.5 n. . .12 Exercises . 15. . . . . 25 (p. . . . . . 15. . . . . . . . 8 (p. . . 560) . . . . . . 560) .9. . . . . . . . . . . . . . . . . . . . . .20 n.9.16. . . .9. . . . . . 15 (p. . . . . . . . . . . . . .8 Bases and dimension . . . . . . . . . 555) . . . . . . 560) .16 n. . . . 567) . . . . . . . . . . . . . . . . . . . . 102 102 102 102 103 103 103 103 103 103 103 104 104 104 104 104 104 105 105 105 105 106 106 106 107 107 108 108 111 111 112 112 112 115 115 115 115 115 116 117 118 . . . . . . . . . . 560) . . 22 (p. . . . 560) . .9. . . . . 555) . . . . . . . . .2 n. . . . . . . .4 n. . . 28 (p. . . . . . . . .9. .9. . 15.9. . . . . . 15. . . . . 15. . . . . . 555) . . . . .16. . . . . . .12. . . .15 n. . . . 560) . 15. . . . 3 (p. . . . . . . . . . .16. . . . . . . . . . . . . . . . . . . . . 560) . . . 11 (p. . . . . .5. . . . .11 Orthogonality in a euclidean space . 567) . . . . . . . . . . . . . . . .14 n. . . . . . . . . . . . . . . . . . . . . . .9. . . 3 (p. . . . . . . . . . . . . . . . . . . . .16 n. . . . . . . . .5. .13 n. . . . . . . . .3 n. . . . . . 2 (p. 15. . . . . . . . . . . . . .
12.4 n. . . . 582) . 16.2 n. . . . . 9 (p. . . . . . . . . . . . . . . .1 n. . . 589) . . .12. . . . . . . . 16 (p. . . . . . . . . . 3 (p. . . . . . . . . . . . 582) . . . . . . . . . . . . .1 n. . . . 16. . . . . . . . . . .11 n. . . . . . . . . . .13 n. 582) . . . . . . . . . . . . 597) . . . .12. . .2 n. . . .10 Matrix representations of linear transformations . . . . 7 (p. . . . .CONTENTS 9 16 Linear transformations and matrices 16. . . . . . .11 Construction of a matrix representation in diagonal 16.4. . .4 n. . . . . . . . 2 (p. 16. . 582) . 16. . . . . . . . 25 (p. . . . . . . . . . . . . . . . . . . . . . . . .4. 16. . . .12 Exercises . . form . . . . . . . . . .3 n. . . . . . . . . . . . 16. . . . . . . . . . . . . . . . 16. . . . 590) . 16. 16.4. . . . . . . . .3 n. .4 Exercises .6 n. . .2 Null space and range . . . . . . . . . . . . . . . . . . . . . . 16. . 16. . 27 (p. . . . . . . . . . 16. . 16. . . . . . . . . . . . . . . . . .14 n. . . . 16. . . . . . . 23 (p. 597) . . . . . . . 16. .15 n. . . . . . . . . . . . . . . . . . 582) . . . . . . . . . . . . . . . . . . . . . .12. . . . . . . . . . . 15 (p. . . . 17 (p. . . . . . . . . 582) . . . .4. . . .2 n.4. 16. . . . . . .4. . . . . . .7 n. . . 16. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4. . . . . . . . . . 582) . . .4. 582) . . . . . 16. 16.4.10 n. .4. . . . . .4. . . . . . . . . . . . 3 (p. . . . . . . . . 596) . . . . . . . .8. . . . . . . . . 16. . . .5 Algebraic operations on linear transformations . . . . . . 16. .3 n. . .9 Linear transformations with prescribed values . . . . . . . . . . . . . . . . . . . . .8 n. 16. . . . . .4. . . . . . . . .4 n. . . . . . . . . . . . 16. . . . . . . . .5 n. . . . . . . . . . . . 582) . . . . . . . . . 17 (p.12 n. . . . 7 (p. . . . . . . . . . . . . . . . . . . . . . . . . . . . 16. . .12. . . . . . . . . . . . . .4. . 582) . . . . . . . . 589) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 (p. . . . . . . . . . .12. . . . . . . . . . . . . . . 596) . . . 582) . . . 16. . . . . . . .6 Inverses . . . . . . . . . . .8 Exercises . . .4. . . .6 n. 16.8. . . . . . . 16. . . . . . . . . . . 16 (p. . . . . . . . . . . . . . . .7 Onetoone linear transformations . . . . 16. . . . . 8 (p. . . .8. . . 4 (p. . . . . 16. . . . . . . . . 596) . . . . . 582) . . . . . . . . . . . . 8 (p. . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . .1 Linear transformations . 16. . . . 1 (p. . . 589) . . . . . . . . . . . . . 5 (p. . . . 4 (p. .8. . . 16. . . . . . . . . . . . . . . . . . . . 582) . . . . .1 n. . . . . . . . . . . . . . . . . 597) . . . . 16. . . . 16. . . . . . . . . . . . . 10 (p. . . . . . 582) . . . . . . . .9 n. . . . . . . . . . . . . . .5 n. . .3 Nullity and rank . . . . 27 (p. . . . . . 121 121 121 121 121 121 122 122 122 123 123 123 123 124 124 125 125 126 126 127 128 128 128 128 128 128 128 129 129 129 129 129 129 131 131 132 133 135 . . . . . . . . . . . 6 (p. . . . 5 (p. . . . . . 16. . . . . . 582) . . . . . . . . . . .
10 CONTENTS .
Part I Volume 1 1 .
.
Chapter 1 CHAPTER 1 .
4 Chapter 1 .
Chapter 2 CHAPTER 2 .
6 Chapter 2 .
Chapter 3 CHAPTER 3 .
8 Chapter 3 .
Chapter 4 CHAPTER 4 .
10 Chapter 4 .
Chapter 5 CHAPTER 5 .
12 Chapter 5 .
Chapter 6 CHAPTER 6 .
14 Chapter 6 .
Chapter 7 CHAPTER 7 .
16 Chapter 7 .
Chapter 8 CHAPTER 8 .
18 Chapter 8 .
Chapter 9 CHAPTER 9 .
20 Chapter 9 .
Chapter 10 CHAPTER 10 .
22 Chapter 10 .
Chapter 11 CHAPTER 11 .
24 Chapter 11 .
4 Exercises 12. (c) a + b − c = (3.2 The vector space of ntuples of real numbers 12. −2) . (1.4. . 7) . (b) a − b = (−3.4.Chapter 12 VECTOR ALGEBRA 12. 0). . 9). (d) 7a − 2b − 3c = (−7. (e) 2a + b − 3c = (0. 21). (4. .2 n. −1. 24. containing all the points (included the starting points A and B.1 Historical introduction 12.3 Geometric interpretation for n ≤ 3 12. 0. 12. 4). 3). 4) . . 450) µ ¶ µ ¶ µ ¶ 7 5 5 11 13 . . I would say).2 . a single picture. 450) (a) a + b = (5. 0. as required. −5) 3 2 2 4 4 The seven points to be drawn are the following: The purpose of the exercise is achieved by drawing. 6.1 n. (0. 2 (p. 1 (p. (3.
(5. (−1. (3. . (−3. the straight line through B − → with direction given by the vector a ≡ OA is obtained. 4) . . the straight line through − → point A with direction given by the vector b ≡ OB is obtained. . 3 (p. 5) . 450) The seven points this time are the following: µ ¶ µ ¶ µ ¶ 5 10 7 5 15 . 1) 3 3 2 2 4 5 4 3 2 1 3 2 1 0 1 2 3 1 2 3 4 5 It can be intuitively seen that. 2) . . .4.26 Vector algebra It can be intuitively seen that. 12. by letting t vary in all R. 2. by letting t vary in all R.3 n.
and it is enough to repeat the construction a few more times to convince oneself that the set © ª xa + yb : (x. 1 . 3 . . 5) 2 4 2 3 3 2 The whole purpose of this part of the exercise is achieved by drawing a single picture. −3) . 4 2 4 2 0 2 4 2 4 (b) It is hard not to notice that all points belong to the same straight line. 4 2 4 4 2 . 450) (a) The seven points to be drawn are the following: µ ¶ µ ¶ µ ¶ µ ¶ 3 5 5 4 7 1 . and B. . and 2. when x = 1 the segment obtained joins the midpoints of 2 the two sides OA and BD. 1. where −→ − → − → OD = OA + OB.4 n. (0.2 . 4 (p. the same construction with the value of x ﬁxed at 1 yields the segment AD. 1 . by the question immediately following. A. indeed. I would say). 5 . (c) If the value of x is ﬁxed at 0 and y varies in [0. −1) . and hence D is the vertex of the parallelogram with three vertices at O. The picture below is made with the value of x ﬁxed at 0.4. Similarly. 4 . (3. 1]. the segment OB is obtained. . containing all the points (included the starting points A and B. (4. This is made clear. 1]2 is matched by the set of all points of the parallelogram OADB. 3 . all the combinations are aﬃne. it seems to me. . y) ∈ [0. as it is going to be clear after the second lecture. .Exercises 27 12.
4.→ II (↑) .6 (a) n.5 n. 451) 1 0 1 x+z d = x 1 + y 1 + z 1 = x + y + z 1 1 0 x+y I x+z =0 II x + y + z = 0 III x+y =0 (c) I x+z =1 II x + y + z = 2 III x+y =3 II − I y=1 II − III z = −1 (↑) . 450) µ ¶ µ ¶ µ ¶ 2 1 c1 +y = x 1 3 c2 I 2x + y = c1 II x + 3y = c2 µ 12.5 2 2. 5 (p.5 1 1.5 3 (d) All the segments in the above construction are substituted by straight lines.28 Vector algebra 4 3 2 1 0 0. of equation 3x − y = 0 and 3x − y = 5 respectively.→ I x = 2 I (↑) . and the resulting set is the (inﬁnite) stripe bounded by the lines containing the sides OB and AD.4.→ III (↑) .→ I x = −z y=0 x=0 z=0 c1 c2 ¶ 3c1 − c2 = 5 µ 3I − II 5x = 3c1 − c2 2II − I 5y = 2c2 − c1 2 1 ¶ 2c2 − c1 + 5 µ 1 3 ¶ (b) . 6 (p. (e) The whole plane. 12.
451) 1 0 1 1 d = x + y 1 1 0 1 I II III IV (c) I II III IV (d) I II III IV x+z =1 x+y+z =2 x+y =3 y=4 x+z =1 x+y+z =5 x+y =3 y=4 IV IV . 451) 1 0 2 x + 2z d = x 1 + y 1 + z 1 = x + y + z 1 1 1 x+y+z I x + 2z = 0 II x + y + z = 0 III x + y + z = 0 I x = −2z (↑) .→ I II (check) y=0 x=0 z=0 0=0 (b) II − I − IV 0 = −3 . 8 (p. 1.8 (a) n.4.→ III (↑) .→ II y = z z ← 1 (−2.→ III (↑) . 7 (p. 1) III − II 0 = 1 (b) (c) I x + 2z = 1 II x + y + z = 2 III x + y + z = 3 12.7 (a) n.→ I II (check) y=4 x = −1 z=2 −1 + 4 + 2 = 5 x+z =0 x+y+z =0 x+y =0 y=0 1 x+z 1 x+y+z +z = 0 x+y 0 y IV IV .Exercises 29 12.4.
a = kd Claim: (∃h ∈ R ∼ {0} . c = hd) ⇔ (∃l ∈ R ∼ {0} . a + b = hd 2. is parallel to d too. 10 (p. If you look carefully. this means that there are two real nonzero numbers α and β such that u = αw and v = βw. kd + b = hd ∃h ∈ R ∼ {0} . 12. (⇒) Since (by I) we have b = c − a. if b is parallel to d.11 n. then (by II) c is the sum of two vectors which are both parallel to d. 451) (b) Here is an illustration of the ﬁrst distributive law (α + β) v = αv + βv . ∃k ∈ R ∼ {0} . ∃k ∈ R ∼ {0} . ⇔ I II ⇔ (b 6= 0) ⇔ l≡h−k ⇔ ∃h ∈ R ∼ {0} . According to the deﬁnition at page 450 (just before the beginning of § 12. 12. 11 (p. it follows that b. (⇐) Since (by I) we have c = a + b. c = hd ∃h ∈ R ∼ {0} .10 n.30 Vector algebra 12. if c is parallel to d. Then µ ¶ v α u = αw = α = v β β that is. b = (h − k) d h 6= k ∃l ∈ R ∼ {0} . is parallel to d too.9 n. b = ld ∃h ∈ R ∼ {0} . they diﬀer only in the phrasing.4. 1.4). then (by II) b is the diﬀerence of two vectors which are both parallel to d. which is nonnull. 451) Let the two vectors u and v be both parallel to the vector w.4. probably).4. u is parallel to v. it follows that c. b = ld) I present two possible lines of reasoning (among others. 451) Assumptions: I c=a+b II ∃k ∈ R ∼ {0} . which is nonnull. 9 (p.
Exercises
31
with v = (2, 1), α = 2, β = 3. The vectors v, αv, βv, αv + βv are displayed by means of representative oriented segments from left to right, in black, red, blue, and redblue colour, respectively. The oriented segment representing vector (α + β) v is above, in violet. The dotted lines are there just to make it clearer that the two oriented segments representing αv + βv and (α + β) v are congruent, that is, the vectors αv + βv and (α + β) v are the same.
8 6 4 2
0
2
4
6
8
10
12
14
tails are marked with a cross, heads with a diamond
An illustration of the second distributive law
α (u + v) = αu + αv
is provided by means of the vectors u = (1, 3), v = (2, 1), and the scalar α = 2. The vectors u and αu are represented by means of blue oriented segments; the vectors v and αv by means of red ones; u + v and α (u + v) by green ones; αu + αv is in violet. The original vectors u, v, u + v are on the left; the “rescaled” vectors αu, αv, α (u + v) on the right. Again, the black dotted lines are there just to emphasize congruence.
32
Vector algebra
8 6 4 2
4
2
0
2
4
6
8
tails are marked with a cross, heads with a diamond 12.4.12 n. 12 (p. 451) The statement to be proved is better understood if written in homogeneous form, with all vectors represented by oriented segments: − → 1− → 1− → OA + AC = OB (12.1) 2 2 Since A and C are opposed vertices, the same holds for O and B; this means that the oriented segment OB covers precisely one diagonal of the parallelogram OABC, and AC precisely covers the other (in the rightwarddownwards orientation), hence what needs to be proved is the following: ³ ´ 1³ ´ → − → − → − → − → 1 − OA + OC − OA = OA + OC 2 2 which is now seen to be an immediate consequence of the distributive properties. − → − → − → − → − → In the picture below, OA is in red, OC in blue, OB = OA + OC in green, and − → − → − → AC = OC − OA in violet.
The dot product
33
The geometrical theorem expressed by equation (12.1) is the following: Theorem 1 The two diagonals of every parallelogram intersect at a point which divides them in two segments of equal length. In other words, the intersection point of the two diagonals is the midpoint of both. Indeed, the lefthand side of (12.1) is the vector represented by the oriented segment OM , where M is the midpoint of diagonal AC (violet), whereas the righthand side is the vector represented by the oriented segment ON , where N is the midpoint of diagonal AB (green). More explicitly, the movement from O to M is described as achieved by the composition of a ﬁrst movement from O to A with a second movement from A towards C, which stops halfway (red plus half the violet); whereas the movement from A to N is described as a single movement from A toward B, −→ − −→ stopping halfway (half the green). Since (12.1) asserts that OM = ON , equality between M and N follows, and hence a proof of (12.1) is a proof of the theorem 12.5 The dot product
12.6
Length or norm of a vector
12.7
Orthogonality of vectors
12.8
Exercises
12.8.1 n. 1 (p. 456) (a) ha, bi = −6 (b) hb, ci = 2 (c) ha, ci = 6 (d) ha, b + ci = 0 (e) ha − b, ci = 4 12.8.2 n. 2 (p. 456) (a) ha, bi c = (2 · 2 + 4 · 6 + (−7) · 3) (3, 4, −5) = 7 (3, 4, −5) = (21, 28, −35) (b) ha, b + ci = 2 · (2 + 3) + 4 · (6 + 4) + (−7) (3 − 5) = 64
456) The statement is false. 1). 1. ci ⇔ ha. 5. (3. . −5α.4 n. bi + y hb. −3α)}α∈R 12. z) must satisfy the two conditions h(2. y. z)i = 0 that is. ha.34 Vector algebra (c) ha + b. 1) c = (1. y) ≡ (−2.8. 0) See also exercise 1. y. bi = 0 ⇔ x ha. (x. (x. 4 . b − ci = 0 ⇔ a ⊥ b − c and the diﬀerence b − c may well be orthogonal to a without being the null vector. −105) (e) a hb. 3 (p.4. Then c = (1.8.−7) 15 = ¡2 7 . Indeed. of coordinates (x. I 2x + y − z = 0 II x − y + 2z = 0 I + II 3x + z = 0 2I + II 5x + y = 0 b = (1. 2) . − 15 15 15 ¢ 12. 5 (p. ci = (2 + 2) · 3 + (4 + 6) · 4 + (−7 + 3) · (−5) = 72 (d) ahb. −1. 5. 456) hc. 6 (p. −7) (2 · 3 + 6 · 4 + 3 · (−5)) = (2.5 n. −4) 6= 0 and h(1. y. z)i = 0 h(1. 2)i = 0. 12. bi = 0 c = xa + yb ¾ ⇒ hxa + yb. bi = ha. −4) . question (d). 2) Thus the set of all such vectors can be represented in parametric form as follows: {(α. The simplest example that I can conceive is in R2 : a = (1. 4. ci = (2. 456) The required vector. 4. 1. bi = 0 ⇔ 7x + 14y = 0 Let (x. 60. −7) 15 = (30. −1) .3 n.ci = (2.8.
bi = 0 and hence 1 (−4. d = a − c = a − αb Thus the second condition becomes hb. for some α to be determined. −8. −1. 2 + 2α) Thus the second condition becomes 1 · (2 − α) + 2 · (−1 − 2α) − 2 (2 + 2α) = 0 that is. ai 2 · 1 + (−1) · 2 + 2 · (−2) 4 = =− 2 + 22 + (−2)2 hb. 8) 9 1 d = (22. −4 − 9α = 0 and hence 1 (−4. a − αbi = 0 that is. 7 (p. ai − α hb. hb. 10) 9 c = α= hb.6 n. −8. d = a − c = (2 − α. 10) 9 c = 2 (same solution.8. 8) 9 1 d = (22. 456) 1 (solution with explicit mention of coordinates) From the last condition. bi 9 1 α=− 4 9 . 2α. −1 − 2α. with no mention of coordinates) From the last condition. c = (α. for some α to be determined. −1. −2α) Substituting in the ﬁrst condition. c = αb Substituting in the ﬁrst condition.Exercises 35 12.
2) or b = (1. Suppose α is nonzero (the situation is completely analogous if β 6= 0 is assumed). (b) b = (2. −1) or b = (−1. Then the ﬁrst equations gives (12.36 Vector algebra 12. (d) b = (−1. If a 6= 0. 456) (a) b = (1. 13 (p. 1). 2). 1) or b = (2.2) gives x = ∓β 2 α2 + β 2 α2 +β 2 α2 = α2 Thus there are only two solutions to the problem. (a) b = (−2.7 n. 1). −1) or b = (1. (c) b = (−3. 456) If b is the required vector.8 n. −1) or b = (2.8. (c) b = (−2. −α) In particular. bi = 0 kbk = kak If a = 0. 1). The above conditions take the form αx + βy = 0 x2 + y 2 = α2 + β 2 Either α or β must be diﬀerent from 0. α) and b = (β. −1). and the coordinates of b by the couple (x. the following conditions must be satisﬁed ha. 12. −a) or b = (−b. y = and hence y = ±α Substituting back in (12. namely. then kak = 0 and hence b = 0 as well. let the coordinates of a be given by the couple (α. y).8. 1). β). a). −2) or b = (3. (d) b = (b. −2) . b = (−β. −1) or b = (−2. 10 (p.2) β x=− y α and substitution in the second equation yields ¶ µ 2 β + 1 y 2 = α2 + β 2 α2 that is. (b) b = (−1.
that is. y. z)]i = 0 that is. 1) . 1) + (3. (x. Then. y. 2 2 2 2 2 (right angle in C. −1. −4) . −4. If the right angle is in C. when rewritten in a more perspicuous way by “completion of the squares” µ ¶2 µ ¶2 µ ¶2 5 3 25 5 x− + y+ + x+ = 2 2 2 4 ¡ ¢ is seen to deﬁne the sphere of center P = 5 . b − xi = 0 or ha. −4) − (x. −1. z)i − h(2. orthogonality of the two vectors − → CA = a − x is required. is unknown). z) . −4. y. (3. same solution. −4.8. [(3. z)i = 0 or x2 + y 2 + z 2 − 5x + 5y + 3z + 6 = 0 The above equation. y. xi + hx. z)] . − 3 and radius 5 . xi = 0 Equivalently. bi . 456) 1 (right angle in C. h(2. 14 (p.Exercises 37 12. 1) − (x. (x. y. − 5 . Thus h[(2.x + ° ° ° 2 ° 2 2 ° °2 ° °2 ° ° ° ° °x − a + b ° = ° a − b ° ° ° 2 ° 2 ° . ha − x. solution with explicit mention of coordinates) Let (x.9 n. hence the vector OC. the two vectors CA and CB must be orthogonal. bi − ha + b. −4)i + h(x. −1. kxk − 2 2 − → CB = b − x ¿ ° ° °2 À ° ° a + b °2 ° ° a+b ° ° = ° a + b ° − ha. with no mention of coordinates) Let − → − → − → a ≡ OA b ≡ OB x ≡ OC − → (stressing that the point C. z) − → − → be the coordinates of C. y.
38 Vector algebra The last characterization of the solution to our problem shows that the solution set is ° ° the locus of all points having ﬁxed distance (° a−b °) from the midpoint of the segment 2 AB. with the notation of point 2. y + 4. 4 (right angle in B. 3. . z + 4). the following must hold D E − − → → 0 = BA. 5) and BC = (x − 1. ha − b. same solution. Indeed. −3. BC = 1 − x + 3y + 12 + 5z + 20 that is. Since BA = − → (−1. x − bi that is. −5). the condition to be required is 0 = ha − b. it is known by elementary geometry that the triangle ACB is rectangle in C. the vectors required to be orthogonal are BA and BC. solution with explicit mention of coordinates) With − → − → − → this approach. 3 (right angle in B. bi Thus the solution plane π is seen to be through B and orthogonal to the segment connecting point B to point A. if π is any plane containing AB. xi = ha − b. x − 3y − 5z = 33 The solution set is the plane π through B and orthogonal to (1. and C is any point of the circle of π having AB as diameter. with no mention of coordinates) Proceeding as in the previous point.
β) = p= 12. 44) 25 I 3α + 4β = 1 II 4α − 3β = 2 1 (11. of equation hb − a. 4α) q = (4β. with α= Thus ¶ 5 5 5 5 p = . we have that p ≡ OP must be equal to αb = − → αOB. 6) 25 The question is identical to the one already answered in exercise 7 of this section.8. −2) 25 (α. 15 (p. . 3) 12. . 2 2 2 2 µ ¶ 3 1 1 3 q = − . −3β) 4II + 3I 25α = 11 4I − 3II 25β = −2 1 (33. − → Recalling solution 2 to that exercise. 456) I c1 − c2 + 2c3 = 0 II 2c1 + c2 − c3 = 0 I + II 3c1 + c3 = 0 I + 2II 5c1 + c2 = 0 c = (−1.10 n.8. xi = hb − a. 12. 17 (p.Exercises 39 5 (right angle in B) It should be clear at this stage that the solution set in this case is the plane π 0 through A and orthogonal to AB. 16 (p. bi 4 . ai It is also clear that π and π 0 are parallel. 456) p = (3α.− . . bi 10 = hb.8.12 n. 456) q= 1 (−8. 2 2 2 2 µ ha. 5.11 n.
You should notice that the above identity has been already used at the end of point 2 in the solution to exercise 14 of the present section. 12.8.13 n. a − bi = ha. 20 (p. 456) It has been quickly seen in class that ka + bk2 = kak2 + kbk2 + 2 ha. bi = 0 ka − bk2 = kak2 + kbk2 − 2 ha. Concerning the geometrical interpretation of the special case of the above identity ka + bk2 = ka − bk2 if and only if ha.8. ka + bk2 − ka − bk2 = 4 ha. 12. ai + 2 ha. a + bi + ha − b. 19 (p. 457) . bi + hb. bi + ha.8. bi it is enough to notice that orthogonality of a and b is equivalent to the property for the parallelogram OACB (in the given order around the perimeter. bi as required. with vertex C opposed to the vertex in the origin O) to be a rectangle.14 n. the sum of the squares of the four sides equals the sum of the squares of the diagonals. bi + hb. bi = 2 kak2 + 2 kbk2 The geometric theorem expressed by the above identity can be stated as follows: Theorem 2 In every parallelogram. that is. 21 (p. and that in such a rectangle ka + bk and ka − bk measure the lengths of the two diagonals. ai − 2 ha. I obtain By subtraction. 456) ka + bk2 + ka − bk2 = ha + b.15 n. bi substituting −b to b.40 Vector algebra 12.
In order to simplify the notation. bi x2 + 4 ha. y). 4ya − 9xbi ¡ ¢ = −9 ha. B. ¢ ¡ ¢ ¡ kuk2 + kvk2 + kwk2 + kzk2 − kck2 + kdk2 = kzk2 − kvk2 − 2 hu. it follows that kbk is equal to 4. and D be the four vertices of the quadrilateral. C.8. wi + 2 hv. 22 (p.Exercises 41 Let A. vi kdk2 = kvk2 + kwk2 + 2 hv. and let M and N be the midpoint of the diagonals AC and DB. thus the condition becomes ¡ ¢ 4 kak2 − 9 kbk2 xy = 0 and choosing now x = y = 1 gives 4 kak2 = 9 kbk2 Since kak is known to be equal to 6. 0 = hxa + yb. −→ w ≡ CD. ui − → v ≡ BC. −→ z ≡ DA = − (u + v + w) kzk2 = kuk2 + kvk2 + kwk2 + 2 hu. wi We are now in the position to prove the theorem. wi = kuk2 + kwk2 + 2 hu. wi kck2 = kuk2 + kvk2 + 2 hu. bi = 0. wi ° ° − °− →°2 = 4 °MN ° 12. vi − 2 hv. bi y 2 + 4 kak2 − 9 kbk2 xy . vi + 2 hu. choosing x = 0 and y = 1 gives ha. let − → u ≡ AB. starting from left − → −→ in clockwise order. Then c −→ − MN −→ − 2MN ° ° − °− →°2 4 °M N ° − → −→ ≡ AC = u + v d ≡ DB = u + z = − (v + w) 1 1 −→ −→ − = AN − AM = u − d − c 2 2 = 2u+ (v + w) − (u + v) = w + u = kwk2 + kuk2 + 2 hw.16 n. 457) Orthogonality of xa + yb and 4ya − 9xb amounts to Since the above must hold for every couple (x.
ka + xbk2 = kak2 + x2 kbk2 + 2x ha. 0 . is true. bi is negative for all x in the open is interval − 2ha. ka + xbk ≥ kak) ⇒ ha.18 n. since a and b have been shown to be orthogonal. In conclusion.17 n. that is. 457) (a) For every x ∈ R. only if it reduces to the second degree term x2 kbk2 . if it is negative.42 Vector algebra Finally. by pure computation. Since the coordinatefree version of the solution procedure is completely independent from the number of coordinates of the vectors involved. 457) This is once again the question raised in exercises 7 and 17.´the trinomial x2 kbk2 + 2x ha. − kbk2 . ai a ha. the full answer to the problem has already been seen to be hb. bi is zero. bi ³ positive. the following biconditional is true: If ha. 24 (p. 3bi = 22 · 62 + 32 · 42 + 2 · 2 · 3 · 0 = 25 · 32 and hence √ k2a + 3bk = 12 2 12. the following biconditional is true: ¡ ¢ (∀x ∈ R. ai d = b− a ha. bi = 0 .8. ka + xbk2 ≥ kak2 ∀x ∈ R. bi ≥ 0 Moreover. ka + xbk2 − kak2 ≥ 0 ⇔ ∀x ∈ R. ai hb. ka + xbk ≥ kak) ⇔ ∀x ∈ R. in the general context of the linear space Rn (where n ∈ N is arbitrary). bi = kak2 + x2 kbk2 if a ⊥ b 2 ≥ kak if a ⊥ b (b) Since the norm of any vector is nonnegative. the trinomial is negative in the open interval 2 ³ ´ kbk 2ha. ai c = 12.bi .8. It follows that the trinomial can be nonnegative for every x ∈ R only if ha.bi 0. I have proved that the conditional (∀x ∈ R. x2 kbk2 + 2x ha. k2a + 3bk2 = k2ak2 + k3bk2 + 2 h2a. 25 (p.
ji 3 c cos a j = = kak kjk 7 ha.11. . ii = kak kik 7 ha.2 n. . Angle between vectors in nspace 43 12. 2 2 2 2 ¶ kbk2 = 4 µ (b) There are just two vectors as required. bi = 10 The projection of a along b is 10 b= 4 12. 9 9 9 ¶ The projection of a along b is 11 b= 9 12. 2 (p. and its opposite: µ ¶ a 6 3 2 u = = .11. the unit direction vector u of a.− kak 7 7 7 µ ¶ a 6 3 2 − = − . Angle between vectors in nspace 12. 3 (p.Projections.3 (a) n. . 460) µ 5 5 5 5 .− .9 Projections.10 The unit coordinate vectors 12.11. 1 (p. . kak 7 7 7 6 ha.11 12. bi = 11 kbk2 = 9 11 22 22 .1 Exercises n. 460) ha. ki 2 c cos a i = =− kak kkk 7 c cos a i = . 460) ha.
AB. The b angles in A. more than that. ri 35 35 41 b cos A = =√ √ = k−qk krk 41 35 41 √ √ hp. 3. −3. AC. The funny thing is in the numerical data of the exercise: it so happens that 1.44 Vector algebra 12. that is. −4) − → r ≡ AB = (−1. −2. OC b ° °° ° an angle of triangle OBC → → °− ° °− ° = cos B OC OB ° °OC ° ° D E − − → → OC.4 Let n. b b C BA. 1) Then B ≡ (1. −6) There is some funny occurrence in this exercise. B.11. cos ACB. 1) − → p ≡ BC = (2. as implicitly argued above. 5 (p. −1. respectively. − → OC. −5) − → q ≡ CA = (−1. This amounts to work. B. which makes a wrong solution apparently correct (or almost correct). −4. one may be led into operate directly with − − − → → → the coordinates of points A. points A. therefore computing D E − − → → OB. Up to this point. there is nothing funny in doing that. C in place of the coordinates of vectors BC. − − → → respectively. OA = BC and OB = AC. −ri 6 6 41 b cos B = =√ √ = kpk k−rk 41 6 41 0 h−p. 5) C ≡ (3. If some confusion is made between points and vectors. C are coplanar with the origin O. −1. − → − → − → − → 2. if one looks only at numerical results. cos C BA. are more precisely described as B AC. ACB. qi b cos C = = √ √ =0 k−pk kqk 6 35 . and a fairly bad one. with the vectors OA. as the three angles of triangle ABC. C. 460) A ≡ (2. OB. OB b ° °° ° an angle of triangle OAB → → °− ° °− ° = cos AOB °OA° °OB ° b b b instead of cos B AC. √ √ h−q. and/or angles. B. being of conceptual type. it’s only a mistake. as a matter of fact. more than that. OA b ° °° ° an angle of triangle OCA → → °− ° °− ° = cos C OA °OC ° °OA° D E − − → → OA.
5 n. but point 2 makes OACB a parallelogram. AB violet It turns out. − → − → − → OA red. if c = −a. that ° ° ° ° ° ° ° ° → → → → °− ° °− ° °− ° °− ° °OA° = °BC ° = kpk °OB ° = °AC ° = k−qk = kqk b b C OA = C BA b b B OC = B AC ° ° ° ° → → °− ° °− ° °OC ° = °AB ° = krk and such a special circumstance leads a wrong solution to yield the right numbers. b 3. bi from ka + c + bk = ka + c − bk it is possible to deduce ha + c. which immediately implies c ac = π 7 c c bc = π − ab = π 8 b b AOB = ACB . OC green. OB blue. and point 3 makes it even a rectangle.11. 460) Since ka + c ± bk2 = ka + ck2 + kbk2 ± 2 ha + c. therefore. as a particular case. bi = 0 This is certainly true. 12.Exercises 45 Point 1 already singles out a somewhat special situation. AOB is a right angle. 6 (p.
7 (a) n. Indeed. µ cos ϑ − 1 sin ϑ − sin ϑ cos ϑ − 1 ¶µ x y ¶ = µ 0 0 ¶ cos ϑ sin ϑ − sin ϑ cos ϑ ¶µ x y ¶ = µ x y ¶ has the trivial solution as its unique solution if ¯ ¯ cos ϑ − 1 sin ϑ ¯ ¯ − sin ϑ cos ϑ − 1 ¯ ¯ ¯ 6= 0 ¯ . the same conclusion holds. 9 π (the second value being superﬂuous). from hc.11. bn i = kak = 6 2 v u n2 (n+1)2 r n(n+1) u 3 n+1 4 cos [ = √ q 2 a bn = t n2 (n+1)(2n+1) = 2 2n + 1 n n(n+1)(2n+1) 6 6 √ 3 π lim [ = lim cos [ = a bn a bn n→+∞ n→+∞ 2 3 12.46 Vector algebra Moreover. ci ha. 10 (p.11.6 n. 8 (p. 461) ha. 460) We have r √ n (n + 1) n (n + 1) (2n + 1) n kbn k = ha. 12. bi = − ha. bi and kck = kak it is easy to check that c cos bc = hb. even if a + c = 0 is not assumed. ci c =− = − cos ab kbk kck kbk kak c c 8 8 and hence that bc = π ± ac = 7 π. bi = cos ϑ sin ϑ − cos ϑ sin ϑ = 0 kak2 = kbk2 = cos2 ϑ + sin2 ϑ = 1 (b) The system µ that is.
11 (p. if ϑ ∈ {2kπ}k∈Z the coeﬃcient matrix of the above system is the identity matrix.8 n. On the other hand. The assumption that OABC is a rhombus is expressed by the equality kak = kck ((rhombus)) .11. 12. 461) Let OABC be a rhombus. Let − → − → − → a ≡ OA (red) b ≡ OB (green) c ≡ OC (blue) Since OABC is a parallelogram. (2k + 1) π)}k∈Z the only vector satisfying the required condition is (0. Thus − → AB = b − → CB = a 6= 6= 6= ∈ / 0 0 1 {2kπ}k∈Z From elementary geometry (see exercise 12 of section 4) the intersection point M of the two diagonals OB (green) and AC (violet) is the midpoint of both. and every vector in R2 satisﬁes the required condition. which I have taken (without loss of generality) with one vertex in the origin O and with vertex B opposed to O. 0). dashed) and OA.Exercises 47 The computation gives 1 + cos2 ϑ + sin2 ϑ − 2 cos ϑ 2 (1 − cos ϑ) cos ϑ ϑ Thus if ϑ ∈ {(−2kπ. and the same is true for CB (red. dashed) and OB are congruent. the oriented segments AB (blue.
The statement to be proved is orthogonality between the diagonals D E − − → → OB. since in every parallelogram ABCD with a = AB and b = AC the − → diagonal vector CB is equal to a − b. 13 (p. too. 12. 461) (a) That the function X Rn → R. c − ai = 0 or kck2 − kak2 + ha. a 7→ ai  i∈n is positive can be seen by the same arguments used for the ordinary norm. AC = 0 ha + c. 17 (p. since norms are nonnegative real numbers.10 n. so that the triangle ABC has side vectors a. Homogeneity is clear.48 Vector algebra that is. nonnegativity is obvious. 461) The equality to be proved is straightforward. too: ∀a ∈ Rn .9 n. The equality in exam can be readily interpreted according to the theorem’s − → − → statement. As a matter of fact.11. ai = 0 and hence (by commutativity) kck2 = kak2 The last equality is an obvious consequence of ((rhombus)). the square of each side is the sum of the squares of the other two sides. ∀α ∈ R. minus their double product multiplied by the cosine of the angle they form. Thus a parallelogram has orthogonal diagonals if and only if it is a rhombus. ci − hc. The “law of cosines” is often called Theorem 3 (Carnot) In every triangle. 12. and a − b. the converse is true. X X X kαak = αai  = α ai  = α ai  = α kak i∈n i∈n i∈n . and strict positivity still relies on the fact that a sum of concordant numbers can only be zero if all the addends are zero. b. The theorem specializes to Pythagoras’ theorem when a ⊥ b.11.
y) ∈ R− : −x − y = 1 ≡ {(x.. the triangle inequality is much simpler to prove for the present norm (sometimes referred to as “the taxicab norm”) than for the euclidean norm: ∀a ∈ Rn . It is homogeneous: ∀a ∈ Rn . y) ∈ R2 : x + y = 1 = S++ ∪ S−+ ∪ S−− ∪ S+− where S++ S−+ S−− S+− © ª ≡ (x. y) ∈ R2 : x + y = 1 + ≡ {(x. but not positive (e.Exercises 49 Finally. ∀b ∈ Rn . with sides parallel to the quadrant bisectrices 2 1 2 1 0 1 1 2 2 (c) The function f : Rn → R. a 7→ X i∈n ai  is nonnegative. f (x. ¯ ¯ ¯ ¯ ¯ ¯ ¯X ¯ ¯ X ¯ ¯X ¯ ¯ ¯ ¯ ¯ ¯ ¯ kαak = ¯ αai ¯ = ¯α ai ¯ = α ¯ ai ¯ = α kak ¯ ¯ ¯ ¯ ¯ ¯ i∈n i∈n i∈n . y) ∈ R+ × R− : x − y = 1} (red) (green) (violet) (blue) Once the lines whose equations appear in the deﬁnitions of the four sets above are drawn. ∀α ∈ R. X X X X ka + bk = ai + bi  ≤ ai  + bi  = ai  + bi  i∈n i∈n i∈n i∈n = kak + kbk (b) The subset of R2 to be described is © ª S ≡ (x. for n = 2. it is apparent that S is a square.g. −x) = 0 ∀x ∈ R). y) ∈ R− × R+ : −x + ª = 1} y © 2 ≡ (x.
y) = − 1 . 2 2 (c) x + y = 3 and y − x = −5 yield (x.1 Exercises n.13 Linear independence 12.14 Bases 12. ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯X ¯ ¯X X ¯ ¯X ¯ ¯X ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯ ka + bk = ¯ (ai + bi )¯ = ¯ ai + bi ¯ ≤ ¯ ai ¯ + ¯ bi ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯ i∈n i∈n i∈n i∈n i∈n = kak + kbk 12.15.12 The linear span of a ﬁnite set of vectors 12. 1 (p. ∀b ∈ Rn . I + III 3x = 9 y = −4 II + III III (check) 3 + 4 = 7 . 467) I 2x + y = 2 II −x + 2y = −11 III x − y = 7 The solution is (x. f is subadditive. too (this is another way of saying that the triangle inequality holds).15. y) = (1. y) = ¡1 ¢ . 12. y) = (3. ∀a ∈ Rn . (a) x + y = 1 and y − x = 0 yield (x. 3 (p.2 n. 467) x (i − j) + y (i + j) = (x + y. Indeed. 1 . 2 2 (d) x + y = 7 and y − x = 5 yield (x. y) = (4.50 Vector algebra Finally. 6).1 . y − x) ¢ ¡ (b) x + y = 0 and y − x = 1 yield (x.15 12. −1). −4).
6 (p. Then.15. γ) Even if parallelism is deﬁned more broadly − see the footnote in exercise 9 of section 4 − α cannot be zero in the present case. 8 (p. 7 (p. for every (α. from (a. at least one between α and β is nonzero. by proving that if a and b are linearly dependent.15. 5 (p. section 8). β. d) has been seen in the lectures (proof of Steinitz’ theorem. 467) (a) The linear combination 1i + 1j + 1k + (−1) (i + j + k) is nontrivial and spans the null vector. it is enough to require (1 + t)2 − (1 − t)2 6= 0 4t 6= 0 t 6= 0 12.15. 467) By the previous exercise.4 n. 467) (a) If there exists some α ∈ R ∼ {0} such that∗ a = αb.Exercises 51 12. The converse is immediate. a and b are linearly dependent. then the linear combination 1a − αb is nontrivial and it generates the null vector. (b) Since. if you prefer). then they are parallel. b) and (c. Thus β αa = −βb. since a and b have been assumed both diﬀerent from the null vector. indeed. If a nontrivial solution (x. part 1) to be equivalent to non existence of nontrivial solutions to the system ax + cy = 0 bx + dy = 0 The system is linear and homogeneous. and hence to each other. 467) Linear independency of the vectors (a. both (a.15. I argue by the principle of counterposition. d) it is immediate to derive ad − bc = 0. ∗ . αi + βj + γk = (α. γ) ∈ R3 . d) must be proportional to (y. and has unknowns and equations in equal number (hence part 2 of the proof of Steinitz’ theorem does not apply). β. c) and (b. and hence a = − α b (or b = − α a. y) exists.3 n. but in the present case they are both nonzero. β 12.5 n. (b) The argument is best formulated by counterposition. c) = h (b. αa + βb = 0 with α = 0 and β 6= 0 implies b = 0. and αa + βb = 0 with β = 0 and α 6= 0 implies a = 0). too 12. −x) (see exercise 10. Let a and b nontrivially generate the null vector: αa + βb = 0 (according to the deﬁnition.6 n. because both a and b have been assumed diﬀerent from the null vector.
(c) Similarly. consider the system I x+y+z =0 II y + z = 0 III 3z = 0 It is immediate to derive that its unique solution is the trivial one. (d) The last argument can be repeated almost verbatim for triples (i. 1. and hence that the given triple is linearly independent. 467) (a) Again from the proof of Steinitz’ theorem. (αi + βj + γk = 0) ⇒ α = β = γ = 0 so that the triple (i. taking into account that αi + β (i + j + k) + γk = (α + β. 0) and ¢ ¡ 0. for every (α.15. part 1. k). 10 (p. γ) ∈ R3 . α + γ) 12. k) and (i + j + k. − 1 .7 n. k) is linearly independent. γ) ∈ R3 . (b) We need to consider the following two systems: I x+y+z =0 II y + z = 1 III 3z = 0 I x+y+z =0 II y + z = 0 III 3z = 1 It is again immediate that the unique solutions to the systems are (−1.52 Vector algebra it is clear that ∀ (α. α + β. j. β + γ. γ) and hence from αi + βj + γ (i + j + k) = 0 it follows α+γ =0 β+γ =0 γ =0 that is. i + j + k) is linearly independent. β. α=β=γ=0 showing that the triple (i. j. 1 respectively. β. β + γ) α (i + j + k) + βj + γk = (α. i + j + k. j. αi + βj + γ (i + j + k) = (α + γ. 3 3 . β.
Thus (a. 3 . Thus the given triple spans R3 . from the fourth. x+z−t x+y+z−t x+y+t y + 3t = = = = 0 0 0 0 Then. b. t. (d) The coordinates of x with respect to {a. that is.15. d) a linearly dependent quadruple. the system I x+y+z = a II y + z = b III 3z = c ¢ ¡ c c has the (unique) solution a − b. −1. For example. are given as the solution (s. c). e}.→ I x = 5 (d) For an arbitrary triple (a. 12. b. v) to the following system: s+u−v s+t+u−v s+t+v t + 3v = = = = 1 2 3 4 . gives y = 0.8 n. Thus (a. x. and hence t. x. 1) (c) Let e ≡ (−1. c) is a linearly independent triple. (b) Any nontrivial linear combination d of the given vectors makes (a. that is.→ II y = − 14 3 (↑) . 1. and z are in turn seen to be equal to 0. and ﬁrst equation in that order. third. c. c. b − 3 . and suppose xa + yb + zc + te = 0. and z are in turn seen to be equal to 0. and ﬁrst equation in that order. 12 (p. third. b. 467) (a) Let xa + yb + zc = 0. from the fourth. x+z x+y+z x+y y = = = = 0 0 0 0 then y. 3. b. d ≡ a + b + c = (2. u. 3). b. c. which has just been seen to be a basis of R4 . subtracting the ﬁrst equation from the second. e) is linearly independent.Exercises 53 (c) The system to study is the following: I x+y+z =2 II y + z = −3 III 3z = 5 III z=5 3 (↑) . 2. and it is linearly independent (as seen at a).
2 a0 ↔ 3 a0 x 0 4a ← 4 a0 − 3 a0 Thus the system has been given the following form: = = = = 1 3 1 3 which is easily solved from the bottom to the top: (s. 2 a0 ← 2 a0 − 1 a0 1 0 1 −1 0 1 0 0 0 1 1 1 0 1 0 3 1 1 0 −1 0 1 1 1 0 0 1 0 0 0 1 3 1 1 0 −1 0 1 1 1 0 0 1 0 0 0 0 3 u+s−v s+t+v t 3v u s t v u s t v u t s v x 1 1 3 4 1 3 1 4 x 1 3 1 3 a2 ↔ a3 . 1. u. and to perform elementary row operations and column exchanges. 1). v) = (1. t. . s t u v x 1 0 1 −1 1 1 1 1 −1 2 1 1 0 1 3 0 1 0 3 4 a1 ↔ a3 .54 Vector algebra It is more direct and orderly to work just with the table formed by the system ¡ ¢ (extended) matrix A x . 1.
t. and a linear dependence in the equation system suggests that it may √ ¢ ¡√ well have nontrivial solutions. 467) √ ¢ ¡√ ¢ ¡ √ ¢ ¡ α 3. 2. Indeed. − 2 . γ) = (0. 1. 1 + γ 0. 1. t) = (0. (b) √ ¢ ¡√ ¢ ¡ √ ¢ ¡ α 2. 0) √ 3α√ β = 0 + I ⇔ II α + √3β + γ = 0 III β + 3γ = 0 √ I 3α + β = 0 √ ⇔ 3II − III − I 2β = 0 √ III β + 3γ = 0 ⇔ (α. 0) I tα + β = 0 ⇔ II α + tβ + γ = 0 III β + tγ = 0 It is clear that t = 0 makes the triple linearly dependent (the ﬁrst and third vector coincide in this case). 2 = (0. 3 = (0. 2. 0) + β (1. 3. 1. t 6= 0. as already noticed. 0 + β 1. From I and III. − 2 . .15. In conclusion. Thus the three given triple of vectors is linearly dependent. 0 + β 1.9 (a) n. 1. (c) α (t. 1. 0) √ I 2α√ β = 0 + ⇔ II α + √2β + γ = 0 III β + 2γ = 0 √ This time the sum of equations I and III (multiplied by 2) is the same as twice equation II. 1. 2 is such a solution. Let us suppose. 0. 0. the given √ ª © √ triple is linearly dependent for t ∈ 0. 0. then. 1) + γ (0. β. 2. −2. Equations II and III then become tβ + 2γ = 0 β + tγ = 0 a 2 by 2 homogeneous system with determinant of coeﬃcient matrix equal to t2 − 2. 0) and the three given vectors are linearly independent. √ ª ©√ Such a system has nontrivial solutions for t ∈ 2.Exercises 55 12. 13 (p. 0. 1 + γ 0. I deduce that α = γ.
β. b + c.56 Vector algebra 12. c) is linearly independent. u − v = e(1) w − z = e(3) v − w = e(2) z = e(4) and (u. (b) Notice that 1 (u + z) = e(1) 2 1 (v − w) = e(3) 2 1 (u − v) = e(2) 2 1 (w − z) = e(4) 2 Since (u. 468) Let a ≡ (0. too (b) On the contrary. w. (c) Similarly. so that v can be dropped. b. α + β + γy.10 n. 14 (p. w. (a) It is clear that v = u + w. it is a basis of R4 . 468) (a) Since the triple (a. a + c) is linearly dependent. (a − b) − (b + c) + (a + c) = 0 it is seen that the triple (a − b. w. every linear combination of u and w has the form (x. v. v. Thus (u. 15 (p.11 n. w. choosing as nontrivial coeﬃcient triple (α. z) is a maximal linearly independent triple. in the order. α (a + b) + β (b + c) + γ (a + c) = 0 ⇔ (α + γ) a + (α + β) b + (β + γ) c = 0 I α+γ =0 α+β =0 ⇔ II III β + γ = 0 I + II − III 2α = 0 ⇔ −I + II + III 2β = 0 I − II + III 2γ = 0 and the triple (a + b. 12. z) spans the four canonical vectors. 1. 17 (p. z) is maximal linearly independent.15. 468) Call as usual u. b. I look for two possible alternative choices of a vector c ≡ (x. a + c) is linearly independent. b + c. w. v. and z the four vectors given in each case.12 n. Since αa + βb + γc = (β + γx. α + β + γz) . 1). z) is maximal linearly independent. −1. y. c) is linearly independent. z) such that the triple (a. y).15. 1). Moreover. v. 12. γ) ≡ (1.15. 1) and b ≡ (1. 1. and cannot be equal to z. y. Thus (u. x.
(0. is closed with respect to formation of linear combinations. 18 (p. since each element of lin T is a linear combination in T . 12.14 n. Let u. 468) (a) It is enough to prove that each element of T belongs to lin S. y. w + z. equations II and III are the same. 1. and then either II or III yields α = 0. 0). 0. Keeping the same notation. since the vectors u and v there coincide with the present ones. a second example is (u. as any subspace of a vector space. I obtain γ (y − z) = 0 my choice of x. v. (1.Exercises 57 Subtracting equation III from equation II. in such a case. (1. if y = z.15. Conversely. and let a and b be the two elements of T (still in the given order). I yields β = 0 (independently of the value assigned to x). w − z). possible choices for c are (0.15. and z must be such to make the following conditional statement true for each (α.13 n. Indeed. and w be the three elements of S (in the given order). γ) ∈ R3 : I β + γx = 0 α=0 II α + β + γy = 0 β=0 ⇒ III α + β + γz = 0 γ=0 Thus any choice of c with y 6= z makes γ = 0 a consequence of IIIII. 12. 468) The ﬁrst example of basis containing the two given vectors is in point (c) of exercise 14 in this section. As an example. v. 0. 19 (p. it is then readily checked that a=u−w b = 2w (b) The converse inclusion holds as well. 0). v =a−b Thus lin S = lin T 1 w= b 2 1 u=v+w = a− b 2 . and system IIII has inﬁnitely many nontrivial solutions α β γ = −γy − β = −γx free provided either x or y (hence z) is diﬀerent from zero. β. 1. 1). and S. 1).
c=u+v d = u + 2v which proves that lin U ⊆ lin S. In conclusion. β) ∈ R2 such that α+β =1 I II 2α + 3β = 0 III 3α + 5β = −1 The above system has the unique solution (3. since from A ⊆ B I can infer A ⊆ lin B. if c and d are the two elements of U . Inverting the above formulas. Indeed. and hence lin A ⊆ lin B. Notice that αc + βd = (α + β. b) is linearly independent. lin S = lin T = lin U 12. B ⊆ lin A and hence.58 Vector algebra Similarly. 20 (p. by part (a) of last exercise. (b) By the last result. −2). u = 2c − d v =d−c It remains to be established whether or not w is an element of lin U . 2α + 3β. b} B ≡ {a + b} lin A ∩ B ⊆ lin B where the couple (a. 468) (a) The claim has already been proved in the last exercise.15 n. from A ∩ B ⊆ A and A ∩ B ⊆ B I infer lin A ∩ B ⊆ lin A which yields lin A ∩ B ⊆ lin A ∩ lin B (c) It is enough to deﬁne A ≡ {a. lin B ⊆ lin lin A = lin A lin A ∩ lin B = lin B On the other hand. 3α + 5β) It follows that w is an element of lin U if and only if there exists (α.15. A∩B =∅ lin A ∩ B = {0} . which proves that lin S ⊆ lin U .
17 Exercises .16 The vector space Vn (C) of ntuples of complex numbers 12.The vector space Vn (C) of ntuples of complex numbers 59 12.
60 Vector algebra .
13. and (e) belong to L.5.5. showing that L is horizontal. Finally. 1 (p. Points (b).Chapter 13 APPLICATIONS OF VECTOR ALGEBRA TO ANALYTIC GEOMETRY 13. because y = 2 requires t = 3. .4 Lines and vectorvalued functions 13. Among the given points. 1). 477) − → A direction vector for the line L is v ≡ 1 P Q = (−2.1 n.3 Some simple properties of straight lines 13. showing that of the three points only (e) belongs to L. Thus a point belongs to L if and only if its second coordinate is equal to 1.2 n. 477) − → A direction vector for the line L is P Q = (4. The parametric equations 2 for L are x = 2 − 2t y = −1 + t If t = 1 I get point (a) (the origin). point (c) does not belong to L.2 Lines in nspace 13. 0). (b). which yields x = −4 6= 1. which requires t = 2.5 Exercises 13.1 Introduction 13. (d). (d) and (e) have the second coordinate equal to 1. This gives x = −2. 2 (p.
I 2h + 2k + 3l = 0 II −2h + 3k + l = 0 III −6h + 4k + l = 0 I − 2II + III 2l = 0 I + II 5k + 4l = 0 2I − III 10h + 5l = 0 the only combination which is equal to the null vector is the trivial one. 477) The parametric equations for L are x = −3 + h y = 1 − 2h z = 1 + 3h The following points belong to L: (c) (h = 1) (d) (h = −1) (e) (h = 5) 13. (d). 5 (p. . −2. 13. Among the given points.5 n. The three points do not belong to the same line.62 Applications of vector algebra to analytic geometry 13. hence the three points do not belong to the same line. 477) I solve each case in a diﬀerent a way. and (e) belong to L. 477) The parametric equations for L are x = −3 + 4k y = 1+k z = 1 + 6h The following points belong to L: (b) (h = −1) (e) µ 1 h= 2 ¶ (f ) µ 1 h= 3 ¶ − → A direction vector for the line L is P Q = (4.5. Thus a point belongs to L if and only if its second coordinate is equal to 1. 3 (p.3 n.5. −2) − → QR = (−1. showing that L is horizontal.4 n. 4 (p. (b). (a) − → P Q = (2.5. 0). 2) The two vectors are not parallel. (b) Testing aﬃne dependence. 0.
and hence they belong to the plane π of equation z = 1. I check that among the twodimensional projections of the oriented segments connecting A with points D to H − → − → − → pxy AD = (−4. 477) The question is easy. First. 13.Exercises 63 (c) The line through P and R has equations x = 2 + 3k y = 1 − 2k z = 1 Trying to solve for k with the coordinates of Q. 6 (p. 8) pxy AH = (12. We can concentrate only on the ﬁrst two coordinates. but it must be answered by following some orderly ¡ ¢ path. as a matter of fact. 2) pxy AE = (−1. 3) − → −→ pxy AG = (−15. Q does not belong to L (P. 1) pxy AF = (−6.6 n. R). to be considered as a picture of the π 15 10 5 15 10 5 0 5 10 15 5 10 15 − → Since the ﬁrst two components of AB are (4. in order to achieve some economy of thought and of computations (there are 8 = 28 2 diﬀerent oriented segments joining two of the eight given points. −2). −7) . all the eight points have their third coordinate equal to 1. we can have a very good hint on the situation by drawing a twodimensional picture. I get k = − 4 from the ﬁrst equation 3 and k = −1 from the second.5. and.
hence they intersect it exactly at one point. and hence no two of them can both belong to a diﬀerent line. Therefore. Direction vectors for these three lines are 1 −→ 1 − → −→ − pxy HE = (−13. it only remains to be checked whether or not the lines through the couples of points (E. b) are determined by the vector equation P + ha = Q + kb which gives P − Q = kb − ha that is. H} 13.8 n. 7 (p. (G. nor D belong to LHE . 477) The coordinates of the intersection point are determined by the following equation system I 1 + h = 2 + 3k II 1 + 2h = 1 + 8k III 1 + 3h = 13k Subtracting the third equation from the sum of the ﬁrst two. C belongs to LEG . G. namely P 1 ≡ {A. C.1) . D. B. 477) (a) The coordinates of the intersection point of L (P . 7) pxy EG = (−7. (H. − → P Q ∈ span {a. I get k = 1.5. F } P 2 ≡ {C. It is seen by direct inspection that. −5) 2 3 and as normal vectors for them I may take nEG ≡ (4. substituting this value in any of the three equations. D. b} (13. 4) pxy GH = (9.5. I end up with their equations as follows LEG : 4x + 7y = 11 LGH : 5x + 9y = 16 LHE : 7x + 13y = 20 All these three lines are deﬁnitely not parallel to LABC . G} P 3 ≡ {F. 9. and point H to the second. Thus there are three (maximal) sets of at least three collinear points. 13. and the two lines intersect at the point of coordinates (5. G) (red). 7) nGH ≡ (5. 9) nHE ≡ (7. F } belong to the same (black) line LABC . The three equations are consistent. I get h = 4. Thus all elements of the set P 1 ≡ {A. H) (blue). and whether or not any elements of P 1 belong to them. F belongs to LGH .7 n. C. E. and neither A nor B. 13) By requiring point E to belong to the ﬁrst and last. 13). within the set P 1 .64 Applications of vector algebra to analytic geometry − → only the ﬁrst and third are parallel to pxy AB. a) and L (Q. E) (green) coincide. B. 8 (p.
477) X (t) = (1 + t. τ ). P i = kP − Qk2 = (λ − %)2 + (µ − σ)2 + (ν − τ )2 The above quadratic polynomial has a second degree term with a positive coeﬃcient.10 n. σ. ν + γt) f (t) ≡ kQ − X (t)k2 = kQ − P − Atk2 = kQk2 + kP k2 + kAk2 t − 2 hQ. the point on L of minimum distance from Q is the orthogonal projection of Q on L. β. Q − X (t0 ) = .9 n. µ. The points of the line L − → through P with direction vector OA are represented in parametric form (the generic point of L is denoted X (t)). . P i + 2 hP − Q. its graph is a parabola with vertical axis and vertex in the point of coordinates ! µ ¶ Ã 2 2 2 b ∆ hQ − P. 2 − 2t. 3 + 2t) (a) d (t) ≡ kQ − X (t)k2 = (2 − t)2 + (1 + 2t)2 + (−2 − 2t)2 = 9t2 + 8t + 9 (b) The graph of ¡ the function t 7→ d (t) ≡ 9t2 + 8t + 9 is a parabola with the ¢ 4 point (t0 . The minimum squared distance is 65 . 65 as vertex.− 9 9 9 9 9 9 hQ − X (t0 ) . 2a 4a kAk2 kAk2 . Ai kAk kP − Qk − hP − Q. 13. µ + βt. Ai = 22 − 2 + 20 =0 9 that is. Ai = α (λ − %) + β (µ − σ) + γ (ν − τ ) 2 c ≡ kQk2 + kP k2 − 2 hQ. d (t0 )) = −√. γ).5. and the 9 9 9 65 minimum distance is 3 . Ai t = at2 + bt + c where a ≡ kAk2 = α2 + β 2 + γ 2 b ≡ hP − Q. P ≡ (λ.5.Exercises 65 13. . (c) ¶ ¶ µ µ 5 26 19 22 1 10 X (t0 ) = . Ai − .− = . ν) and Q ≡ (%. 9 (p. Then X (t) ≡ P + At = (λ + αt. 477) (a) Let A ≡ (α. 10 (p.
Q − P = ca and equation (13. or − → P Q ∈ span {a} that is. a) and L (Q. Ai hA. Ai = hQ − P. a) is P + ha = Q + ka It gives P − Q = (h − k) a and there are two cases: either − → P Q ∈ span {a} / and equation (13. 477) The vector equation for the coordinates of an intersection point of L (P . ∃c ∈ R.5. they coincide). (b) Q − X (t0 ) = Q − (P + At0 ) = (Q − P ) − hQ − X (t0 ) . Ai α (% − λ) + β (σ − β) + γ (τ − ν) = 2 α2 + β 2 + γ 2 kAk where hQ − P. i. 11 (p.2) has no solution (the two lines are parallel). Ai A kAk2 hQ − P.2) . The two expressions P + ha and Q + ka are seen to provide alternative parametrizations of the same line. k) such that k − h = c (the two lines intersect in inﬁnitely many points.2) is satisﬁed by all couples (h.e. Ai − hQ − P. Ai = 0 kAk2 13.11 n.. For each given point of L.66 Applications of vector algebra to analytic geometry Thus the minimum value of this polynomial is achieved at t0 ≡ and it is equal to kAk2 kP − Qk2 − kAk2 kP − Qk2 cos2 ϑ = kP − Qk2 sin2 ϑ 2 kAk −\ → − → ϑ ≡ OA. (13. QP is the angle formed by direction vector of the line and the vector carrying the point Q not on L to the point P on L. the transformation h 7→ k (h) ≡ h + c identiﬁes the parameter change which allows to shift from the ﬁrst parametrization to the second.
2. z) to 4. 2. equating (x. so that the point with coordinates 4.6 Planes in euclidean nspaces 13. 1 belongs to the 2 2 plane.Planes in euclidean nspaces 67 13.8 Exercises 13. 0. 3 belongs to the 2 2 plane. 3) − → P R = (2. so that the point with coordinates 2. proceeding in the same way with the triple (−3.1 n. y. 2 (p. 477) 13. 0. (c) Again. y.3) ¡ ¢ (a) Equating (x. we get 2s + 2t = −4 2s − 2t = 0 3s − t = −2 . −1).3) we get 2 ¡ ¢ yielding s = 1 and t = 0. 2.5. −2. 3 in (13.8. 1 in (13. 12 (p.7 Planes and vectorvalued functions 13. 482) We have: − → P Q = (2. −1) so that the parametric equations of the plane are x = 1 + 2s + 2t y = 1 + 2s − 2t z = −1 + 3s − t (13.12 n.3) we get 2 2s + 2t = 3 2s − 2t = −1 1 3s − t = 2 2s + 2t = 1 2s − 2t = 1 3 3s − t = 2 ¡ ¢ yielding s = 1 and t = 1. z) to 2. 1. ¡ ¢ (b) Similarly.
contradicting the 2 third equation). 1. 1. 2) does not belong to the plane. because the system 2s + 2t = 2 2s − 2t = 0 3s − t = 4 is inconsistent (from the ﬁrst two equations we get s = t = 1. so that the point with coordinates (−3. 3 (p. 0) = (1. 4) x = s+t y = 1+s z = s + 4t . 0.8. 1) v = (1.2 (a) n. 0. the point of coordinates (0. (e) Finally. 4) − (0. (d) The point of coordinates (3. 1. 1. −1) belongs to the plane. 2. 0) = (1. 1. 1) − (0.68 Applications of vector algebra to analytic geometry yielding s = t = −1. contradicting the third equation). 5) does not belong to the plane. because the system 2s + 2t = −1 2s − 2t = −1 3s − t = 1 is inconsistent (the ﬁrst two equations yield s = − 1 and t = 0. 482) x = 1+t y = 2+s+t z = 1 + 4t (b) u = (1. 13. 1.
4 n.. 1. 4 (p. deﬁning p ≡ 1 − (q + r). 0) does not belong to the plane M . 1) Finally. with the third point I get I s − 2t + 1 = 2 II s + 4t + 2 = −3 III 2s + t = −3 I + II − III 2III − I − II check on I check on II check on III t+3=2 2s − 3 = −5 −1 + 2 + 1 = 2 −1 − 4 + 2 = −3 −2 − 1 = −3 (13. 4.8. −3. 482) This exercise is a replica of material already presented in class and inserted in the notes.4) and (2. 482) (a) Solving for s and t with the coordinates of the ﬁrst point. 0) + s (1. 1) 13. and R. 5 (p.3 n. The second point is directly seen to belong to M from the parametric equations (1. 0) a = (1. 4.Exercises 69 13. −3) belongs to M . 0. there exist real numbers q and r such that S = P + q (Q − P ) + r (R − P ) Then. (a) If p + q + r = 1.4): P ≡ (1. then pP + qQ + rR = P − (1 − p) P + qQ + rR = P + (q + r) P + qQ + rR = P + q (Q − P ) + r (R − P ) and the coordinates of pP + qQ + rR satisfy the parametric equations of the plane M through P . 1. 2. 2) b = (−2. I get I s − 2t + 1 = 0 II s + 4t + 2 = 0 III 2s + t = 0 I + II − III t + 3 = 0 2III − I − II 2s − 3 = 0 3 check on I − 6 + 1 6= 0 2 and (0.8. (b) If S is a point of M. 2. 2) + t (−2. (b) The answer has already been given by writing equation (13. Q. S = (1 − q − r) P + qQ + rR = pP + qQ + rR .
better. 6 (p. I obtain the system 2a + 3b + c + d = 0 −2a − b − 3c + d = 0 4a + 3b − c + d = 0 which I solve by elimination µ 0 1 2 0 2 3 1 1 −2 −1 −3 1 4 3 −1 1 2a ← ( 2a + 1a ) 0 0 3a ← 3a − 2 1a 0 0 0 ¶ 2 3 1 1 0 1 −1 1 0 −3 −3 −1 2 3 1 1 0 1 −1 1 0 0 −3 1 ¡ 3a ← 1 ( 3 a0 + 3 1 a0 ) 2 ¢ . 4z = 4 + x + y − z − 4 − 3x + 3y + 3z − 6 I obtain a cartesian equation for π 1 x − 2y + z + 3 = 0 (b) I consider the four coeﬃcients (a. b.5 n. (a) The parametric equations for the ﬁrst plane π 1 are I x = 2 + 3h − k II y = 3 + 2h − 2k III z = 1 + h − 3k Eliminating the parameter h (I − II − III) I get 4k = x − y − z + 2 Eliminating the parameter k (I + II − III) I get 4h = x + y − z − 4 Substituting in III (or.8. 482) I use three diﬀerent methods for the three cases.70 Applications of vector algebra to analytic geometry 13. c. d) of the cartesian equation of π 2 as unknown. in 4 · III). and I require that the three given points belong to π 2 .
2) Thus π 3 coincides with π2 . At any rate (just to show how the method applies in general). they have the same normal direction. however. 1. 482) (a) Only the ﬁrst two points belong to the given plane.Exercises 71 From the last equation (which reads −3c + d = 0) I assign values 1 and 3 to c and d. −5. in order to obtain the third coordinate of a point of M . I have thought it better to show its application here already. −2) × (1. 0. and I obtain P = (1. The cartesian equation of π 2 is x − 2y + z + 3 = 0 (notice that π 1 and π 2 coincide) (c) This method requires the vector product (called cross product by Apostol ). and from the ﬁrst (which is unchanged) I get a = 1. 1) = (2. The plane π3 and the given plane being parallel. (0. −1. Such a normal is n = (2. 1) 2−6+1+d =0 yielding x − 2y + z + 3 = 0 13. 1) (I may have as well taken one of the ﬁrst two points of part a). 0) . 1) and form a linearly independent couple can be chosen as direction vectors for M . 7 (p. 0). a cartesian equation for π 3 is x − 2y + z + d = 0 and the coeﬃcient d is determined by the requirement that π 3 contains (2. Any two vectors which are orthogonal to the normal n = (3. and I get dI = (−1.6 n. Thus I assign values arbitrarily to d2 and d3 in the orthogonality condition 3d1 − 5d2 + d3 = 0 say. 3. (b) I assign the values −1 and 1 to y and z in the cartesian equation of the plane M . 3. 3) The parametric equations for M are x = 1 − s + 5t y = −1 + 3t z = 1 + 3s dII = (5. 0. which is presented in the subsequent section. 3) and (3. since it has the same normal and has a common point with it. −4. from the second (which reads b − c + d = 0) I get b = −2.8. respectively.
− 5 is all right. 2.. Substituting in the lefthand side of (13. just to check on the computations.  1  2  0 s t h k  const. and the constant terms on the righthand one. I proceed to ﬁnish up the elimination.8. − 2 .5). respectively. which is a trace of the parametric equations of M . This is already enough to get the ﬁrst point from the parametric equation of M 0 . On the contrary. and to s = −2 from the ﬁrst. 8 (p.. 5 5 5 5 5 ¡ 2 1¢ and the check by means of (s. which is apparent (though incomplete) from the righthand side of (13. which is likely to have no solutions. This gives R = 2 .  2  5  16 µ 0 0 0 3r ← 3r + 2 2r 0 0 1r ↔ 2r ¶ A handy solution to the last equation (which reads −19h − 21k = 16) is obtained by assigning values 8 and −8 to h and k. 9 (p. thereby suggesting the wrong answer that M and M 0 are parallel. 3) × (3.72 Applications of vector algebra to analytic geometry 13. which leads to t = 11 from the second row. −1 0 −2 −2  2 0 −1 −5 −7  5 0 2 −9 −7  6 s t h k −1 0 −2 −2 0 −1 −5 −7 0 0 −19 −21  const. −4) . 482) The question is formulated in a slightly insidious way (perhaps Tom did it on purpose. the equation system for the coordinates of points in M ∩ M 0 has four unknowns: I 1 + 2s − t = 2 + h + 3k II 1 − s = 3 + 2h + 2k III 1 + 3s + 2t = 1 + 3h + k (13. 17). and I proceed by elimination: s t h k 2 −1 −1 −3 −1 0 −2 −2 3 2 −3 −1 ← 1 r0 + 2 2 r0 3 r0 ← 3 r0 + 3 2 r0 0 0 1r ↔ 2r 1r 0  const.5) I take all the variables on the lefthand side. Another handy solution to¡the equation corresponding¡ the last to ¢ ¢ row of the ﬁnal elimination table is (h. 2. 8.7 n. 3. − 3 . 13.8. I do get indeed (−14. Thus Q = (−14. k) = − 2 . t) = − 5 . 17). 482) (a) A normal vector for M is n = (1. 3. This may easily lead to an incorrect attempt two solve the problem by setting up a system of three equations in the two unknown s and t. 7 . because the parametric equations of the two planes M and M 0 are written using the same names for the two parameters. 1) = (−4.5).).8 n. However.
2 . 0 and R = 0. ¡ ¡ 3 3 the ¢ Q = − 2 . with π parallel to the yaxis. − 3 4 2 13. namely. 3 . and the coordinates of the point P ∈ M do not satisfy the equation of M 0 . −2. 10 (p. 3 r . . the line L can be represented by the simpler system ½ 2x + 2z = −3 4y = 3 as the intersection of two diﬀerent planes π and π 0 . Coordinate of points on L are now easy to produce.Exercises 73 whereas the coeﬃcient vector in the equation of M 0 is (1. and π 0 parallel to ¢ xzplane. 2 ¡ = s5= −7 ¢ Thus L ∩ M consists of a single point. (b) A cartesian equation for M is x − 2y + z + d = 0 From substitution of the coordinates of P . − 2 . 1). The coordinates of the points of the intersection line L ≡ M ∩ M 00 satisfy the system ½ x − 2y + z + 3 = 0 x + 2y + z = 0 By sum and subtraction.9 n. from which the third and second equation give t = 3. the point of coordinates −2. the coeﬃcient d is seen to be equal to 3.. the two planes are parallel. 483) The parametric equations for the line L are x = 1 + 2r y = 1−r z = 1 + 3r and the parametric equations for the plane M are x = 1 + 2s y = 1+s+t z = −2 + 3s + t The coordinates of a point of intersection between L and M must satisfy the system of equations 2r − 2s = 0 r+s+t = 0 3r − 3s − t = −3 The ﬁrst equation yields r = s.g. e. 4 . Since the latter is parallel to the former.8.
−1. −2) and P R = (2. 11 (p. b} is a linearly independent set. 2 (c) Since a normal vector to π c has for components the coeﬃcients of the unknowns in the equation of π c . 3): h(2. −2). because subtraction of the third equation from the second gives y = −2. b as columns ¯ ¯ ¯ ¯ ¯ ¯ 2 2 3 ¯ ¯ 4 0 −5 ¯ ¯ 5 ¯ 4 ¯ 4 ¯ ¯ ¯ ¯ ¯ −1 1 1 ¯ = ¯ −1 1 1 ¯ = ¯ 4 − 4 ¯ = − 1 ¯ ¯ ¯ ¯ ¯ 6 −2 ¯ 2 ¯ 3 3 1 ¯ ¯ 6 0 −2 ¯ and the three equations are again inconsistent. 2. . hence x = 2.g. − → (b) By computing two independent directions for the plane π b . 1. −1. 1. 1 . −1) − (1. L is not parallel to π b . 3) . computing the determinant of the matrix having v. 483) The parametric equations for L are x = 1 + 2t y = 1−t z = 1 + 3t and a direction vector for L is v = (2. e. Thus I consider the system 4 3 2x + y = 2 4 x + y = −1 3x + y = 3 shows that {v. a. 2) − − → (1. 5. it suﬃces to check orthogonality between v and (1. 1.74 Applications of vector algebra to analytic geometry 13. a. With either reasoning. whereas subtraction of the ﬁrst from the third yields x = 1 . 4. contradicting the ﬁrst. I am reduced to the previous case. 3) and 3 . these two values contradicting all equations. 2.. 1. The system to be studied now is 2x + y = 2 4x + 3y = −1 4x + y = 3 where subtraction of the second equation from the third gives 2x = 4. 3) (a) L is parallel to the given plane (which I am going to denote π a ) if v is a linear ¡ ¢ combination of (2. this yields y = −3 in the last two equations. Alternatively. it is seen that L is not parallel to πa . (1. P Q = (3.10 n. 3)i = 9 6= 0 L is not parallel to π c .8.
zQ ) be a point of L. and (2. .6) is satisﬁed.8. 13. 14 (p. 483) Let d be a direction vector for the line L. yP . M contains the points having coordinates equal to (1. 3). t = 1.11 n.. 3) + t (1. 483) Since every point of L belongs to M . 2.13 n.6) If S belongs to the line through P and Q. 5). obtaining x−y+1 =0 13. other than P or Q.8. and let Q ≡ (xQ . by subtracting the ﬁrst equaiton from the second. (1. there exists a real number p such that S = P + p (Q − P ) Then. zP ) is the plane π through Q with − → direction vectors d and QP x = xQ + hd1 + k (xP − xQ ) y = yQ + hd2 + k (yP − yQ ) z = zQ + hd3 + k (zP − zQ ) L belongs to π because the parametric equation of π reduces to that of L when k is assigned value 0. 1. 3. π is the only plane containing L and P .Exercises 75 13. 1) for each t ∈ R. A point S then belongs to M if and only if there exists real numbers q and r such that S = P + q (Q − P ) + r (R − P ) (13. 12 (p. the parametric equations for M are x = 1+r+s y = 2+r+s z = 3 + r + 2s It is possible to eliminate the two parameters at once. by deﬁning q≡p r≡0 it is immediately seen that condition (13. 13 (p.8. e. 2. yQ . Since any two distinct points on L and P determine a unique plane. A plane containing L and the point P ≡ (xP . 483) Let R be any point of the given plane π. Choosing.12 n.g. and P belongs to π because the coordinates of P are obtained from the parametric equation of π when h is assigned value 0 and k is assigned value 1.
9 The cross product 13. 487) A×B (a) kA×Bk = − √4 i + √3 j + 26 26 (b) (c) A×B kA×Bk A×B kA×Bk 41 = − √2054 i − 1 = − √6 i − √1 k 26 A×B or − kA×Bk = √4 i 26 − √3 j 26 − + √1 k.76 Applications of vector algebra to analytic geometry 13. −3) × (3. (e) (A × B) × C = 8i + 3j − 7k.2 n. 0) = (3. 13. 487) − → − → AB × AC = (2.11. −6. 26 √ 18 j 2054 √ 18 j 2054 + √7 k 2054 A×B or − kA×Bk = 1 √ i 6 √ 41 i 2054 − √ 7 k. (g) (A × C) × B = −2i − 8j − 12k. (d) A × (C × A) = 8i + 10j + 4k.1 n. 3 (p. 1 (p. (c) C × A = 4i − 4j + 2k. −1.11 Exercises 13. 15) √ ° 1 °− → ° → − ° 3 35 area ABC = °AB × AC ° = 2 2 (b) . 9. 2054 2 √ j 6 − 1 √ k 6 A×B or − kA×Bk = + 2 √ j 6 + 1 √ k. (h) (A + B) × (A − C) = 2i − 2j. (f) A × (B × C) = 10i + 11j + 5k.10 The cross product expressed as a determinant 13.3 (a) n. 2 (p. (b) B × C = 4i − 5j + 3k. 3) × (3. −2.11. 2.11. 6 13. (i) (A × B) × (A × C) = −2i + 4k. −2) = (10. 487) (a) A × B = −2i + 3j − k. 10) ° 1 °− → ° → − ° 15 area ABC = °AB × AC ° = 2 2 − → − → AB × AC = (3. −5.
Moreover. 4 (p. and kck = 5. ci = hb.11. 487) Let a ≡ (l. m.11.π bc 2 c . ci < 0 kbk c cos bc = − <0 kck i i c ∈ π. (b × a)i + hb. 6 (p. −1) ° √3 1 °− → ° → − ° area ABC = °AB × AC ° = 2 2 n.4 − → − → CA × AB = (−i + 2j − 3k) × (2j + k) = 8i − j − 2k 13. −bi = − kbk2 because b × a is orthogonal to b.6 (a) n. 1. . bi = 0 13. Then ka × bk2 = (mr − nq)2 + (np − lr)2 + (lq − mp)2 = m2 r2 + n2 q 2 − 2mnrq + n2 p2 + l2 r2 −2lnpr + l2 q 2 + m2 p2 − 2lmpq ¡2 ¢¡ ¢ kak2 kbk2 = l + m2 + n2 p2 + q2 + r2 = l2 p2 + l2 q 2 + l2 r2 + m2 p2 + m2 q 2 +m2 r2 + n2 p2 + n2 q 2 + n2 r2 ¢ ¡ (ka × bk = kak kbk) ⇔ ka × bk2 = kak2 kbk2 ⇔ (lp + mq + nr)2 = 0 ⇔ ha. 487) ha. q. b × ai = 0 because b × a is orthogonal to a. Thus hb. 487) 13.5 n. √ (c) By the formula above. (b × a) − bi = hb.Exercises 77 (c) − → − → AB × AC = (0. b + ci = ha. n) and b ≡ (p. 1) × (1. 5 (p.11. r) for the sake of notational simplicity. (b) hb. kck2 = 22 + 12 . bc = −π is impossible. 1. b) is linearly independent and kb × ak > 0. because \ kck2 = kb × ak2 + kbk2 − 2 kb × ak kbk cos (b × a) b = kb × ak2 + kbk2 > kbk2 since (a. 1) = (1. 0.
The proof that (a × b) × b = −a is identical. (b) By Lagrange’s identity again.de). bi = 0. 7 (p. (a × b)×b is either equal to a or to −a.11. ai2 = 1 (c) And again. Since a is a unit vector. . (a × b) × a is either equal to b or to −b. Both the righthand rule and the lefthand rule yield now (a × b) × a = b (d) (a3 b1 − a1 b3 ) a3 − (a1 b2 − a2 b1 ) a2 (a × b) × a = (a1 b2 − a2 b1 ) a1 − (a2 b3 − a3 b2 ) a3 (a2 b3 − a3 b2 ) a2 − (a3 b1 − a1 b3 ) a1 2 (a2 + a2 ) b1 − a1 (a3 b3 + a2 b2 ) 3 = (a2 + a2 ) b2 − a2 (a1 b1 + a3 b3 ) 1 3 (a2 + a2 ) b3 − a3 (a1 b1 + a2 b2 ) 1 2 (a2 + a2 ) b1 + a1 (a1 b1 ) 2 3 (a × b) × a = (a2 + a2 ) b2 + a2 (a2 b2 ) 1 3 (a2 + a2 ) b3 + a3 (a3 b3 ) 1 2 b1 (a × b) × a = b2 b3 (a × b) × b = −a Since a and b are orthogonal.12. kck2 = ka × bk2 kak2 − ha × b. a × b are mutually orthogonal either by assumption or by the properties of the vector product (theorem 13.7 n. by Lagrange’s identity (theorem 13. Similarly.f) ka × bk2 = kak2 kbk2 − ha.12. bi2 = 1 k(a × b) × ak2 = kck2 = 1 Since the direction which is orthogonal to (a × b) and to b is spanned by a. 488) (a) Since kak = kbk = 1 and ha. b. bi2 = 1 so that a × b is a unit vector as well. The three vectors a.78 Applications of vector algebra to analytic geometry 13. k(a × b) × bk2 = ka × bk2 kbk2 − ha × b.
that is. either h kbk = 0 or k kak = 0. From a×b = kak kbk sin ϑ. and observe that a and c are orthogonal. bi = 1 are: I II III IV −2q − r = 3 2p − 2r = 4 p + 2q = −1 2p − q + 2r = 1 Standard manipulations yield 2II + 2IV + III (↑) .bi a of b along a and the projecting vector kak2 (of length kb×ak ) are null. b − ci = 0 imply that b − c = 0. the unique solution is 9p = 9 2 − 2r = 4 −2q + 1 = 3 2+1−2 =1 1−2 =1 b = (1.→ II (↑) .8 n. 488) (a) From a × b = 0. 9 (p. suppose that a 6= 0. Geometrically.11. Then both the projection ha. kak (b) From the previous point. 13. In particular. bi = k kak2 .→ I check on IV check on III that is.9 n. 488) c (a) Let ϑ ≡ ab. in order to satisfy the condition a×b= c kck Thus kak ≤ kbk < +∞. in a×c c×a which case it has to be equal to kak2 or to kak2 . Thus two solutions to the problem are µ ¶ 7 8 11 ± .11. and its norm must depend on ϑ according to the relation kck kbk = (13. In the second case. or there exists some k ∈ R such that b = ka. bi = h kbk2 or ha.7) kak sin ϑ (b) Let (p.− . a × (b − c) = 0. r) be the coordinates of b.Exercises 79 13.− 9 9 9 the vector b must be orthogonal to c. 8 (p. q. either there exists some h ∈ R such that a = hb. Then either ha. b can be taken orthogonal to a as well. and ha. or kbk = 0 (which is equivalent to b = 0). either h = 0 (which means a = 0). the hypotheses a 6= 0. −1. either k = 0 (which means b = 0). the conditions a × b = c and ha. or kak = 0 (which is equivalent to a = 0). −1) The solution to this exercise given by Apostol at page 645 is wrong. . In the ﬁrst case. which can only happen if b = 0.
1. 1.10 n. a × c + αai = 1 that is.8) is © ª x ∈ R3 : ∃α ∈ R.8). and hence it does not meet the additional requirement ha. (a × c) × a = c Therefore. .11 (a) Let n. 0) − → v ≡ BC = (3. 10 (p.9) We are bound to look for other solutions to (13.11. xi = 1 a × (x − z) = a × x − a × z = c − c = 0 which implies that x − z is parallel to a. If x and z both solve (13.8) However.9). α=− Thus the unique solution to the problem is b ≡ a × c− 13.11. by skewsimmetry of the vector product.8). x = a × c + αa From condition (13. 11 (p. c × a is orthogonal to a. 488) − → u ≡ AB = (−2. 488) Replacing b with c in the ﬁrst result of exercise7.80 Applications of vector algebra to analytic geometry 13. −1) Each side of the triangle ABC can be one of the two diagonals of the parallelogram to be determined. −2. a kak2 1 kak2 (13. Thus the set of all solutions to equation (13. ha. 1) − → w ≡ CA = (−1. it is seen that the c × a is a solution to the equation a×x=c (13.
−2. − ¯ −2 0 ¯ ¯ −1 −1 ¯ ¯ ¯¶ ¯ ¯ −2 1 ¯ ¯. In this case E = B + v − u = “A + C − B” = (4. the other is AD.Exercises 81 − → − → − → If one of the diagonals is BC. the other is BE. where BE = BC + BA = v − u. 0. where AD = AB + AC = u − w. the other is CF . where CF = CA + CB = −v + w. 0) (b) µ¯ ¯ 1 0 ¯ u×w = ¯ 1 −1 √ 6= ku × wk = ¯ ¯ ¯ ¯ ¯ . −2. 2. In this case D = A + u − w = “B + C − A” = (0. In this case F = A − v + w = “A + B − C” = (−2.¯ ¯ ¯ ¯ −1 1 ¯ = (−1. −1) √ 6 area (ABC) = 2 . 2) − → − → − → If one of the diagonals is CA. 2) − → − → − → If one of the diagonals is AB.
12 n.b}. Thus z = 0. 488) b + c = 2 (a × b) − 2b ha. 12 (p.82 Applications of vector algebra to analytic geometry 13. since the only vector / which is orthogonal to itself is the null vector. 488) (a) It is true that. (a × b) is nonnull. x+y =0 and x−y =0 (13.11. because it is orthogonal to a and b. then the triple (a + b. (a × b) is orthogonal to both a + b and a − b (as well as to any vector in lin {a. b + ci = −2 ha. bi 1 c cos ab = = kak kbk 2 2 2 2 c ka × bk = kak kbk sin2 ab = 12 13. a × bi − 3 kbk2 = 48 √ 3 hb. if the couple (a. b). since every ntuple having the null vector among its components is linearly dependent. ci c cos bc = = kbk kck 2 . and (13. Indeed. Fourth. b) is linearly independent. in particular. suppose that x (a + b) + y (a − b) + z (a × b) = 0 Then z cannot be diﬀerent from zero. b}).11.13 n. a − b. b) is linearly independent. (a × b) ∈ lin {a. in the ﬁrst place a and b are nonnull. b}. too. too. Third. 13 (p.10) becomes x (a + b) + y (a − b) = 0 which is equivalent to (x + y) a+ (x − y) b = 0 By linear independence of (a. because (a. bi = −4 ha. bi + 9 kbk2 = 192 √ kck = 8 3 hb. since otherwise x y (a × b) = − (a + b) − (a − b) z z x+y y−x a+ b = − z z would belong to lin {a. a × b) is linearly independent. Secondly. ci = 2 hb.10) kck2 = 4 ka × bk2 − 12 ha × b.
(a + b) × (a − b) = a × a − a × b + b × a − b × b = −2a × b and the triple (a. b) is so. The system x+y = 0 x+z = 0 y+z = 0 has the trivial solution as the only solution. b. − → Otherwise.11. then A and C coincide and it is clear that the line through them and B contains all the three points. a + a × b. b) is linearly independent. yielding y− − → → AB = − AC x . a × b) is linearly independent when the couple (a. if AC is null.14 n. too. AC is linearly dependent. y + z must be null. 14 (p. 488) → − → − (a) The cross product AB × AC equals the null vector if and only if the couple ³ ´ − − → → − → AB. y. too. then the triple (a. then the triple (a + b. Indeed. In such a case. let (x. if AC is nonnull. b) is linearly independent. and this in turn implies that both x + y and x + z are also null. b. then in the nontrivial null combination − → − → xAB + y AC = 0 x must be nonnull. Indeed. arguing as in the fourth part of the previous point. if the couple (a. (b) It is true that. (c) It is true that. if the couple (a. (a + b) × (a − b)) is linearly independent. z) be such that x (a + b) + y (a + a × b) + z (b + a × b) = 0 which is equivalent to (x + y) a + (x + z) b + (y + z) a × b = 0 Since.Exercises 83 and hence x = y = 0. 13. b + a × b) is linearly independent.
p × bi + hp. 13.11) 1 = kak2 = kp × bk2 + kpk2 = 2 kpk2 that is. b) deﬁne the same orientation. Since the triples (p. there exists some h ∈ R such that (p × b) × b = hp Thus. ai hp. p. p) deﬁnes the opposite one. pi = hb. kpk = 22 . and (p × b. and P belong to the same line. pi = 0 that is. that is. (b) By the previous point. (b) Since p. h = −1. the set n o − → − → P : AP × BP = 0 is the set of all points P such that A. (c) Since p is orthogonal both to p × b (deﬁnition of vector product) and to b (point a). and p×b are pairwise orthogonal. B. p × bi + hb. 15 (p. 488) (a) From the assumption (p × b) + p = a. p × b) is linearly independent. ai hb. −1}. hb. (p. b. b. taking into account (13. b. y− → B = A − AC x which means that B belongs to the line through A and C. hp. This gives kp × bk = kpk kbk = kpk and hence √ (13. ai = kpk2 = 1 2 1 2 p × (p × b) + p × p = p × a kp × ak = kpk2 kbk = . b. b.84 Applications of vector algebra to analytic geometry that is.11). pi = hp. ¯ π¯ ¯ ¯ h kpk = kp × bk kbk ¯sin ¯ = kpk 2 and h ∈ {1. p is orthogonal to b. p × b} is a basis of R3 . (d) Still from the assumption (p × b) + p = a.11.15 n. the set of all points P belonging to the line through A and B. and {p. p × b) and (p × b.
13 Cramer’s rule for solving systems of three linear equations 13. respectively. I have ﬁnally obtained 1 1 p = a − (b × a) 2 2 13. 4 2 4 with p. the vectors a = p + b × p. and b × a form angles of π . b × p. It follows 4 that k = −1.12 The scalar triple product 13. which is orthogonal to a.12) requires that the angle formed with p by q is − π .15 Normal vectors to planes 13. π . there exists some k ∈ R such that q = k (b × a) Now c kqk2 = 4 kpk2 + kak2 − 4 kpk kak cos pa √ √ 2 2 = 2+1−4 =1 2 2 kqk = 1 k = 1 If on the plane orthogonal to b the mapping u 7→ b × u rotates counterclockwise.16 Linear cartesian equations for planes . so that q is discordant with b × a. On the other hand.The scalar triple product 85 Thus 1 1 p= a+ q 2 2 (13.14 Exercises 13. Since a normal vector to π is b. and 3π .12) where q = 2p − a is a vector in the plane π generated by p and a. the decomposition of p obtained in (13.
− 3 3 3 7 0. . 0 2 ¶ Zaxis : (c) The distance from the origin is 7 . 2.2 (a) n. (x. 9 9 9 . 2 or 7x − 2y + 2z = 9 or 7x − 2y + 2z = 0 n = knk (b) The three intersection points are: Xaxis : (−7. 0.1 n.86 Applications of vector algebra to analytic geometry 13. − . y.17. 496) µ ¶ µ ¶ 7 0. 3 (d) Intersecting with π the line through the origin which is directed by the normal to π x=h y = 2h z = −2h x + 2y − 2z + 7 = 0 yields h + 4h + 4h + 7 = 0 7 h = − µ9 ¶ 7 14 14 (x. 0) Y axis : µ 1 2 2 . (1. z)i = hn. 496) (a) Out of the inifnitely many vectors satisfying the requirement. 0. a distinguished one is n ≡ (2i + 3j − 4k) × (j + k) = 7i − 2j + 2k (b) hn. y. 3)i 13. y.− . (x. . 2 (p.17 Exercises 13. 1 (p.17. z)i = 0 (c) hn. z) = − .
tD = 7 . and 9 6 ° ° 19 √ °−→° 19 kn4 k = 6 °CD° = 18 18 Alternatively. CD = 6 − 9 n4 .17. rewriting the equation of π 4 with normal vector n2 3x − 6y + 3z = 21 d2 − d4  19 =√ kn2 k 54 dist (π. which can be determined by the equations 3tC − 6 (−2tC ) + 3tC = 2 tD − 2 (−2tD ) + tD = 7 −→ ¡ 7 1 ¢ Thus tC = 1 . 2. (b) The straight line through the origin having direction vector n4 has equations x = t y = −2t z = t and intersects π 2 and π 4 at points C and D. 3 (p.4 n. respectively.3 n. n3 i = 0. π 1 and π3 are orthogonal because hn1 . π 0 ) = . −3) and is parallel to the plane of equation 3x − y + 2z = 4 is 3 (x − 1) − (y − 2) + 2 (z + 3) = 0 or 3x − y + 2z + 5 = 0 The distance between the two planes is √ 4 − (−5) 9 14 √ = 14 9+1+4 13.Exercises 87 13. 496) A cartesian equation for the plane which passes through the point P ≡ (1. 4 (p. 496) π1 π2 π3 π4 : : : : x + 2y − 2z = 5 3x − 6y + 3z = 2 2x + y + 2z = −1 x − 2y + z = 7 (a) π 2 and π 4 are parallel because n2 = 3n4 .17.
17. 1. −2) is − → − → n ≡ P Q × QR = (2. a normal vector for the plane through the points P ≡ (1. Q ≡ (3. −2) is − → − → n ≡ P Q × RQ = (1. 2. 4. 2). 9) A cartesian equation for the plane is (x − 2) + 2 (y − 3) + 9 (z + 7) = 0 or x + 2y + 9z = 55 13. 8 (p. −1. 3). 5 (p. 8. 496) A direction vector for the line is just the normal vector of the plane. 7.17. −8) (b) A cartesian equation for π is x + 2y − 2z = 5 (c) The distance of π from the origin is 5 . −3. −1).88 Applications of vector algebra to analytic geometry 13. 3) × (0. 2. 4). 9 (p. Thus the parametric equations are x = 2 + 4h y = 1 − 3h z = −3 + h . 1) × (3. −4. 2. −4) = (4. Q ≡ (2. −4. 3. 496) Proceeding as in points (a) and (b) of the previous exercise. R ≡ (−1. 3. 6) = (10.17. 6 (p.8 n. 496) (a) A normal vector for the plane π through the points P ≡ (1. −7) and a cartesian equation for it is 10x − 3y − 7z + 17 = 0 13. 496) A normal vector to the plane is given by any direction vector for the given line n ≡ (2.7 n. 3 13.17.5 n.6 n. R ≡ (3. 3) = (1. 12) − (1. 1. 2.
3) is −d. 1).17. the moving point has coordinates (−2. −4. 5). (d) At time t = 3. −7.9 n.Exercises 89 13. Thus −1 · −1 − 3 · −4 + 2 · 3 + d = 0 d = −19 A cartesian equation for π 00 is (x + 1) + 3 (y + 4) − 2 (z − 3) = 0 or x + 3y − 2z + 19 = 0 . Substituting. (b) The direction vector of the line L is d = (−1. 5) is 2x + 3y + 2z + 15 = 0 (e) At time t = 2. 2) (c) The hitting instant is determined by substituting the coordinates of the moving point into the equation of the given plane π 2 (1 − t) + 3 (2 − 3t) + 2 (−1 + 2t) + 1 = 0 −7t + 7 = 0 yielding t = 1. A normal vector for the plane π 00 which is orthogonal to L and contains (−1. 2 · (−2) + 3 · (−7) + 2 · 5 + d = 0 d = 15 A cartesian equation for the plane π 0 which is parallel to π and contains (−2. the moving point has coordinates (−1. 496) (a) The position of the point at time t can be written as follows: x= 1−t y = 2 − 3t z = −1 + 2t which are just the parametric equations of a line. −4. −3. Hence the hitting point is (0. 3). −7. 10 (p. −1.
the other values ´ easily obtained: (l.10 From n.17. 1 . 14 (p. 0).12 n.17. 15 (p. 0. 2. 496) I work directly on the coeﬃcient matrix for the equation system: 3 1 1  5 3 1 5  7 1 −1 3  3 With obvious manipulations 0 4 −8  −4 0 0 4  2 1 −1 3  3 ¡ ¢ I obtain a unique solution (x.13 n. y. √8 . m) = − 7 . 496) First I ﬁnd a vector of arbitrary norm which satisﬁes the given conditions: l + 2m − 3n = 0 l − m + 5n = 0 13. 0. z) = 3 .11 n.90 Applications of vector algebra to analytic geometry 13. 2 2 ¡ ¢ Assigning value 1 to n. 8 . 11 (p.17. a cartesian equation for π is x−y+z = 2 13. √3 the required vector is − 122 122 122 . 13 (p. 496) c π ni = 3 c π nj = 4 π c nk = 3 we get A cartesian equation for the plane in consideration is √ (x − 1) + 2 (y − 1) + (z − 1) = 0 or x+ √ √ 2y + z = 2 + 2 n 1³ √ ´ = 1.17. 496) A normal vector for the plane π which is parallel to both vectors i + j and j + k is n ≡ (i + j) × (j + k) = i − j + k Since the intercept of π with the Xaxis is (2. Then are 3 3 ³ √7 . 1 knk 2 13.
which have as normal vectors n1 ≡ (1. 2. −1) yields 6 − 2 − 2 + 4 6 − 2 − 2 + d = 3 3 that is. 3) = (1. 1) Since ` goes through the point P ≡ (1. parametric eqautions for ` are the following: x=1+t y = 2 − 2t z =3+t 13. 4) × (1. 3.17.20 Polar equations for conic sections 13. 4). 20 (p.21 Exercises . 2.19 Eccentricity of conic sections 13.18 The conic sections 13. 2. 3) and n2 ≡ (2. 3). 17 (p. 2.17. a direction vector for ` is d ≡ (2. Thus the required equation is 2x − y + 2z = 8 13. −2.The conic sections 91 13. 497) If the line ` under consideration is parallel to the two given planes. 3.15 n. 497) A cartesian equation for the plane under consideration is 2x − y + 2z + d = 0 The condition of equal distance from the point P ≡ (3.14 n. 2 + d = 6 The two solutions of the above equation are d1 = 4 (corresponding to the plane already given) and d2 = −8.
23 Cartesian equations for the conic sections 13.92 Applications of vector algebra to analytic geometry 13.22 Conic sections symmetric about the origin 13.24 Exercises 13.25 Miscellaneous exercises on conic sections .
Chapter 14 CALCULUS OF VECTORVALUED FUNCTIONS .
94 Calculus of vectorvalued functions .
3 Examples of linear spaces 15.5 Exercises 15.1 Introduction 15. Q. 1 (p. β) = (0. This shows that the set in question is closed with respect to the two linear space operations of function sum and function multiplication by a scalar.2 The deﬁnition of a linear space 15. and S be any four real polynomials. Of course it can also be seen directly that the identically null function x 7→ 0 is a rational function. Indeed. let P . and let f : x 7→ P (x) Q (x) g : x 7→ R (x) S (x) Then for every two real numbers α and β αf + βg : x 7→ αP (x) S (x) + βQ (x) R (x) Q (x) S (x) is a well deﬁned real rational function. −1) respectively. 555) The set of all real rational functions is a real linear space. R. β) = (0. From this the other two existence axioms (of the zero element and of negatives) also follow as particular cases. as the quotient of the constant polynomials P : x 7→ 0 . 0) and for (α. for (α.1 n.Chapter 15 LINEAR SPACES 15.4 Elementary consequences of the axioms 15.5.
Similarly. allows to conclude. that the other linear space axioms hold is a straightforward consequence of the general properties of the operations of function sum and function multiplication by a scalar.2 n. . Indeed.96 Linear spaces and Q : x 7→ 1. 555) The set of all real rational functions having numerator of degree not exceeding the degree of the denominator is a real linear space. and which achieve at 0 the double value they achieve at 1 is a real linear space. and using the same notation. The remaining linear space axioms are immediately seen to hold.5. Indeed. 2 (p. A ﬁnal remark concerning the other axioms.3 n. Indeed. taking into account the general properties of the operations of function sum and function multiplication by a scalar 15. 15. and which have the same value at 0 and 1. 555) The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0 and 1.5. 555) The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0 and 1. since for every two real numbers α and β deg [αP S + βQR] ≤ max {deg P S. 15. and it is anyway clear that the identically null function x 7→ 0 achieves the same value at 0 and 1. it only needs to be proved that if deg P ≤ deg Q and deg R ≤ deg S. for every two real numbers α and β. the two existence axioms follow as particular cases. It may be also noticed that the degree of both polynomials P and Q occurring in the representation of the identically null function as a rational function is zero. (αf + βg) (0) = αf (0) + βg (0) = α2f (1) + β2g (1) = 2 (αf + βg) (1) so that the closure axioms hold. of a type which has become usual at this point. (αf + βg) (0) = αf (0) + βg (0) = αf (1) + βg (1) = (αf + βg) (1) so that the closure axioms hold.5. for every two real numbers α and β. 4 (p.4 n. is a real linear space. deg QR} deg P S = deg P deg S ≤ deg Q deg S deg QR = deg Q deg R ≤ deg Q deg S so that the closure axioms hold. Again. 3 (p. taking into account exercise 1. then deg [αP S + βQR] ≤ deg QS This is clear.
Indeed.Exercises 97 15. that the set in question is an aﬃne subspace of the real linear space). 1]. 1] with σ 0 = 0 and σ m = 1.5. is not a real linear space. let α and β be any two real numbers such that α + β 6= 1.5 n. and some ntuple (δ s )s∈{0}∪n of real numbers∗ . for any two real numbers α and β. Then (αf + βg) (1) = αf (1) + βg (1) = α [1 + f (0)] + β [1 + g (0)] = α + β + αf (0) + βg (0) = α + β + (αf + βg) (0) 6= 1 + (αf + βg) (0) and the two closure axioms fail to hold (the above shows. The other failing axioms are: existence of the zero element (the identically null function has the same value at 0 and at 1). some increasing (n + 1)tuple (τ s )s∈{0}∪n of elements of [0. 1] is a real linear space. and existence of negatives f (1) = 1 + f (0) ⇒ (−f) (1) = −1 + (−f) (0) 6= 1 + (−f ) (0) 15. (Js )s∈n are the partitions of (0. some (m + 1)tuple (γ r )r∈{0}∪m of real numbers. 1] with τ 0 = 0 and τ m = 1. This is necessary in order to allow for “point steps”. Indeed. 1] associated to (σ r )r∈{0}∪m . σ] ≡ {σ}. however. 5 (p. 555) The set of all real valued step functions which are deﬁned on [0. 555) The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0 and 1. χC is the characteristic function of C ¿ 1 if x ∈ C χC ≡ x 7→ 0 if x ∈ C / and (Ir )r∈m. αf + βg = (αγ 0 + βδ 0 ) χ{0} + ∗ (r ∈ m) (s ∈ n) X (αγ r + βδ s ) χKrs (r. and which have value at 1 which exceeds the value at 0 by 1.s)∈m×n I am using here the following slight abuse of notation for degenerate intervals: (σ. 6 (p.5. . let f and g be any two such functions. τ s ] Then.6 n. some increasing (m + 1)tuple (σ r )r∈{0}∪m of elements of [0. so that for some nonnegative integers m and n. f ≡ γ 0 χ{0} + X γ r χIr g ≡ δ 0 χ{0} + X s∈n δ s χJs r∈m where for each subset C of [0. (τ s )s∈{0}∪n Ir ≡ (σ r−1 . σ r ] Js ≡ (τ s−1 .
The axiom of existence of negatives fails. 13 (p. Indeed. and any two real numbers α and β. 11 (p. with the same number of steps. Thus. 15. 7 γ 0 = γ 1 = 0). All the other linear space axioms hold. for some (r. Z 1 Z 1 Z 1 (αf + βg) (x) dx = α f (x) dx + β g (x) dx = 0 0 0 0 Some potential steps may “collapse” if. 1]. Final remark concerning the other linear space axioms.9 n. is a real linear space. 1] → R. since the identically null function is weakly increasing (and. for that matter.8 n. x→+∞ lim (αf + βg) (x) = α lim f (x) + β lim g (x) x→+∞ x→+∞ = 0 so that the closure axioms hold. depending on whether monotonicity is meant in the weak or in the strict sense. by classical theorems in the theory of limits. 555) The set of all real valued functions which are deﬁned and integrable on [0. as every constant function) but not strictly increasing.5. with no more† than m + n steps on [0. s) ∈ m × n) which shows that αf + βg is a step function too. The ﬁrst closure axiom holds. The second closure axiom does not hold. 1] equal to zero. † . weakly decreasing too. Indeed. It may also be noticed independently that the identically null function [0. 7 (p. σr−1 = τ s−1 and ασr + βτ s = ασr−1 + βτ s−1 . for every two such functions f and g. because the function αf is decreasing if f is increasing and α is a negative real number. the set of all real valued and (weakly) increasing functions of a real variable is a convex cone.5. 1]. for any two such functions f and g. 15. as in the previous examples. and a ﬁnal. because the sum of two increasing functions is an increasing function too. It may be noticed independently that the identically null function [0. for that matter. too. with just one step (m = 1. 1] 7→ R. at any x0 ∈ R∪{−∞}). The axiom of existence of the zero element holds or fails. 15. s) ∈ m × n.s)∈m×n is the “meet” partition of (Ir )r∈m and (Js )s∈n Krs ≡ Ir ∩ Js ( (r. 555) The set of all real valued and increasing functions of a real variable is not a real linear space. with the integral over [0. Thus the closure axioms hold.7 n. usual remark concenrning the other linear space axioms applies. x 7→ 0 is indeed a step function. and that the opposite of a step function is a step function too. x 7→ 0 indeed converges to 0 at +∞ (and. and for any two real numbers α and β.5.98 Linear spaces where (Krs )(r. however. 555) The set of all real valued functions which are deﬁned on R and convergent to 0 at +∞ is a real linear space. in the former case.
1]. is a convex cone. but not a linear space. . At any rate. 14 (p. P (n−1) (0) = (n − 1)!p1 P (n) (0) = n!p0 n X i=0 n X f (j) (0) j=0 R1 (αf ) (x) dx < 0.5. The discussion of the above statement is a bit complicated by the fact that nothing is said concerning the point where our Taylor polynomials are to be centered. : x 7→ n!p0 x + (n − 1)!p1 : x 7→ n!p0 P 0 (0) = pn−1 P 00 (0) = 2pn−2 . For any two such functions f and g. .5. . 16 (p. with nonnegative integral over [0. since it coincides with the set of all real polynomials of degree less than or equal to n. . . (αf ) (x) dx > 0.11 n. . so that j! xj pi xn−i . and any two nonnegative real numbers α and β. P (j) (0) = j!pn−j . . Let P : x 7→ Then P0 P 00 .10 n. The set of all real Taylor polynomials of degree less than or equal to n (including the zero polynomial) is a real linear space. 555) The set of all real valued functions which are deﬁned and integrable on [0. 555) First solution. .Exercises 99 15. . . If the center is taken to be 0. P (j) . . if one thinks at the motivation for the deﬁnition of Taylor polynomials: best ndegree polynomial approximation of a given function. This is immediate. as a real function of a real variable). . which is already known to be a real linear space. as it is. the issue is really easy: every polynomial is the Taylor polynomial centered at 0 of itself (considered. then 15. if one takes the standard formula as a deﬁnition Taylorn (f ) at 0 ≡ x 7→ here are the computations. 1]. Z 1 (αf + βg) (x) dx = α R1 0 0 Z 1 f (x) dx + β 0 Z 1 0 g (x) dx ≥ 0 0 It is clear that if α < 0 and axioms 2 and 6 fail to hold. P (n−i)! : x 7→ n−j (n−i−j)! pi xn−j−i i=0 . P (n−1) P (n) P : x 7→ n−1 (n − i) pi xn−1−i Pi=0 : x 7→ n−2 (n − i) (n − (i + 1)) pi xn−2−i i=0 .
100
Linear spaces
and hence
n X P (j) (0) j=0
j!
xj =
On the other hand, if the Taylor polynomials are meant to be centered at some x0 6= 0 Taylorn (f ) at x0 ≡ x 7→
n X f (j) (x0 ) j=0
n X j!pn−j j=0
j!
xj =
n X i=0
pi xn−i = P (x)
j!
(x − x0 )j
it must be shown by more lengthy arguments that for each polynomial P the following holds: n n X P (j) (x0 ) X (x − x0 )j = pi xn−i j! j=0 i=0 Second solution. The set of all real Taylor polynomials (centered at x0 ∈ R) of degree less than or equal to n (including the zero polynomial) is a real linear space. Indeed, let P and Q be any two such polynomials; that is, let f and g be two real functions of a real variable which are m and n times diﬀerentiable in x0 , and let P ≡ Taylorm (f ) at x0 , Q ≡ Taylorn (g) at x0 , that is, P : x 7→
m X f (i) (x0 ) i=0
i!
(x − x0 )
i
Q : x 7→
Suppose ﬁrst that m = n. Then for any two real numbers α and β · (k) ¸¾ m X ½ · f (k) (x0 ) ¸ g (x0 ) +β (x − x0 )k αP + βQ = α k! k!
k=0 m X (αf + βg)(k) (x0 ) = (x − x0 )k k! k=0
n X g (j) (x0 ) j=0
j!
(x − x0 )j
= Taylorm (αf + βg) at x0
n X αf (k) (x0 ) + βg (k) (x0 ) k=0
Second, suppose (without loss of generality) that m > n. In this case, however, αP + βQ = k!
m X αf (k) (x0 ) (x − x0 ) + (x − x0 )k k! k k=n+1
a polynomial which can be legitimately considered the Taylor polynomial of degree n at x0 of the function αf + βg only if all the derivatives of g of order from n + 1 to m are null at x0 . This is certainly true if g itself a polynomial of degree n. In fact, this is true only in such a case, as it can be seen by repeated integration. It is hence necessary, in addition, to state and prove the result asserting that each Taylor polynomial of any degree and centered at any x0 ∈ R can be seen as the Taylor polynomial of itself.
Exercises
101
15.5.12 n. 17 (p. 555) The set S of all solutions of a linear secondorder homogeneous diﬀerential equation where P and Q are given everywhere continuous real functions of a real variable, and (a, b) is some open interval to be determined together with the solution, is a real linear space. First of all, it must be noticed that S is nonempty, and that its elements are indeed real functions which are everywhere deﬁned (that is, with (a, b) = R), due to the main existence and solution continuation theorems in the theory of diﬀerential equations. Second, the operator where RR is the set of all the real functions of a real variable, and D2 is the subset of RR of all the functions having a second derivative which is everywhere deﬁned, is linear: ∀y ∈ D2 , ∀z ∈ D2 , ∀α ∈ R, ∀β ∈ R, L (αy + βz) = (αy + βz)00 + P (αy + βz)0 + Q (αy + βz) = α (y 00 + P y 0 + Qy) + β (z 00 + P z 0 + Qz) = αL (y) + βL (z) Third, D2 is a real linear space, since ∀y ∈ D2 , ∀z ∈ D2 , ∀α ∈ R, ∀β ∈ R, αy + βz ∈ D2 L : D2 → RR , y 7→ y 00 + P y 0 + Qy ∀x ∈ (a, b) , y 00 (x) + P (x) y 0 (x) + Q (x) y (x) = 0
by standard theorems on the derivative of a linear combination of diﬀerentiable functions, and the usual remark concerning the other linear space axioms. Finally, S = ker L is a linear subspace of D2 , by the following standard argument: and the usual remark. 15.5.13 n. 18 (p. 555) The set of all bounded real sequences is a real linear space. Indeed, for every two real sequences x ≡ (xn )n∈N and y ≡ (yn )n∈N and every two real numbers α and β, if x and y are bounded, so that for some positive numbers ε and η and for every n ∈ N the following holds: then the sequence αx + βy is also bounded, since for every n ∈ N Thus the closure axioms hold, and the usual remark concerning the other linear space axioms applies. αxn + βyn  ≤ α xn  + β yn  < α ε + β η xn  < ε yn  < η L (y) = 0 ∧ L (z) = 0 ⇒ L (αy + βz) = αL (y) + βL (z) = 0
102
Linear spaces
15.5.14 n. 19 (p. 555) The set of all convergent real sequences is a real linear space. Indeed, for every two real sequences x ≡ (xn)n∈N and y ≡ (yn )n∈N and every two real numbers α and β, if x and y are convergent to x and y respectively, then the sequence αx + βy converges to αx + βy. Thus the closure axioms hold. Usual remark concerning the other linear space axioms. 15.5.15 n. 22 (p. 555) The set U of all elements of R3 with their third component equal to 0 is a real linear space. Indeed, by linear combination the third component remains equal to 0; U is the kernel of the linear function R3 → R, (x, y, z) 7→ z.
15.5.16 n. 23 (p. 555) The set W of all elements of R3 with their second or third component equal to 0 is not a linear subspace of R3 . For example, (0, 1, 0) and (0, 0, 1) are in the set, but their sum (0, 1, 1) is not. The second closure axiom and the existence of negatives axiom fail too. The other axioms hold. W is not even an aﬃne subspace, nor it is convex; however, it is a cone. 15.5.17 n. 24 (p. 555) The set π of all elements of R3 with their second component which is equal to the third multiplied by 5 is a real linear space, being the kernel of the linear function R3 → R, (x, y, z) 7→ 5x − y.
15.5.18 n. 25 (p. 555) The set ` of all elements (x, y, z) of R3 such that 3x + 4y = 1 and z = 0 (the line through the point P ≡ (−1, 1, 0) with direction vector v ≡ (4, −3, 0)) is an aﬃne subspace of R3 , hence a convex set, but not a linear subspace of R3 , hence not a linear space itself. Indeed, for any two triples (x, y, z) and (u, v, w) of R3 , and any two real numbers α and β 3x + 4y = 1 ½ z=0 3 (αx + βu) + 4 (αy + βv) = α + β ⇒ 3u + 4v = 1 z+w =0 w=0 Thus both closure axioms fail for `. Both existence axioms also fail, since neither the null triple, nor the opposite of any triple in `, belong to `. The associative and distributive laws have a defective status too, since they make sense in ` only under quite restrictive assumptions on the elements of R3 or the real numbers appearing in them.
15.5.19 n. 26 (p. 555) The set r of all elements (x, y, z) of R3 which are scalar multiples of (1, 2, 3) (the line through the origin having (1, 2, 3) as direction vector) is a real linear space. For any
3k) = (αh + βk) (1. 0. its dimension is 2. 555) The set of solutions of the linear homogenous system of equations A (x. 2h.9 15. y. z) 7→ A (x. 3h) + β (k. the linear combination α (h. 15. (0. 1 (p.6 Subspaces of a linear space 15. 3h) and (k. It is immediate to check that every linear combination of linear combinations of a and b is again a linear combination of a and b. A standard basis for it is {(0. b}. 3x − z).Subspaces of a linear space 103 two elements (h. 27 (p. 15.22 above).7 Dependent and independent sets in a linear space 15. 0. (x. 1) . r is the kernel of the linear function R3 → R2 . S1 is the coordinate Y Zplane. z) 7→ (2x − y.20 n. y. 0) .5. and that for every real numbers y and z (0. 2k.9. 2k. 2.8 Bases and dimension 15. 15. y. and for any two real numbers α and β.5.1 Exercises n. 0) + z (0. y. 1. z)0 and hence a real linear space. 555) The subset of Rn of all the linear combinations of two given vectors a and b is a vector subspace of Rn . namely span {a. z)0 = 0 is the kernel of the linear function R3 → R.21 n. 3) belongs to r. 2h. 1. It is clear that the two vectors are linearly independent. (x. 3k) of r. 560) The set S1 of all elements of R3 with their ﬁrst coordinate equal to 0 is a linear subspace of R3 (see exercise 5. 1)}. z) = y (0. 28 (p. y.
A basis for it is {(1. 0) .9. y and z such that x + y = 0 (x.5 n. −1. −1. 1. 0) . its dimension is 2. 560) The set S3 of all elements of R3 with the coordinates summing up to 0 is a linear subspace of R3 (S3 is the plane through the origin and normal vector n ≡ (1. It is clear that the two vectors are linearly independent. −1. 1. −1.104 Linear spaces 15. 3 (p. z) = x (1. and that for every three real numbers x. −1. y.3 n. It is clear that the two vectors are linearly independent. 560) The set S6 of all elements of R3 with the ﬁrst coordinate equal either to the second or to the third is not a linear subspace of R3 (S6 is the union of the plane S4 of exercise 4 and the plane containing the Y axis and the bisectrix of the oddnumbered quadrants of the XZplane). 0. 0. y and z such that x + y + z = 0 (x. 1) does not. but their sum (2. 1. 0) + z (0. 1)}. 560) The set S7 of all elements of R3 with the ﬁrst and second coordinates having identical square is not a linear subspace of R3 (S7 is the union of the two planes containing the Zaxis and one of the bisectrices of the odd and evennumbered quadrants of the XY plane). 1) 15. (1. 7 (p. y. y and z such that x = y (x. A basis for it is {(1. 0) + z (0. 1) both belong to S6 . 0. 0) and (1. 5 (p.7 n. its dimension is 2. 1. and that for every three real numbers x. 1)}. its dimension is 2.9. −1. 0) and (1. but their sum . y.9. 0.9. 0) both belong to S7 . A basis for it is {(1. 1)). 1. (0. (0. 1) 15.9. and that for every three real numbers x. 0) + z (0. For example. 6 (p. −1. 1.2 n. z) = x (1. 4 (p. 1. 0) .6 n. 560) The set S4 of all elements of R3 with the ﬁrst two coordinates equal is a linear subspace of R3 (S4 is the plane containing the Zaxis and the bisectrix of the oddnumbered quadrants of the XY plane). 15. 2 (p.9. 1)}. It is clear that the two vectors are linearly independent. 560) The set S2 of all elements of R3 with the ﬁrst and second coordinate summing up to 0 is a linear subspace of R3 (S2 is the plane containing the Zaxis and the bisectrix of the evennumbered quadrants of the XY plane). (1. 0. For example. (0.4 n. A basis for it is {d}. z) = x (1. its dimension is 1. 1) 15. 560) The set S5 of all elements of R3 with all the coordinates equal is a linear subspace of R3 (S5 is the line through the origin and direction vector d ≡ (1. 15. 1)).
X P : t 7→ p0 + ph th (15. 1. for every α 6= 1.11 n. and it is even closed with respect to the external product (multiplication by arbitrary real numbers). y. 1. 10 (p. and hence a linear space. 8 (p. a linear combination of B is the null polynomial. 0) or their semisum (1.9. 0) do not. w) ∈ R3 . Moreover. 1. then (x. and α (0.558). 0) do not belong to S8 . 0. α (1. 15. since ∀ (x. 15. A base for it is {(0. ∀α ∈ R (x + y = 1) ∧ (u + v = 1) ⇒ [(1 − α) x + αu] + [(1 − α) y + αv] = 1 15. y. p.9 n. 0.9. 0) and Q ≡ (0. 9 (p. 0. . p. 1.9. If (x. ∀ (u. 0)) For example.Exercises 105 (2.9. 560) See exercise 26. A basis for S11 is ª © B ≡ t 7→ th h∈n Indeed. and it is not even convex. It follows that dim S11 = n. 0. v. and taking value 0 at 0 is a linear subspace of the set of all polynomials of degree not exceeding n. It is clear that any linear combination of polynomials in S11 takes value 0 at 0. y.1) h∈n then P belongs to S11 if and only if p0 = 0. z) = y (0. if and only if all its coeﬃcients are null. 0) . 11 (p. If P is any polynomial of degree not exceeding n. a subspace of R3 of dimension 1. 560) The set S10 of all elements (x. S7 is not an aﬃne subspace of R3 . S7 is a cone. y. −1) 15. by the general principle of polynomial identity (which is more or less explicitly proved in example 6.10 n. z) ∈ R3 . However. z) is any element of S10 . z) of R3 such that x+y+z = 0 x−y−z = 0 is a line containing the origin. S8 is an aﬃne subspace of R3 . −1)}. 560) The set S11 of all polynomials of degree not exceeding n.8 n. every polynomial belonging to S11 is a linear combination of B. 560) The set S8 of all elements of R3 with the ﬁrst and second coordinates summing up to 1 is not a linear subspace of R3 (S8 is the vertical plane containing the line through the points P ≡ (1.555.
for any two suach polynomials P and Q. If P and Q are any two polynomials in S.12 n.9. is a linear subspace of the set of all polynomials of degree not exceeding n.1) belongs to S13 if and only if p2 = 0. A basis for S is n¡ o ¢ B ≡ t 7→ th+2 h∈n−2 . Indeed. is a linear subspace of the set of all polynomials of degree not exceeding n. Thus dim S12 = n. (αP + βQ) (0) + (αP + βQ)0 (1) = αP (0) + βQ (0) + αP 0 (0) + βQ0 (0) = α [P (0) + P 0 (0)] + β [Q (0) + Q0 (0)] = α·0+β·0 A basis for S14 is n o ¡ ¢ h+1 B ≡ t 7→ 1 − t. and any two real numbers α and β. (αP + βQ)0 (0) = αP 0 (0) + βQ0 (0) = α · 0 + β · 0 = 0 A polynomial P as in (15. 560) The set S12 of all polynomials of degree not exceeding n.1) belongs to S12 if and only if p1 = 0. with ﬁrst derivative taking value at 0 which is the opposite of the polynomial value at 0. 12 (p. (t 7→ t) That B is a basis can be seen as in the exercise 11. A basis for S is n¡ o ¢ h+1 B ≡ t 7→ t . Thus dim S13 = n. with ﬁrst derivative taking value 0 at 0. (t 7→ 1) . 15. This is proved exactly as in the previous exercise (just replace 0 with 00 ).14 n.9. 15. 14 (p. 560) The set S13 of all polynomials of degree not exceeding n. is a linear subspace of the set of all polynomials of degree not exceeding n. then X R : t 7→ α1 − α1 t + αh+1 th+1 h∈n−1 .106 Linear spaces 15. t 7→ t h∈n−1 Indeed. with second derivative taking value 0 at 0. and α and β are any two real numbers.13 n. A polynomial P as in (15. (t 7→ 1) h∈n−1 That B is a basis can be seen as in the previous exercise. A polynomial P as in (15. 560) The set S14 of all polynomials of degree not exceeding n.1) belongs to S14 if and only if p0 + p1 = 0. 13 (p. if α is the ntuple of coeﬃcients of a linear combination R of B.9.
This can be seen exactly as in exercise 3. and hence that dim S = n. 16 (p. 15. 560) The set S16 of all polynomials of degree not exceeding n. 560) The set S15 of all polynomials of degree not exceeding n. and hence a linear space. the combination must be the trivial one. If P is any polynomial such that p0 + p1 = 0. A polynomial P belongs to S16 if and only if X p0 = p0 + 2h ph h∈n n o ¡ ¢ B ≡ (t 7→ 1) . 15 (p. t 7→ (1 − t) th h∈n−1 that is. if and only if X h∈n 2h ph = 0 . the ntuple α of coeﬃcients of any linear combination of B spanning the null vector must satisfy the conditions α1 = 0 αh+1 = 0 (h ∈ n − 1) that is. if and only if X h∈n ph = 0 A basis for S15 is and the dimension of S15 is n.9. since by the position α1 = p0 = −p1 αh+1 = ph+1 (h ∈ n − 1) the linear combination of B with α as ntuple of coeﬃcients is equal to P . and taking the same value at 0 and at 1 is a linear subspace of the set of all polynomials of degree not exceeding n.555. P belongs to span B. It follows that B is a basis of S14 .555. and taking the same value at 0 and at 2 is a linear subspace of the set of all polynomials of degree not exceeding n.9. 15.Exercises 107 By the general principle of polynomial identity.16 n. p. A polynomial P belongs to S15 if and only if X p0 = p0 + ph h∈n that is. This can be seen exactly as in exercise 3.15 n. p. and hence a linear space.
(e) If S and T are subspaces of V .17 n. 0) . (c) If S = lin S. and hence S ∩ T = lin S ∩ lin T Since the intersection of any family of subspaces is a subspace. 22 (p. (g) Let V ≡ R2 . v = X i∈n αi ui ) hence if T is a subspace. R → R. S ≡ {(1. 0)} L (S) = L (T ) = R2 L (S ∩ T ) = {(x. t 7→ (2 − t) th h∈n−1 and the dimension of S16 is n. y) : y = 0} 15. 1)}. then by point c above lin S = S and lin T = T . 560) (b) It has alredy been shown in the notes that \ lin S ≡ W W subspace of V S⊆W = ( v : ∃n ∈ N. R → R. (0. S ∩ T is a subspace of V . then S is one among the subspaces appearing in the deﬁnition of lin S.108 Linear spaces A basis for S16 is n o ¡ ¢ B ≡ (t 7→ 1) . and hence lin S ⊆ S. (1. 15. x 7→ 1 x 7→ eax x 7→ ebx x 7→ 0 and let (u. ∃u ≡ (ui )i∈n ∈ V n . Conversely. R → R. then T contains lin S. 23 (p. T ≡ {(1. then of course S is a subspace of V because lin S is so. 1)}.9.9. ∃a ≡ (αi )i∈n ∈ F n . Then S ∩ T = {(1. and it contains S. w) ∈ R3 be such that uf + vg + wh = 0 .18 (a) Let n. it follows that S = lin S. if S is a subspace of V . 560) f g h 0 : : : : R → R. 0) . v. since S ⊆ lin S always.
g) is linearly dependent if and only if g is constant. we have u+v+w = 0 u + ea v + eb w = 0 u + e2a v + e2b w = 0 The determinant of the ¯ ¯ 1 1 ¯ ¯ 1 ea ¯ ¯ 1 e2a coeﬃcient matrix of the above system is ¯ ¯ ¯ ¯ 1 ¯ 1 ¯ 1 1 ¯ ¯ ¯ eb ¯ = ¯ 0 ea − 1 eb − 1 ¯ ¯ ¯ ¯ ¯ 0 e2a − 1 e2b − 1 ¯ e2b ¯ ¡ ¢£ ¤ = (ea − 1) eb − 1 eb + 1 − (ea + 1) ¡ ¢¡ ¢ = (ea − 1) eb − 1 eb − ea (15. This yields v=0⇒u=0 and hence v 6= 0.2) has only the trivial solution. similarly. too. if (f. let (u. at most one of the two numbers can be equal to zero. then dim lin {f. 0. −1). If none of them is null. v. For any function g : X → R. w) ≡ (1. w) ≡ (1. It is convenient to state and prove the following (very) simple Lemma 4 Let X be a set containing at least two distinct elements. yf 6= 0. then h = f . v) ∈ R2 ∼ (0. g) is linearly dependent. Then the function yg f − yf g is the null function. the triple (f. g. and dim lin {f. then g = f . respectively). system (15. v. 0) or (u.Exercises 109 In particular.2) Since by assumption a 6= b. Thus if either a or b is null then (f. h) is linearly dependent (it suﬃces to take (u. g (x) = − yf v and g is constant. uyf + vg (x) = 0 Since f is nonnull. g. g. h} = 2. If g is constant. Thus ∀x ∈ X. Conversely. If a = 0. h) is linearly independent. for x = 0. and x = 2. It immediately follows from the above lemma (and from the assumption a 6= b) that if either a = 0 or b = 0. (b) The two functions f : x 7→ eax g : x 7→ xeax . Then u ∀x ∈ X. if b = 0. −1. Proof. 0) be such that uf + vg is the null function. g. h} = 3. and let f : X → R be a nonnull constant function. let Im f ≡ {yf } and Im g ≡ {yg }. the couple (f. too. x = 1.
g. w) = (0. sin (γ + x) = 0 . since [∀x ∈ R. √ α 2 sin x + √ β 2 cos x = 0 2 2 α +β α +β m ∀x ∈ R. g. w) ∈ R3 be such that uf + vg + wh = 0 where f : x 7→ 1 g : x 7→ eax h : x 7→ xeax In particular. and dim lin {f. α sin x + β cos x = 0 m ∀x ∈ R. and x = −1. ∀x ∈ R. 0. Thus dim lin {f. (α + βx) eax = 0] ⇔ [∀x ∈ R. yielding (u. g. h} = 2. we have u+v+w = 0 u + ea v + ea w = 0 u + e−a v − e−a w = 0 a homogeneous system of equations whose coeﬃcient matrix has determinant ¯ ¯ ¯ ¯ ¯ 1 1 ¯ 1 ¯ 1 ¯ 1 1 ¯ ¯ ¯ ¯ a a a a ¯ 1 e e ¯ = ¯ 0 e −1 e −1 ¯ ¯ ¯ ¯ ¯ ¯ 1 e−a −e−a ¯ ¯ 0 e−a − 1 −e−a − 1 ¯ ¡ ¢ = − (ea − 1) e−a + 1 + e−a − 1 = 2e−a (1 − ea ) Thus if a 6= 0 the above determinant is diﬀerent from zero. h) is linearly independent. x = 1. so that dim lin {f. 0). The triple (f. for every two real numbers α and β which are not both null. let (u. v.3) are linearly independent. v. (f) The two functions f : x 7→ sin x g : x 7→ cos x (15. (c) Arguing as in point a. for x = 0. (α + βx) = 0] ⇔ α=β=0 Notice that the argument holds even in the case a = 0. Indeed. if a = 0 then f = g and h = IdR . h} = 3. On the other hand. g} = 2. αex + βxeax = 0] ⇔ [∀x ∈ R.110 Linear spaces are linearly independent.
Norms 111 where α α2 + β 2 γ ≡ sign β arccos p since the sin function is not identically null.11 Orthogonality in a euclidean space . g. dim lin {f. the last condition cannot hold. Euclidean spaces. g.Inner products. 15. x → cos x} = 2. 7 (h) From the trigonometric addition formulas ∀x ∈ R cos 2x = cos2 x − sin2 x = 1 − 2 sin2 x Let then f : x 7→ 1 g : x 7→ cos x The triple (f. Thus dim lin {x 7→ sin x. since g : x 7→ sin2 x f − g + 2h = 0 By the lemma discussed at point a. h} = 2. Euclidean spaces. Norms 15.10 Inner products. h) is linearly dependent.
567) hu1 . u3 i = hu3 . u1 i = ku1 k2 = ku2 k2 = ku3 k2 = 1 cos u2 u3 = q q = [ 2 2 8 3 3 ¯1 t2 ¯ 1 1 t dt = ¯ = − = 0 ¯ 2 −1 2 2 −1 ¯1 µ ¶ Z 1 t3 ¯ 2 ¯ = 1 − −1 = 2 t + t dt = 0 + ¯ 3 −1 3 3 3 −1 Z 1 1 + t dt = t1 + 0 = 1 − (−1) = 2 −1 −1 Z 1 √ 1 dt = 2 ku1 k = 2 −1 r Z 1 2 2 ku2 k = t2 dt = 3 3 −1 r Z 1 2 8 1 + 2t + t2 dt = 2 + ku3 k = 3 3 −1 Z 1 2 3 π u2 u3 = [ 3 √ 2 3 cos u1 u3 = √ q = [ 2 2 8 3 π u1 u3 = [ 6 15. 567) (a) Let for each n ∈ N In ≡ Then Z +∞ −t Z +∞ e−t tn dt 0 I0 = = x→+∞ = 1 e dt = lim e−t dt x→+∞ 0 0 ¯ −t ¯x lim −e 0 = lim − e−x + 1 x→+∞ Z x .112 Linear spaces 15. u2 i = hu2 . 11 (p.1 Exercises n.12 15.2 n.12.12. 9 (p.
since for each n ∈ N lim e−x xn = 0 x→+∞ It follows that ∀n ∈ N.Exercises 113 and. of degree m and n respectively. Then the product f g is a real polynomial of degree m + n. j∈n i+j=k . In+1 = e t dt = lim e−t tn+1 dt x→+∞ 0 0 ½ ¾ Z x ¯ −t n+1 ¯x −t n = lim −e t + (n + 1) e t dt 0 x→+∞ 0 ½Z x ¾ © −x n+1 ª −t n = lim −e x − 0 + (n + 1) lim e t dt x→+∞ x→+∞ 0 Z +∞ = 0 + (n + 1) e−t tn dt 0 Z +∞ −t n+1 Z x = (n + 1) In The integral involved in the deﬁnition of In is always convergent. Let now X i∈m In = n! f : t 7→ α 0 + αi ti g : t 7→ β 0 + X i∈n β j tj be two real polynomials. for each n ∈ N. containing for each k ∈ m + n a monomial of degree k of the form αi ti β j tj = αi β j ti+j whenever i + j = k. Thus f g = α0 β 0 + k∈m+n X X αi β j tk i∈m.
j∈n i+j=k i∈m. xn i = Z +∞ e−t tm tn dt 0 = Im+n = (m + n)! (c) If f : t 7→ (t + 1)2 then t 7→ t4 + 2t3 + 2t2 + 2t + 1 Z +∞ ¡ ¢ hf. j∈n i+j=k 0 = α0 β 0 + k∈m+n X X αi β j k! i∈m. gi = e−t t4 + 2t3 + 2t2 + 2t + 1 dt fg : 0 g : t 7→ t2 + 1 = I4 + 2I3 + 2I2 + 2I1 + I0 = 4! + 2 · 3! + 2 · 2! + 2 · 1! + 1 = 43 .114 Linear spaces and the scalar product of f and g is results in the sum of m + n + 1 converging integrals Z +∞ X X hf. j∈n i+j=k (b) If for each n ∈ N xn : t 7→ tn then for each m ∈ N and each n ∈ N hxm . j∈n i+j=k +∞ e−t tk dt = α0 β 0 I0 + k∈m+n X X αi β j Ik i∈m. gi = e−t α0 β 0 + αi β j tk dt 0 k∈m+n = α0 β 0 Z +∞ 0 Z X X e dt + αi β j −t k∈m+n i∈m.
x3 } = 2. gi = 0 ⇔ (α. x3 } = π and dim lin {x1 . x2 .16 Exercises 15. x2 . −4x. The GramSchmidt process 15.13 Construction of orthogonal sets. β) = (2γ. 2) 3 Every vector belonging to the line π ∩ perp {v} must have the form (x. Since no two of them are parallel. The GramSchmidt process 115 (d) If f : t 7→ t + 1 then f g : t 7→ αt2 + (α + β) t + β hf. A unit vector in π is given by v≡ 1 (2. 1. projections 15.Construction of orthogonal sets. −3γ) (γ ∈ R) g : t 7→ αt + β that is. it is seen that they all belong to the plane π of equation x−z = 0.1 n. 576) (a) By direct inspection of the coordinates of the three given vectors. the set of all polynomials of degree less than or equal to 1 which are orthogonal to f is P1 ∩ perp {f } = {gγ : t 7→ 2γt − 3γ}γ∈R 15. x) .14 Orthogonal complements. lin {x1 . 1 (p.16. gi = αI2 + (α + β) I1 + βI0 = 3α + 2β and hf.15 Best approximation of elements in a euclidean space by elements in a ﬁnitedimensional subspace 15.
y1 i y1 − − hx3 . −3. 3. 1. 1) 2 x3 − hx3 . 1. 0)° √ √ is the hyperplane of R4 through the origin. y. 1.0 ≡ (x. x2 . 0.116 Linear spaces √ with norm 3 2 x. 0 6 = q2 2 = (−1. 1) 6 The required orthonormal basis for lin {x1 . y1 i y1 k °(0. −1. x2 } / 0 = x1 − x2 + x3 − x4 Thus dim W ≡ dim lin {x1 . 1. 0) ° ° (1. 1. −4.16. 2. 0. y2 i y2 = ≡ kx3 − hx3 . 0)° y3 Second solution. z. . 0. 0) 6 1 + 1 +1 4 4 √ √ 2 2 2 2 √ √ 2 2 2 2 (1. 1. 1) − 0 − 2 66 66 (−1. 1. y1 i y1 ≡ =° ° kx2 − hx2 . x3 } is {v. w}. 1. 0) − x2 − hx2 . is √ 2 w≡ (1. 2 (p. 15. 2. x2 . 1. 0. x2 . (b) The answer is identical to the one just given for case a. 576) (a) First solution. 0. x4 } = 3. to directly exhibit two mutually orthogonal unit vectors in lin {x1 . t) ∈ R4 : x − y + z − t = 0 √ 2 u≡ (1. 1) − 0 − 2 66 66 (−1. 1. it suﬃces to realize that © ª lin {x1 . 0) − ¢ ¡ 1 1 √ − . and which is orthogonal to v. 1. 1. 1. 1. 1. 0) ° ° √ √ ° ° °(0. 1 3 = q3 = (1.−1). 0) y1 ≡ kx1 k 2 y2 (0. −1. −1) as normal vector. 3) 6 1 + 1 + 1 +1 9 9 9 (0. 0. with (1.−1. 1.2 n. 0) 2 √ 2 v≡ (0. x4 } . x3 . y1 i y1 − − hx3 . 0. x4 } = H(1. x3 .1. x3 . Thus a second unit vector which together with v spans π. and three vectors are required for any basis of W . y2 i y2 k ¡1 1 1 ¢ √ . 2. More easily. It is easily checked that x2 ∈ lin {x1 } / x3 ∈ lin {x1 . x2 . The following vectors form an orthonormal one: √ x1 2 = (1. 1.
Exercises
117
It remains to ﬁnd a unit vector in H(1,−1,1,−1),0 ∩ perp {u, v}. The following equations characterize H(1,−1,1,−1),0 ∩ perp {u, v}: x−y+z−t = 0 x+y = 0 z+t = 0 yielding (by addition) 2 (x + z) = 0, and hence 1 1 (1, −1, −1, 1) or w ≡ − (1, −1, −1, 1) 2 2 Thus an orthonormal basis for lin {x1 , x2 , x3 , x4 } is {u, v, w}. (b) It is easily checked that w≡ x2 ∈ lin {x1 } / 0 = 2x1 − x2 − x3 Thus dim W ≡ dim lin {x1 , x2 , x3 } = 2, and two vectors are required for any basis of W . The following vectors form an orthonormal one: √ x1 3 y1 ≡ = (1, 1, 0, 1) kx1 k 3 y2 (1, 0, 2, 1) − 2 33 33 (1, 1, 0, 1) x2 − hx2 , y1 i y1 ° =° ≡ √ √ ° ° kx2 − hx2 , y1 i y1 k °(0, 1, 1, 0) − 2 33 33 (1, 1, 0, 1)° ¡1 2 ¢ √ , − 3 , 2, 1 42 3 = q3 (1, −2, 6, 1) = 42 1 + 4 +4+ 1 9 9 9 Z
π √ √
15.16.3
n. 3 (p. 576)
Let
1 1 1 √ √ dt = π = 1 π π π 0 r Z πr Z 2 2 2 π hyn , yn i = cos2 nt dt cos nt cos nt dt = π π π 0 0 hy0 , y0 i = In ≡ Z
π
cos ntdt =
2
0
integrating by parts In
Z
π
cos nt cos nt dt
0
¯π Z π sin nt ¯ sin nt ¯ − = cos nt −n sin nt dt ¯ n 0 n 0 Z π Z π 2 = sin nt dt = 1 − cos2 nt dt
0 0
= π − In
118
Linear spaces
Thus π 2 hyn , yn i = 1 In = and every function yn has norm equal to one. To check mutual orthogonality, let us compute ¯π Z π 1 sin nt ¯ 1 ¯ =0 √ cos nt dt = √ hy0 , yn i = π π n ¯0 0 Z π hym , yn i = cos mt cos nt dt 0 ¯ Z π sin nt ¯π sin nt ¯ − = cos mt −m sin mt dt ¯ n 0 n 0 Z m π sin mt sin nt dt = n 0 µ ¶¯π µ ¶ Z π m cos nt ¯ cos nt ¯ −m = sin mt − dt m cos mt − ¯ n n n 0 n 0 ³ m ´2 = hym , yn i n ¡ ¢2 Since m and n are distinct positive integers, m is diﬀerent from 1, and the equation n ³ m ´2 hym , yn i = hym , yn i n
can hold only if hym , yn i = 0. That the set {yn }+∞ generates the same space generated by the set {xn }+∞ n=0 n=0 is trivial, since, for each n, yn is a multiple of xn . 15.16.4 n. 4 (p. 576) We have Z 1 hy0 , y0 i = 1dt = 1 0 Ã ¯1 ! Z 1 ¯ ¡ 2 ¢ 4 3 hy1 , y1 i = 3 4t − 4t + 1 dt = 3 t − 2t2 + t¯ ¯ 3 0 0 = 1 Z 1 ¡ ¢ hy2 , y2 i = 5 36t4 − 72t3 + 48t2 − 12t + 1 dt 0 Ã ¯1 ! ¯ 36 5 = 5 t − 18t4 + 16t3 − 6t2 + t¯ ¯ 5 0 = 1
Exercises
119
which proves that the three given functions are unit vectors with respect to the given inner product. Moreover, Z 1√ √ ³ ¯1 ´ hy0 , y1 i = 3 (2t − 1) dt = 3 t2 − t¯0 = 0 0 Z 1√ √ ³ ¯1 ´ ¡ ¢ hy0 , y2 i = 5 6t2 − 6t + 1 dt = 5 2t3 − 3t2 + t¯0 = 0 Z0 1 √ ¡ ¢ hy1 , y2 i = 15 12t3 − 18t2 + 8t − 1 dt 0 √ ³ 4 ¯1 ´ 3 2 15 3t − 6t + 4t − t¯0 = 0 = which proves that the three given functions are mutually orthogonal. Thus {y1 , y2 , y3 } is an orthonormal set, and hence linearly independent. Finally, Ã √ √ ! √ √ 1 3 5 3 5 y1 y1 + y1 + y2 x0 = y0 x1 = y0 + x2 = 1 − 2 2 30 3 30 which proves that lin {y1 , y2 , y3 } = lin {x1 , x2 , x3 }
120 Linear spaces .
αy + βv)] (αy + βv. y) ∈ R2 .1 Exercises n. y) is orthogonal to r. T (x. 0)}. .3 Nullity and rank 16. v)] The null space of T is the trivial subspace {(0. y) ∈ R2 . for every β ∈ R. which is (x + y. x + y). αx + βu) α (y. and the midpoint of [(x. y)] + βT [(u. y) − (x. y)]. T [α (x. v) ∈ R2 . x) + β (v. T (x. y) + β (u. and hence rank T = 2 nullity T = 0 T is the orthogonal symmetry with respect to the bisectrix r of the ﬁrst and third quadrant. for every (u.1 Linear transformations 16. 582) T is linear.4. v)] = = = = T [(αx + βu.Chapter 16 LINEAR TRANSFORMATIONS AND MATRICES 16. for every (x. the range of T is R2 . since for every (x. 1 (p. since. for every α ∈ R.4 16. y) . lies on r. u) αT [(x.2 Null space and range 16.
x) + β (u. and for every β ∈ R. and hence rank T = nullity T = 1 . 2 (p.122 Linear transformations and matrices 16. for every (u. the range of T is the Xaxis. y) ∈ R2 . v)] = = = = T [(αx + βu. T [α (x. u) αT [(x. 0) αT [(x. v)] The null space of T is the trivial subspace {(0. v) ∈ R2 . T [α (x. αx + βu) α (x. 0) + β (u. 582) T is linear . 0)}. 16.3 n.4. since for every (x. αy + βv)] (αx + βu. αx + βu) α (y. 0) α (x. for every (u.4. v)] The null space of T is the Y axis. for every α ∈ R. u) αT [(x. v)] = = = = T [(αx + βu. y) + β (u. and hence rank T = 2 nullity T = 0 T is the orthogonal symmetry with respect to the Xaxis. y)] + βT [(u. the range of T is R2 . the range of T is the bisectrix of the I and III quadrant. y) ∈ R2 . y)] + βT [(u. v)] = = = = T [(αx + βu. y)] + βT [(u. 4 (p. y) + β (u. 582) T is linear. 16. and for every β ∈ R. and hence rank T = 1 nullity T = 1 T is the orthogonal projection on the Xaxis. αy + βv)] (αy + βv.4. 582) T is linear. 3 (p. x) + β (v. v)] The null space of T is the Y axis. y) ∈ R2 . v) ∈ R2 . αy + βv)] (αx + βu. for every (u.4 n. for every α ∈ R. since for every (x. for every α ∈ R. since for every (x. v) ∈ R2 . T [α (x. and for every β ∈ R. y) + β (u.2 n.
y) + β (u. when x = y = 0 and u = v = α = β = 1. for every (x. for every α ∈ R.7 n. 1) α (x. for every (x. for every α ∈ R. for every (u. and for every β ∈ R. 0) + (1. y) ∈ R2 . αy + βv)] (αx + βu. 7 (p. 5 (p. y) + β (u.6 n. for every (u.5 n. 0) = x2 + 2xy + u2 .. T [(0. v)] = = αT [(x. 1) + β (u. Indeed.4. for every (x. (ey )α · (ev )β αT [(x. v)] = = 16. y + 1) + β (u + 1. for every α ∈ R. αy + βv + α + β) .g. v)] = = αT [(x.4. v) ∈ R2 .8 n.4. αy + βv)] ¡ ¢ = eαx+βu . αey + βev ) so that. 16. y) ∈ R2 . 0)] + T [(1. ey ) + β (eu . 8 (p. 1)] = (1 + e. T [α (x. 582) T [(αx + βu. 582) T is not an aﬃne function. for x and u both diﬀerent from 0. 1) (αx + βu. y)] + βT [(u. 582) T is not linear.. y) + β (u. αy + βv + 1) α (x + 1. for every β ∈ R. 1)] = (e. eαy+βv i h = (ex )α · (eu )β .4. y)] + βT [(u. 0 T is not a linear function. indeed. 0) + T (u. v) ∈ R2 . but not linear. v)] = α (ex . y)] + βT [(u. 0) = x2 + u2 . v + 1) (αx + βu + α + β. v) ∈ R2 . indeed. for every (u. ev ) = (αex + βeu . T [α (x. but it is not linear. v)] = = T [(αx + βu. y) ∈ R2 . and for every β ∈ R. 0 ¡ ¢ T (x. v)] = T [(αx + βu. e) T [(0. e. 1 + e) 16. αy + βv)] (αx + βu + 1. 582) ¢ ¡ T (x + u. α + β) T is aﬃne. since. 6 (p.g. e.Exercises 123 16. T [α (x.
αx + βu + αy + βv) α (x − y. αy + βv)] (2 (αx + βu) − (αy + βv) . and for every β ∈ R. for every α ∈ R. y) + β (u. T [α (x. 9 (p.124 Linear transformations and matrices 16. y)] + βT [(u. 10 (p. 0)}. and hence rank T = 2 The matrix representing T is A≡ the characteristic polynomial of A is λ2 − 3λ + 3 · 2 −1 1 1 ¸ nullity T = 0 . 16. x + y) + β (2u − v.9 n. v)] = = = = T [(αx + βu.4. v) ∈ R2 . y) ∈ R2 . y) ∈ R2 . 582) T is linear. v)] The null space of T is the trivial subspace {(0. x + y) + β (u − v. T [α (x. for every (u. v) ∈ R2 . u + v) αT [(x. (αx + βu) + (αy + βv)) α (2x − y. the range of T is R2 . v)] The null space of T is the trivial subspace {(0. and for every β ∈ R. for every (u. the range of T is R2 . for every α ∈ R. 0)}. u + v) αT [(x. y)] + βT [(u. since for every (x.10 n. 582) T is linear. since for every (x. and hence rank T = 2 The matrix representing T is A ≡ 1 −1 1 1 ¸· · √ 2 √ 0 cos = sin 0 2 · ¸ √ = 2 " √ √ 2 − 22 2 √ √ 2 2 2 2 π − sin π 4 4 π cos π 4 4 nullity T = 0 # ¸ Thus T is the composition of a counterclockwise rotation by an angle of π with a 4 √ homothety of modulus 2. αy + βv)] (αx + βu − αy − βv.4. v)] = = = = T [(αx + βu. y) + β (u.
y. y. 0) α (x. and hence 16. w)] nark T = 0 · √ ¸· 3 √ 0 cos π − sin 6 = sin π cos π 3 0 6 6 √ 3 2 3 2 # √ = 3 " √ 3 2 1 2 −1 √2 3 2 ¸ π 6 # The null space of T is the trivial subspace {(0. y. v. since for every (x. 0) αT [(x. the range of T is the XY plane. for every (u. for every (u. v. 16.Exercises 125 and the eigenvalues of A are √ 3 + 3i λ1 ≡ 2 An eigenvector associated to λ1 is µ z≡ √ 3 − 3i λ2 ≡ 2 2 √ 1 − 3i ¶ The matrix representing T with respect to the basis ½µ ¶ µ ¶¾ 0 2 √ {Im z. αy + βv. y. z)] + βT [(u. x) + β (w. 0) + β (u. since for every (x.12 n. z) ∈ R3 . z) ∈ R3 . w)] nark T = 1 The null space of T is the Zaxis. αy + βv. the range of T is R3 . αy + βv. v. 16 (p. w) ∈ R3 . T [α (x. for every α ∈ R. y. v. 582) T is linear. for every α ∈ R.4. w) ∈ R3 . and hence . z)] + βT [(u. αz + βw)] (αz + βw. 0)}.11 n. y. u) αT [(x. αx + βu) α (z. 1 − 3 is B ≡ " 3 2 √ 3 2 − Thus T is the composition of a counterclockwise rotation by an angle of π with a 6 √ homothety of modulus 3. 0. 17 (p.4. w)] = = = = rank T = 2 T [(αx + βu. v. αy + βv. w)] = = = = rank T = 3 T [(αx + βu. for every β ∈ R. v. z) + β (u. Re z} = . z) + β (u. y. v. for every β ∈ R. T [α (x. αz + βw)] (αx + βu. v. y. 582) T is linear (as every projection on a coordinate hyperplane).
+. 0. f 7→ (x 7→ xf 0 (x)) then T is a linear operator. 1) or. w) αT [(x. 1). α (x + y)] + [β (u + w) . Indeed. v. If T : D1 → R(−1.13 n. f 0 (x) = 0} By an important though sometimes overlooked theorem in calculus. y.4. 582) Let D1 (−1.126 Linear transformations and matrices 16. of parametric equations x=t y = −t z = −t the range of T is the XZplane. ker T = {f ∈ D1 : ∀x ∈ (−1. z) + β (u. y. 582) T is linear. w)] = = = = = T [(αx + βu. and hence rank T = 2 nark T = 1 16. 25 (p. w) ∈ R3 .14 n. v. z) + β (u. v. 0. αy + βv. for every β ∈ R. provided lim f 0 (x) exists and is ﬁnite. xf 0 (x) = 0} = {f ∈ D1 : ∀x ∈ (−1. every function f which is diﬀerentiable in (−1. −)orthant and of the (−. ¡ ¢ T (f + g) = x 7→ x [f + g]0 (x) = (x 7→ x [f 0 (x) + g 0 (x)]) = (x 7→ xf 0 (x) + xg 0 (x)) = (x 7→ xf 0 (x)) + (x 7→ +xg 0 (x)) = T (f ) + T (g) Moreover. for every α ∈ R. y. β (u + v)] α (x. since for every (x. αx + βu + αy + βv) [α (x + z) . is diﬀerentiable in 0 as well. T [α (x. in which case x→0 f 0 (0) = lim f 0 (x) x→0 . v. more shortly. −. z) ∈ R3 .4. 1) . 23 (p. 1) . w)] The null space of T is the axis of central symmetry of the (+. 0) ∪ (0. for every (u. y. D1 be the linear space of all real functions of a real variable which are deﬁned and everywhere diﬀerentiable on (−1. 0) ∪ (0. 1) and continuous in 0. αz + βw)] (αx + βu + αz + βw. z)] + βT [(u. +)orthant. 0.1) .
y0 ) ∈ R2 there exists a unique solution to equation (16.4. If f belongs to ker T . by the classical theorem due to Lagrange. 1). u and v are linearly independent. y 7→ y 00 + P y 0 + Q ∀x ∈ R. from βv (x) = 0 for each x ∈ R and v 0 (0) = 1 it is easily deduced that β = 0 as well. v 0 (0)) = (0. Since T (D1 ) contains. in such a case. and nullity T = 1. u0 (0)) = (1. y 00 (x) + P (x) y 0 (x) + Q (x) y (x) = 0 (16. for each (y0 . It follows that a basis of ker T is given by the constant function (x 7→ 1).1) such 0 that y (0) = y0 and y 0 (0) = y0 . 582) Let D2 be the linear space of all real functions of a real variable which are deﬁned and everywhere diﬀerentiable on R. and let v be the solution to equation (16. 0). If where P and Q are real functions of a real variable which are continuous on R. In other words.Exercises 127 Thus (here 0 is the identically null function deﬁned on (−1. ϕ−1 ((y0 . let u be the solution to equation (16. v} . all the restrictions to (−1. ker L = span {u. the function 0 y : x 7→ y0 u (x) + y0 v (x) ker T = {f ∈ D1 : f 0 = 0} L : D2 → RR .g. it has been shown in the solution to exercise 17. and Finally.1) such that (u (0) . p. x) f (x) = f (0) + xf 0 (ϑx ) and hence f is constant on (−1. that L is a linear operator. y 7→ y 0 (0) is injective and surjective. 16. e. 1)). since u (0) = 1 and v (0) = 0. 1) of the polynomials with vanishing zerodegree monomial. αu (x) + βv (x) = 0 for each x ∈ R can only hold (by evaluating at x = 0) if α = 0. Then for each (y0 .15 n.1). 27 (p.1) such that 0 (v (0) . Hence the function µ ¶ y (0) 2 ϕ : ker L → R .1) is a solution to equation (16. the dimension of T (D1 ) is inﬁnite.555. Thus ker L is the set of all solutions to the linear diﬀerential equation of the second order By the uniqueness theorem for Cauchy’s problems in the theory of diﬀerential equa0 tions. and by direct inspection it is seen that y (0) = y0 and 0 0 y 0 (0) = y0 . indeed. Thus nullity L = 2. 1). and it is already known that the linear space of all polynomials has an inﬁnite basis. since ker L is a linear subspace of D2 . ∃ϑx ∈ (0.. y0 ) ∈ R2 . y0 )). ∀x ∈ (−1. 1) . Moreover.
y . v. z) = (x0 . since x = x0 y = y0 T (x. since x + 1 = x0 + 1 0 0 0 T (x. y .8. z 0 ) x + y + z = x0 + y 0 + z 0 0 0 0 T −1 (u. z) = T (x . z ) ⇔ 3z = 3z 0 ³ v w´ T −1 (u. w) = (u − 1. 2 3 16. y 0 .8. z) = (x0 . z ) ⇔ y + 1 = y 0 + 1 ⇔ (x. y. y 0 .2 n. v. 17 (p.3 n.5 Algebraic operations on linear transformations 16.7 Onetoone linear transformations 16. y.8.6 Inverses 16. z 0 ) z − 1 = z0 − 1 T −1 (u.128 Linear transformations and matrices 16. 16 (p. v. y. y. y 0 . 589) T is injective (or one to one). w − u − v) 16. z) = T (x . y . since x = x0 0 0 0 2y = 2y 0 ⇔ (x. z ) ⇔ ⇔ (x. w + 1) . z) = (x0 . 589) T is injective. . v. y. w) = u.1 n. 15 (p. z 0 ) T (x. v − 1. w) = (u. y.8 Exercises 16. 589) T is injective. z) = T (x .
i.9 Linear transformations with prescribed values 16.12.e. 596) (a) Since T (i) = i + j and T (j) = 2i − j. and ker T D is the set of all constant polynomials..8.4 Let n.1 n. 16. T (3i − 4j) = = 2 T (3i − 4j) = = 3T (i) − 4T (j) = 3 (i + j) − 4 (2i − j) −5i + 7j T (−5i + 7j) = −5T (i) + 7T (j) −5 (i + j) + 7 (2i − j) = 9i − 12j .11 Construction of a matrix representation in diagonal form 16. 590) p = x 7→ p0 + p1 x + p2 x2 + · · · + pn−1 xn−1 + pn xn · Z x 0 DT (p) = D [T (p)] = D x 7→ = x 7→ p (x) ¸ p (t) dt the last equality being a consequence of the fundamental theorem of integral calculus.Linear transformations with prescribed values 129 16. zerodegree polynomials.10 Matrix representations of linear transformations 16. 27 (p. Im T D is equal to W .12 Exercises 16. £ ¤ T D (p) = T [D (p)] = T x 7→ p1 + 2p2 x + · · · (n − 1) pn−1 xn−2 + npn xn−1 Z x = x→ 7 p1 + 2p2 t + · · · (n − 1) pn−1 tn−2 + npn tn−1 dt 0 ¯x = x 7→ p1 t + p2 t2 + · · · pn−1 tn−1 + pn tn ¯0 = x 7→ p1 x + p2 x2 + · · · pn−1 xn−1 + pn xn = p − p0 Thus T D acts as the identity map only on the subspace W of V containing all polynomials having the zero degree monomial (p0 ) equal to zero. 3 (p.
j}. e2 } to basis {i. The operation of the matrix B representing T with respect to the basis {e1 . j} we have T (e1 ) = = T (e2 ) = = T (i − j) = T (i) − T (j) = (i + j) − (2i − j) −i + 2j T (3i + j) = 3T (i) + T (j) = 3 (i + j) + (2i − j) 5i + 2j .r. The matrix B representing T with respect to the basis {e1 . and P −1 is the matrix of coordinate change from basis {i. 0) and (0.r. β) and (γ. e2 } (that is. e2 } can be described as the combined eﬀect of the following three actions: 1) coordinate change from coordinates w. to the basis {i. e2 }. δ) with respect to the basis {e1 . multiplication by matrix P ). e2 }.130 Linear transformations and matrices (b) The matrix of T with respect to the basis {i.r. j} (that is. j} is · ¸ £ ¤ 1 2 A ≡ T (i) T (j) = 1 −1 and the matrix of T 2 with respect to the same basis is · ¸ 3 0 2 A ≡ 0 3 transforms the (canonical) coordinates (1. Thus · ¸· ¸· ¸ 1 1 −3 1 2 1 3 −1 B = P AP = 1 −1 −1 1 4 1 1 · ¸· ¸ 1 −2 5 1 3 = 2 1 −1 1 4 · ¸ 1 −7 −1 = 1 7 4 Second solution (matrix for T ). on the other hand. to the basis {i. multiplication by P −1 ). e2 } is £ ¤ B ≡ T (e1 ) T (e2 ) (c) First solution (matrix for T ). with respect to the basis {i. to basis {e1 . multiplication by A). to the basis {i. If e1 = i − j and e2 = 3i + j. 3) coordinate change from coordinates w. e2 } into coordinates w. 1) of e1 and e2 with respect to the basis {e1 . 2) transformation according to T as expressed by the matrix representing T w. −1) and (3. to basis {e1 . hence P is the matrix of coordinate change from basis {e1 . Since. e2 } into their coordinates (1. j} (that is.r.r. the matrix · ¸ £ ¤ 1 3 P ≡ e1 e2 = −1 1 provided T (e1 ) and T (e2 ) are meant as coordinate vectors (α. 1) with respect to the basis {i. j} into coordinates w. j} to basis {e1 . j}.
e2 }coordinates (α. 2y) T (i + 2j + 3k) = T (k) + T (j + k) + T (i + j + k) = (2i + 3j + 5k) + i + (j − k) = 3i + 4j + 4k . y) 7→ (−2x.Exercises 131 it suﬃces to ﬁnd the {e1 . Thus we want to solve the two equation systems (in the unknowns (α. δ) = 1 (−1. β) and (γ. so that 4 4 1 B= 7 −7 −1 1 7 Continuation (matrix for T 2 ). it is represented by the same matrix with respect to every basis (indeed. since scalar diagonal matrices commute with every matrix of the same order.12. δ) which correspond to the {i. 596) ≡ µ 4 0 0 4 ¶ 7→ (−x. 596) T : (x. 5 (p. P −1 DP = D for every scalar diagonal matrix D). j}coordinates (−1. 4 (p.12. 1) and (γ.2 n. 2) and (5. β) = 1 (−7.3 a) n. to the Y axis) (length doubling) Thus T may be represented by the matrix µ ¶ −2 0 AT ≡ 0 2 and hence T 2 by the matrix A2 T 16. β) and (γ. 2).r. y) (reﬂection w. 16. respectively) αe1 + βe2 = −i + 2j γe1 + δe2 = 5i + 2j that is. ½ α + 3β = −1 −α + β = 2 · ½ γ + 3δ = 5 −γ + δ = 2 ¸ The (unique) solutions are (α. 7). δ). Since T 2 is represented by a scalar diagonal matrix with respect to the initially given basis.
the linear combination xT (k) + yT (j + k) + zT (i + j + k) = x (2i + 3j + 5k) + yi + z (j − k) = (2x + y) i + (3x + z) j + (5x − z) k spans the null vector if and only if 2x + y = 0 3x + z = 0 5x − z = 0 which yields (II + III) x = 0. k} of the image vectors T (i). T (j) = = T (i) = = T (j + k − k) = T (j + k) − T (k) −i − 3j − 5k T (i + j + k − j − k) = T (i + j + k) − T (j + k) −i + j − k −1 −1 2 AT ≡ 1 −3 3 −1 −5 3 16. The last one is given in the problem statement. 7 (p. T (k)} is a linearly independent set. T (i + j + k) form a linearly independent triple. ker T = {0} (b) µ 0 1 1 0 1 −1 ¶ rank T = 2 AT = . 597) T (4i − j + k) = 4T (i) − T (j) + T (k) = (0. and hence (by substitution in I and II) y = z = 0. j.132 Linear transformations and matrices The three image vectors T (k).r. T (j). Thus the range space of T is R3 . 0)}. −2) Since {T (j) . Indeed. T (k).12. T (j + k). b) The matrix of T with respect to the basis {i. to {i. and nark T is 0. It follows that the null space of T is the trivial subspace {(0. 0. j. but the ﬁrst two need to be computed.4 (a) n. k} is obtained by aligning as columns the coordinates w. and rank T is 3.
597) (a) I shall distinguish the canonical unit vectors of R2 and R3 by marking the former ¡¢ with an underbar. If w1 = i + j and w2 = i + 2j. k} is 1 1 ¡¢ ¤ £ A ≡ T (i) T j = 0 0 1 −1 (c) First solution. 8 (p. −1)}. k. ¡ ¢ ¡¢ T 2i − 3j = 2T (i) − 3T j = 2 (i + k) − 3 (i − k) = −i + 5k For any two real numbers x and y. j. ¡ ¢ ¡¢ T xi + yj = xT (i) + yT j = x (i + k) + y (i − k) = (x + y) i + (x − y) k It follows that Im T = lin {i. Since T (i) = i + k and T j = −i + k. Then µ ¶ 1 0 0 (AT )B. k} rank T = 2 nullity T = 2 − 2 = 0 © ª (b) The matrix of T with respect to the bases i.C = 0 1 0 16. i} and C = {w1 .Exercises 133 (c) Let C = (w1 .12. j and {i. the matrix · ¸ £ ¤ 1 1 P ≡ w1 w2 = 1 2 µ 0 1 −1 2 0 0 3 2 ¶ . w2 ) T (i) = 0w1 + 0w2 T (j) = 1w1 + 0w2 1 3 T (k) = − w1 + w2 2 2 (AT )C = (d) Let B ≡ {j.5 n. 1) . w2 } = {(1. (1.
j. j to basis {w1 . j . 2) with respect to the©basis © ª ª i. j. j. w2 } into their coordinates (1. that is. v2 . multiplication by A). to the basis i. 0) and (0. j (that is. The matrix B representing T with respect to the bases {w1 . w2 } to basis i.r. w2 }. 1) and (1. k} (that is. k} is £ ¤ B ≡ T (w1 ) T (w2 ) T (w1 ) = = T (w2 ) = = Thus ¡ ¡¢ ¢ T i + j = T (i) + T j = (i + k) + (i − k) i ¡ ¡¢ ¢ T i + 2j = T (i) + 2T j = (i + k) + 2 (i − k) 3i − k 1 3 B= 0 0 0 −1 where if and only if the following holds (c) The matrix C representing T w. to bases {u1 . j .r. j and {i. w2 } into coordinates w. w2 } and {i. u2 } and {v1 . γ 11 0 C = 0 γ 22 0 0 T (u1 ) = γ 11 v1 T (u2 ) = γ 22 v2 . 1) of w1 and w2 with respect to the basis {w1 .r. Thus · ¸ 1 1 1 1 B = AP = 0 0 1 2 1 −1 2 3 = 0 0 0 −1 Second solution (matrix for T ). 2) transformation according to T as expressed by the © ª matrix representing T w. v3 } is diagonal. © ª and P −1 is the matrix of coordinate change from basis i. multiplication by matrix P ).r. w2 } and {i. The operation of the matrix B of T with respect to the bases {w1 . hence P is the matrix of coordinate change from basis {w1 . to basis {w1 .134 Linear transformations and matrices transforms the (canonical) coordinates (1. k} can be described as the combined eﬀect of the following two actions: 1) coordinate ª © change from coordinates w. to the bases i.
id · sin. v3 } is lin.Exercises 135 There are indeed very many ways to achieve that.6 n. 597) We have D (sin) = cos D (cos) = − sin D (id · sin) = sin + id · cos D (id · cos) = cos − id · sin D2 (sin) = D (cos) = − sin D2 (cos) = D (− sin) = − cos D2 (id · sin) = D (sin + id · cos) = 2 cos − id · sin D2 (id · cos) = D (cos − id · sin) = −2 sin − id · cos and hence the matrix representing the diﬀerentiation operator D and its square D2 acting on lin {sin. cos. id · sin. 16 (p. In the present situation the simplest way seems to me to be given by deﬁning u1 ≡ i u2 ≡ j v1 ≡ T (u1 ) = i + k v2 ≡ T (u2 ) = i − k v3 ≡ any vector such that {v1 . independent 1 0 C = 0 1 0 0 thereby obtaining 16. cos.12. id · cos} is 0 −1 1 0 −1 0 0 −2 1 0 0 1 0 −1 2 0 A2 = A= 0 0 0 −1 0 0 −1 0 0 0 1 0 0 0 0 −1 . v2 . id · cos} with respect to the basis {sin.
This action might not be possible to undo. Are you sure you want to continue?