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) . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . .4. . . . 24 (p.11. . . . . . . . . . .4 n. . . . . . . . . . . . . . . . . . . 11 (p. . . . . . . 12. . . . . . . . . . . .16 n. . . . . . . . . . . . . . . . . . . . . 12. . . . . 456) . 12. . .8. . . 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 . . . . . . . . . . .11 n. . . . . . . . . 3 (p. . . . . . . . . . . . . . . . . . . . . 12. . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 n. . . . . . . .1 n. . . . 456) . . .6 n.5 n. . . .6 n. . . . . . . 25 (p. . . . . . . 12. . . 5 (p. . . .8 Exercises . . . . . . . . 12. . . .1 n. . . . .10 The unit coordinate vectors . . . . . . .11. . . . . . . . . . . . . . . . . . . 9 (p. 12. . . 451) .11 n. . . . . . . 14 (p. . . Angle between vectors in nspace 12. . .4. . 22 (p. . . . . . . . 456) . . . . . 12.8. .4 n. . . 461) . . . . . . . . . . . . . . . . . . . . . .12 The linear span of a ﬁnite set of vectors . . .12 n. . . . . . . . 12. . . . 1 (p. 6 (p. . . . . . .2 n. . . . . . . . . .8. .11. . . .2 n. . . . . .5 The dot product . . . . . . . . . . . . 15 (p. . 12. . .11. . . . . . . . . . . . .8. . . .7 Orthogonality of vectors .8. . .8 n. 456) . . . . . . . . . . . .8. . . . . . .3 n. . . . .8. 10 (p. . . . . . . . . . . . 456) . . . . . . . . . . . . 12. . . . . . . . . . . . . . 13 (p. . . 457) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 Length or norm of a vector . . 17 (p. . . . . . . . . . . . 12. . . . . 12. . . . . . . . . . . 13 (p. . . . 20 (p. 5 (p. . . .8 n. . . . .11. . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . 457) . . . 460) . . . . . . . . . . 456) . . . . . . . . . 456) . .3 n. . . . 12. . . . . . .5 n. .9 n. . . . . . . . . . .7 n. . . . 10 (p. .12 n. . . . . . . . . . . . . . . . . . . . . . . . 12. . . . . . 12. . . . . 460) . . 12. . . . . . . .11 Exercises . . . . . 12. .17 n. . 457) . . . . . . . . . . . . . . . . . .8. . . . . . .7 n. . . . . . . . . . . . . . . . . . . . 456) . 12. . . . . . . . . . .8.8. . . . 12. . . . . . . 456) . . . . . . . . . . . . 16 (p. . . . . . . . . . . . . . . . . 21 (p. . .9 n. 11 (p. . . . . . .8. . 12 (p. . . . . . . . . . . . . 456) . . . . . . . 8 (p. . . . . . . . . 456) . . . . . . 451) . .8 n. . . . 451) .8. . . 12. . 12. . . . . . . . . . 12. . . 12. . . . . . . . . . . . .14 n. . . . . . . . . . . . . . . . . . . . 457) . . . . . 12. . 451) . . .10 n. . . . . . 17 (p. . . . . . . . . . . . . . . . . .15 n. . . . . . . . . 12. 12. . 12. . . . . 460) . . .11.9 Projections. .8.4. . . 19 (p. 460) . . . . . 12. . . . .9 n. . . . . . . . . . . . . . . . . . . . . . . .7 n. . . . . . . . . . .18 n. . . . . . . . . . . . . . . . . . . . . . . . 461) . .11. . . .11. . . . . . . .8. . . . . 12. . . . . . . . . 7 (p. . . . . . 460) . .8. . . . . . . . . . . . . . . 12. . . . . . 461) . . . . 12. . . . . . . . . . . . .4. . . .8. . . . . . . 451) . . . . . . 8 (p. . 3 (p. . . 456) . . . . . . 12. . 6 (p. . 10 (p. 461) . . . . . . . 12. .11. . . 2 (p. . . . 456) .8. . . . . . . . 12. 460) .10 n. . 12. . . . . . . 12. . . . . . .4 CONTENTS 12. . . . . . . . . . . . 2 (p. . 12.11. . . .4.10 n. .8. . . . . . . . . . . . . . . . . . . . . . . . . 12. . 451) . . . . . . . . 12. . . 1 (p. . . . . 7 (p. . . . . . . 456) . . .
. . . . . . . . . . . . . . . . . . 13. . . . . . . . . . . . . . 18 (p. .5.5. . . . . . . .15. . . . . . . . . . . . . 12. . . . . . . . . . . . . . . . . . . . . . . . . . . 467) . . . . . . . . . . . . . . . .7 Planes and vectorvalued functions . . . . . . . . .15. . . 12. . . 477) . . . . . . . 477) .15. . . . .5. . . . . 12 (p. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. . . . . . . . . . . . . . . . . . . . . . .1 Introduction . . . . . . .5. . . . . . . . . . 468) . 5 (p. . . . . . . . . . . . . .1 n. . . . . 10 (p. . . . . . . . .4 n. 13. . . . . . . . . .3 n. . . . . . . . . . . . . 13. . . . . .5. . . . . . 13. . . 467) . . .11 n. 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. .15. . . . .6 Planes in euclidean nspaces . . . . . . . . . . . . . . . . . . . . . 1 (p. . . . . .17 Exercises . 13. . . 13. . . . . . . . 12. . . . . 477) . . . . . . . 477) . . . 468) . . . . . . . . . . . . 12. . . .11 n. 11 (p. . . . . . . . . . . . 467) . . . . . . . . . .8 Exercises . . . . . . . . . 13. . . . . . . . . . 12. 468) . . . . .15. . . .6 n. . . . . . .15 Exercises . . 15 (p. . . . . . . . . .14 Bases . . . . . . . . 468) . . . . . . . . . . 12. . . . . . 477) . . . .12 n. . . .7 n. . . . . . . . . . . . . . . . 12. . . 12. . . . .2 n. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 (p. . . . . . . . .15. . . . . . . . . . 13. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477) . . . . . . . . . . .15. . . . . . 9 (p. . . . . . . . 467) .12 n. . . . . . . . . . . . . . . . . . . . . . . . 8 (p. . . . . . 3 (p. . . . .CONTENTS 5 12. . . . . . . . . . . . . . . . . . .5 Exercises . . . . . . . . 12 (p. . . . . . . . . . . . . .15.4 Lines and vectorvalued functions . . . . . . . . . . . . . . . . . . . . . 467) . . . .15. . . . . . . 468) . . 1 (p. .8 n. . . . . 8 (p. . . . . . 13. . . . . . . . . . 10 (p. . 477) . . . . . . .9 n. . . . . .13 Linear independence . . . of complex numbers . . .5. . 13. 12. 13. . . 477) . . . . . . . . . . 12. . . . . . 12. . . . . . . . . . 12. . .5. . . . . . . . . . . . . . . . . . . 12. . . 477) . . . . . . . . 13. . . . 7 (p. . . . 13. . . . . . . .5. . . . . . . . . .5. . . . . . . . . . . . . . . . . . 3 (p.15. . . . . 477) . . . . . . . . . . . . . . . . . . .15. 13. . . . . . . . . . . . 467) . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Lines in nspace . . . . . . . . . . 19 (p. . .3 n. . .10 n. . . . . . . . 14 (p. . . . . . 13. . . . 5 (p. . . . .1 n. . . . .5 n. . . . . . . 6 (p. . .10 n. . . 12.15 n. . . . 13. . . . . . . . . . . . . . . . . . . . . . . 467) . .16 The vector space Vn (C) of ntuples 12. . . 13. .5. 2 (p.15. . . . . 13 (p. . . . . . . . . . 467) . . . 20 (p. . . . . . . . 4 (p.15. . . . . . .5 n. . . . 17 (p. . 477) . . .8 n. . . . . . . . . . . .14 n. . . . . . . . . .15. . . . . . . . . .13 n. 467) . .5.2 n. . . .5. . . 12. . . . . . . . . . 468) . . . . . . . . . . . . . . . . . .15. . 13. . . . . . . . . . . . . . . . . . . 6 (p. 12.3 Some simple properties of straight lines . . 477) . . .7 n. . . . . . . . . . . . . . 13. . . . . . 12. .9 n.6 n. . . .4 n. .
. . .6 n. . . . . . . . . . . . . . . . 483) . . . . 2 (p. . 13. . . . . . . 13. . . . . . . . . 4 (p. . . . . . . 13. . . . .8. . . .8. . . . . . . . . . 15 (p. . . . . . . 13.8. . . . . . . . . . . . . . . . . . . . . . . . . . 6 (p. . 5 (p. . . . . . . . . .9 n. . . . . . . . . . . 487) . . . . . . .11. . . . . . . . . . . 13. . . . . . . . 482) . .17. . . . . 3 (p. . . . . . . . . . . . . . . . . . .8 n. . . . . . . . . . . 13. 13. . . . . . 12 (p. .2 n.4 n. 13. . . . . . . 2 (p. . . . . . . .9 The cross product . . . . .11. . . . . . . . 13. . . .8 n. . . . . . . . . 482) . . . . . . . . . . . . 8 (p. . . . . . . . . . . . .12 n. . . . . . . . . . . . . . .1 n. . . . . . . . . . . . . . . . .7 n. . . . . . . 6 (p. 3 (p. . 7 (p. . . .6 n. . 13. . 488) . . . . . . . . . . . . . . . 496) . . . . . . . . . . 482) . . . . 13 (p. . . . . . . . . 487) . .17. . . . . 13. . 13. . . . . . . . . . . . . . 13. . . . . . . .4 n. . . . . . . . . . . . . . .11. . . 13 (p.8. 1 (p. . . . . . . . . . . . . 13. . . . . . . . .11. . . . . . . . . . 488) . . . .8. . . . . . .11.11. . . . . . . . . . . . 13. . . . . . . . . . . . . . . . . . . . .8.17 Exercises . . . . . . . . . . . . . . . . . 487) . . . . . . . . . . 13. . . 10 (p. 13. . .10 The cross product expressed as a determinant . . . . . .3 n. . . . . . . . . . 13.6 CONTENTS 13. . . .11. . . . . . . . . . . . . . . . . . . 13.8. . . . . . . . . . . . . . . . . . . . . . . . . . 482) . 488) . . . . . . . . 14 (p. . . . . . . . . . . . . . . 4 (p. 11 (p. . equations . . . . . . . . . . .11. .2 n. . . . . . . . 482) . . . . . . .17. . .11.13 n. . . . . 496) . . . . 488) . . . . . . . . . . . . . .5 n. . . . . . . . . . 487) . . .5 n. .14 Exercises . 13. 488) . . . . . . . . . 3 (p. . . . . . 4 (p. . . . 1 (p. . . . . . .8. . . . . . . . . . . 13. . . . . . . 13. . . .11. . . 496) . . . . . . . . . . . . . . . 5 (p. . . . . . . . . . . .13 Cramer’s rule for solving systems of three linear 13. . . . . . . . . . . . . . . . . .14 n. . 482) . . .11 Exercises . . . 483) . . . .7 n. . 13. . .12 n. . . . 487) . . . . . 496) . . . . . 13. . . . . . 488) . . . . . 488) . . . . . . . . . . .10 n. . .11.16 Linear cartesian equations for planes . . . . . 13. . . . 13. . . . . 483) .12 The scalar triple product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13. . . .4 n. . . . .2 n. . 482) . . . . . . . . . . . . . . . . . . . . 496) . . 487) . 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 . . .17. 13. . . . . . . . . . . . .13 n. . 14 (p. . . . . .8. . 9 (p.3 n. . . . . . . . .8. . .15 Normal vectors to planes . . . . 488) . . . . . 11 (p. . . . . . . . . . . 13. .11. . . . .8. . . . . 483) . . . . . 13. . . . . . . . 483) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 482) . . . . . 10 (p. 9 (p. . .11 n. . .11.1 n. 13. . . . . . . . . . .5 n. . . 488) . . . . . . . . . . 13. . . . . .11. . . . . 12 (p. . . 13. . .1 n. . . . . . . . . 13. . .3 n. . . . . . . . .9 n. 8 (p. . . 7 (p.8. . . . . . . . 13. . . . . . .11. 13. .8. 2 (p. . . . 13. . . . . . . . . .17. .11 n. 13. . . . . . . . . .15 n. . . 13. 13. . . . . . . . . . . . . . . . . . . . 5 (p. . . . .
. . . 555) . . . . . . . . 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 . . . . . . . . .14 n. . 13 (p. . . . 496) . . . . . . . . . . . .13 n. . . . .17. . . . . . . 555) . . . . . 13. . 15. . .5. 6 (p. . . . 15. . . . . .15 n. . . . . . . . .6 n. . . . . . . . . . . . . 555) .9 n. .5. . . . . . . 15. . . . 15. . . . . 555) .25 Miscellaneous exercises on conic sections . 15. 555) . 7 (p. . 15. . . . . . . . . . . . . . . . . . . 6 (p. . . . . . . 496) . . 555) . .7 n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 (p. . 555) . . . . 5 (p. . . 19 (p. . . . . . . . . . . . . . . . . . 496) . . . . . . . . . . . . . . . .6 n.11 n. . . . . . . . . . . . 555) . . . . . . . . . . . . . . . . . . . 496) . . . . . . . . . . . . . . . . . . . . 13. . . . . . . . . . . . . . . .13 n. . . . . .8 n. . . . . . . 3 (p. . 15. . . 13. . . . . . . . 15. . . . . . . . 15. . . . . . 8 (p. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13. . . . .5. . . . . . . . . . . . . 15. . . . . . . .5. . . . . . 555) . . . . . . . . . . . . . 2 (p. . . . . . . . . 555) . . . .17. . . . . . . . . . . . . . . .19 Eccentricity of conic sections . . . . . . . . . . . . 15. . . . . 496) . . . . . . . . . . . . . . .23 Cartesian equations for the conic sections 13. . . . . . . . . . . . . . .5. . 15. . . . . . . . . 13. . . . . . 555) . . .4 n. . . . . . . . 18 (p. . . . . . . . . . . . . . . . 14 Calculus of vectorvalued functions 15 Linear spaces 15. . .17.2 n. . . . . . . . . . . 4 (p. . . . . . . . .5. . . . . . . . . . . . . . . . . . .5. 496) . . . 555) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 (p. . . 13. 13. . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . .11 n. . . . 13. . . . . .5. .24 Exercises . . . . . . . . . . . . 15. . . .5. . . . . . . . . . .10 n. . .17. . . . . . . . .5. 496) .17. . . . .5. . . .18 The conic sections .12 n. . . . . . . . . . . . . . . . . . . . . . . . . . 14 (p. . . . . . . . . . . . 497) . 15 (p. . . . . . .17. 11 (p. . . 13. . . . . . . . . . . .20 Polar equations for conic sections . . . 496) .9 n. . . . 16 (p. . . . . . . . . . . 17 (p. . 13. . . 1 (p. . . . . 22 (p. . . .17.12 n. .14 n. . . . . . . . .10 n. . . . .15 n. . . . . . . . . . . . . . 14 (p. . . 497) . .3 n. .5. . . . . . . . . . . . 13. . . . . . .1 Introduction . . . . . . . . 15. . .7 n. . . . . . . . . 555) . . 17 (p. . . . . . . 15. . . . . .8 n. . . . . . . . . . . . . . . . .CONTENTS 7 13. . . . . . . . . . . . . . .3 Examples of linear spaces . . . . . . 10 (p. . . . .5.5 n. . .17. . . . . . 13 (p. .4 Elementary consequences of the axioms 15. 15. . . . . .22 Conic sections symmetric about the origin 13. . . . . . . . . . . 555) . . . 9 (p. . .2 The deﬁnition of a linear space . . . . . . . . . . . . . . . .5 Exercises . . . . . . . . . . . .21 Exercises . 15.5. . . . . 13. . . . . . . . . . . . . . . . . . . . . . . . 15. . . . 13. .1 n. . . . . . . . . .17. . . . . . .17. . 13. . . 555) . . . . . . . . . . 13. . . . .
. . . . . . . . . . . 576) . . . 4 (p. . . .8 CONTENTS 15. . . . . . . . . .5. . . . . . . . . . . . . . . 15. . . . . . . . . . . . . . . . . . . . . . 15. . . .18 n. . 560) . . . . . 15. . . . . 11 (p. . . . . . . . . 555) . . . . . . . . . . . . . . . .15 Best approximation of elements in a euclidean space by elements in a ﬁnitedimensional subspace . . . . . . . 560) . . 560) . . . . . . .2 n. . . . . 15. 15. . . . . . 560) . . . . . 8 (p. . . . 1 (p. . . . . .9 Exercises . . . . . . . . . Euclidean spaces. . . . . . . . . . . . . . .1 n. . . . . . . . 15. . . . . . . . . . . . . 15. . 15. . . 2 (p. . . . . . . . . . . . . . . . 15. . . . . .9. 15. . . . . . 15. . . . . . 3 (p. . . . 15. . . . . . . . . . 560) . . . . . . . 576) . . . . . . . . . . 555) . . . . . The GramSchmidt process . . . . . . 1 (p. . . . . . . . . 2 (p. .18 n. 560) . . .12 n. . . . . . .16. 15. . . . . . . . . . . .14 n. . . . .3 n. . . . . . . 15. . . . . . . . . . . . . . . . . . . . . 15. . . . .9. . . . . . .10 Inner products. . . .5.9. . . . 560) . . . . . . . 13 (p. . . .9. . . . . . . . 555) . . . . . 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 . 15. . . . projections . 15.5. . . 15. . . . . . 560) . . Norms . 26 (p. . . . . . . . . . . 10 (p. . .9. . . . .9. . .5. . . .2 n. . . . . . . . . . . . . . . . 6 (p. . . . 15. . . . . . . 27 (p. . . . . . . . . . . . . . . . 560) . . . . . . . . 15. . . . . . . . . . . . . . . . .8 n. . . . 15.12 Exercises . . . . .9. . 560) . .12. . . . . . . . . . . . . . . . . . . . . . . . 15. 560) . . 15. . . . . 7 (p. . . .4 n.16 n. . . . . . 15. . . . . . . . . . . . 560) . . .9. . . . . 22 (p.1 n. . . . . . . . . . . . . 567) . . . . . . . . . . .15 n. . . . . . . . . . . .16.14 Orthogonal complements. . . 15. . . 15. . . . . 3 (p. . . .5 n. . . . . . 567) . 24 (p. .9. . . .5. .20 n. . . 560) . . . . . . . .21 n. . . . . . . . . . . . . . . .2 n. . . . . . . . . . 15. 15 (p. .9. . . . . . . . . .10 n. . . 555) . . . . . . . . .13 Construction of orthogonal sets.5. . 9 (p. . 15. 16 (p. . 560) . . . . . . . .1 n. 14 (p. . . . . . . . . . . . . . . . . . . . . 23 (p. . . . .13 n. . . . . . . . . 15. .4 n. . . . .9. . . . . .11 n. . . .9. 15. . . . . . . . 15.9. . . . . .9 n. .9. . . . . . 9 (p. 12 (p. .8 Bases and dimension . .19 n. . . 555) . . . . . . . . . . 555) . . . . . . . . . . . . . . . . 25 (p. . . . . . . . . . . . . . . .9.12. . . . . . . . . . . 560) . . . . . 28 (p. . . .9. . . . . .6 Subspaces of a linear space . 15. . . . . 4 (p. . . . . . . . . . . . . 576) . . . . . . . . . . . . . . . . . . . . . . . . . 15. . . . . . . . . . .17 n. . . . . .17 n. . . . . . . . 15.16. . 560) . . . . . . . . . . . . . . . . . . . . . . .11 Orthogonality in a euclidean space . . . . . . . . .16. . . . .7 n. . . . . . . . . . . .9. . . 5 (p. . . . . . . . . . . . . 23 (p. 15. .6 n. . . . . . . . . . . 15. . . . . . . . . . . . 15. . . . . . . . . . . . . . .7 Dependent and independent sets in a linear space . . . . . . . . . . . . . . . . . . . . . . . . .3 n. . . . . . . . . . . . . . . . . 15. . . . 15. 15. . . . . . . . . . . . . . . . . . . . . . . 576) . . . . . .16 n. . . .9. . 11 (p. . . . . . . . 560) . . 560) . . . . . . . . . . . . . . . . . .16 Exercises . . . . .
7 n. . . .4. . . . . . 16 (p. . . .5 n. . . . . . . 582) . . . . 16. . . . . . . . . 582) . . . . . . . . . . 16. . . . .2 n. . 582) . .1 n. . . . . . . 5 (p. . . . . . 16. . . . . . . . . . 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 . . . . . 16. . . . . . . . . . . . . 1 (p. . . . .CONTENTS 9 16 Linear transformations and matrices 16. . . . . . . . . . . . . . . . 16. . . . 15 (p. 582) . . .2 n. . . . . . . . .12. .10 n. . . . . . . . . . 5 (p. . . 16. . . . . . . . . . . . . . . . . . . . . . . 582) . . . . . . . . .4. . . . . . .8. . . . . . . . . . . . . . 589) . . 16. . . .5 Algebraic operations on linear transformations . . . . 16. . . . . . . . . . . . . . . . . . . 27 (p. . . . . . . . . . . . . . .4. 16. . . . . . . .9 n. . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . 16. . 16. . . . . . . . . . .4 n. . . 10 (p. . .14 n. . . . . . . . . . . . . . . . . . . . 16. . . . 597) . . . .12 n. . . . . . 582) . . . . . . . . . . . .6 Inverses . 8 (p. . . . . . 16. . . .4. . . . . . . .4. . 582) . 589) . . . . . . . . . . . . . . . . . . . .4. . . 7 (p. . . . . . . 596) . . . . . . . . . . . . . . . 16. . . .5 n. .4. . . . . . .4. . 16. . . . . . 596) . . . . .6 n. 16.4 Exercises . 17 (p. . 16. . . . 582) . 2 (p. . . . . .2 n. . . . . . . . . . 3 (p. . . . . . . . .15 n. .2 Null space and range .3 n. . . . . . . . . . . . . . . . . . . . . 23 (p. 25 (p. . . . . . . . . . . . . 582) . . . . . . . 4 (p. . . . .1 n. . . . . . . 16. . . . . . . . . . . . . . . . . . . . . . . . 16. . . form . 16. . 16. . . . . . . . 16. . . . . 590) . . . 16. . . . . . . . . . . . .13 n. . . . . 16. . .1 n. . . . .12. . 9 (p. . 597) . . . . . . . . . . .3 n. . . . . . 17 (p. . . . . . 16. . . . . . . 16. . . . . . . . . 3 (p. . . . . 597) . . .8.12. . . . . . . . . . . . . . . . . . . . . . . .1 Linear transformations . . . . . . .11 n. . 16.4. . . . . 596) . . 16. . . . 16. . . . . . .12. . .3 n. . .8. 582) . . .6 n. . . . . . .7 Onetoone linear transformations . . . . . . 16. . .4. . . . .3 Nullity and rank . . .4. . . . 27 (p. . . . . . . . .11 Construction of a matrix representation in diagonal 16. . . . . . .12. . 16. .10 Matrix representations of linear transformations . 589) . .8. . .9 Linear transformations with prescribed values . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . 582) . . . . . . .8 n. . .8 Exercises . . . . . . . . . . . . 582) . . . . . . 7 (p. . . . . . 16 (p. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 (p. . . . . . 16. . . . . . . . 16. . 8 (p. . . . . . . . .4. . . . . . . . 16 (p. . . . .4 n. .4 n. . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 Exercises . 16. .12. . . . . 582) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16. . 582) . . . 4 (p. . . . 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 .
−1.4 Exercises 12. 12. 4). (b) a − b = (−3. . 7) . .1 Historical introduction 12. 21).2 n. a single picture.Chapter 12 VECTOR ALGEBRA 12. 450) µ ¶ µ ¶ µ ¶ 7 5 5 11 13 . (3.3 Geometric interpretation for n ≤ 3 12. (e) 2a + b − 3c = (0. −5) 3 2 2 4 4 The seven points to be drawn are the following: The purpose of the exercise is achieved by drawing. 1 (p.2 . 24. 6.4.4. . 0. −2) .1 n. 4) . I would say). (1. 3). (0. 0). containing all the points (included the starting points A and B. (4. 0. as required.2 The vector space of ntuples of real numbers 12. 2 (p. . . 9). 450) (a) a + b = (5. (d) 7a − 2b − 3c = (−7. (c) a + b − c = (3.
. 3 (p. (−3. (5. (3. . 12. 450) The seven points this time are the following: µ ¶ µ ¶ µ ¶ 5 10 7 5 15 . 4) . 5) .4. . by letting t vary in all R. . the straight line through − → point A with direction given by the vector b ≡ OB is obtained. 2) . 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. .26 Vector algebra It can be intuitively seen that. the straight line through B − → with direction given by the vector a ≡ OA is obtained.3 n. by letting t vary in all R. (−1.
1. 4 2 4 4 2 . (c) If the value of x is ﬁxed at 0 and y varies in [0. −3) . all the combinations are aﬃne.2 . by the question immediately following. as it is going to be clear after the second lecture.Exercises 27 12. .4. and hence D is the vertex of the parallelogram with three vertices at O. containing all the points (included the starting points A and B. indeed. I would say). the segment OB is obtained. where −→ − → − → OD = OA + OB. y) ∈ [0. 4 . . the same construction with the value of x ﬁxed at 1 yields the segment AD. when x = 1 the segment obtained joins the midpoints of 2 the two sides OA and BD. A. 1 . it seems to me. 5) 2 4 2 3 3 2 The whole purpose of this part of the exercise is achieved by drawing a single picture. (3. 450) (a) The seven points to be drawn are the following: µ ¶ µ ¶ µ ¶ µ ¶ 3 5 5 4 7 1 . Similarly. 1]. . 1 . and B. and 2. −1) .4 n. 4 (p. (0. 5 . (4. 3 . and it is enough to repeat the construction a few more times to convince oneself that the set © ª xa + yb : (x. This is made clear. The picture below is made with the value of x ﬁxed at 0. 1]2 is matched by the set of all points of the parallelogram OADB. 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. . . 3 .
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 (↑) . of equation 3x − y = 0 and 3x − y = 5 respectively. 12.5 1 1.4. 5 (p. 450) µ ¶ µ ¶ µ ¶ 2 1 c1 +y = x 1 3 c2 I 2x + y = c1 II x + 3y = c2 µ 12.5 3 (d) All the segments in the above construction are substituted by straight lines.→ II (↑) .5 2 2.→ I x = 2 I (↑) .5 n.28 Vector algebra 4 3 2 1 0 0.→ 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) .4. (e) The whole plane. and the resulting set is the (inﬁnite) stripe bounded by the lines containing the sides OB and AD. 6 (p.6 (a) n.→ III (↑) .
→ II y = z z ← 1 (−2.7 (a) n.→ III (↑) .8 (a) n.→ III (↑) .→ I II (check) y=0 x=0 z=0 0=0 (b) II − I − IV 0 = −3 .→ 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 . 8 (p. 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 (↑) .4. 7 (p.Exercises 29 12.4. 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 . 1. 1) III − II 0 = 1 (b) (c) I x + 2z = 1 II x + y + z = 2 III x + y + z = 3 12.
1. ⇔ I II ⇔ (b 6= 0) ⇔ l≡h−k ⇔ ∃h ∈ R ∼ {0} . If you look carefully. if b is parallel to d. kd + b = hd ∃h ∈ R ∼ {0} . if c is parallel to d. 12. u is parallel to v. c = hd) ⇔ (∃l ∈ R ∼ {0} . 451) Let the two vectors u and v be both parallel to the vector w.4.30 Vector algebra 12. ∃k ∈ R ∼ {0} . which is nonnull. it follows that c. probably). this means that there are two real nonzero numbers α and β such that u = αw and v = βw. 11 (p. 10 (p.4).4. b = ld ∃h ∈ R ∼ {0} .11 n. 451) Assumptions: I c=a+b II ∃k ∈ R ∼ {0} .9 n. they diﬀer only in the phrasing.10 n. then (by II) b is the diﬀerence of two vectors which are both parallel to d. (⇐) Since (by I) we have c = a + b. a = kd Claim: (∃h ∈ R ∼ {0} . 451) (b) Here is an illustration of the ﬁrst distributive law (α + β) v = αv + βv . a + b = hd 2. ∃k ∈ R ∼ {0} . is parallel to d too. 12.4. 9 (p. (⇒) Since (by I) we have b = c − a. is parallel to d too. which is nonnull. According to the deﬁnition at page 450 (just before the beginning of § 12. Then µ ¶ v α u = αw = α = v β β that is. b = ld) I present two possible lines of reasoning (among others. then (by II) c is the sum of two vectors which are both parallel to d. b = (h − k) d h 6= k ∃l ∈ R ∼ {0} . c = hd ∃h ∈ R ∼ {0} . it follows that b.
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
z) must satisfy the two conditions h(2. −4) . bi = 0 ⇔ x ha. 456) The statement is false. −5α. ha. −7) (2 · 3 + 6 · 4 + 3 · (−5)) = (2. 5 (p. bi = 0 ⇔ 7x + 14y = 0 Let (x. 2) Thus the set of all such vectors can be represented in parametric form as follows: {(α. 1. 12. −1) .−7) 15 = ¡2 7 . −7) 15 = (30. 3 (p.5 n. 4. Then c = (1. 6 (p. −105) (e) a hb. 456) The required vector. bi = 0 c = xa + yb ¾ ⇒ hxa + yb. y.34 Vector algebra (c) ha + b.3 n. y) ≡ (−2. y. of coordinates (x. 4. b − ci = 0 ⇔ a ⊥ b − c and the diﬀerence b − c may well be orthogonal to a without being the null vector.4 n. (3. −3α)}α∈R 12. −4) 6= 0 and h(1. 2)i = 0. 1. bi + y hb. question (d). 4 .4. ci = (2. bi = ha. 5. 2) . 0) See also exercise 1.8. − 15 15 15 ¢ 12. . 456) hc.8. The simplest example that I can conceive is in R2 : a = (1. 1) c = (1. 5.8. (x. −1. 1). z)i = 0 h(1. z)i = 0 that is. (x. I 2x + y − z = 0 II x − y + 2z = 0 I + II 3x + z = 0 2I + II 5x + y = 0 b = (1.ci = (2. y. ci ⇔ ha. 60. Indeed. ci = (2 + 2) · 3 + (4 + 6) · 4 + (−7 + 3) · (−5) = 72 (d) ahb.
10) 9 c = α= hb. c = αb Substituting in the ﬁrst condition. −8.8. hb. 8) 9 1 d = (22. −8. bi 9 1 α=− 4 9 . 8) 9 1 d = (22. −4 − 9α = 0 and hence 1 (−4. for some α to be determined. ai − α hb. for some α to be determined. c = (α. −2α) Substituting in the ﬁrst condition. 7 (p.Exercises 35 12. d = a − c = a − αb Thus the second condition becomes hb. −1.6 n. 2 + 2α) Thus the second condition becomes 1 · (2 − α) + 2 · (−1 − 2α) − 2 (2 + 2α) = 0 that is. −1. bi = 0 and hence 1 (−4. a − αbi = 0 that is. with no mention of coordinates) From the last condition. 2α. d = a − c = (2 − α. 456) 1 (solution with explicit mention of coordinates) From the last condition. 10) 9 c = 2 (same solution. ai 2 · 1 + (−1) · 2 + 2 · (−2) 4 = =− 2 + 22 + (−2)2 hb. −1 − 2α.
456) (a) b = (1. −1) or b = (2. (c) b = (−2. 10 (p. (d) b = (−1. bi = 0 kbk = kak If a = 0. −a) or b = (−b.8. If a 6= 0. 12. Then the ﬁrst equations gives (12. namely.2) β x=− y α and substitution in the second equation yields ¶ µ 2 β + 1 y 2 = α2 + β 2 α2 that is. The above conditions take the form αx + βy = 0 x2 + y 2 = α2 + β 2 Either α or β must be diﬀerent from 0. 1). −1). (c) b = (−3. 456) If b is the required vector. b = (−β. 2). −1) or b = (1. then kak = 0 and hence b = 0 as well. 1) or b = (2. Suppose α is nonzero (the situation is completely analogous if β 6= 0 is assumed). (b) b = (−1. 13 (p. 1). let the coordinates of a be given by the couple (α. y). the following conditions must be satisﬁed ha. 1). α) and b = (β. −2) . a).8 n. (b) 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. and the coordinates of b by the couple (x. −1) or b = (−2. −2) or b = (3. 1). y = and hence y = ±α Substituting back in (12.8.7 n. −1) or b = (−1. (d) b = (b. 2) or b = (1.36 Vector algebra 12. β). −α) In particular.
xi = 0 Equivalently. −4)i + h(x. z)]i = 0 that is. orthogonality of the two vectors − → CA = a − x is required. b − xi = 0 or ha. y. h(2. −4. ha − x. xi + hx. bi − ha + b. y. (x. the two vectors CA and CB must be orthogonal. −4) . −1. same solution. y. z)i = 0 or x2 + y 2 + z 2 − 5x + 5y + 3z + 6 = 0 The above equation. 14 (p. solution with explicit mention of coordinates) Let (x. bi . z) . − 3 and radius 5 . hence the vector OC. If the right angle is in C. − 5 . 1) − (x. z) − → − → be the coordinates of C.8. 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 . −1. 2 2 2 2 2 (right angle in C. kxk − 2 2 − → CB = b − x ¿ ° ° °2 À ° ° a + b °2 ° ° a+b ° ° = ° a + b ° − ha. y. Then.x + ° ° ° 2 ° 2 2 ° °2 ° °2 ° ° ° ° °x − a + b ° = ° a − b ° ° ° 2 ° 2 ° . is unknown). that is. −4.9 n. (x. [(3. z)] . with no mention of coordinates) Let − → − → − → a ≡ OA b ≡ OB x ≡ OC − → (stressing that the point C. −1. y. y. Thus h[(2. z)i − h(2.Exercises 37 12. −4. 1) + (3. 456) 1 (right angle in C. −4) − (x. (3. 1) .
the condition to be required is 0 = ha − b. 3 (right angle in B. solution with explicit mention of coordinates) With − → − → − → this approach.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. −3. with the notation of point 2. −5). 3. . BC = 1 − x + 3y + 12 + 5z + 20 that is. and C is any point of the circle of π having AB as diameter. y + 4. ha − b. x − 3y − 5z = 33 The solution set is the plane π through B and orthogonal to (1. with no mention of coordinates) Proceeding as in the previous point. Indeed. bi Thus the solution plane π is seen to be through B and orthogonal to the segment connecting point B to point A. x − bi that is. same solution. Since BA = − → (−1. xi = ha − b. if π is any plane containing AB. the vectors required to be orthogonal are BA and BC. it is known by elementary geometry that the triangle ACB is rectangle in C. z + 4). 4 (right angle in B. the following must hold D E − − → → 0 = BA. 5) and BC = (x − 1.
5. 456) I c1 − c2 + 2c3 = 0 II 2c1 + c2 − c3 = 0 I + II 3c1 + c3 = 0 I + 2II 5c1 + c2 = 0 c = (−1. 16 (p. 6) 25 The question is identical to the one already answered in exercise 7 of this section. we have that p ≡ OP must be equal to αb = − → αOB. 3) 12. − → Recalling solution 2 to that exercise. 2 2 2 2 µ ¶ 3 1 1 3 q = − .11 n. 15 (p. β) = p= 12. 456) p = (3α.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. 2 2 2 2 µ ha. bi 10 = hb. 17 (p. . −2) 25 (α. of equation hb − a. 12.8. 456) q= 1 (−8. bi 4 .8. −3β) 4II + 3I 25α = 11 4I − 3II 25β = −2 1 (33. ai It is also clear that π and π 0 are parallel. . . xi = hb − a. 44) 25 I 3α + 4β = 1 II 4α − 3β = 2 1 (11. with α= Thus ¶ 5 5 5 5 p = . 4α) q = (4β.8.10 n.− .12 n.
21 (p.14 n. bi + ha. I obtain By subtraction. ai + 2 ha.8.15 n.13 n. a + bi + ha − b. bi = 2 kak2 + 2 kbk2 The geometric theorem expressed by the above identity can be stated as follows: Theorem 2 In every parallelogram. 12. with vertex C opposed to the vertex in the origin O) to be a rectangle. bi substituting −b to b. 457) .8. that is.8. ai − 2 ha. bi as required. 20 (p. and that in such a rectangle ka + bk and ka − bk measure the lengths of the two diagonals. 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. ka + bk2 − ka − bk2 = 4 ha. 456) It has been quickly seen in class that ka + bk2 = kak2 + kbk2 + 2 ha. 19 (p. bi + hb. a − bi = ha. 12. 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. the sum of the squares of the four sides equals the sum of the squares of the diagonals. 456) ka + bk2 + ka − bk2 = ha + b. 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. bi + hb.40 Vector algebra 12.
choosing x = 0 and y = 1 gives ha. 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. −→ w ≡ CD. and let M and N be the midpoint of the diagonals AC and DB. let − → u ≡ AB. wi kck2 = kuk2 + kvk2 + 2 hu. bi = 0. 457) Orthogonality of xa + yb and 4ya − 9xb amounts to Since the above must hold for every couple (x. starting from left − → −→ in clockwise order. 22 (p. wi = kuk2 + kwk2 + 2 hu. 4ya − 9xbi ¡ ¢ = −9 ha. ¢ ¡ ¢ ¡ kuk2 + kvk2 + kwk2 + kzk2 − kck2 + kdk2 = kzk2 − kvk2 − 2 hu. and D be the four vertices of the quadrilateral. vi − 2 hv. −→ z ≡ DA = − (u + v + w) kzk2 = kuk2 + kvk2 + kwk2 + 2 hu. 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. ui − → v ≡ BC. B. wi + 2 hv. bi x2 + 4 ha. C. vi + 2 hu.16 n. 0 = hxa + yb. vi kdk2 = kvk2 + kwk2 + 2 hv. bi y 2 + 4 kak2 − 9 kbk2 xy . it follows that kbk is equal to 4. wi ° ° − °− →°2 = 4 °MN ° 12.Exercises 41 Let A. In order to simplify the notation.8. y). wi We are now in the position to prove the theorem.
42 Vector algebra Finally.´the trinomial x2 kbk2 + 2x ha. 24 (p. bi = 0 . bi ³ positive. ai d = b− a ha. ka + xbk2 = kak2 + x2 kbk2 + 2x ha. bi is negative for all x in the open is interval − 2ha.17 n. bi ≥ 0 Moreover. ai hb. 3bi = 22 · 62 + 32 · 42 + 2 · 2 · 3 · 0 = 25 · 32 and hence √ k2a + 3bk = 12 2 12.8. in the general context of the linear space Rn (where n ∈ N is arbitrary).8. 0 . ai c = 12. only if it reduces to the second degree term x2 kbk2 . ai a ha. ka + xbk ≥ kak) ⇒ ha. the full answer to the problem has already been seen to be hb.bi 0. k2a + 3bk2 = k2ak2 + k3bk2 + 2 h2a. x2 kbk2 + 2x ha. the following biconditional is true: If ha. ka + xbk2 ≥ kak2 ∀x ∈ R. bi = kak2 + x2 kbk2 if a ⊥ b 2 ≥ kak if a ⊥ b (b) Since the norm of any vector is nonnegative.18 n. bi is zero. 25 (p. that is. the following biconditional is true: ¡ ¢ (∀x ∈ R. 457) (a) For every x ∈ R. ka + xbk2 − kak2 ≥ 0 ⇔ ∀x ∈ R. if it is negative. since a and b have been shown to be orthogonal. − kbk2 . I have proved that the conditional (∀x ∈ R. 457) This is once again the question raised in exercises 7 and 17. is true. the trinomial is negative in the open interval 2 ³ ´ kbk 2ha. ka + xbk ≥ kak) ⇔ ∀x ∈ R. It follows that the trinomial can be nonnegative for every x ∈ R only if ha. In conclusion. Since the coordinatefree version of the solution procedure is completely independent from the number of coordinates of the vectors involved.bi . by pure computation.
kak 7 7 7 6 ha.− kak 7 7 7 µ ¶ a 6 3 2 − = − . 460) ha.2 n. 2 2 2 2 ¶ kbk2 = 4 µ (b) There are just two vectors as required. ki 2 c cos a i = =− kak kkk 7 c cos a i = . .9 Projections.3 (a) n. . 2 (p. 3 (p. Angle between vectors in nspace 43 12.11. bi = 10 The projection of a along b is 10 b= 4 12.− .11. the unit direction vector u of a. 1 (p.11. bi = 11 kbk2 = 9 11 22 22 . ii = kak kik 7 ha. 460) ha.10 The unit coordinate vectors 12. . ji 3 c cos a j = = kak kjk 7 ha.Projections. . 460) µ 5 5 5 5 . Angle between vectors in nspace 12. and its opposite: µ ¶ a 6 3 2 u = = . 9 9 9 ¶ The projection of a along b is 11 b= 9 12.11 12.1 Exercises n.
If some confusion is made between points and vectors. more than that. if one looks only at numerical results. OC b ° °° ° an angle of triangle OBC → → °− ° °− ° = cos B OC OB ° °OC ° ° D E − − → → OC. −2. 1) Then B ≡ (1. C are coplanar with the origin O. − − → → respectively. respectively. 5) C ≡ (3. B. −4) − → r ≡ AB = (−1. This amounts to work. √ √ h−q. ACB. −6) There is some funny occurrence in this exercise. C. with the vectors OA. there is nothing funny in doing that. −3.4 Let n. one may be led into operate directly with − − − → → → the coordinates of points A. cos ACB. −5) − → q ≡ CA = (−1. B. 5 (p. 460) A ≡ (2. 1) − → p ≡ BC = (2. as a matter of fact. being of conceptual type. b b C BA. B. − → OC. which makes a wrong solution apparently correct (or almost correct). −1. and a fairly bad one. OA b ° °° ° an angle of triangle OCA → → °− ° °− ° = cos C OA °OC ° °OA° D E − − → → OA. 3. OB b ° °° ° an angle of triangle OAB → → °− ° °− ° = cos AOB °OA° °OB ° b b b instead of cos B AC. more than that. AB. points A. AC. as implicitly argued above.11. that is. cos C BA. qi b cos C = = √ √ =0 k−pk kqk 6 35 .44 Vector algebra 12. −ri 6 6 41 b cos B = =√ √ = kpk k−rk 41 6 41 0 h−p. and/or angles. − → − → − → − → 2. Up to this point. it’s only a mistake. OA = BC and OB = AC. The funny thing is in the numerical data of the exercise: it so happens that 1. OB. as the three angles of triangle ABC. −4. The b angles in A. ri 35 35 41 b cos A = =√ √ = k−qk krk 41 35 41 √ √ hp. C in place of the coordinates of vectors BC. −1. therefore computing D E − − → → OB. are more precisely described as B AC.
6 (p. as a particular case. OB blue. 460) Since ka + c ± bk2 = ka + ck2 + kbk2 ± 2 ha + c. 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. which immediately implies c ac = π 7 c c bc = π − ab = π 8 b b AOB = ACB . AB violet It turns out.Exercises 45 Point 1 already singles out a somewhat special situation.11. AOB is a right angle. if c = −a. OC green. bi from ka + c + bk = ka + c − bk it is possible to deduce ha + c. bi = 0 This is certainly true. therefore. b 3. 12. − → − → − → OA red. but point 2 makes OACB a parallelogram. and point 3 makes it even a rectangle.5 n.
ci ha.11. the same conclusion holds. even if a + c = 0 is not assumed. 12. 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 ¯ . 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. 461) ha.11.46 Vector algebra Moreover. 8 (p. from hc. ci c =− = − cos ab kbk kck kbk kak c c 8 8 and hence that bc = π ± ac = 7 π.6 n. 9 π (the second value being superﬂuous).7 (a) n. 10 (p. Indeed. bi = cos ϑ sin ϑ − cos ϑ sin ϑ = 0 kak2 = kbk2 = cos2 ϑ + sin2 ϑ = 1 (b) The system µ that is. bi and kck = kak it is easy to check that c cos bc = hb. 460) We have r √ n (n + 1) n (n + 1) (2n + 1) n kbn k = ha. bi = − ha.
(2k + 1) π)}k∈Z the only vector satisfying the required condition is (0. The assumption that OABC is a rhombus is expressed by the equality kak = kck ((rhombus)) .Exercises 47 The computation gives 1 + cos2 ϑ + sin2 ϑ − 2 cos ϑ 2 (1 − cos ϑ) cos ϑ ϑ Thus if ϑ ∈ {(−2kπ. the oriented segments AB (blue.11. and the same is true for CB (red. 0). On the other hand.8 n. 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. 12. 11 (p. dashed) and OB are congruent. dashed) and OA. 461) Let OABC be a rhombus. if ϑ ∈ {2kπ}k∈Z the coeﬃcient matrix of the above system is the identity matrix. 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.
461) (a) That the function X Rn → R. b. 17 (p.9 n. c − ai = 0 or kck2 − kak2 + ha. since in every parallelogram ABCD with a = AB and b = AC the − → diagonal vector CB is equal to a − b. The statement to be proved is orthogonality between the diagonals D E − − → → OB. As a matter of fact. and strict positivity still relies on the fact that a sum of concordant numbers can only be zero if all the addends are zero. too: ∀a ∈ Rn .11. since norms are nonnegative real numbers. ci − hc. nonnegativity is obvious. Homogeneity is clear. the square of each side is the sum of the squares of the other two sides. ∀α ∈ R. 461) The equality to be proved is straightforward. the converse is true. ai = 0 and hence (by commutativity) kck2 = kak2 The last equality is an obvious consequence of ((rhombus)). too.48 Vector algebra that is. minus their double product multiplied by the cosine of the angle they form.11. The equality in exam can be readily interpreted according to the theorem’s − → − → statement. Thus a parallelogram has orthogonal diagonals if and only if it is a rhombus.10 n. 13 (p. 12. The “law of cosines” is often called Theorem 3 (Carnot) In every triangle. The theorem specializes to Pythagoras’ theorem when a ⊥ b. a 7→ ai  i∈n is positive can be seen by the same arguments used for the ordinary norm. 12. X X X kαak = αai  = α ai  = α ai  = α kak i∈n i∈n i∈n . and a − b. so that the triangle ABC has side vectors a. AC = 0 ha + c.
it is apparent that S is a square. y) ∈ R2 : x + y = 1 + ≡ {(x. 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. f (x.g. −x) = 0 ∀x ∈ R). with sides parallel to the quadrant bisectrices 2 1 2 1 0 1 1 2 2 (c) The function f : Rn → R. ¯ ¯ ¯ ¯ ¯ ¯ ¯X ¯ ¯ X ¯ ¯X ¯ ¯ ¯ ¯ ¯ ¯ ¯ kαak = ¯ αai ¯ = ¯α ai ¯ = α ¯ ai ¯ = α kak ¯ ¯ ¯ ¯ ¯ ¯ i∈n i∈n i∈n .. 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 . a 7→ X i∈n ai  is nonnegative.Exercises 49 Finally. It is homogeneous: ∀a ∈ Rn . ∀b ∈ Rn . y) ∈ R− × R+ : −x + ª = 1} y © 2 ≡ (x. but not positive (e. 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. y) ∈ R− : −x − y = 1 ≡ {(x. y) ∈ R2 : x + y = 1 = S++ ∪ S−+ ∪ S−− ∪ S+− where S++ S−+ S−− S+− © ª ≡ (x.
2 2 (c) x + y = 3 and y − x = −5 yield (x.15. 3 (p.1 . 1 (p. ∀a ∈ Rn . 467) I 2x + y = 2 II −x + 2y = −11 III x − y = 7 The solution is (x. y) = (1.15. −4).2 n. I + III 3x = 9 y = −4 II + III III (check) 3 + 4 = 7 . y) = ¡1 ¢ . 1 . y) = (4. 12. 2 2 (d) x + y = 7 and y − x = 5 yield (x. ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯ ¯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 12.12 The linear span of a ﬁnite set of vectors 12.14 Bases 12. y) = − 1 . y − x) ¢ ¡ (b) x + y = 0 and y − x = 1 yield (x. too (this is another way of saying that the triangle inequality holds).50 Vector algebra Finally. Indeed.1 Exercises n.13 Linear independence 12. 467) x (i − j) + y (i + j) = (x + y. −1). y) = (3. 6). f is subadditive.15 12. (a) x + y = 1 and y − x = 0 yield (x. ∀b ∈ Rn .
at least one between α and β is nonzero. if you prefer).6 n. it is enough to require (1 + t)2 − (1 − t)2 6= 0 4t 6= 0 t 6= 0 12. indeed. 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.15. 6 (p. 467) By the previous exercise.3 n. d) must be proportional to (y. too 12. a and b are linearly dependent. for every (α. but in the present case they are both nonzero. c) = h (b. 8 (p. ∗ . 467) (a) The linear combination 1i + 1j + 1k + (−1) (i + j + k) is nontrivial and spans the null vector. and has unknowns and equations in equal number (hence part 2 of the proof of Steinitz’ theorem does not apply). d) has been seen in the lectures (proof of Steinitz’ theorem. The converse is immediate. then the linear combination 1a − αb is nontrivial and it generates the null vector. 7 (p. y) exists. Then. c) and (b. β. and hence to each other. If a nontrivial solution (x. and αa + βb = 0 with β = 0 and α 6= 0 implies a = 0). from (a. 5 (p. β. then they are parallel. since a and b have been assumed both diﬀerent from the null vector. −x) (see exercise 10. and hence a = − α b (or b = − α a. by proving that if a and b are linearly dependent. (b) Since.Exercises 51 12. αi + βj + γk = (α.15. γ) ∈ R3 . αa + βb = 0 with α = 0 and β 6= 0 implies b = 0.5 n. because both a and b have been assumed diﬀerent from the null vector. 467) Linear independency of the vectors (a.4 n.15. (b) The argument is best formulated by counterposition. 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. β 12. section 8).15. Thus β αa = −βb. both (a. b) and (c. d) it is immediate to derive ad − bc = 0. I argue by the principle of counterposition. Let a and b nontrivially generate the null vector: αa + βb = 0 (according to the deﬁnition. 467) (a) If there exists some α ∈ R ∼ {0} such that∗ a = αb.
467) (a) Again from the proof of Steinitz’ theorem. j. β + γ) α (i + j + k) + βj + γk = (α. 1 respectively. α + γ) 12. (c) Similarly. for every (α. k) is linearly independent. γ) ∈ R3 .52 Vector algebra it is clear that ∀ (α. (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. 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. γ) ∈ R3 . (αi + βj + γk = 0) ⇒ α = β = γ = 0 so that the triple (i.7 n. β. − 1 . i + j + k) is linearly independent. α=β=γ=0 showing that the triple (i. 3 3 . k) and (i + j + k. 10 (p. j. β. αi + βj + γ (i + j + k) = (α + γ. 1. (d) The last argument can be repeated almost verbatim for triples (i. j. i + j + k.15. part 1. 0) and ¢ ¡ 0. and hence that the given triple is linearly independent. β. taking into account that αi + β (i + j + k) + γk = (α + β. γ) and hence from αi + βj + γ (i + j + k) = 0 it follows α+γ =0 β+γ =0 γ =0 that is. k). β + γ. α + β.
u. which has just been seen to be a basis of R4 . d ≡ a + b + c = (2. c. d) a linearly dependent quadruple. b. and ﬁrst equation in that order. c. are given as the solution (s. Thus the given triple spans R3 . b. and ﬁrst equation in that order.→ I x = 5 (d) For an arbitrary triple (a. from the fourth. subtracting the ﬁrst equation from the second. 3 . and it is linearly independent (as seen at a). 3. 1. t. c). 467) (a) Let xa + yb + zc = 0. For example.→ II y = − 14 3 (↑) . the system I x+y+z = a II y + z = b III 3z = c ¢ ¡ c c has the (unique) solution a − b. and z are in turn seen to be equal to 0. Thus (a. b. x+z−t x+y+z−t x+y+t y + 3t = = = = 0 0 0 0 Then. 12 (p. third. that is. 2. e}.15. Thus (a.8 n. c) is a linearly independent triple. x+z x+y+z x+y y = = = = 0 0 0 0 then y. that is. b. and hence t.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 (↑) . third. 3). 1) (c) Let e ≡ (−1. b. from the fourth. c. (d) The coordinates of x with respect to {a. gives y = 0. v) to the following system: s+u−v s+t+u−v s+t+v t + 3v = = = = 1 2 3 4 . (b) Any nontrivial linear combination d of the given vectors makes (a. −1. 12. x. e) is linearly independent. and z are in turn seen to be equal to 0. and suppose xa + yb + zc + te = 0. x. b − 3 .
u. v) = (1. . 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). 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.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 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 a1 ↔ a3 . 1.
0. (c) α (t. − 2 . t) = (0. Indeed. 3 = (0.Exercises 55 12. 2. 1. 1. From I and III. and a linear dependence in the equation system suggests that it may √ ¢ ¡√ well have nontrivial solutions. 2. 2 is such a solution. − 2 . 467) √ ¢ ¡√ ¢ ¡ √ ¢ ¡ α 3. 1 + γ 0. 3. then. 0) + β (1. I deduce that α = γ. 2 = (0. 1) + γ (0. −2. 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. t. 0 + β 1. 0 + β 1. the given √ ª © √ triple is linearly dependent for t ∈ 0. γ) = (0. (b) √ ¢ ¡√ ¢ ¡ √ ¢ ¡ α 2. 1 + γ 0. . 1. 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. β. 13 (p. 0. 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. 1.15. 0. 1. Thus the three given triple of vectors is linearly dependent. 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). 0. as already noticed. 1.9 (a) n. √ ª ©√ Such a system has nontrivial solutions for t ∈ 2. In conclusion. 2. 0) and the three given vectors are linearly independent. t 6= 0. Let us suppose.
(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. 15 (p. (c) Similarly. 1) and b ≡ (1. z) is maximal linearly independent. b. u − v = e(1) w − z = e(3) v − w = e(2) z = e(4) and (u. in the order. Thus (u. 14 (p. c) is linearly independent. v. w. and cannot be equal to z. w. choosing as nontrivial coeﬃcient triple (α. 468) Let a ≡ (0. z) is a maximal linearly independent triple. α + β + γy. a + c) is linearly dependent. −1.56 Vector algebra 12. 1). Moreover. 12. x.11 n. β. α + β + γz) . 468) (a) Since the triple (a. v. a + c) is linearly independent. b + c. I look for two possible alternative choices of a vector c ≡ (x. w.15. b + c. and z the four vectors given in each case.10 n. 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.15. v. 12.15. w.12 n. 1). Since αa + βb + γc = (β + γx. 468) Call as usual u. (a − b) − (b + c) + (a + c) = 0 it is seen that the triple (a − b. 1. v. c) is linearly independent. (a) It is clear that v = u + w. it is a basis of R4 . Thus (u. y). so that v can be dropped. too (b) On the contrary. 17 (p. y. z) such that the triple (a. 1. y. w. z) spans the four canonical vectors. b. every linear combination of u and w has the form (x. γ) ≡ (1. z) is maximal linearly independent.
and then either II or III yields α = 0. since the vectors u and v there coincide with the present ones. w + z. β. 0). Keeping the same notation. Let u.Exercises 57 Subtracting equation III from equation II. 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. 19 (p. 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. As an example. 12. it is then readily checked that a=u−w b = 2w (b) The converse inclusion holds as well. (0. 468) The ﬁrst example of basis containing the two given vectors is in point (c) of exercise 14 in this section. 18 (p. 1. as any subspace of a vector space. and z must be such to make the following conditional statement true for each (α. y. a second example is (u. v. 12. w − z). 0. I yields β = 0 (independently of the value assigned to x). v =a−b Thus lin S = lin T 1 w= b 2 1 u=v+w = a− b 2 . and S. since each element of lin T is a linear combination in T .13 n. v. and let a and b be the two elements of T (still in the given order). in such a case. equations II and III are the same.15. and w be the three elements of S (in the given order). if y = z. is closed with respect to formation of linear combinations. 0. 1. I obtain γ (y − z) = 0 my choice of x. 1). (1. 1). possible choices for c are (0.14 n. 468) (a) It is enough to prove that each element of T belongs to lin S. 0). Conversely.15. (1. Indeed.
Notice that αc + βd = (α + β.15 n. B ⊆ lin A and hence. by part (a) of last exercise. (b) By the last result. and hence lin A ⊆ lin B. b} B ≡ {a + b} lin A ∩ B ⊆ lin B where the couple (a. u = 2c − d v =d−c It remains to be established whether or not w is an element of lin U . 2α + 3β. 3α + 5β) It follows that w is an element of lin U if and only if there exists (α. since from A ⊆ B I can infer A ⊆ lin B. b) is linearly independent. 20 (p. lin S = lin T = lin U 12. lin B ⊆ lin lin A = lin A lin A ∩ lin B = lin B On the other hand. if c and d are the two elements of U .58 Vector algebra Similarly. In conclusion. c=u+v d = u + 2v which proves that lin U ⊆ lin S. 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. −2). Indeed.15. 468) (a) The claim has already been proved in the last exercise. Inverting the above formulas. A∩B =∅ lin A ∩ B = {0} . which proves that lin S ⊆ lin U . β) ∈ R2 such that α+β =1 I II 2α + 3β = 0 III 3α + 5β = −1 The above system has the unique solution (3.
17 Exercises .The vector space Vn (C) of ntuples of complex numbers 59 12.16 The vector space Vn (C) of ntuples of complex numbers 12.
60 Vector algebra .
showing that L is horizontal.5 Exercises 13.2 Lines in nspace 13. .Chapter 13 APPLICATIONS OF VECTOR ALGEBRA TO ANALYTIC GEOMETRY 13. because y = 2 requires t = 3.5. 13. (b). (d) and (e) have the second coordinate equal to 1.5. point (c) does not belong to L.2 n. Points (b). 1 (p. This gives x = −2. 477) − → A direction vector for the line L is v ≡ 1 P Q = (−2. 477) − → A direction vector for the line L is P Q = (4.1 n.3 Some simple properties of straight lines 13.4 Lines and vectorvalued functions 13. Thus a point belongs to L if and only if its second coordinate is equal to 1.1 Introduction 13. Finally. Among the given points. 1). 2 (p. which requires t = 2. and (e) belong to L. showing that of the three points only (e) belongs to L. (d). 0). which yields x = −4 6= 1. The parametric equations 2 for L are x = 2 − 2t y = −1 + t If t = 1 I get point (a) (the origin).
and (e) belong to L.5.5.3 n. 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. Thus a point belongs to L if and only if its second coordinate is equal to 1. (d). 477) I solve each case in a diﬀerent a way. Among the given points. 13. 2) The two vectors are not parallel. 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. The three points do not belong to the same line. −2. 5 (p. showing that L is horizontal.5. −2) − → QR = (−1. 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.4 n. 3 (p.62 Applications of vector algebra to analytic geometry 13. (b). (a) − → P Q = (2. 4 (p. hence the three points do not belong to the same line. (b) Testing aﬃne dependence.5 n. . 0). 0.
1) pxy AF = (−6. 3) − → −→ pxy AG = (−15.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. all the eight points have their third coordinate equal to 1.5. 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. and hence they belong to the plane π of equation z = 1. 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). 477) The question is easy. as a matter of fact. R). I check that among the twodimensional projections of the oriented segments connecting A with points D to H − → − → − → pxy AD = (−4. 6 (p. −7) . 13. and. we can have a very good hint on the situation by drawing a twodimensional picture. We can concentrate only on the ﬁrst two coordinates. 2) pxy AE = (−1. I get k = − 4 from the ﬁrst equation 3 and k = −1 from the second. First. Q does not belong to L (P.6 n. but it must be answered by following some orderly ¡ ¢ path. 8) pxy AH = (12.
G) (red). Direction vectors for these three lines are 1 −→ 1 − → −→ − pxy HE = (−13.64 Applications of vector algebra to analytic geometry − → only the ﬁrst and third are parallel to pxy AB. 4) pxy GH = (9. E. −5) 2 3 and as normal vectors for them I may take nEG ≡ (4. b) are determined by the vector equation P + ha = Q + kb which gives P − Q = kb − ha that is. hence they intersect it exactly at one point. b} (13. 7) nGH ≡ (5. 13). C. F belongs to LGH . F } belong to the same (black) line LABC . H} 13. substituting this value in any of the three equations. Thus there are three (maximal) sets of at least three collinear points. 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 . nor D belong to LHE . D. and the two lines intersect at the point of coordinates (5. and point H to the second. G} P 3 ≡ {F. within the set P 1 . and whether or not any elements of P 1 belong to them. B.1) . C belongs to LEG . 7 (p. It is seen by direct inspection that. Therefore. 7) pxy EG = (−7. D. I get h = 4. namely P 1 ≡ {A. 477) (a) The coordinates of the intersection point of L (P . I get k = 1. Thus all elements of the set P 1 ≡ {A. B. 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. G. 9) nHE ≡ (7.5. − → P Q ∈ span {a. (G. 8 (p.5. 9. C. The three equations are consistent. E) (green) coincide. 13) By requiring point E to belong to the ﬁrst and last. 13. (H. F } P 2 ≡ {C. it only remains to be checked whether or not the lines through the couples of points (E. a) and L (Q. and hence no two of them can both belong to a diﬀerent line. H) (blue).8 n.7 n. and neither A nor B.
and the 9 9 9 65 minimum distance is 3 . Ai kAk kP − Qk − hP − Q. Ai t = at2 + bt + c where a ≡ kAk2 = α2 + β 2 + γ 2 b ≡ hP − Q.− = . µ. 477) (a) Let A ≡ (α. γ). the point on L of minimum distance from Q is the orthogonal projection of Q on L. ν) and Q ≡ (%. τ ). .5. its graph is a parabola with vertical axis and vertex in the point of coordinates ! µ ¶ Ã 2 2 2 b ∆ hQ − P. d (t0 )) = −√. 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)). Ai = 22 − 2 + 20 =0 9 that is. σ. 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. 2a 4a kAk2 kAk2 . Then X (t) ≡ P + At = (λ + αt.5. 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 . Q − X (t0 ) = . Ai − . 9 (p. 2 − 2t. µ + βt. Ai = α (λ − %) + β (µ − σ) + γ (ν − τ ) 2 c ≡ kQk2 + kP k2 − 2 hQ.Exercises 65 13.9 n.− 9 9 9 9 9 9 hQ − X (t0 ) . β. 65 as vertex. 10 (p. ν + γt) f (t) ≡ kQ − X (t)k2 = kQ − P − Atk2 = kQk2 + kP k2 + kAk2 t − 2 hQ. 13. P ≡ (λ. . P i + 2 hP − Q.10 n. 477) X (t) = (1 + t. (c) ¶ ¶ µ µ 5 26 19 22 1 10 X (t0 ) = .
11 (p. ∃c ∈ R. Ai hA. k) such that k − h = c (the two lines intersect in inﬁnitely many points. Q − P = ca and equation (13. (13.66 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. 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. Ai = hQ − P. 477) The vector equation for the coordinates of an intersection point of L (P .11 n.2) is satisﬁed by all couples (h. Ai α (% − λ) + β (σ − β) + γ (τ − ν) = 2 α2 + β 2 + γ 2 kAk where hQ − P.e.2) has no solution (the two lines are parallel). (b) Q − X (t0 ) = Q − (P + At0 ) = (Q − P ) − hQ − X (t0 ) . For each given point of L. Ai A kAk2 hQ − P. i. a) and L (Q.5.2) . they coincide). 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. or − → P Q ∈ span {a} that is.. Ai = 0 kAk2 13. Ai − hQ − P. the transformation h 7→ k (h) ≡ h + c identiﬁes the parameter change which allows to shift from the ﬁrst parametrization to the second. The two expressions P + ha and Q + ka are seen to provide alternative parametrizations of the same line.
8 Exercises 13. 2. ¡ ¢ (b) Similarly. z) to 4.1 n. 1 belongs to the 2 2 plane. −2. 3) − → P R = (2.8. 3 belongs to the 2 2 plane.3) we get 2 ¡ ¢ yielding s = 1 and t = 0. we get 2s + 2t = −4 2s − 2t = 0 3s − t = −2 . 482) We have: − → P Q = (2.3) 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 = 1.6 Planes in euclidean nspaces 13. 3 in (13. 0. 477) 13.Planes in euclidean nspaces 67 13. proceeding in the same way with the triple (−3. −1) so that the parametric equations of the plane are x = 1 + 2s + 2t y = 1 + 2s − 2t z = −1 + 3s − t (13. 2.7 Planes and vectorvalued functions 13.3) ¡ ¢ (a) Equating (x. so that the point with coordinates 4. y. equating (x. 2. 1 in (13. 0. 1. 2 (p. so that the point with coordinates 2. 12 (p. (c) Again. z) to 2.12 n.5. y. −1).
0. 4) − (0. contradicting the third equation). 482) x = 1+t y = 2+s+t z = 1 + 4t (b) u = (1. the point of coordinates (0. −1) belongs to the plane. (e) Finally. 0.2 (a) n. 1. 1. 0) = (1. so that the point with coordinates (−3. (d) The point of coordinates (3. 2) does not belong to the plane.68 Applications of vector algebra to analytic geometry yielding s = t = −1. because the system 2s + 2t = −1 2s − 2t = −1 3s − t = 1 is inconsistent (the ﬁrst two equations yield s = − 1 and t = 0. 4) x = s+t y = 1+s z = s + 4t . 1.8. contradicting the 2 third equation). 3 (p. 1) v = (1. because the system 2s + 2t = 2 2s − 2t = 0 3s − t = 4 is inconsistent (from the ﬁrst two equations we get s = t = 1. 1. 13. 0) = (1. 1. 1. 2. 5) does not belong to the plane. 1) − (0.
0) + s (1.4): P ≡ (1.. 1) Finally.Exercises 69 13. −3. (a) If p + q + r = 1.8. 0.4) and (2. there exist real numbers q and r such that S = P + q (Q − P ) + r (R − P ) Then. S = (1 − q − r) P + qQ + rR = pP + qQ + rR . 4 (p. 0) a = (1. deﬁning p ≡ 1 − (q + r).4 n. 2. 2) + t (−2. 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. 0) does not belong to the plane M . 1. 5 (p. 1. 4. 1) 13. and R. 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. 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 . (b) The answer has already been given by writing equation (13. The second point is directly seen to belong to M from the parametric equations (1. 4. 2) b = (−2. Q.3 n.8. 482) This exercise is a replica of material already presented in class and inserted in the notes. −3) belongs to M . 482) (a) Solving for s and t with the coordinates of the ﬁrst point. (b) If S is a point of M.
and I require that the three given points belong to π 2 .70 Applications of vector algebra to analytic geometry 13. 6 (p. better. c. (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. in 4 · III).8. d) of the cartesian equation of π 2 as unknown. 482) I use three diﬀerent methods for the three cases. 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. 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 ¢ .
0. −2) × (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 ). 1) and form a linearly independent couple can be chosen as direction vectors for M . 3) and (3. (0. Thus I assign values arbitrarily to d2 and d3 in the orthogonality condition 3d1 − 5d2 + d3 = 0 say. 2) Thus π 3 coincides with π2 . −1. and from the ﬁrst (which is unchanged) I get a = 1. which is presented in the subsequent section. and I get dI = (−1. I have thought it better to show its application here already. 3. 1) (I may have as well taken one of the ﬁrst two points of part a). −4.6 n. they have the same normal direction.8. 1. and I obtain P = (1.Exercises 71 From the last equation (which reads −3c + d = 0) I assign values 1 and 3 to c and d. The plane π3 and the given plane being parallel. Any two vectors which are orthogonal to the normal n = (3. respectively. 0. since it has the same normal and has a common point with it. 7 (p. however. from the second (which reads b − c + d = 0) I get b = −2. 0). 482) (a) Only the ﬁrst two points belong to the given plane. 3. 0) . −5. (b) I assign the values −1 and 1 to y and z in the cartesian equation of the plane M . 1) = (2. At any rate (just to show how the method applies in general). 1) 2−6+1+d =0 yielding x − 2y + z + 3 = 0 13. in order to obtain the third coordinate of a point of M . Such a normal is n = (2. 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) The parametric equations for M are x = 1 − s + 5t y = −1 + 3t z = 1 + 3s dII = (5.
 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. just to check on the computations.72 Applications of vector algebra to analytic geometry 13. 8 (p. which leads to t = 11 from the second row. 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.8. I do get indeed (−14. 3) × (3.5) I take all the variables on the lefthand side. 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. 3. −4) .7 n.). 8. − 3 .. 2. which is apparent (though incomplete) from the righthand side of (13.8. t) = − 5 . − 5 is all right.. 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 5 5 5 5 ¡ 2 1¢ and the check by means of (s. Thus Q = (−14. 3.8 n.  1  2  0 s t h k  const. − 2 . 9 (p. thereby suggesting the wrong answer that M and M 0 are parallel. I proceed to ﬁnish up the elimination. 17). 482) (a) A normal vector for M is n = (1. k) = − 2 . and to s = −2 from the ﬁrst. respectively. 13. 7 . 482) The question is formulated in a slightly insidious way (perhaps Tom did it on purpose.5). and the constant terms on the righthand one. Substituting in the lefthand side of (13. which is a trace of the parametric equations of M . −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. because the parametric equations of the two planes M and M 0 are written using the same names for the two parameters. Another handy solution to¡the equation corresponding¡ the last to ¢ ¢ row of the ﬁnal elimination table is (h. 2. However. 1) = (−4. On the contrary. This gives R = 2 .5). which is likely to have no solutions. This is already enough to get the ﬁrst point from the parametric equation of M 0 .
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. with π parallel to the yaxis. (b) A cartesian equation for M is x − 2y + z + d = 0 From substitution of the coordinates of P . 2 ¡ = s5= −7 ¢ Thus L ∩ M consists of a single point. 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 . 3 r . 10 (p. .g. − 3 4 2 13. from which the third and second equation give t = 3. 0 and R = 0. ¡ ¡ 3 3 the ¢ Q = − 2 . 4 . Since the latter is parallel to the former. the point of coordinates −2.9 n. Coordinate of points on L are now easy to produce. −2. the two planes are parallel. and π 0 parallel to ¢ xzplane. − 2 . namely. e..8.Exercises 73 whereas the coeﬃcient vector in the equation of M 0 is (1. 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. 3 . the coeﬃcient d is seen to be equal to 3. and the coordinates of the point P ∈ M do not satisfy the equation of M 0 . 1). 2 .
this yields y = −3 in the last two equations. −2) and P R = (2. hence x = 2. 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. it is seen that L is not parallel to πa . I am reduced to the previous case. a. these two values contradicting all equations. 3)i = 9 6= 0 L is not parallel to π c . 1. 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. computing the determinant of the matrix having v. With either reasoning. 2.g. 5. 3): h(2. − → (b) By computing two independent directions for the plane π b . a.. (1. because subtraction of the third equation from the second gives y = −2. 4. it suﬃces to check orthogonality between v and (1. Alternatively. L is not parallel to π b . 11 (p. 2 (c) Since a normal vector to π c has for components the coeﬃcients of the unknowns in the equation of π c . −1) − (1. Thus I consider the system 4 3 2x + y = 2 4 x + y = −1 3x + y = 3 shows that {v. −1. 2) − − → (1. −2). b} is a linearly independent set. −1. whereas subtraction of the ﬁrst from the third yields x = 1 .8. 2. 1. 1 . P Q = (3. .74 Applications of vector algebra to analytic geometry 13. 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. 1.10 n. contradicting the ﬁrst. 1. e. 3) . 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. 3) and 3 .
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. 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. and let Q ≡ (xQ . 483) Let d be a direction vector for the line L. 13 (p.12 n. 483) Since every point of L belongs to M .6) is satisﬁed. 1) for each t ∈ R.Exercises 75 13. π is the only plane containing L and P . 2. by deﬁning q≡p r≡0 it is immediately seen that condition (13.8. 3. 483) Let R be any point of the given plane π. Choosing. 5). yQ .8. . obtaining x−y+1 =0 13.13 n. t = 1. 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. and (2. 14 (p. 3) + t (1. zQ ) be a point of L.g. 2. e. 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. M contains the points having coordinates equal to (1.. A plane containing L and the point P ≡ (xP . by subtracting the ﬁrst equaiton from the second. there exists a real number p such that S = P + p (Q − P ) Then.6) If S belongs to the line through P and Q. 1. 12 (p.11 n. 3). 13. other than P or Q.8. Since any two distinct points on L and P determine a unique plane. yP .
3 (p. (i) (A × B) × (A × C) = −2i + 4k. 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. −5. 10) ° 1 °− → ° → − ° 15 area ABC = °AB × AC ° = 2 2 − → − → AB × AC = (3. 26 √ 18 j 2054 √ 18 j 2054 + √7 k 2054 A×B or − kA×Bk = 1 √ i 6 √ 41 i 2054 − √ 7 k. (e) (A × B) × C = 8i + 3j − 7k.11 Exercises 13.2 n. (g) (A × C) × B = −2i − 8j − 12k.76 Applications of vector algebra to analytic geometry 13. 487) − → − → AB × AC = (2. −2. −1.1 n. 2054 2 √ j 6 − 1 √ k 6 A×B or − kA×Bk = + 2 √ j 6 + 1 √ k.11. −3) × (3. 6 13. −2) = (10. 1 (p. (f) A × (B × C) = 10i + 11j + 5k. 3) × (3. 2 (p. (c) C × A = 4i − 4j + 2k.11. 13. (b) B × C = 4i − 5j + 3k.3 (a) n. 0) = (3. (h) (A + B) × (A − C) = 2i − 2j.9 The cross product 13.11. 487) (a) A × B = −2i + 3j − k.10 The cross product expressed as a determinant 13. 15) √ ° 1 °− → ° → − ° 3 35 area ABC = °AB × AC ° = 2 2 (b) . −6. 9. (d) A × (C × A) = 8i + 10j + 4k. 2.
b) is linearly independent and kb × ak > 0. m. 0. b + ci = ha. Thus hb. 1) = (1. 487) 13. b × ai = 0 because b × a is orthogonal to a. 1. −1) ° √3 1 °− → ° → − ° area ABC = °AB × AC ° = 2 2 n. (b × a) − bi = hb. n) and b ≡ (p.11. kck2 = 22 + 12 . . 1) × (1. bc = −π is impossible.11.5 n. because \ kck2 = kb × ak2 + kbk2 − 2 kb × ak kbk cos (b × a) b = kb × ak2 + kbk2 > kbk2 since (a.11. bi = 0 13. 487) Let a ≡ (l. ci < 0 kbk c cos bc = − <0 kck i i c ∈ π.4 − → − → CA × AB = (−i + 2j − 3k) × (2j + k) = 8i − j − 2k 13. 487) ha. 6 (p.Exercises 77 (c) − → − → AB × AC = (0.π bc 2 c . r) for the sake of notational simplicity. 5 (p. 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. 4 (p. √ (c) By the formula above. (b) hb.6 (a) n. q. 1. (b × a)i + hb. ci = hb. Moreover. and kck = 5. −bi = − kbk2 because b × a is orthogonal to b.
b. Similarly. (a × b)×b is either equal to a or to −a. bi2 = 1 k(a × b) × ak2 = kck2 = 1 Since the direction which is orthogonal to (a × b) and to b is spanned by a. The three vectors a. k(a × b) × bk2 = ka × bk2 kbk2 − ha × b. 7 (p. (a × b) × a is either equal to b or to −b. Since a is a unit vector. ai2 = 1 (c) And again.7 n. The proof that (a × b) × b = −a is identical. bi2 = 1 so that a × b is a unit vector as well. 488) (a) Since kak = kbk = 1 and ha.12. by Lagrange’s identity (theorem 13. kck2 = ka × bk2 kak2 − ha × b. bi = 0. a × b are mutually orthogonal either by assumption or by the properties of the vector product (theorem 13.78 Applications of vector algebra to analytic geometry 13.11. 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. . (b) By Lagrange’s identity again.de).12.f) ka × bk2 = kak2 kbk2 − ha.
either h kbk = 0 or k kak = 0. 8 (p. 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 (↑) . Geometrically. the hypotheses a 6= 0. in order to satisfy the condition a×b= c kck Thus kak ≤ kbk < +∞. kak (b) From the previous point. q. the unique solution is 9p = 9 2 − 2r = 4 −2q + 1 = 3 2+1−2 =1 1−2 =1 b = (1.Exercises 79 13. and ha. In the second case. . and its norm must depend on ϑ according to the relation kck kbk = (13.9 n.→ I check on IV check on III that is. Thus two solutions to the problem are µ ¶ 7 8 11 ± . either k = 0 (which means b = 0). suppose that a 6= 0.− . 9 (p. Then either ha. b − ci = 0 imply that b − c = 0. a × (b − c) = 0. bi = h kbk2 or ha. or kak = 0 (which is equivalent to a = 0). In particular. the conditions a × b = c and ha. −1. In the ﬁrst case. or there exists some k ∈ R such that b = ka. From a×b = kak kbk sin ϑ. bi = k kak2 . in a×c c×a which case it has to be equal to kak2 or to kak2 .→ II (↑) .bi a of b along a and the projecting vector kak2 (of length kb×ak ) are null. 488) (a) From a × b = 0. either h = 0 (which means a = 0). b can be taken orthogonal to a as well. Then both the projection ha. either there exists some h ∈ R such that a = hb. −1) The solution to this exercise given by Apostol at page 645 is wrong. 13. that is. or kbk = 0 (which is equivalent to b = 0).7) kak sin ϑ (b) Let (p.8 n. r) be the coordinates of b. 488) c (a) Let ϑ ≡ ab.− 9 9 9 the vector b must be orthogonal to c. which can only happen if b = 0.11.11.
11 (a) Let n. If x and z both solve (13. and hence it does not meet the additional requirement ha.80 Applications of vector algebra to analytic geometry 13.8).9). . x = a × c + αa From condition (13.10 n. 1. by skewsimmetry of the vector product. a kak2 1 kak2 (13. c × a is orthogonal to a. a × c + αai = 1 that is. α=− Thus the unique solution to the problem is b ≡ a × c− 13. 10 (p. 1) − → w ≡ CA = (−1. 488) Replacing b with c in the ﬁrst result of exercise7. ha. it is seen that the c × a is a solution to the equation a×x=c (13.8). Thus the set of all solutions to equation (13.11. (a × c) × a = c Therefore. −1) Each side of the triangle ABC can be one of the two diagonals of the parallelogram to be determined.8) is © ª x ∈ R3 : ∃α ∈ R. 488) − → u ≡ AB = (−2. 11 (p. xi = 1 a × (x − z) = a × x − a × z = c − c = 0 which implies that x − z is parallel to a. 0) − → v ≡ BC = (3. −2. 1.11.9) We are bound to look for other solutions to (13.8) However.
2. the other is CF . where AD = AB + AC = u − w. In this case E = B + v − u = “A + C − B” = (4. In this case D = A + u − w = “B + C − A” = (0. the other is AD. 0) (b) µ¯ ¯ 1 0 ¯ u×w = ¯ 1 −1 √ 6= ku × wk = ¯ ¯ ¯ ¯ ¯ . In this case F = A − v + w = “A + B − C” = (−2. where CF = CA + CB = −v + w. −2.¯ ¯ ¯ ¯ −1 1 ¯ = (−1. − ¯ −2 0 ¯ ¯ −1 −1 ¯ ¯ ¯¶ ¯ ¯ −2 1 ¯ ¯. −1) √ 6 area (ABC) = 2 . 2) − → − → − → If one of the diagonals is AB.Exercises 81 − → − → − → If one of the diagonals is BC. where BE = BC + BA = v − u. 0. the other is BE. 2) − → − → − → If one of the diagonals is CA. −2.
Third.11. b). in the ﬁrst place a and b are nonnull. a × bi − 3 kbk2 = 48 √ 3 hb.b}. b}). x+y =0 and x−y =0 (13. suppose that x (a + b) + y (a − b) + z (a × b) = 0 Then z cannot be diﬀerent from zero.10) kck2 = 4 ka × bk2 − 12 ha × b. 13 (p. b + ci = −2 ha. 488) (a) It is true that. bi 1 c cos ab = = kak kbk 2 2 2 2 c ka × bk = kak kbk sin2 ab = 12 13. bi + 9 kbk2 = 192 √ kck = 8 3 hb. 488) b + c = 2 (a × b) − 2b ha. since every ntuple having the null vector among its components is linearly dependent. b}. Indeed. a × b) is linearly independent. Thus z = 0. ci = 2 hb. bi = −4 ha. too.13 n. if the couple (a. a − b. then the triple (a + b. b) is linearly independent. and (13. (a × b) ∈ lin {a. because (a. in particular. 12 (p.82 Applications of vector algebra to analytic geometry 13. (a × b) is nonnull. 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. Fourth.12 n. Secondly. because it is orthogonal to a and b. since the only vector / which is orthogonal to itself is the null vector. too. b) is linearly independent.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. (a × b) is orthogonal to both a + b and a − b (as well as to any vector in lin {a. ci c cos bc = = kbk kck 2 .11.
b + a × b) is linearly independent. y. y + z must be null. (c) It is true that. too. let (x. 14 (p. and this in turn implies that both x + y and x + z are also null. yielding y− − → → AB = − AC x . b) is linearly independent. then A and C coincide and it is clear that the line through them and B contains all the three points.11. (b) It is true that. (a + b) × (a − b) = a × a − a × b + b × a − b × b = −2a × b and the triple (a. 13.14 n. if AC is null. then the triple (a. b. a × b) is linearly independent when the couple (a. if the couple (a. then in the nontrivial null combination − → − → xAB + y AC = 0 x must be nonnull. a + a × b. Indeed. − → Otherwise. The system x+y = 0 x+z = 0 y+z = 0 has the trivial solution as the only solution. b) is linearly independent.Exercises 83 and hence x = y = 0. 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. too. b. AC is linearly dependent. if AC is nonnull. (a + b) × (a − b)) is linearly independent. if the couple (a. then the triple (a + b. Indeed. 488) → − → − (a) The cross product AB × AC equals the null vector if and only if the couple ³ ´ − − → → − → AB. arguing as in the fourth part of the previous point. In such a case. b) is so.
B. −1}. (p. b. h = −1. the set of all points P belonging to the line through A and B. p × bi + hb. ¯ π¯ ¯ ¯ h kpk = kp × bk kbk ¯sin ¯ = kpk 2 and h ∈ {1.84 Applications of vector algebra to analytic geometry that is. 13. and {p. and P belong to the same line. ai hb. ai hp. 488) (a) From the assumption (p × b) + p = a. This gives kp × bk = kpk kbk = kpk and hence √ (13. and p×b are pairwise orthogonal. pi = hp.11) 1 = kak2 = kp × bk2 + kpk2 = 2 kpk2 that is. b. b. b) deﬁne the same orientation. 15 (p. p is orthogonal to b. y− → B = A − AC x which means that B belongs to the line through A and C. pi = 0 that is.15 n. p × b) is linearly independent. the set n o − → − → P : AP × BP = 0 is the set of all points P such that A. p × b) and (p × b. that is. and (p × b. (c) Since p is orthogonal both to p × b (deﬁnition of vector product) and to b (point a). b. hb. kpk = 22 . Since the triples (p. (d) Still from the assumption (p × b) + p = a. p. p) deﬁnes the opposite one. p × bi + hp. (b) By the previous point.11). hp. (b) Since p. pi = hb. p × b} is a basis of R3 . ai = kpk2 = 1 2 1 2 p × (p × b) + p × p = p × a kp × ak = kpk2 kbk = . taking into account (13. there exists some h ∈ R such that (p × b) × b = hp Thus.11. b.
so that q is discordant with b × a.16 Linear cartesian equations for planes . It follows 4 that k = −1. π .14 Exercises 13. b × p.15 Normal vectors to planes 13. 4 2 4 with p. the vectors a = p + b × p. and 3π . the decomposition of p obtained in (13. and b × a form angles of π . I have ﬁnally obtained 1 1 p = a − (b × a) 2 2 13. which is orthogonal to a. On the other hand.The scalar triple product 85 Thus 1 1 p= a+ q 2 2 (13. Since a normal vector to π is b.13 Cramer’s rule for solving systems of three linear equations 13. respectively. 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.12) requires that the angle formed with p by q is − π .12 The scalar triple product 13.12) where q = 2p − a is a vector in the plane π generated by p and a.
1 (p. a distinguished one is n ≡ (2i + 3j − 4k) × (j + k) = 7i − 2j + 2k (b) hn.17.2 (a) n. z)i = hn. 9 9 9 . 496) (a) Out of the inifnitely many vectors satisfying the requirement.1 n. 0) Y axis : µ 1 2 2 . 0 2 ¶ Zaxis : (c) The distance from the origin is 7 . y. 2. 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. y. z) = − .86 Applications of vector algebra to analytic geometry 13. 3)i 13. . 2 (p. (x. 496) µ ¶ µ ¶ 7 0.− 3 3 3 7 0.− . .17. 0. (1. z)i = 0 (c) hn. (x. y. − . 0.17 Exercises 13. 2 or 7x − 2y + 2z = 9 or 7x − 2y + 2z = 0 n = knk (b) The three intersection points are: Xaxis : (−7.
17. 2. respectively.17. which can be determined by the equations 3tC − 6 (−2tC ) + 3tC = 2 tD − 2 (−2tD ) + tD = 7 −→ ¡ 7 1 ¢ Thus tC = 1 . −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. and 9 6 ° ° 19 √ °−→° 19 kn4 k = 6 °CD° = 18 18 Alternatively. (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. rewriting the equation of π 4 with normal vector n2 3x − 6y + 3z = 21 d2 − d4  19 =√ kn2 k 54 dist (π. 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 . 4 (p. π 0 ) = . π 1 and π3 are orthogonal because hn1 . 3 (p.4 n. CD = 6 − 9 n4 . 496) A cartesian equation for the plane which passes through the point P ≡ (1. n3 i = 0.Exercises 87 13.3 n. tD = 7 .
−4. 5 (p. −8) (b) A cartesian equation for π is x + 2y − 2z = 5 (c) The distance of π from the origin is 5 . −7) and a cartesian equation for it is 10x − 3y − 7z + 17 = 0 13.7 n. 496) Proceeding as in points (a) and (b) of the previous exercise. R ≡ (3. −4) = (4.17. −1. 496) (a) A normal vector for the plane π through the points P ≡ (1. 2.5 n.8 n. 9) A cartesian equation for the plane is (x − 2) + 2 (y − 3) + 9 (z + 7) = 0 or x + 2y + 9z = 55 13. 3). 4). 4. 3. −2) is − → − → n ≡ P Q × QR = (2. 3) × (0. 2). 2. Q ≡ (3. Q ≡ (2. 8. 8 (p. −3.17. Thus the parametric equations are x = 2 + 4h y = 1 − 3h z = −3 + h . 3. 3) = (1. 2. 7. 496) A direction vector for the line is just the normal vector of the plane. 1.6 n. a normal vector for the plane through the points P ≡ (1. 6 (p.88 Applications of vector algebra to analytic geometry 13. 1) × (3. 12) − (1. −2) is − → − → n ≡ P Q × RQ = (1. R ≡ (−1. 9 (p. −1). 3 13. 2. 496) A normal vector to the plane is given by any direction vector for the given line n ≡ (2.17. 1.17. 6) = (10. −4.
(b) The direction vector of the line L is d = (−1. −7. −1. 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 = 1.17. 3) is −d. 10 (p. 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 . −7.9 n. −3. A normal vector for the plane π 00 which is orthogonal to L and contains (−1. Hence the hitting point is (0. 3). 5) is 2x + 3y + 2z + 15 = 0 (e) At time t = 2. −4. 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. the moving point has coordinates (−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 (−2. −4. (d) At time t = 3.Exercises 89 13. Substituting. 5). 1).
the other values ´ easily obtained: (l. 15 (p.17.17.17.10 From n. 0). √8 . 0. 2.13 n. 1 knk 2 13. 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. z) = 3 . a cartesian equation for π is x−y+z = 2 13. 1 . Then are 3 3 ³ √7 . 8 . m) = − 7 .90 Applications of vector algebra to analytic geometry 13. 11 (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. 496) First I ﬁnd a vector of arbitrary norm which satisﬁes the given conditions: l + 2m − 3n = 0 l − m + 5n = 0 13. √3 the required vector is − 122 122 122 .12 n. 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. 0.11 n. 13 (p.17. y. 2 2 ¡ ¢ Assigning value 1 to n. 14 (p.
4) × (1.17.14 n. 2 + d = 6 The two solutions of the above equation are d1 = 4 (corresponding to the plane already given) and d2 = −8. 2. 497) If the line ` under consideration is parallel to the two given planes. 3. 3) = (1. 2. Thus the required equation is 2x − y + 2z = 8 13.19 Eccentricity of conic sections 13.21 Exercises . 17 (p. 1) Since ` goes through the point P ≡ (1. 3. 4).15 n. 2.18 The conic sections 13.The conic sections 91 13. −2. −1) yields 6 − 2 − 2 + 4 6 − 2 − 2 + d = 3 3 that is.20 Polar equations for conic sections 13. 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. parametric eqautions for ` are the following: x=1+t y = 2 − 2t z =3+t 13. 2. a direction vector for ` is d ≡ (2. 3). 20 (p. 3) and n2 ≡ (2. which have as normal vectors n1 ≡ (1.17.
22 Conic sections symmetric about the origin 13.24 Exercises 13.23 Cartesian equations for the conic sections 13.92 Applications of vector algebra to analytic geometry 13.25 Miscellaneous exercises on conic sections .
Chapter 14 CALCULUS OF VECTORVALUED FUNCTIONS .
94 Calculus of vectorvalued functions .
From this the other two existence axioms (of the zero element and of negatives) also follow as particular cases. −1) respectively. Indeed. 0) and for (α.Chapter 15 LINEAR SPACES 15.1 Introduction 15. R. Of course it can also be seen directly that the identically null function x 7→ 0 is a rational function. for (α. 555) The set of all real rational functions is a real linear space. 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.1 n. 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.2 The deﬁnition of a linear space 15.4 Elementary consequences of the axioms 15. β) = (0.3 Examples of linear spaces 15. Q. as the quotient of the constant polynomials P : x 7→ 0 .5 Exercises 15. β) = (0. let P . 1 (p. and S be any four real polynomials.5.
it only needs to be proved that if deg P ≤ deg Q and deg R ≤ deg S. taking into account exercise 1. and which have the same value at 0 and 1. of a type which has become usual at this point. the two existence axioms follow as particular cases. . and which achieve at 0 the double value they achieve at 1 is a real linear space. 555) The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0 and 1. 2 (p. taking into account the general properties of the operations of function sum and function multiplication by a scalar 15.96 Linear spaces and Q : x 7→ 1.3 n. Indeed. Similarly. for every two real numbers α and β. 15. and it is anyway clear that the identically null function x 7→ 0 achieves the same value at 0 and 1. and using the same notation.5. 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. 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. 555) The set of all real rational functions having numerator of degree not exceeding the degree of the denominator is a real linear space. 555) The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0 and 1. then deg [αP S + βQR] ≤ deg QS This is clear. A ﬁnal remark concerning the other axioms. 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. Indeed. allows to conclude. (αf + βg) (0) = αf (0) + βg (0) = αf (1) + βg (1) = (αf + βg) (1) so that the closure axioms hold. 15.4 n.5. is a real linear space.5. Indeed. Again. since for every two real numbers α and β deg [αP S + βQR] ≤ max {deg P S. 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. 3 (p. 4 (p. The remaining linear space axioms are immediately seen to hold.
1] with τ 0 = 0 and τ m = 1. 555) The set of all real valued step functions which are deﬁned on [0. 1].5 n. and which have value at 1 which exceeds the value at 0 by 1. some (m + 1)tuple (γ r )r∈{0}∪m of real numbers. 5 (p. for any two real numbers α and β.Exercises 97 15. so that for some nonnegative integers m and n.s)∈m×n I am using here the following slight abuse of notation for degenerate intervals: (σ. 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. 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. (Js )s∈n are the partitions of (0. χC is the characteristic function of C ¿ 1 if x ∈ C χC ≡ x 7→ 0 if x ∈ C / and (Ir )r∈m. and existence of negatives f (1) = 1 + f (0) ⇒ (−f) (1) = −1 + (−f) (0) 6= 1 + (−f ) (0) 15. 1] is a real linear space. 6 (p. let f and g be any two such functions. Indeed. and some ntuple (δ s )s∈{0}∪n of real numbers∗ .5. σ] ≡ {σ}. σ r ] Js ≡ (τ s−1 . τ s ] Then. αf + βg = (αγ 0 + βδ 0 ) χ{0} + ∗ (r ∈ m) (s ∈ n) X (αγ r + βδ s ) χKrs (r. 555) The set of all real valued functions which are deﬁned on a ﬁxed domain containing 0 and 1. Indeed.5. . (τ s )s∈{0}∪n Ir ≡ (σ r−1 . let α and β be any two real numbers such that α + β 6= 1. 1] with σ 0 = 0 and σ m = 1. This is necessary in order to allow for “point steps”. that the set in question is an aﬃne subspace of the real linear space). some increasing (m + 1)tuple (σ r )r∈{0}∪m of elements of [0. some increasing (n + 1)tuple (τ s )s∈{0}∪n of elements of [0. is not a real linear space.6 n. however. The other failing axioms are: existence of the zero element (the identically null function has the same value at 0 and at 1). 1] associated to (σ r )r∈{0}∪m .
as in the previous examples. 1] → R. for every two such functions f and g. The second closure axiom does not hold. depending on whether monotonicity is meant in the weak or in the strict sense. Indeed. however. 1]. and a ﬁnal. † . 7 (p. with no more† than m + n steps on [0. s) ∈ m × n) which shows that αf + βg is a step function too. in the former case. 1].5. by classical theorems in the theory of limits.s)∈m×n is the “meet” partition of (Ir )r∈m and (Js )s∈n Krs ≡ Ir ∩ Js ( (r. because the sum of two increasing functions is an increasing function too. for that matter. Final remark concerning the other linear space axioms. Thus the closure axioms hold. s) ∈ m × n. All the other linear space axioms hold.8 n. The axiom of existence of negatives fails. for any two such functions f and g. with the same number of steps. too. 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. is a real linear space. x 7→ 0 is indeed a step function. 15. usual remark concenrning the other linear space axioms applies. 555) The set of all real valued functions which are deﬁned and integrable on [0. 13 (p. 555) The set of all real valued functions which are deﬁned on R and convergent to 0 at +∞ is a real linear space. as every constant function) but not strictly increasing. 1] equal to zero. Indeed. 1] 7→ R.5. x→+∞ lim (αf + βg) (x) = α lim f (x) + β lim g (x) x→+∞ x→+∞ = 0 so that the closure axioms hold.98 Linear spaces where (Krs )(r. with the integral over [0.7 n. and any two real numbers α and β. σr−1 = τ s−1 and ασr + βτ s = ασr−1 + βτ s−1 . at any x0 ∈ R∪{−∞}). and for any two real numbers α and β. since the identically null function is weakly increasing (and. The axiom of existence of the zero element holds or fails. with just one step (m = 1. because the function αf is decreasing if f is increasing and α is a negative real number. Thus. weakly decreasing too. the set of all real valued and (weakly) increasing functions of a real variable is a convex cone. 15. and that the opposite of a step function is a step function too. 15. It may be noticed independently that the identically null function [0.5. It may also be noticed independently that the identically null function [0. The ﬁrst closure axiom holds. for that matter. for some (r. 555) The set of all real valued and increasing functions of a real variable is not a real linear space. x 7→ 0 indeed converges to 0 at +∞ (and. 7 γ 0 = γ 1 = 0). 11 (p.9 n.
Let P : x 7→ Then P0 P 00 .5. 14 (p. . : x 7→ n!p0 x + (n − 1)!p1 : x 7→ n!p0 P 0 (0) = pn−1 P 00 (0) = 2pn−2 . . with nonnegative integral over [0.5. and any two nonnegative real numbers α and β. . If the center is taken to be 0. For any two such functions f and g. P (j) . P (j) (0) = j!pn−j . 1]. . . This is immediate. since it coincides with the set of all real polynomials of degree less than or equal to n. . but not a linear space. . . the issue is really easy: every polynomial is the Taylor polynomial centered at 0 of itself (considered. which is already known to be a real linear space. as it is. 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 .11 n. as a real function of a real variable). 555) First solution. . P (n−i)! : x 7→ n−j (n−i−j)! pi xn−j−i i=0 . so that j! xj pi xn−i . 1]. 555) The set of all real valued functions which are deﬁned and integrable on [0. if one thinks at the motivation for the deﬁnition of Taylor polynomials: best ndegree polynomial approximation of a given function. 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. 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. . . then 15.10 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. At any rate. 16 (p. if one takes the standard formula as a deﬁnition Taylorn (f ) at 0 ≡ x 7→ here are the computations. . is a convex cone.Exercises 99 15. 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. (αf ) (x) dx > 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
and that for every real numbers y and z (0.8 Bases and dimension 15. 3k) of r. and for any two real numbers α and β. 1)}. (0. 3x − z). y. 1. 0) . 2h.Subspaces of a linear space 103 two elements (h.1 Exercises n. z) 7→ (2x − y. 1 (p.20 n. 15. y. z)0 = 0 is the kernel of the linear function R3 → R.9.7 Dependent and independent sets in a linear space 15.5. 28 (p. namely span {a.5. (x.6 Subspaces of a linear space 15. 1) .9 15. 1. It is clear that the two vectors are linearly independent. 15. 3h) and (k. b}. 3) belongs to r. r is the kernel of the linear function R3 → R2 . y. 555) The set of solutions of the linear homogenous system of equations A (x. 0) + z (0. z) 7→ A (x. 0. its dimension is 2. 2h. y. z)0 and hence a real linear space. the linear combination α (h. 0. 27 (p. A standard basis for it is {(0.22 above). y.21 n. (x. 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. 3k) = (αh + βk) (1. S1 is the coordinate Y Zplane. 555) The subset of Rn of all the linear combinations of two given vectors a and b is a vector subspace of Rn . z) = y (0. 15. 2k. 2. 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. 2k. 3h) + β (k.
z) = x (1. −1. 1. 0. −1. its dimension is 2.6 n.3 n. 0) . 0) and (1.9. 0. 1)).9.104 Linear spaces 15. its dimension is 2. 1. It is clear that the two vectors are linearly independent.9. its dimension is 1. its dimension is 2. 1) 15. and that for every three real numbers x. 1) does not. 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. 1. −1. 0) .9. 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). 15. 3 (p. z) = x (1. 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). 0) + z (0. y and z such that x = y (x. −1. y. (1. It is clear that the two vectors are linearly independent. 0. y. 0) and (1. 0.4 n. 0) . 1)). 1)}.9. −1. 0) + z (0. 2 (p. (0. 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. 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). 0. 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. A basis for it is {(1. 7 (p. For example.2 n. but their sum (2. 15. 5 (p. It is clear that the two vectors are linearly independent. (0. 0) + z (0. y and z such that x + y = 0 (x. but their sum .7 n. 1)}. A basis for it is {(1. and that for every three real numbers x.9. 1) 15.5 n. 1. −1. 1)}. For example. −1. 1. A basis for it is {(1. y and z such that x + y + z = 0 (x. A basis for it is {d}. 0) both belong to S7 . 1. (1. (0. 6 (p. 4 (p. y. 1) both belong to S6 . and that for every three real numbers x. z) = x (1.
. 9 (p.558). X P : t 7→ p0 + ph th (15. 0) and Q ≡ (0. y.11 n. since ∀ (x. every polynomial belonging to S11 is a linear combination of B. 1.9. 15.1) h∈n then P belongs to S11 if and only if p0 = 0. −1) 15. p. 560) The set S10 of all elements (x. y. and taking value 0 at 0 is a linear subspace of the set of all polynomials of degree not exceeding n. 560) See exercise 26. A basis for S11 is ª © B ≡ t 7→ th h∈n Indeed. It is clear that any linear combination of polynomials in S11 takes value 0 at 0. 0. a linear combination of B is the null polynomial. 560) The set S11 of all polynomials of degree not exceeding n. S7 is not an aﬃne subspace of R3 . Moreover. It follows that dim S11 = n.9. 10 (p. 0) or their semisum (1. and α (0.Exercises 105 (2. w) ∈ R3 . 0. If (x. 0)) For example. −1)}. p. 1.555. 0. v. then (x. a subspace of R3 of dimension 1. 0) do not. and hence a linear space. 1.9. 11 (p. α (1. 0) . z) is any element of S10 . 1. y.9 n. 0.10 n. if and only if all its coeﬃcients are null. z) ∈ R3 . S7 is a cone. S8 is an aﬃne subspace of R3 . and it is even closed with respect to the external product (multiplication by arbitrary real numbers). 0) do not belong to S8 . 8 (p. for every α 6= 1. and it is not even convex. y. by the general principle of polynomial identity (which is more or less explicitly proved in example 6. ∀α ∈ R (x + y = 1) ∧ (u + v = 1) ⇒ [(1 − α) x + αu] + [(1 − α) y + αv] = 1 15. A base for it is {(0. 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. ∀ (u. z) = y (0.9. However.8 n. If P is any polynomial of degree not exceeding n. 15. z) of R3 such that x+y+z = 0 x−y−z = 0 is a line containing the origin.
(t 7→ 1) h∈n−1 That B is a basis can be seen as in the previous exercise. then X R : t 7→ α1 − α1 t + αh+1 th+1 h∈n−1 . t 7→ t h∈n−1 Indeed. Indeed.9. with ﬁrst derivative taking value 0 at 0. 12 (p.13 n. and any two real numbers α and β. 14 (p. A polynomial P as in (15. 13 (p.1) belongs to S12 if and only if p1 = 0. with second derivative taking value 0 at 0. If P and Q are any two polynomials in S. with ﬁrst derivative taking value at 0 which is the opposite of the polynomial value at 0. (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. is a linear subspace of the set of all polynomials of degree not exceeding n. is a linear subspace of the set of all polynomials of degree not exceeding n. 560) The set S14 of all polynomials of degree not exceeding n. and α and β are any two real numbers. (t 7→ 1) . A basis for S is n¡ o ¢ B ≡ t 7→ th+2 h∈n−2 . is a linear subspace of the set of all polynomials of degree not exceeding n. for any two suach polynomials P and Q.14 n. A basis for S is n¡ o ¢ h+1 B ≡ t 7→ t .1) belongs to S13 if and only if p2 = 0. if α is the ntuple of coeﬃcients of a linear combination R of B.12 n. 15.9. Thus dim S13 = n. 15. This is proved exactly as in the previous exercise (just replace 0 with 00 ).106 Linear spaces 15. 560) The set S12 of all polynomials of degree not exceeding n. Thus dim S12 = n. 560) The set S13 of all polynomials of degree not exceeding n. (αP + βQ)0 (0) = αP 0 (0) + βQ0 (0) = α · 0 + β · 0 = 0 A polynomial P as in (15. (t 7→ t) That B is a basis can be seen as in the exercise 11.1) belongs to S14 if and only if p0 + p1 = 0. A polynomial P as in (15.9.
560) The set S16 of all polynomials of degree not exceeding n. p. the combination must be the trivial one. A polynomial P belongs to S15 if and only if X p0 = p0 + ph h∈n that is. if and only if X h∈n 2h ph = 0 .555. and hence that dim S = n. if and only if X h∈n ph = 0 A basis for S15 is and the dimension of S15 is n. It follows that B is a basis of S14 . 560) The set S15 of all polynomials of degree not exceeding n. 16 (p. This can be seen exactly as in exercise 3. 15 (p. A polynomial P belongs to S16 if and only if X p0 = p0 + 2h ph h∈n n o ¡ ¢ B ≡ (t 7→ 1) . 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. 15. and hence a linear space.16 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. 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 . p.9. 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.15 n.Exercises 107 By the general principle of polynomial identity.9. 15. t 7→ (1 − t) th h∈n−1 that is. If P is any polynomial such that p0 + p1 = 0. This can be seen exactly as in exercise 3. and hence a linear space. P belongs to span B.555.
∃a ≡ (αi )i∈n ∈ F n . Conversely. it follows that S = lin S. t 7→ (2 − t) th h∈n−1 and the dimension of S16 is n. then by point c above lin S = S and lin T = T . v = X i∈n αi ui ) hence if T is a subspace. Then S ∩ T = {(1. v. 560) f g h 0 : : : : R → R. then of course S is a subspace of V because lin S is so. and it contains S. 1)}. since S ⊆ lin S always. (c) If S = lin S. (e) If S and T are subspaces of V . 0) . S ∩ T is a subspace of V . S ≡ {(1. 560) (b) It has alredy been shown in the notes that \ lin S ≡ W W subspace of V S⊆W = ( v : ∃n ∈ N. (0.108 Linear spaces A basis for S16 is n o ¡ ¢ B ≡ (t 7→ 1) .18 (a) Let n.9. 15. x 7→ 1 x 7→ eax x 7→ ebx x 7→ 0 and let (u. R → R. 0) . 23 (p. and hence S ∩ T = lin S ∩ lin T Since the intersection of any family of subspaces is a subspace. 0)} L (S) = L (T ) = R2 L (S ∩ T ) = {(x. T ≡ {(1. if S is a subspace of V . y) : y = 0} 15. (g) Let V ≡ R2 . and hence lin S ⊆ S. R → R. 22 (p. R → R. w) ∈ R3 be such that uf + vg + wh = 0 . then S is one among the subspaces appearing in the deﬁnition of lin S.17 n. (1.9. then T contains lin S. 1)}. ∃u ≡ (ui )i∈n ∈ V n .
and let f : X → R be a nonnull constant function.2) has only the trivial solution. let (u. v) ∈ R2 ∼ (0. w) ≡ (1. if b = 0. If g is constant. if (f. then h = f . It immediately follows from the above lemma (and from the assumption a 6= b) that if either a = 0 or b = 0. For any function g : X → R. let Im f ≡ {yf } and Im g ≡ {yg }. g. at most one of the two numbers can be equal to zero. g (x) = − yf v and g is constant. the couple (f. h) is linearly dependent (it suﬃces to take (u. v. −1). g. h) is linearly independent. 0) be such that uf + vg is the null function. then dim lin {f. too. If a = 0. 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. 0) or (u. respectively). g. −1. If none of them is null. Then the function yg f − yf g is the null function. uyf + vg (x) = 0 Since f is nonnull. v. Then u ∀x ∈ X. Conversely. too. 0. h} = 2. g. and x = 2. system (15. and dim lin {f. the triple (f. Thus ∀x ∈ X. g) is linearly dependent. h} = 3. g) is linearly dependent if and only if g is constant. (b) The two functions f : x 7→ eax g : x 7→ xeax . This yields v=0⇒u=0 and hence v 6= 0. Proof. similarly. It is convenient to state and prove the following (very) simple Lemma 4 Let X be a set containing at least two distinct elements. x = 1. Thus if either a or b is null then (f.2) Since by assumption a 6= b.Exercises 109 In particular. for x = 0. w) ≡ (1. then g = f . yf 6= 0.
On the other hand. 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. ∀x ∈ R. (f) The two functions f : x 7→ sin x g : x 7→ cos x (15. g. h} = 2. and dim lin {f. h) is linearly independent. v. g. since [∀x ∈ R. (α + βx) = 0] ⇔ α=β=0 Notice that the argument holds even in the case a = 0. w) = (0. 0). x = 1. √ α 2 sin x + √ β 2 cos x = 0 2 2 α +β α +β m ∀x ∈ R. g} = 2. and x = −1. w) ∈ R3 be such that uf + vg + wh = 0 where f : x 7→ 1 g : x 7→ eax h : x 7→ xeax In particular. Thus dim lin {f. h} = 3. Indeed. g. (c) Arguing as in point a. (α + βx) eax = 0] ⇔ [∀x ∈ R. yielding (u. v. 0.3) are linearly independent. αex + βxeax = 0] ⇔ [∀x ∈ R. for x = 0. let (u. α sin x + β cos x = 0 m ∀x ∈ R. The triple (f.110 Linear spaces are linearly independent. for every two real numbers α and β which are not both null. sin (γ + x) = 0 . if a = 0 then f = g and h = IdR . so that dim lin {f.
the last condition cannot hold. Thus dim lin {x 7→ sin x. Euclidean spaces. Euclidean spaces. h} = 2. h) is linearly dependent. x → cos x} = 2. g.Inner products. dim lin {f.11 Orthogonality in a euclidean space .10 Inner products. 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. Norms 15. g. 15. since g : x 7→ sin2 x f − g + 2h = 0 By the lemma discussed at point a. Norms 111 where α α2 + β 2 γ ≡ sign β arccos p since the sin function is not identically null.
12.112 Linear spaces 15. u2 i = hu2 . 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 15.1 Exercises n. 11 (p.2 n. u3 i = hu3 . 567) hu1 .12. 9 (p.12 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 .
since for each n ∈ N lim e−x xn = 0 x→+∞ It follows that ∀n ∈ N. Then the product f g is a real polynomial of degree m + n. j∈n i+j=k . 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. Thus f g = α0 β 0 + k∈m+n X X αi β j tk i∈m. 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. 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. for each n ∈ N.Exercises 113 and. of degree m and n respectively.
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.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. 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. 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 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 . j∈n i+j=k 0 = α0 β 0 + k∈m+n X X αi β j k! i∈m. j∈n i+j=k +∞ e−t tk dt = α0 β 0 I0 + k∈m+n X X αi β j Ik i∈m.
2) 3 Every vector belonging to the line π ∩ perp {v} must have the form (x. x2 . lin {x1 . x3 } = 2. The GramSchmidt process 115 (d) If f : t 7→ t + 1 then f g : t 7→ αt2 + (α + β) t + β hf. β) = (2γ. −3γ) (γ ∈ R) g : t 7→ αt + β that is. x) . gi = αI2 + (α + β) I1 + βI0 = 3α + 2β and hf. gi = 0 ⇔ (α. 1 (p. The GramSchmidt process 15. Since no two of them are parallel. A unit vector in π is given by v≡ 1 (2.Construction of orthogonal sets.13 Construction of orthogonal sets. projections 15. x3 } = π and dim lin {x1 . 1.15 Best approximation of elements in a euclidean space by elements in a ﬁnitedimensional subspace 15.14 Orthogonal complements.16. 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 15.1 n. x2 . it is seen that they all belong to the plane π of equation x−z = 0. −4x.16 Exercises 15.
−1. 1. x4 } = H(1.1. x2 . 0) y1 ≡ kx1 k 2 y2 (0. 0) − x2 − hx2 . x4 } = 3. 576) (a) First solution. (b) The answer is identical to the one just given for case a. x3 . z. 0. 1. 0) 6 1 + 1 +1 4 4 √ √ 2 2 2 2 √ √ 2 2 2 2 (1. 1. −3. 1. y1 i y1 ≡ =° ° kx2 − hx2 . 1) 6 The required orthonormal basis for lin {x1 . 1. 0 6 = q2 2 = (−1. x4 } . 0. x2 . 1) 2 x3 − hx3 . x3 . 1. 0) − ¢ ¡ 1 1 √ − . x2 } / 0 = x1 − x2 + x3 − x4 Thus dim W ≡ dim lin {x1 . 3. −1. y1 i y1 k °(0. and which is orthogonal to v. 1. 1. It is easily checked that x2 ∈ lin {x1 } / x3 ∈ lin {x1 . More easily. y. 1 3 = q3 = (1. 15.116 Linear spaces √ with norm 3 2 x. 1. 1. The following vectors form an orthonormal one: √ x1 2 = (1. 1. 0)° y3 Second solution. 1) − 0 − 2 66 66 (−1. y1 i y1 − − hx3 . with (1. w}. 2 (p. 0) 2 √ 2 v≡ (0. 2. x2 . 1. −4. 0)° √ √ is the hyperplane of R4 through the origin. Thus a second unit vector which together with v spans π. t) ∈ R4 : x − y + z − t = 0 √ 2 u≡ (1. 0. 2. 1. to directly exhibit two mutually orthogonal unit vectors in lin {x1 .0 ≡ (x. 0. x3 } is {v. y2 i y2 = ≡ kx3 − hx3 . 0.2 n. is √ 2 w≡ (1. it suﬃces to realize that © ª lin {x1 . 0. . 3) 6 1 + 1 + 1 +1 9 9 9 (0.−1. −1) as normal vector. and three vectors are required for any basis of W . x3 . 1. y1 i y1 − − hx3 . y2 i y2 k ¡1 1 1 ¢ √ . 1. 0. x2 .−1). 0) ° ° √ √ ° ° °(0. 1. 1) − 0 − 2 66 66 (−1. 0) ° ° (1. 2. 1.16.
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 .
y) + β (u. x + y). for every β ∈ R. lies on r.1 Exercises n. since. αx + βu) α (y.4. y) . v)] = = = = T [(αx + βu. 1 (p. which is (x + y. αy + βv)] (αy + βv.Chapter 16 LINEAR TRANSFORMATIONS AND MATRICES 16. .2 Null space and range 16. 582) T is linear.1 Linear transformations 16. y) ∈ R2 . T (x. x) + β (v. y) − (x. 0)}. y) ∈ R2 . y) is orthogonal to r. for every α ∈ R. T (x. for every (x. v)] The null space of T is the trivial subspace {(0. v) ∈ R2 . and the midpoint of [(x. 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. u) αT [(x.4 16. T [α (x. y)]. the range of T is R2 .3 Nullity and rank 16. since for every (x. y)] + βT [(u. for every (u.
for every α ∈ R. x) + β (v. and hence rank T = nullity T = 1 . y) + β (u. for every α ∈ R. since for every (x. the range of T is R2 . 2 (p. for every (u. and hence rank T = 2 nullity T = 0 T is the orthogonal symmetry with respect to the Xaxis. y) ∈ R2 . 0) αT [(x. αx + βu) α (y.4. u) αT [(x. y) + β (u.3 n. αy + βv)] (αx + βu. 0)}. 582) T is linear. 16. for every α ∈ R. y) ∈ R2 . y) ∈ R2 . the range of T is the bisectrix of the I and III quadrant. u) αT [(x. and for every β ∈ R. 582) T is linear. 3 (p. T [α (x.2 n. x) + β (u. the range of T is the Xaxis. for every (u. v)] The null space of T is the Y axis.4. v)] The null space of T is the Y axis. v)] = = = = T [(αx + βu. since for every (x. 0) + β (u. αy + βv)] (αx + βu. y) + β (u. v)] = = = = T [(αx + βu. T [α (x. 0) α (x.122 Linear transformations and matrices 16.4 n. 582) T is linear . y)] + βT [(u. and for every β ∈ R. 16. v) ∈ R2 . y)] + βT [(u. αy + βv)] (αy + βv. and for every β ∈ R. and hence rank T = 1 nullity T = 1 T is the orthogonal projection on the Xaxis. αx + βu) α (x. v) ∈ R2 .4. T [α (x. 4 (p. y)] + βT [(u. since for every (x. for every (u. v)] The null space of T is the trivial subspace {(0. v) ∈ R2 . v)] = = = = T [(αx + βu.
for every α ∈ R. e) T [(0. v)] = T [(αx + βu. 1)] = (e.g. v)] = = αT [(x. v) ∈ R2 . 16. ey ) + β (eu . y)] + βT [(u. T [(0.6 n. y) ∈ R2 . when x = y = 0 and u = v = α = β = 1. ev ) = (αex + βeu . αy + βv + 1) α (x + 1. for every (u. y)] + βT [(u.. but not linear. indeed. αy + βv)] ¡ ¢ = eαx+βu .4.4.7 n. 0 T is not a linear function. y + 1) + β (u + 1. for every β ∈ R. for every (u. and for every β ∈ R. 1 + e) 16. v)] = = 16. 7 (p. v)] = α (ex . (ey )α · (ev )β αT [(x. y) + β (u. 1) α (x. for every (x. α + β) T is aﬃne. 0) = x2 + 2xy + u2 .8 n.. 582) T is not linear. y) ∈ R2 . 6 (p. αy + βv)] (αx + βu + 1. 0)] + T [(1. y) ∈ R2 . αy + βv)] (αx + βu. 582) T [(αx + βu. v)] = = αT [(x. for every (u. for every (x. but it is not linear. for every (x. 582) ¢ ¡ T (x + u. eαy+βv i h = (ex )α · (eu )β . 0) = x2 + u2 . αey + βev ) so that. 1) (αx + βu. y) + β (u. v)] = = T [(αx + βu.4.Exercises 123 16. 582) T is not an aﬃne function. for x and u both diﬀerent from 0. 5 (p. indeed. 8 (p.4. 1)] = (1 + e. Indeed. T [α (x. e. 0) + (1. v + 1) (αx + βu + α + β. T [α (x.g. 0) + T (u. and for every β ∈ R. 0 ¡ ¢ T (x. y) + β (u. αy + βv + α + β) . v) ∈ R2 . v) ∈ R2 . e. 1) + β (u. for every α ∈ R. y)] + βT [(u. for every α ∈ R. since.5 n. T [α (x.
y) + β (u. v)] = = = = T [(αx + βu. u + v) αT [(x. y) + β (u. for every α ∈ R. u + v) αT [(x. 582) T is linear. v) ∈ R2 .124 Linear transformations and matrices 16. and for every β ∈ R. x + y) + β (u − v.10 n. αy + βv)] (2 (αx + βu) − (αy + βv) . since for every (x. 0)}. 0)}. y)] + βT [(u. for every (u. the range of T is R2 . 582) T is linear. αy + βv)] (αx + βu − αy − βv. 16. the range of T is R2 . y)] + βT [(u. v)] = = = = T [(αx + βu.4. x + y) + β (2u − v. 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. and for every β ∈ R. y) ∈ R2 . y) ∈ R2 . since for every (x. (αx + βu) + (αy + βv)) α (2x − y. 9 (p. 10 (p. v) ∈ R2 . T [α (x. v)] The null space of T is the trivial subspace {(0. for every α ∈ R. 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 . αx + βu + αy + βv) α (x − y. for every (u. T [α (x. v)] The null space of T is the trivial subspace {(0.4.9 n.
and hence . for every α ∈ R. 0) α (x. v. 16. since for every (x. 582) T is linear (as every projection on a coordinate hyperplane). 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. αz + βw)] (αz + βw. v. 17 (p. x) + β (w. αz + βw)] (αx + βu. 0. for every β ∈ R. y. w) ∈ R3 . v. αy + βv.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. for every α ∈ R. for every β ∈ R. 0)}. y.11 n. w)] nark T = 1 The null space of T is the Zaxis. z) + β (u. 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. αy + βv. v. z)] + βT [(u. T [α (x. u) αT [(x. αy + βv. 582) T is linear. z) ∈ R3 . 0) + β (u. 16 (p. for every (u. v. the range of T is R3 .4. the range of T is the XY plane. y.4. v. y. w)] = = = = rank T = 2 T [(αx + βu. v. z) ∈ R3 . w) ∈ R3 . w)] = = = = rank T = 3 T [(αx + βu. y. v. Re z} = . y. since for every (x.12 n. αx + βu) α (z. y. and hence 16. for every (u. z) + β (u. αy + βv. T [α (x. y. z)] + βT [(u. 0) αT [(x.
and hence rank T = 2 nark T = 1 16. −)orthant and of the (−. D1 be the linear space of all real functions of a real variable which are deﬁned and everywhere diﬀerentiable on (−1.1) . for every β ∈ R. w)] = = = = = T [(αx + βu. 1) . 1) . 25 (p. w)] The null space of T is the axis of central symmetry of the (+. Indeed. −.4. 23 (p. 0) ∪ (0. +)orthant. z) + β (u. 0. 1). 0. T [α (x. v. v. z) ∈ R3 . αy + βv. since for every (x. ¡ ¢ 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. 1) or. β (u + v)] α (x. v. w) αT [(x. f 0 (x) = 0} By an important though sometimes overlooked theorem in calculus. y. f 7→ (x 7→ xf 0 (x)) then T is a linear operator. every function f which is diﬀerentiable in (−1. z) + β (u. αx + βu + αy + βv) [α (x + z) . ker T = {f ∈ D1 : ∀x ∈ (−1. for every α ∈ R. in which case x→0 f 0 (0) = lim f 0 (x) x→0 . xf 0 (x) = 0} = {f ∈ D1 : ∀x ∈ (−1. +. 0) ∪ (0. 582) Let D1 (−1. is diﬀerentiable in 0 as well. w) ∈ R3 . more shortly. y.13 n.126 Linear transformations and matrices 16. 582) T is linear. for every (u. 0. v. α (x + y)] + [β (u + w) . αz + βw)] (αx + βu + αz + βw. 1) and continuous in 0.4. y. provided lim f 0 (x) exists and is ﬁnite. If T : D1 → R(−1. of parametric equations x=t y = −t z = −t the range of T is the XZplane. y. z)] + βT [(u.14 n.
If where P and Q are real functions of a real variable which are continuous on R. and let v be the solution to equation (16. 16. and by direct inspection it is seen that y (0) = y0 and 0 0 y 0 (0) = y0 . v 0 (0)) = (0. v} . in such a case. 1) .1) such that 0 (v (0) . Hence the function µ ¶ y (0) 2 ϕ : ker L → R . ker L = span {u. Since T (D1 ) contains. 1) of the polynomials with vanishing zerodegree monomial.g. from βv (x) = 0 for each x ∈ R and v 0 (0) = 1 it is easily deduced that β = 0 as well. ∃ϑx ∈ (0. Then for each (y0 .1) is a solution to equation (16. since ker L is a linear subspace of D2 . all the restrictions to (−1. y 00 (x) + P (x) y 0 (x) + Q (x) y (x) = 0 (16. Moreover. 1). and it is already known that the linear space of all polynomials has an inﬁnite basis. for each (y0 . since u (0) = 1 and v (0) = 0. If f belongs to ker T .1). 1). that L is a linear operator. x) f (x) = f (0) + xf 0 (ϑx ) and hence f is constant on (−1. It follows that a basis of ker T is given by the constant function (x 7→ 1). ∀x ∈ (−1.1) such 0 that y (0) = y0 and y 0 (0) = y0 . 0). y0 )). let u be the solution to equation (16. 27 (p. y 7→ y 00 + P y 0 + Q ∀x ∈ R. e. ϕ−1 ((y0 . indeed.15 n. u and v are linearly independent. by the classical theorem due to Lagrange. the dimension of T (D1 ) is inﬁnite.4. In other words. αu (x) + βv (x) = 0 for each x ∈ R can only hold (by evaluating at x = 0) if α = 0. 1)). and nullity T = 1.1) such that (u (0) . the function 0 y : x 7→ y0 u (x) + y0 v (x) ker T = {f ∈ D1 : f 0 = 0} L : D2 → RR . it has been shown in the solution to exercise 17. y0 ) ∈ R2 there exists a unique solution to equation (16.555. p.Exercises 127 Thus (here 0 is the identically null function deﬁned on (−1. y0 ) ∈ R2 . 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. 582) Let D2 be the linear space of all real functions of a real variable which are deﬁned and everywhere diﬀerentiable on R. and Finally. u0 (0)) = (1.. y 7→ y 0 (0) is injective and surjective. Thus nullity L = 2.
w + 1) . z) = T (x .3 n. w − u − v) 16. y. z) = (x0 . z ) ⇔ 3z = 3z 0 ³ v w´ T −1 (u.1 n. 15 (p.8. w) = (u − 1. v − 1.8. z) = T (x . y 0 .5 Algebraic operations on linear transformations 16.6 Inverses 16.128 Linear transformations and matrices 16.8 Exercises 16. y. z 0 ) z − 1 = z0 − 1 T −1 (u. 2 3 16.2 n. y. y . 589) T is injective (or one to one). 17 (p.8. z ) ⇔ y + 1 = y 0 + 1 ⇔ (x.7 Onetoone linear transformations 16. y . v. 16 (p. since x + 1 = x0 + 1 0 0 0 T (x. z) = (x0 . z) = T (x . since x = x0 0 0 0 2y = 2y 0 ⇔ (x. since x = x0 y = y0 T (x. 589) T is injective. v. 589) T is injective. z) = (x0 . y . w) = u. z ) ⇔ ⇔ (x. w) = (u. z 0 ) x + y + z = x0 + y 0 + z 0 0 0 0 T −1 (u. v. y 0 . y 0 . . y. z 0 ) T (x. y. y. v.
16. and ker T D is the set of all constant polynomials.. Im T D is equal to W . 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 .Linear transformations with prescribed values 129 16. 27 (p.12. 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.11 Construction of a matrix representation in diagonal form 16. 3 (p. 596) (a) Since T (i) = i + j and T (j) = 2i − j. £ ¤ 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.1 n.9 Linear transformations with prescribed values 16.10 Matrix representations of linear transformations 16.12 Exercises 16. zerodegree polynomials.e.8.4 Let n. i.
r. j} into coordinates w. e2 } is £ ¤ B ≡ T (e1 ) T (e2 ) (c) First solution (matrix for T ).r. e2 } to basis {i. 1) with respect to the basis {i. e2 } (that is. to the basis {i. −1) and (3. Since. β) and (γ. 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 ). e2 } into coordinates w. to basis {e1 . 2) transformation according to T as expressed by the matrix representing T w. 0) and (0.r. j} to basis {e1 . 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. If e1 = i − j and e2 = 3i + j. j} (that is. e2 } can be described as the combined eﬀect of the following three actions: 1) coordinate change from coordinates w. 3) coordinate change from coordinates w. j}.130 Linear transformations and matrices (b) The matrix of T with respect to the basis {i. multiplication by matrix P ). the matrix · ¸ £ ¤ 1 3 P ≡ e1 e2 = −1 1 provided T (e1 ) and T (e2 ) are meant as coordinate vectors (α. e2 }. e2 }. with respect to the basis {i. j}. The operation of the matrix B representing T with respect to the basis {e1 . to the basis {i. to the basis {i. 1) of e1 and e2 with respect to the basis {e1 . and P −1 is the matrix of coordinate change from basis {i. The matrix B representing T with respect to the basis {e1 . hence P is the matrix of coordinate change from basis {e1 .r. multiplication by A). δ) with respect to the basis {e1 . j} (that is. e2 } into their coordinates (1. 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 . multiplication by P −1 ).r. on the other hand. to basis {e1 .
β) and (γ. respectively) αe1 + βe2 = −i + 2j γe1 + δe2 = 5i + 2j that is. P −1 DP = D for every scalar diagonal matrix D). 596) T : (x.Exercises 131 it suﬃces to ﬁnd the {e1 . e2 }coordinates (α.12. δ) = 1 (−1. 596) ≡ µ 4 0 0 4 ¶ 7→ (−x. so that 4 4 1 B= 7 −7 −1 1 7 Continuation (matrix for T 2 ). δ) which correspond to the {i. y) (reﬂection w. Since T 2 is represented by a scalar diagonal matrix with respect to the initially given basis. j}coordinates (−1. 5 (p. it is represented by the same matrix with respect to every basis (indeed. 16. 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. 2) and (5. 2y) T (i + 2j + 3k) = T (k) + T (j + k) + T (i + j + k) = (2i + 3j + 5k) + i + (j − k) = 3i + 4j + 4k . ½ α + 3β = −1 −α + β = 2 · ½ γ + 3δ = 5 −γ + δ = 2 ¸ The (unique) solutions are (α. β) = 1 (−7.r. since scalar diagonal matrices commute with every matrix of the same order.2 n. δ). 7).3 a) n. y) 7→ (−2x. β) and (γ. 2).12. 4 (p. Thus we want to solve the two equation systems (in the unknowns (α. 1) and (γ.
Indeed. 0)}. but the ﬁrst two need to be computed. j. and rank T is 3. 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.r. and hence (by substitution in I and II) y = z = 0. It follows that the null space of T is the trivial subspace {(0. Thus the range space of T is R3 .4 (a) n. j. to {i. k} of the image vectors T (i). −2) Since {T (j) . 597) T (4i − j + k) = 4T (i) − T (j) + T (k) = (0.132 Linear transformations and matrices The three image vectors T (k). T (j + k). b) The matrix of T with respect to the basis {i. T (j). T (k). T (k)} is a linearly independent set. The last one is given in the problem statement. T (i + j + k) form a linearly independent triple. ker T = {0} (b) µ 0 1 1 0 1 −1 ¶ rank T = 2 AT = .12. 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. 7 (p. and nark T is 0. k} is obtained by aligning as columns the coordinates w. 0.
the matrix · ¸ £ ¤ 1 1 P ≡ w1 w2 = 1 2 µ 0 1 −1 2 0 0 3 2 ¶ . k} rank T = 2 nullity T = 2 − 2 = 0 © ª (b) The matrix of T with respect to the bases i.12. −1)}. ¡ ¢ ¡¢ 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. j. 597) (a) I shall distinguish the canonical unit vectors of R2 and R3 by marking the former ¡¢ with an underbar. w2 } = {(1. 1) . Since T (i) = i + k and T j = −i + k. i} and C = {w1 . 8 (p. If w1 = i + j and w2 = i + 2j. ¡ ¢ ¡¢ T 2i − 3j = 2T (i) − 3T j = 2 (i + k) − 3 (i − k) = −i + 5k For any two real numbers x and y.C = 0 1 0 16. j and {i. k. (1. k} is 1 1 ¡¢ ¤ £ A ≡ T (i) T j = 0 0 1 −1 (c) First solution.Exercises 133 (c) Let C = (w1 . Then µ ¶ 1 0 0 (AT )B. 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.
j to basis {w1 . k} can be described as the combined eﬀect of the following two actions: 1) coordinate ª © change from coordinates w. © ª and P −1 is the matrix of coordinate change from basis i.r.r. multiplication by matrix P ). γ 11 0 C = 0 γ 22 0 0 T (u1 ) = γ 11 v1 T (u2 ) = γ 22 v2 . 2) transformation according to T as expressed by the © ª matrix representing T w. w2 }. to basis {w1 . j (that is. The operation of the matrix B of T with respect to the bases {w1 . Thus · ¸ 1 1 1 1 B = AP = 0 0 1 2 1 −1 2 3 = 0 0 0 −1 Second solution (matrix for T ). 1) and (1. u2 } and {v1 . 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 . to the basis i. j . w2 } into their coordinates (1. 2) with respect to the©basis © ª ª i. j . j and {i. to the bases i.r. w2 } and {i. w2 } into coordinates w. v2 . multiplication by A). v3 } is diagonal. j. w2 } and {i. 1) of w1 and w2 with respect to the basis {w1 . that is. hence P is the matrix of coordinate change from basis {w1 . j.r. k} (that is. 0) and (0. w2 } to basis i.134 Linear transformations and matrices transforms the (canonical) coordinates (1. The matrix B representing T with respect to the bases {w1 . j.
independent 1 0 C = 0 1 0 0 thereby obtaining 16.6 n. v2 .Exercises 135 There are indeed very many ways to achieve that.12. 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. v3 } is lin. id · sin. id · sin. cos. cos. id · cos} with respect to the basis {sin. 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 . 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 . 16 (p.
This action might not be possible to undo. Are you sure you want to continue?