The present book is based on lectures given by the author over a number of years to students of various engineering and
269 100 21MB
English Pages 816 Year 1972
Table of contents :
Front Cover
Title Page
Preface 5
Contents
Introduction 19
1. The Subject of Mathematics 19
2. The Importance of Mathematics and Mathematical Education 20
3. Abstractness 20
4. Characteristic Features of Higher Mathematics 22
5. Mathematics in the Soviet Union 23
CHAPTER I. VARIABLES AND FUNCTIONS 25
§ 1. Quantities 25
1. Concept of a Quantity 25
2. Dimensions of Quantities 25
3. Constants and Variables 26
4. Number Scale. Slide Rule 27
5. Characteristics of Variables 29
§ 2. Approximate Values of Quantities 32
6. The Notion of an Approximate Value 32
7. Errors 32
8. Writing Approximate Numbers 33
9. Addition and Subtraction of Approximate Numbers 34
10. Multiplication and Division of Approximate Numbers Remarks 36
§ 3. Functions and Graphs 39
11. Functional Relation 39
12. Notation 40
13. Methods of Representing Functions 42
14. Graphs of Functions 45
15. The Domain of Definition of a Function 47
16. Characteristics of Behaviour of Functions 48
17. Algebraic Classification of Functions 51
18. Elementary Functions 53
19. Transforming Graphs 54
20. Implicit Functions 56
21. Inverse Functions 58
§ 4. Review of Basic Functions 60
22. Linear Function 60
23. Quadratic Function 62
24. Power Function 63
25. LinearFractional Function 66
26. Logarithmic Function 68
27. Exponential Function 69
28. Hyperbolic Functions 70
29. Trigonometric Functions 72
30. Empirical Formulas 75
CHAPTER II. PLANE ANALYTIC GEOMETRY 78
§ 1. Plane Coordinates 78
1. Cartesian Coordinates 78
2. Some Simple Problems Concerning Cartesian Coordinates 79
3. Polar Coordinates 81
§ 2. Curves in Plane 82
4. Equation of a Curve in Cartesian Coordinates 82
5. Equation of a Curve in Polar Coordinates 84
6. Parametric Representation of Curves and Functions 87
7. Algebraic Curves 90
8. Singular Cases 92
§ 3. FirstOrder and SecondOrder Algebraic Curves 94
9. Curves of the First Order 94
10. Ellipse 96
11. Hyperbola 99
12. Relationship Between Ellipse, Hyperbola and Parabola 102
13. General Equation of a Curve of the Second Order 105
CHAPTER III. LIMIT. CONTINUITY 109
§ 1. Infinitesimal and Infinitely Large Variables 109
1. Infinitesimal Variables 109
2. Properties of Infinitesimals 111
3. Infinitely Large Variables 112
§ 2. Limits 113
4. Definition 113
5. Properties of Limits 115
6. Sum of a Numerical Series 117
§ 3. Comparison of Variables 121
7. Comparison of Infinitesimals 121
8. Properties of Equivalent Infinitesimals 122
9. Important Examples 122
10. Orders of Smallness 124
11. Comparison of Infinitely Large Variables 125
§ 4. Continuous and Discontinuous Functions 125
12. Definition of a Continuous Function 125
13. Points of Discontinuity 126
14. Properties of Continuous Functions 129
15. Some Applications 131
CHAPTER IV. DERIVATIVES, DIFFERENTIALS, INVESTIGATION OF THE BEHAVIOUR OF FUNCTIONS 134
§ 1. Derivative 134
1. Some Problems Leading to the Concept of a Derivative 134
2. Definition of Derivative 136
3. Geometrical Meaning of Derivative 137
4. Basic Properties of Derivatives 139
5. Derivatives of Basic Elementary Functions 142
6. Determining Tangent in Polar Coordinates 146
§ 2. Differential 148
7. Physical Examples 148
8. Definition of Differential and Its Connection with Increment 149
9. Properties of Differential 152
10. Application of Differentials to Approximate Calculations 153
§ 3. Derivatives and Differentials of Higher Orders 155
11. Derivatives of Higher Orders 155
12. HigherOrder Differentials 156
§ 4. L'Hospital's Rule 158
13. Indeterminate Forms of the Type $\dfrac{0}{0}$ 158
14. Indeterminate Forms of tl1e Type $\dfrac{\infty}{\infty}$ 160
§ 5. Taylor's Formula and Series 161
15. Taylor's Formula 161
16. Taylor's Series 163
§ 6. Intervals of Monotonicity. Exrtremum 165
17. Sign of Derivative 165
18. Points of Extremum 166
19. The Greatest and the Least Values of a Function 168
§ 7. Constructing Graphs of Functions 173
20. Intervals of Convexity of a Graph and Points of Inflection 173
21. Asymptotes of a Graph 174
22. General Scheme for Investigating a Function and Constructing Its Graph 175
CHAPTER V. APPROXIMATING ROOTS OF EQUATIONS. INTERPOLATION 179
§ 1. Approximating Roots of Equations 179
1. Introduction 179
2. CutandTry Method. Method of Chords. Method of Tangents 181
3. Iterative Method 185
4. Formula of Finite Increments 187
5*. Small Parameter Method 189
§ 2. Interpolation 191
6. Lagrange's Interpolation Formula 191
7. Finite Differences and Their Connection with Derivatives 192
8. Newton's Interpolation Formulas 196
9. Numerical Differentiation 198
CHAPTER VI. DETERMINANTS AND SYSTEMS OF LINEAR ALGEBRAIC EQUATIONS 200
§ 1. Determinants 200
1. Definition 200
2. Properties 201
3. Expanding a Determinant in Minors of Its Row or Column 203
§ 2. Systems of Linear Algebraic Equations 206
4. Basic Case 206
5. Numerical Solution 208
6. Singular Case 209
CHAPTER VII. VECTORS 212
§ 1. Linear Operations on Vectors 212
1. Scalar and Vector Quantities 212
2. Addition of Vectors 213
3. Zero Vector and Subtraction of Vectors 215
4. Multiplying a Vector by a Scalar 215
5. Linear Combination of Vectors 216
§ 2. Scalar Product of Vectors 219
6. Projection of Vector on Axis 219
7. Scalar Product 220
8. Properties of Scalar Product 221
§ 3. Cartesian Coordinates in Space 222
9. Cartesian Coordinates in Space 222
10. Some Simple Problems Concerning Cartesian Coordinates 223
§ 4. Vector Product of Vectors 227
11. Orientation of Surface and Vector of an Area 227
12. Vector Product 228
13. Properties of Vector Products 230
14*. Pseudovectors 233
§ 5. Products of Three Vectors 235
15. Triple Scalar Product 235
16. Triple Vector Product 236
§ 6. Linear Spaces 237
17. Concept of Linear Space 237
18. Examples 239
20. Concept of Euclidean Space 244
21. Orthogonality 245
§ 7. Vector Functions of Scalar Argument. Curvature 248
22. Vector Variables 248
23. Vector Functions of Scalar Argument 248
24. Some Notions Related to the Second Derivative 251
25. Osculating Circle 252
26. Evolute and Evolvent 255
CHAPTER VIII. COMPLEX NUMBERS AND FUNCTIONS 259
§ 1. Complex Numbers 259
1. Complex Plane 259
2. Algebraic Operations on Complex Numbers 261
3. Conjugate Complex Numbers 263
4. Euler's Formula 264
5. Logarithms of Complex Numbers 266
§ 2. Complex Functions of a Real Argument 267
6. Definition and Properties 267
7*. Applications to Describing Oscillations 269
§ 3. The Concept of a Function of a Complex Variable 271
8. Factorization of a Polynomial 271
9*. Numerical Methods of Solving Algebraic Equations 273
10. Decomposition of a Rational Fraction into Partial Rational Fractions 277
11*. Some General Remarks on Functions of a Complex Variable 280
CHAPTER IX. FUNCTIONS OF SEVERAL VARIABLES 283
§ 1. Functions of Two Variables 283
1. Methods of Representing 283
2. Domain of Definition 286
3. Linear Function 287
4. Continuity and Discontinuity 288
5. Implicit Functions 291
§ 2. Functions of Arbitrary Number of Variables 291
6. Methods of Representing 291
7. Functions of Three Arguments 292
8. General Case 292
9. Concept of Field 293
§ 3. Partial Derivatives and Differentials of the First Order 294
10. Basic Definitions 294
11. Total Differential 296
12. Derivative of Composite Function 298
13. Derivative of Implicit Function 300
§ 4. Partial Derivatives and Differentials of Higher Orders 303
14. Definitions 303
15. Equality of Mixed Derivatives 304
16. Total Differentials of Higher Order 305
CHAPTER X. SOLID ANALYTIC GEOMETRY 307
§ 1. Space Coordinates 307
1. Coordinate Systems in Space 307
2*. Degrees of Freedom 309
4. Cylinders, Cones and Surfaces of Revolution 314
5. Curves In Space 316
6. Parametric Representation of Surfaces in Space. Parametric Representation of Functions of Several Variables 317
§ 3. Algebraic Surfaces of the First and of the Second Orders 319
7. Algebraic Surfaces of the First Order 319
8. Ellipsoids 322
9. Hyperboloids 324
10. Paraboloids 326
11. General Review of the Algebraic surfaces of the second order 327
CHAPTER XI. MATRICES AND THEIR APPLICATIONS 329
§ 1. Matrices 329
1. Definitions 329
2. Operations on Matrices 331
3. Inverse Matrix 333
4. Eigenvectors and Eigenvalues of a Matrix 335
7. Transformation of the Matrix of a Linear Mapping When the Basis Is Changed 347
8. The Matrix of a Mapping Relative to the Basis Consisting of Its Eigenvectors 350
9. Transforming Cartesian Basis 352
10. Symmetric Matrices 353
§ 3. Quadratic Forms 355
11. Quadratic Forms 355
12. Simplification of Equations of SecondOrder Curves and Surfaces 357
§ 4. NonLinear Mappings 358
13*. General Notions 358
14*. Non Linear Mapping in the Small 360
15*. Functional Relation Between Functions 362
CHAPTER XII. APPLICATIONS OF PARTIAL DERIVATIVES 365
§ 1. Scalar Field 365
1. Directional Derivative. Gradient 365
2. Level Surfaces 368
3. Implicit Functions of Two Independent Variables 370
4. Plane Fields 371
5. Envelope of OneParameter Family of Curves 372
§ 2. Extremum of a Function of Several Variables 374
6. Taylor's Formula for a Function of Several Variables 374
7. Extremum 375
8. The Method of Least Squares 380
9*. Curvature of Surfaces 381
10. Conditional Extremum 384
11. Extremum with Unilateral Constraints 388
12*. Numerical Solution of Systems of Equations 390
CHAPTER XIII. INDEFINITE INTEGRAL 393
§ 1. Elementary Methods of Integration 393
1. Basic Definitions 393
2. The Simplest Integrals 394
3. The Simplest Properties of an Indefinite Integral 397
4. Integration by Parts 399
5. Integration by Change of Variable (by Substitution) 402
§ 2. Standard Methods of Integration 404
6. Integration of Rational Functions 405
7. Integration of Irrational Functions Involving Linear and LinearFractional Expressions 407
8. Integration of Irrational Expressions Containing Quadratic Trinomials 408
9. Integrals of Binomial Differentials 411
lO. Integration of Functions Rationally Involving Trigonometric Functions 412
11. General Remarks 415
CHAPTER XIV. DEFINITE INTEGRAL 417
§ 1. Definition and Basic Properties 417
1. Examples Lending to the Concept of Definite Integral 417
3. Relationship Between Definite Integral and Indefinite Integral 426
4. Basic Properties of Definite Integral 433
5. Integrating Inequalities 436
§ 2. Applications of Definite Integral 436
6. Two Schemes of Application 436
7. Differential Equations with Variables Separable 437
8. Computing Areas of Plane Geometric Figures 443
9. The Arc Length of a Curve 445
10. Computing Volumes of Solids 447
11. Computing Area of Surface of Revolution 448
§ 3. Numerical Integration 448
12. General Remarks 448
13. Formulas of Numerical Integration 450
§ 4. Improper Integrals 454
14. Integrals with Infinite Limits of Integration 455
15. Basic Properties of Integrals with Infinite Limits 464
16. Other Types of Improper Integral 468
17*. Gamma Function 468
18*. Beta Function 471
19*. Principal Value of Divergent Integral 473
§ 5. Integrals Dependent on Parameters 474
20*. Proper Integrals 474
21*· Improper Integrals 476
§ 6. Line Integrals of Integration 478
22. Line Integrals of the First Type 482
23. Line Integrals of the Second Type 484
§ 7. The Concept of Generalized Function 488
25*. Delta Function 488
26*. Application to Constructing Influence Function 492
27*. Other Generalized Functions 495
CHAPTER XV. DIFFERENTIAL EQUATIONS 497
§ 1. General Notions 497
1. Examples 497
2. Basic Definitions 498
§ 2. FirstOrder Differential Equations 500
3. Geometric Meaning 500
4. Integrable Types of Equations 503
5*. Equation for Exponential Function 506
6. Integrating Exact Differential Equations 509
7. Singular Points and Singular Solutions 512
8. Equations Not Solved for the Derivative 516
9. Method of Integration by Means of Differentiation 517
§ 3. HigherOrder Equations and Systems of Differential Equations 519
10. HigherOrder Differential Equations 519
11*. Connection Between HigherOrder Equations and Systems of FirstOrder Equations 521
12*. Geometric Interpretation of System of FirstOrder Equations 522
13*. First Integrals 526
§ 4. Linear Equations of General Form 528
14. Homogeneous Linear Equations 528
15. NonHomogeneous Equations 530
16*. BoundaryValue Problems 535
§ 5. Linear Equations with Constant Coefficients 541
17. Homogeneous Equations 541
18. NonHomogeneous Equations with RightHand Sides of Special Form 545
19. Euler's Equations 548
20*. Operators and the Operator Method of Solving Differential Equations 549
§ 6. Systems of Linear Equations 553
21. Systems of Linear Equations 553
22*. Applications to Testing Lyapunov Stability of Equilibrium State 558
§ 7. Approximate and Numerical Methods of Solving Differential Equations 562
23. Iterative Method 562
24*. Application of Taylor's Series 564
25. Application of Power Series with Undetermined coefficients 565
26*. Bessel's Functions 566
27*. Small Parameter Method 569
28*. General Remarks on Dependence of Solutions on Parameters 572
29*. Methods of Minimizing Discrepancy 575
30*. Simplification Method 576
31. Euler's Method 578
32. RungeKutta Method 580
33. Adams Method 582
34. Milne's Method 583
CHAPTER XVI. Multiple Integrals 585
§ 1. Definition and Basic Properties of Multiple Integrals 585
1. Some Examples Leading to the Notion of a Multiple Integral 585
2. Definition of a Multiple Integral 586
3. Basic Properties of Multiple Integrals 587
4. Methods of Applying Multiple Integrals 589
5. Geometric Meaning of an Integral Over a Plane Region 591
§ 2. Two Types of Physical Quantities 592
6*. Basic Example. Mass and Its Density 592
7*. Quantities Distributed in Space 594
§ 3. Computing Multiple Integrals in Cartesian Coordinates 596
8. Integral Over Rectangle 596
9. Integral Over an Arbitrary Plane Region 599
10. Integral Over an Arbitrary Surface 602
11. Integral Over a ThreeDimensional Region 604
§ 4. Change of Variables in Multiple Integrals 605
12. Passing to Polar Coordinates in Plane 605
13. Passing to Cylindrical and Spherical Coordinates 606
14*. Curvilinear Coordinates in Plane 608
15*. Curvilinear Coordinates in Space 611
16*. Coordinates on a Surface 612
§ 5. Other Types of Multiple Integrals 615
17*. Improper Integrals 615
18*. Integrals Dependent on a Parameter 617
19*. Integrals with Respect to Measure. Generalized Functions 620
20*. Multiple Integrals of Higher Order 622
§ 6. Vector Field 626
21*. Vector Lines 626
22*. The Flux. of a Vector Through a Surface 627
23*. Divergence 629
24*. Expressing Divergence in Cartesian Coordinates 632
25. Line Integral and Circulation 634
26*. Rotation 634
27. Green's Formula. Stokes' Formula 638
28*. Expressing Differential Operations on Vector Fields in a Curvilinear Orthogonal Coordinate System 641
29*. General Formula for Transforming Integrals 642
CHAPTER XVII. SERIES 645
§ 1. Number Series 645
1. Positive Series 645
2. Series with Terms of Arbitrary Signs 650
3. Operations on Series 652
4*. Speed of Convergence of a Series 654
5. Series with Complex, Vector and Matrix Terms 658
6. Multiple Series 659
§ 2. Functional Series 661
7. Deviation of Functions 661
8. Convergence of a Functional Series 662
9. Properties of Functional Series 664
§ 3. Power Series 666
10. Interval of Convergence 666
11. Properties of Power Series 667
12. Algebraic Operations on Power Series 671
13. Power Series as a Taylor Series 675
14. Power Series with Complex Terms 676
15*. Bernoullian Numbers 677
16*. Applying Series to Solving Difference Equations 678
17*. Multiple Power Series 680
18*. Functions of Matrices 681
19*. Asymptotic Expansions 685
§ 4. Trigonometric Series 686
20. Orthogonality 686
21. Series in Orthogonal Functions 689
22. Fourier Series 690
23. Expanding a Periodic Function 695
24*. Example. Bessel's Functions as Fourier Coefficients 697
25. Speed of Convergence of a Fourier Series 698
26. Fourier Series in Complex Form 702
27*. Parseval Relation 704
28*. Hilbert Space 706
29*. Orthogonality with Weight Function 708
30*. Multiple Fourier Series 710
31*. Application to the Equation of Oscillations of a String 711
§ 5. Fourier Transformation 713
32*. Fourier Transform 713
33*. Properties of Fourier Transforms 717
34*. Application to Oscillations of Infinite String 719
CHAPTER XVIII. ELEMENTS OF THE THEORY OF PROBABILITY 721
§ 1. Random Events and Their Probabilities 721
1. Random Events 721
2. Probability 722
3. Basic Properties of Probabilities 725
4. Theorem of Multiplication of Probabilities 727
5. Theorem of Total Probability 729
6*. Formulas for the Probability of HyPotheses 730
7. Disregarding LowProbability Events 731
§ 2. Random Variables 732
8. Definitions 732
9. Examples of Discrete Random Variables 734
10. Examples of Continuous Random Variables 736
11. Joint Distribution of Several Random Variables 737
12. Functions of Random Variables 739
§ 3. Numerical Characteristics of Random Variables 741
13. The Mean Value 741
14. Properties of the Mean Value 742
15. Variance 744
16*. Correlation 746
17. Characteristic Functions 748
§ 4. Applications of the Normal Law 750
18. The Normal Law as the Limiting One 750
19. Confidence Interval 752
20. Data Processing 754
CHAPTER XIX. COMPUTERS 757
§ 1. Two Classes of Computers 757
1. Analogue Computers 758
2. Digital Computers 762
§ 2. Programming 764
4. Representing Numbers in a Computer 766
5. Instructions 769
6. Examples of Programming 772
Appendix. Equations of Mathematical Physics 780
1*. Derivation of Some Equations 780/750,Black,notBold,notItalic,open,FitPage 2*. Some Other Equations 783
3*. Initial and Boundary Conditions 784/754,Black,notBold,notItalic,open,FitPage § 2. Method of Separation of Variables 786
4*. Basic Example 786
5*. Some Other Problems 791
Bibliography 796
Name Index 798
Subject Index 8OO
List of Symbols 815
INTRODUCTORY MATHEMATICS FOR ENGINEERS Lectures In Higher Mathematics A. D. MYSKIS
M ir Publishers Moscow
A.
MLIIUKHC •
JIEKIJHH n o BBICfflEK MATEMATHKE
H3JtATEHbCTBO «HAyKA» MOCKBA
A. D. MYSKIS
INTRODUCTORY MATHEMATICS FOR ENGINEERS Lectures in Higher Mathematicf Translated from, the Russiuu by
V. M. YOLOSOY, D.Sc.
UDC 510 (022) = 20
First published 1972 Revised from the 1969 Russian edition
Ha aMAUUCKOM R3blKe
Preface
The present book is based on lectures given by the author over a number of years to students of various colleges studying engineering and physics. The book includes some optional material which can be skipped for the first reading. The corresponding items in the table of contents are marked by an asterisk. In designing this course the author tried to select the most impor tant mathematical facts and present them so that the reader could acquire the necessary mathematical conception and apply mathema tics to other branches of science. Therefore in most cases tlio author did not give rigorous formal proofs of the theorems and intentionally simplified their statements referring the reader to characteristic particular cases and obvious examples. The rigorousness of a proof often fails to be fruitful and therefore it is usually ignored in practical applications. Some purely mathematical stipulations are made in the book only in the cases when they help the reader to avoid mis conception in theory and application. Mathematical facts and objects which can be regarded as exceptional from the point of view of applied science are not even mentioned in the book. (For instance, when we speak about “all functions” we do not include the functions which are not Lcbesgue measurable and even such functions as the everywhere discontinuous Dirichlet function and the like.) We tried to demonstrate the meaning of the basic mathe matical concepts and to give a convincing explanation of the most important mathematical facts on the basis of intuitive notions. It is the author’s belief that in applied mathematics an explanation of this kind should be regarded as a proof. Such an approach is characteristic of applied mathematics whose main purpose is to provide an adequate qualitative description of a phenomenon and obtain the numerical solution of the corresponding problem in the most economical manner without exerting unnecessary effort. This approach essentially differs from that of pure mathematics whose cornerstone is the logical consistency of all the considerations based
6
INTRODUCTORY
MATHEMATICS FOR
ENG INEERS
only on the concepts which have an exhaustive logical foundation. The author is sure that it is the aspects of applied mathematics that must determine the character of mathematical education of an engineer and physicist. (But of course a teacher of mathematics should have a goodcommand both of pure and applied mathematics.) These ideas of the author concerning mathematical education (represented in greater detail in his article on applied mathematics published in the journal Vestnik Vysshei Shkoly, 1967, No. 4, pp. 7480) are still difficult to realize consistently. Therefore the author will be grateful to the readers for any advice and criticism. The book is composed in such a way that it is possible to use it both for studying in a college under the guidance of a teacher and for selfeducation. The subject matter of the book is divided into small sections so that the reader could study the material in suitable order and to any extent depending on the profession and the needs of the reader. It is also intended that the book can be used by students taking a correspondence course and by the readers who have some prerequisites in higher mathematics and want to perfect their knowledge by reading some chapters of the book. For this purpose we sometimes refer the reader to supplementary books (the bibliography is placed at the end of this course; the references are indicated by numbers in square brackets). We also supply the book with the name index, subject index and the list of symbols which enable the reader to find a desired definition, term or symbol. In some colleges analytic geometry and linear algebra are studied as independent courses. The structure of the book facilitates such a separation: the fundamentals of analytic geometry and linear algebra are given in Chapters II, VI, VII, X and XI. Some attention should be paid to the way of the numeration of the formulas and sections in the book. The sections entering into each chapter are numerated in succession beginning with the first number. In references inside each chapter we omit the number of the chapter. For instance, the expression “formula (2)” placed in the text of Chapter VI means “formula (2) of Chapter VFV But when formula (2) of Chapter VI is mentioned in some other chapter we write “formula (VI.2)”. Similarly, “§ II.3” means “§ 3 of Chapter II” but we simply write “§ 3” when § 3 of Chapter II is referred to in this chapter; the expression “Sec. V.6” means “Sec. 6 of Chapter V” and so on. Studying the theoretical material should be followed by solving problems and doing exercises. For this purpose we can recommend the wellknown collections of problems [21, [4], [26] and [47]. But it should be noted that some divisions of applied mathematics are not treated to a sufficient extent in these collections and therefore it is advisable that a teacher of mathematics should add some inte resting and instructive problems concerning these divisions.
PREFACE
7
The book can be of use to readers of various professions dealing with applications of mathematics in their current work. Modern applied mathematics contains, of course, many important special divisions which are not included in this book. The author intends to write another hook devoted to some supplementary topics such as the theory of functions of a complex argument, variational calculus, mathematical physics, some special questions of the theory of ordi nary diSerential equations and so on. • When preparing the book for the second edition the author con siderably revised the text and added some new material including the chapter on the theory of probability*. Besides, the author has taken into account valuable advice and criticism received from many mathematicians, in particular from the members of the Mos cow mathematical society where the book was discussed. Some sections of the book were written or revised under the influence of ideas and useful comments of L. M. Altshuler, Ya. B. Zeldovich and B. 0 . Solonouts. To all of them the author expresses his warmest gratitude. A. D. Myskis April 19, 1966 * This edition is the English translation of tho second Russian edition of the book. The first Russian edition contained a chapter in which n brief review of basic equations of mathematical physics was given. The chapter was excluded from the second edition because of some changes in the syllabus of tech nical colleges. We have included the matorinl of this chaptor in this English edition as the Appendix at the end of the book. Tho present translation incorpo rates suggestions made by tho author.—Tr.
Contents (The starred items in the tabic ot contents Indicate those sections which contain some optional material that may he omitted for the first reading of the boot.)
In tro d u c tio n ...................................................................................... 1. The Subject of M athem atics............................. • • • • • ; • • • 2. The Importance of Mathematics and Mathematical Education . . 3. A b stractn e ss.................... • . ..................................... 4. Characteristic Features of Higher M athem atics............................... 5. Mathematics in the Soviet U n io n .................................................
19 19 20 20 22 23
CHAPTER I. VARIABLES AND FU N C TIO N S.................................. § 1. Quantities .......................................................................................... 1. Concept of a Q u a n t i t y .................................................................... 2. Dimensions of Q u a n titie s ................................................................ 3. Constants and V a r ia b le s ................................................................ 4. Number Scale. Slide R u l e ............................................................. 5. Characteristics of Variables ............................................................. § 2. Approximate Values of Quantities..................................................... 6. The Notion of an Approximate V a l u e .......................................... 7. Errors .................................................................................................. S. Writing Approximate Numbers .................................................... 9. Addition and Subtraction of Approximate N u m b e r s ................... 10. Multiplication and Division of Approximate Numbers. General R e m a rk s.............................................................................................. § 3. Functions and G r a p h s ........................................................................ 11. Functional Relation ........................................................................ 12. N o ta tio n .............................................................................................. 13. Methods of Representing F u n c tio n s .............................................. 14. Graphs of F u n c tio n s ........................................................................ 15. The Domain of Definition of a F u n c ti o n ...................................... 16. Characteristics of Behaviour of F u n c tio n s .................................. 17. Algebraic Classification of F u n c tio n s ............................................. 18. Elementary Functions........................................................................ 19. Transforming G r a p h s ............................................................! . . ! 20. Implicit F u n c tio n s ............................................................................ 21. Inverse Functions................... ’ § 4. Review of Basic Functions................................................................ 22. Linear F unction................................................................................... 23. Quadratic F u n c tio n ............................. ! ! ! ! ! ! ! ! ! . ! ! ! 24. Power F unction..................................................... "
25 25 25 25 26 27 29 32 32 32 33 34 30 39 39 40 42 45 47 48 51 53 54 56 58 00 60 62 63
10 25. 26. 27. 28. 29. 30.
INTRODUCTORY MATHEMATICS FOR ENGINEERS
LinearFractional F u n c t io n ....................................................................... Logarithmic F u n c t i o n ............................................................................... Exponential F u n c t io n ................................................................................ Hyperbolic F u n c t io n s ................................................................................ Trigonometric F u n c tio n s ........................................................................... Empirical F o r m u la s....................................................................................
66 68 69 70 72 75
CHAPTER II. PLANE ANALYTIC G E O M E T R Y .................................... $ 1. Plane C o o rd in a te s....................................................................................... 1. Cartesian C o o rd in a tes................................................................................ 2. Some Simple Problems Concerning Cartesian Coordinates . . . 3. Polar C oord in ates........................................................................................ $ 2. Curves in P l a n e ........................................................................................... 4. Equation of a Curve in Cartesian C o o r d in a te s.................................. 5. Equation of a Curve in Polar C o o r d in a te s ...................................... 6. Parametric Representation of Curves and F u n c t io n s ................. 7. Algebraic C u r v e s ........................................................................................ 8. Singular Cases ............................................................................................ 3. FirstOrder and SecondOrder Algebraic C u rv e s..................................... 9. Curves of the First O r d e r ........................................................................ 10. E l l i p s e ............................................................................................................. 11. H y p e r b o la ..................................................................................................... 12. Relationship Between Ellipse, Hyperbola and Parabola . . . . 13. General Equation of a Curve of the Second O r d e r .........................
78 78 78 79 81 82 82 84 87 90 92 94 94 96 99 102 105
CHAPTER III. LIMIT. C O N T IN U IT Y ...................................................... $ 1. Infinitesimal and Infinitely Large V a ria b les.......................................... 1. Infinitesimal Variables ............................................................................ 2. Properties of Infinitesimals ................................................................... 3. Infinitely Large V a r i a b l e s ....................................................................... § 2. L i m i t s ............................................................................................................. 4. Definition ..................................................................................................... 5. Properties of L i m i t s ................................................................................ 6. Sum of a Numerical S e r i e s ................................................................... § 3. Comparison of V a ria b les............................................................................ 7. Comparison of Infinitesimals ......................................................... 8. Properties of Equivalent Infinitesimals .......................................... 9. Important E x a m p le s.................................................................................... 10. Orders of S m a l l n e s s .................................................................................... 11. Comparison of Infinitely Large V a r i a b l e s .................' ...................... § 4. Continuous and Discontinuous F u n ctio n s.............................................. 12. Definition of a Continuous Function ................................................... 13. Points of Discontinuity ....................................................................... 14. Properties of Continuous F u n c t io n s .................................................. 15. Some Applications ....................................................................................
109 109 109 HI 112 119 113 115 117 121 121 122 122 124 125 125 125 126 129 131
CHAPTER IV. DERIVATIVES, DIFFERENTIALS, INVESTIGATION OF THE BEHAVIOUR OF F U N C T IO N S .............................................. § 1. Derivative ..................................................................................................... 1. Some Problems Leading to the Concept of a D e r i v a t i v e ................. 2. Definition of Derivative .......................................................................
134 134 • 134 136
11
CONTENTS
3. Geometrical Meaning o! D e r iv a tiv e ................................................. 4. Basic Properties ol Derivatives ..................................................... 5. Derivatives ol Basic Elementary F u n c tio n s ........................... (;[ Determining Tangent in Polar C o o rd in a tes.............................. ... § 2. D ifferential..................................* .................................................... 7. Physical Examples *T* # ' \ * m * 8. Definition of Differential and Its Connection with Increment . . 9. Properties of D iffe re n tia l................ . • • • • • • • ................... 10. Application of Differentials to Approximate Calculations . . . § 3. Derivatives and Differentials of Higher Orders.................................. 11. Derivatives of Higher O r d e r s ..................................................... 12. HigherOrder D ifferen tials................................................................ § 4. VHospital*$ R u l e ............................................................................... 0 13. Indeterminate Forms of the Type  g  ............................................. 14. Indeterminate Forms of the Type ^ § 5. 15. 16. § 6. 17. 18. 19. § 7. 20. 21. 22.
..........................................
Taylor7s Formula and S e r i e s ............................................................. Taylor’s F o r m u la ............................................................................... Taylor’s S e r i e s ................................................................................... Intervals of Monotonicity. Extrem um ................................................. Sign of D e riv a tiv e ............................................................... Points of E x tre m u m ..................................... The Greatest and the Least Values of a F u n c tio n ........................... Constructing Graphs of Functions ............................................. . Intervals of Convexity of a Graph and Points of Inflection . . . Asymptotes of a G r a p h ........................................................ General Scheme for Investigating a Function and Constructing Its Graph .............................................................................................. CHAPTER V. APPROXIMATING ROOTS OF EQUATIONS. INTERPOLATION....................................................................................... § 1. Approximating Roots of Equations..................................................... 1. In tro d u c tio n ....................................................................................... 2. CutandTry Method. Method of Chords. Method of Tangents . . 3. Iterative M e th o d ............................................................................... 4. Formula of Finite In c re m e n ts ......................................................... 5*. Small Parameter M e th o d ................................................................. § 2. Interpolatio n....................................................................................... 6. Lagrange’s Interpolation F o r m u l a ................................................. 7. Finite Differences and Their Connection with Derivatives . . . . 8. Newton’s Interpolation F o r m u la s ................................................. 9. Numerical D ifferentiation................................................................ CHAPTER VI. DETERMINANTS AND SYSTEMS OF LINEAR ALGEBRAIC EQUATIONS .................................................................... § 1. D eterm inants....................................................................................... 1. Definition .......................................................................................... 2. Properties ................................................I . . * . ! ! ! ! ! ! ! ! 3. Expanding a Determinant in Minors of Its Row or Coiumn ! .
137 139 142 146 148 148 149 152 153 155 155 156 158 158 160 *61 161 163 165 165 166 168 173 173 174 175 179 179 179 181 185 187 189 191 101 192 196 198 200 200 200 201 203
12 § 2. 4. 5. 6.
INTRODUCTORY MATHEMATICS FOR ENGINEERS
Systems of Linear Algebraic E q u a tio n s.................................................. Basic C a s e .................................................................................................... Numerical S o l u t i o n .................................................................................... Singular C a s e ................................................................................................
206 206 208 209
CHAPTER VII. V E C T O R S ............................................................................... § 1. Linear Operations on V ecto rs........................................* ....................... 1. Scalar and Vector Q u a n titie s................................................................... 2. Addition of Vectors ................................................................................ 3. Zero Vector and Subtraction of V e c t o r s .............................................. 4. Multiplying a Vector by a S c a l a r ........................................................... 5. Linear Combination of V e c t o r s ........................................................... § 2. Scalar Product of V e cto rs........................................................................... 6. Projection of Vector on Axis ................................................................... 7. Scalar P r o d u c t ............................................................................................ 8. Properties of Scalar P r o d u c t ............................................................... § 8. Cartesian Coordinates in S p a c e .............................................................. 9. Cartesian Coordinates in S p a c e ............................................................. 10. Some Simple Problems Concerning Cartesian Coordinates . . . § 4. Vector Product of V e cto rs............................................................................ 11. Orientation of Surface and Vector of an A r e a ...................................... 12. Vector Product ........................................................................................ 13. Properties of Vector P r o d u c t ............................................................... 14*. Pseudo vectors ............................................................................................ § 5. Products of Three V e cto rs........................................................................... 15. Triple Scalar P r o d u c t ................................................................................ 16. Triple Vector Product ............................................................................ § d. Linear S p a c e s ................................................................................................ 17. Concept of Linear Space ........................................................................... 18. Examples ..................................................................................................... 19. Dimension of Linear S p a c e ................................................................... 20. Concept of Euclidean Space ................................................................... 21. Orthogonality ............................................................................................ § 7. Vector Functions of Scalar Argument.C u rv a tu re................................. 22. Vector V a r i a b l e s ........................................................................................ 23. Vector Functions of Scalar A r g u m e n t .............................................. 24. Some Notions Related to the SecondD e r i v a t i v e ............................ 25. Osculating C i r c l e ........................................................................................ 26. Evolute and E v o lv e n t ................................................................................
212 212 212 213 215 215 216 219 219 220 221 222 222 223 227 227 228 230 233 235 235 236 237 237 239 241 244 245 248 248 248 251 252 255
CHAPTER VIII. COMPLEX NUMBERS AND FUNCTIONS . . . § 1. Complex N u m b e r s ....................................................................................... 1. Complex P l a n e ............................................................................................ 2. Algebraic Operations on Complex N u m b e rs...................................... 3. Conjugate Complex N u m b e rs................................................................... 4. Euler’s Formula ........................................................................................ 5. Logarithms of Complex N u m b e r s........................................................... § 2. Complex Functions of a Real A r g u m e n t............................................. 6. Definition and Properties ........................................................................ 7*. Applications to Describing O sc illa tio n s..............................................
259 259 259 261 263 264 266 267 267 269
13
CONTENTS
§ 3. The Concept of a Function of a Complex V a ria b le .............................. 8.
F a cto riza tio n o f a P o ly n o m ia l
. . .
• • • • ..............................
o* Numerical Methods of Solving Algebraic Equations . ■ • • • , • to! Decomposition of a Rational Fraction into Partial Rational
271 271 273
11*. Some General Remarks on Functions of a Complex Variable . . .
277 280
CHAPTER IX. FUNCTIONS OF SEVERAL VARIABLES . . . . § 1. Functions of Tu>o Variables................................................................ 1. Methods of R ep resen tin g ................................................................ 2. Domain of D e fin itio n ........................................................................ 3. Linear Function ................................................................................ 4. Continuity and D isc o n tin u ity ......................................................... 5. Implicit F u n c tio n s ........................................................................... § 2. Functions of Arbitrary Number of Variables...................................... 6. Methods of R ep resen tin g ................................................................ 7. Functions of Three A rg u m e n ts..................................................... 8. General C a s e ....................................................................................... 9. Concept of F i e l d ............................................................................... § 3. Partial Derivatives and Differentials of the First O rd er................... 10. Basic D efin itio n s............................................................................... 11. Total Differential ............... ............................................................ 12. Derivative of Composite Function .................................................. 13. Derivative of Implicit Function ..................................................... § 4. Partial Derivatives and Differentials of Higher O rders................... 14. D efin itio n s.......................................................................................... 15. Equality of Mixed D e riv a tiv e s ...................................................... 16. Total Differentials of Higher O r d e r ............................................. •
283 283 283 286 287 288 291 291 291 292 292 293 291 294 296 298 300 303 303 304 305
CHAPTER X. SOLID ANALYTIC G E O M E T R Y ................................... § 1. Space Coordinates............................................................................... 1. Coordinate Systems in Space ........................................................ 2*. Degrees of F re e d o m ........................................................................ § 2. Surfaces and Curves in S p a c e ............................................................ 3. Surfaces In S p a c e ................................................................................ 4. Cylinders, Cones and Surfaces of R e v o lu tio n ................................ 5. Curves in S p a c e ................................................................................ 6. Parametric Representation of Surfaces in Space. Parametric Repre sentation of Functions of Several V a r ia b le s .................................. § 3. Algebraic Surfaces of the First and of theSecond Orders.................. 7. Algebraic Surfaces of the First O r d e r ............................................... 8. Ellipsoid ........................................................................................... 9. Hyperboloids ................................................................................... 10. Paraboloids ...................................................................................... . 11. General Review of Algebraic Surfaces of the Second Order . .
307 307 307 309 313 313 314 316
CHAPTER XI. MATRICES AND THEIR APPLICATIONS . . . .
329
§ 1. M a t r i c e s .................................................
1. D efinitions................................................................................... 2. Operations on M a tr ic e s ................................. 1 " . ! ! ! ! ! ! * 3. Inverse M a t r i x ........................................................... ! . ! . ’ *
317 319 319 322 324 326 327 329
329 331 333
14
INTRODUCTORY MATHEMATICS FOR ENGINEERS
4. 5. § 2. 6. 7. 8. 9. 10. § 3. 11. 12. § 4. 13*. 14*. 15*.
Eigenvectors and Eigenvalues of a M a t r ix ...................................... The Rank of a M a tr ix ................................................................................ Linear M a p p i n g s ........................................................................................ Linear Mapping and Its M a t r ix ............................................................... Transformation of the Matrix of a Linear Mapping When the Basis Is C h an ged .................................................................................................... The Matrix of a Mapping Relative to the Basis Consisting of Its Eigenvectors ................................................................................................ Transforming Cartesian Basis ............................................................... Symmetric M a tr ic e s.................................................................................... Quadratic F o r m s ............................................................................................ Quadratic Forms ........................................................................................ Simplification of Equations of SecondOrder Curves and Surfaces . NonLinear M a p p i n g s ................................................................................ General Notions ........................................................................................ NonLinear Mapping in the S m a l l ....................................................... Functional Relation Between F u n c tio n s..............................................
335* 337 339“ 339 347 350 352 353355 355 357 358358 360 362
yC H A P T E R X II. APPLICATIONS OF PARTIAL DERIVATIVES . . § 1. Scalar F i e l d ................................................................................................. 1. Directional Derivative. G r a d ie n t........................................................... 2. Level Surfaces ............................................................................................ 3. Im plicit Functions of Two Independent V a ria b le s.......................... 4. Plane F i e l d s ................................................................................................. 5. Envelope of OneParameter Family of C u r v es.................................. § 2. Extremum of a Function of Several V a ria b les...................................... 6. Taylor’s Formula for a Function of Several V a ria b le s................. 7. E x t r e m u m ..................................................................................................... 8. The Method of Least S q u a r e s................................................................... 9*. Curvature of S u r fa c es................................................................................ 10. Conditional Extremum ............................................................................ 11. Extremum with Unilateral C onstraints................................................... 12*. Numerical Solution of Systems of E q u a t io n s ..................................
365 365 365 360 370 371 372 374 374 375 380 381 384 388 390
CHAPTER XIII. INDEFINITE IN T E G R A L .......................................... § 7. Elementary Methods of In teg ra tio n ........................................................... 1. Basic Definitions .................................................................................... 2. The Simplest Integrals ........................................................................... 3. The Simplest Properties of an Indefinite I n t e g r a l ......................... 4. Integration by Parts ................................................................................ 5. Integration by Change of Variable (by S u b s titu tio n ) .................
393 393 393 394 397 390 402
§ 2. Standard Methods of Integration ............................................................... 6. Integration of Rational Functions ............................................... 7. Integration of Irrational Functions Involving Linear and LinearFractional E x p ression s............................................................... 8. Integration of Irrational Expressions Containing Quadratic Trino mials ............................................................................................................... 9. Integrals of Binomial Differentials 10. Integration of Functions Rationally Involving Trigonometric Functions ..................................................................................................... 11. General R e m a r k s ........................................................................................
404 405 407 408 411 412 415
CONTENTS
CHAPTER XIV. DEFINITE INTEGRAL § 1. Definition and Basic Properties............... ... ................................. 1. Examples Leading to the Concept of Definite Integral . . . . 2. Basic Definition.................................   • j . ‘ .i ’ 3. Relationship Between Definite Integral and Indefinite Integral . 4. Basic Properties of Definite I n te g r a l.......................................... 5. Integrating In e q u a litie s ..................................................... ... § 2. Applications of Definite In teg ra l........................... 6. Two Schemes of A pplication............................. • • 7. Differential Equations with Variables Separable 8. Computing Areas of Plane Geometric Figures . . 9. The Arc Length of ai C u rv o ...................................... 10. Computing Volumes of Solids .................................. 11. Computing Area of Surface of Revolution . . . . § 3. Numerical I n te g r a tio n .....................
12. 13. § 4. 14. 15. 16. 17*. IS*. 19*. § 5. 20*. 21*. § 6. 22. 23. 24. § 7. 25*. 26*. 27*.
General R e m a rk s ..................... ; Formulas of Numerical Integration Improper Integrals ............................................................................ Integrals with Infinite Limits of Integration . . . . . . . . . Basic Properties of Integrals with Infinite Limits of Integration Other Types of Improper In te g ra l.................................................. Gamma F un ctio n............................................................................... Beta F u n c tio n ................................................................................... Principal Value of Divergent In te g ra l.............................................. Integrals Dependent on Parameters..................................................... Proper Integrals ............................................................................... Improper I n te g r a ls ............................................................................ Line Integrals....................................................................................... Line Integrals of the First T y p e ..................................................... Line Integrals of the Second T y p e .................................................. Conditions for a Lino Integral of the Second Type to Be Independ ent of the Path of In teg ratio n ......................................................... The Concept of Generalized F unction................................................. Delta F u n c tio n ................................................................................... Application to Constructing Influence F u n ctio n ........................... Other Generalized F unctions.............................................................
CHAPTER XV. DIFFERENTIAL EQ U A TIO N S.................................. § 1. General N otions................................................................................... 1. Examples .............................................................................. * * * * 2. Basic D efin itio n s...............................................................................* § 2. FirstOrder Differential E q u a tio n s..................................................... 3. Geometric Meaning ....................................................................... [ 4. Integrable Types of E q u atio n s..................................................... 5*. Equation for Exponential F u n c tio n .................................................. __5. Integrating Exact Differential E q u atio n s.................... * . ................. 7*. Singular Points and Singular S olutions.......................................... 8*. Equations Not Solved for the Derivative .......................... . . ! 9*. Method of Integration by Means of Differentiation........................... § 3. HigherOrder Equations and Systems of Differential Equations . . . 10. HigherOrder Differential E q u atio n s................................................
15 417 417 417 419 423 426 433 436 436 437 439 443 445 447 448 448 450 454 455 458 464 468 471 473 474 474 476 478 478 482 484 488 488 492 495 497 497 497 498 500 500 503 506 509 512 516 517 519 519
16
INTRODUCTORY MATHEMATICS FOR ENGINEERS
11*. Connection Between HigherOrder Equations and Systems of FirstOrder E q u a tio n s.................................................................. 12*. Geometric Interpretation of System of FirstOrder Equations . . 13*. First I n t e g r a l s .............................................................* ...........................
52i 522 526
§ 4. Linear Equations of General F o r m ....................................................... 528 14. Homogeneous Linear E q u a tio n s........................................................ 528 15. NonHomogeneous E q u a tio n s............................................................ 530 16*. BoundaryValue P r o b le m s ................................................................. 535 § 5. Linear Equations with Constant Coefficients................................... 541 17. Homogeneous E q u a tio n s.................................................. . 541 18. NonHomogeneous Equations with RightHand Sides of Special Form ............................................................................................................. 545 19*. Euler’s E q u a tio n s.................................................................. 548 20*. Operators and the Operator Method of Solving Differential Equations 549 § 6. 21. 22*. § 7.
Systems of Linear E qu ation s.................................................. . . . . Systems of Linear E q u a tio n s......................................... Applications to Testing Lyapunov Stability of Equilibrium State Approximate and Numerical Methods of Solving Differential Equations
553 553 558 562
23. 24"s. 25. 26*. 27*. 28*. 29*. 30*. 31. 32. 33. 34.
Iterative Method ........................................................................................ 562 Application of Taylor’s S e r i e s ........................................# .................... 564 Application of Power Series with Undetermined Coefficients . . . 555 Bessel’s Functions .................................................................................... 566 Small Parameter Method ........................................................................ 569 General Remarks on Dependence of Solutions on Parameters . . 572 Methods of Minimizing D iscrep an cy................................................ 575 Simplification M e th o d ......................................................................... 576 Euler’s M e th o d ..................................................................................... 578 RungeKutta Method ................................................................................ 580 Adams M e t h o d ..................................................................................... 582 Milne’s M e th o d ..................................................................................... 583
CHAPTER XVI. Multiple In te g r a ls............................................................... 585 § 1. Definition and Basic Properties of M ultiple In te g r a ls .......... 585 1. Some Examples Leading to the Notion of a Multiple Integral . . 585 2. Definition of a Multiple I n te g r a l.................................................... 586 3. Basic Properties of Multiple I n te g r a ls.................... 587 4. Methods of Applying Multiple Integrals . . . . . . .............. 589 5. Geometric Meaning of an Integral Over a Plane R e g io n .......... 591 § 2. Two Types of Physical Q u a n tities............................. 592 6*. Basic Example. Mass and Its D e n s it y ....................................... 592 7*. Quantities Distributed in S p a c e ............................. 594 § 3. Computing M ultiple Integrals in Cartesian C oordinates.......... 596 8. Integral Over R e c ta n g le .................................................. 596 9. Integral Over an Arbitrary Plane Region . . . .............................. 599 10. Integral Over an Arbitrary S u r fa c e ....................................................... 602 11. Integral Over a ThreeDimensional Region . . .............................. 604 § 4. 12. 13. 14*.
Change of Variables in M ultiple Integrals . . . .............................. Passing to Polar Coordinates in P l a n e ................................................... Passing to Cylindrical and Spherical Coordinates .............................. Curvilinear Coordinates in P l a n e ...........................................................
605 605 606 608
CONTENTS
17
15*. 16*. § 5. 17*. 18*. 19*. 20*. § 6. 21*. 22*. 23*. 24*. 25. 26*. 27. 28*.
Curvilinear Coordinates in S p a c e ................................................. . Coordinates on a Surface ............................................................ . Other Types of Multiple Integrals................................................ * • Improper I n t e g r a l s ............................................. .............................. Integrals Dependent on a Parameter ................ Integrals with Respect to Measure. Generalized Functions . . . Multiple Integrals of Higher O rd e r.............................................. Vector Field ............................................................ ... ...................... Vector L i n e s ....................................................................................... The Flux of a Vector Through a S u rface...................................... Divergence ............................. . . .• ......................................... Expressing Divergence in Cartesian Coordinates........................... Line Integral and C irculation........................................................... Rotation .......................................................................................... Green’s Formula. Stokes’ F o rm u la.................................................... Expressing Differential Operations on Vector Fields in a Curvilinear Orthogonal Coordinate System ............................ .......................... 29*. General Formula for Transforming In te g ra ls ...............................
611 612 615 615 617 620 622 626 626 627 629 632 634 634 638
CHAPTER XVII. SERIES ............................................. '§'■1. Number S e r i e s ............................................................ 1. Positive S e r i e s .................................................... 2. Series with Terms of Arbitrary S ig n s ....................... 3. Operations on S e r i e s ................................................. 4*. Speed of Convergence of a S e rie s .............................. 5. Series with Complex, Vector and Matrix Terms . . 6. Multiple Series ........................................................ § 2. Functional S e r i e s ......................................................... 7. Deviation of F u n c tio n s .............................................. 8. Convergence of a Functional S e rie s........................... 9. Properties of Functional S e rie s.................................. § 3. Power Series ................................................................ 10. Interval of Convergence............................................. 11. Properties of Power S eries.......................................... 12. Algebraic Operations on Power S e rie s ....................... 13. Power Series as a Taylor S e rie s .............................. 14. Power Series with Complex T e rm s ........................... 15*. Bernoullian Numbers ............................................. 16*. Applying Series to Solving Difference Equations . . 17*. Multiple Power Series ............................................. 18*. Functions of Matrices ............................................. 19*. Asymptotic Expansions .......................................... § i. Trigonometric Series ................................................. 20. O rthogonality................................................................ 21. Series in Orthogonal F u n ctio n s.............................. 22. Fourier S e rie s.......................... ... ................................. 23. Expanding a Periodic F u n ctio n .................................. 24*. Example. Bessel’s Functions as Fourier Coefficients 25. Speed of Convergence of a Fourier S e rie s ............... 26. Fourier Series in Complex F o r m .............................. 27*. Parseval R e la tio n ........................................................* 28*. Hilbert Space ........................................................! !! 29*. Orthogonality with Weight Function . . . . . . . 30*. Multiple Fourier S e r i e s .........................................! .!
645 645 645 650 652 654 658 659 661 661 662 664 666 666 667 671 675 676 677 678 680 681 685 686 686 689 690 695 697 698 702 704 706 708 710
641 642
18
INTRODUCTORY MATHEMATICS FOR ENGINEERS
31*. Application to the Equation of Oscillations of a S t r in g ................... § 5.. Fourier Transformation ............................................................................ 32*. Fourier T r a n s fo r m ....................................................................................... 33*. Properties of Fourier T ransform s........................................................... 34*. Application to Oscillations of Infinite S t r in g ..................................
711 713 713 717 719
CHAPTER XVIII. ELEMENTS OF THE THEORY OF PROBABI LITY .......................................................................................................................... § 1, Random Events and Their P ro b a b ilitie s............................................... 1. Random E v e n t s ............................................................................................ 2. P r o b a b ility ..................................................................................................... 3. Basic Properties of P ro b a b ilities........................................................... 4. Theorem of Multiplication of P ro b a b ilities.......................................... 5. Theorem of Total P r o b a b ility ............................................................... 6*. Formulas for the Probability of H y p o th e se s.................................. 7. Disregarding LowProbability E v e n t s ................................................... § 2. Random V a ria b les........................................................................................ 8. D e f in itio n s ..................................................................................................... 9. Examples of Discrete Random V a r ia b le s.............................................. 10. Examples of Continuous Random V a r ia b le s...................................... 11. Joint Distribution of Several Random V a r ia b le s.............................. 12. Functions of Random V a r ia b le s........................................................... § 3. Numerical Characteristics of Random V a ria b les.................................. 13. Tbe Mean V a lu e ........................................................................................ 14. Properties of tbe Mean V a l u e ............................................................... 15. V a r ia n c e ......................................................................................................... 16*. C o rr ela tio n ..................................................................................................... 17. Characteristic F u n c tio n s............................................................................ § 4. Applications of the Normal L a w ............................................................ 18. The Normal Law as tbe Limiting O n e ............................................... 19. Confidence I n t e r v a l .................................................................................... 20. Data P ro cessin g .............................................................................................
721 721 721 722 725 727 729 730 731 732 732 734 736 737 739 741 741 742 744 746 748 750 750 752 754
CHAPTER XIX. C O M P U T E R S .................................................................... 757 § 1. Two Classes of C o m p u ters........................................................................ 757 1. Analogue C om p uters.................................................................................... 758 2. Digital C om puters........................................................................................ 762 § 2. P ro g ra m m in g ................................................................................................. 764 3. Number Systems ......................................................................................... 764 4. Representing Numbers in a C o m p u te r ................................................... 766 5. Instructions ................................................................................................. 769 6. Examples of Program m ing........................................................................ 772 Appendix. Equations of Mathematical P h y s ic s ...................................... 780 1*. Derivation of Some Equations . . . ....................................................780 2*. Some Other E q u a tio n s.......................................... 783 3*. Initial and Boundary C o n d it io n s ....................................................... 784 § 2, Method of Separation of V a ria b les........................................................... 786 4*. Basic Example ......................................................................................... 786 5*. Some Other P r o b le m s .................................................................... ’ . . . 781 Bibliography ......................................................................................................... 796 Name I n d e x .............................................................................................................. 798 Subject I n d e x ......................................................................................................... 800 List of S y m b o ls ..................................................................................................... 815
Introduction
1. Hie Subject of Mathematics. Numerical calculations are pene trating into the fields of work of physicists, chemists and engineers of various specialities. The modem development of science and engineering makes it necessary to deduce and apply still more com plicated laws, to solve very complicated problems and perform extensive calculations. All such calculations are based on mathematics, the science which treats of relations existing between spatial forms, quantities and magnitudes of the real world. All the basic notions of mathematics emerged and were developed in connection with the demands of natural sciences (physics, mechanics, astronomy etc.) and enginee ring. The appearance of more complicated problems led to the crea tion of more sophisticated mathematical methods of investigation (i.e. mathematical rules, techniques, formulas and the like) and, in particular, to the foundation of higher mathematics. It is there fore not accidental that the fundamentals of higher mathematics were created in the 17th and 18th centuries, i.e. at the beginning of an intensive development of industry, although some elements of higher mathematics appeared as early as antiquity in the works of the great Greek mathematician and mechanician Archimedes (287212 B.C.). Higher mathematics was founded in the works of the prominent French philosopher, physicist, mathematician and physiologist R. Descartes (15961650), the great English physicist, mechanician, astronomer and mathematician I. Newton (16421727), the great German mathematician and philosopher G. Leibniz (16461716), the great mathematician, mechanician and physicist L. Euler (17071783) and many other famous scientists. In their works diffe rent divisions of mathematics were created for investigating phe nomena of nature and solving engineering problems. In mathematics, as in other sciences, practical work is the main source of scientific discoveries. Another important source is the need of mathematics
20
INTRODUCTORY MATHEMATICS FOR E N G IN E E R S
itself to systematize the facts discovered, to investigate their inter relations and so on. 2. The Importance of Mathematics and Mathematical Education. Mathematics and, in particular, higher mathematics plays a very important role in modern natural science and engineering. Mathema tics lies in the foundation of all divisions of physics, mechanics and many divisions of other natural sciences, engineering and some other branches of knowledge. Designing the construction^ of an airplane or a dam of a hydroelectric power station, investigating complicated processes involved in deformation of metals, propa gation of radiowaves, diffusion of neutrons in an atomic reactor etc. •cannot be performed without systematic application of mathe matics. The high level of development of computational methods in the USSR is one of the main factors that led to the triumphant achieve ments in launching the first artificial satellites of the Earth, space rockets and spacecraft. The creation of highspeed electronic compu ters and other mathematical automatic devices leads to further extension of the application of higher mathematics and facilitates the introduction of computational methods into many new fields. In particular, this is the case in such fields as economics, manage ment and control of industry, elaborating optimal (i.e. the best) plans of capital investments or construction, transportation problems, controlling technological processes, dispatching and so on. The application of mathematical methods in these fields has already proved to be very effective and profitable. In recent years mathema tics has been penetrating into such traditionally “nonmathematical” fields as biology, physiology, geography etc. Therefore nowadays the requirements for mathematical educa tion of an engineer are very high. An engineer must know the basic principles of higher mathematics and be able to apply them to con crete problems. Then mathematics will become a powerful tool in his hands. Besides, a great deal of scientific literature and many special technical subjects are saturated with mathematical techni ques and formulas. Without sufficient knowledge of mathematics much effort is needed to understand all these formulas which may hinder the reader’s work and mislead him. Mathematics also faci litates a better understanding of many questions related to other sciences (the theory of vibrations, mechanics of continuous media and so on). 3. Abstractness. Mathematics itself is not a technical subject and therefore a course in mathematics for engineers and scientists must not treat any special technical questions. Its aim is to provide the necessary mathematical education. Therefore a student may sometimes feel that the questions treated in a course of higher mathe matics are too abstract. But the abstractness of mathematics is one
INTRODUCTION
2i
of its most essential features. This does not at all mean that mathe matics has little to do with practical activities. On the contrary, it is the possibility to apply mathematics to various kinds of acti vities that makes its abstractness so important. For instance, m geometry we consider an “abstract” cylinder and find its volume. This immediately enables us to compute the volume of any concrete cylinder no matter whether it is a component of a mechanism or a column or a portion of space occupied by an electric Bold. Simi larly, in higher mathematics we deduce some general abstract laws whose statements are not directly connected with a particular form of practical activity, a natural science or engineering, but the con crete applications and realizations of these laws (which are studied as examples in a mathematical course) are always related to various phenomena of the real world. Thus, mathematics considers pure, ideal (schematized) forms, relations, processes etc. whose realization serves only as an appro ximation to reality. For instance, a real cylinder can never he a perfect cylinder from the mathematical point of view. Here we see the manifestation of a distinguishing feature characteristic of any kind of human cognition: when considering a real object or process we always select a number of basic properties from an infinite variety of properties of the object or process and investigate these most essential properties abstracting them from inessential ones. B ut it may sometimes happen that an assumption, hypothesis, that all the properties except those chosen as basic ones are inessential is not true and then we can arrive at a contradiction between our mathematical inferences and reality. Such a possibility must never be forgotten! Because of the abstractness of forms and relations the logical consistency of inferences in mathematics is extremely important, more important than in other sciences, this being well known even from elementary mathematics. In higher mathematics too, all the assertions must be completely clear and logically justified so that it should be possible to regard them as objective laws adequate to reality. In mathematics, and particularly in higher mathematics, we also sometimes draw certain conclusions from experiment, obser vation and analogy but nevertheless such a situation is rarer in mathematics than in other sciences. • There is a characteristic tendency in mathematics to deduce all the assertions from a few basic principles (called axioms). This is the socalled deductive method. But in our introductory course which is intended for those who are mainly interested in applications we shall not rigorously follow this method in all cases. The reader interested in theory may find some inferences in our course to be imperfect from the point of view of logic. If he wants to get a better understanding of some exceptions to general rules and to study higher
22
INTRODUCTORY MATHEMATICS
FOR
E N G IN EE R S
mathematics more thoroughly, he should study a course wri for mathematicians, for instance, [14]. To consider the same ques from diSerent points of view it is advisable to take other co\ intended for technical colleges (for instance, [5], [37], [441 and we particularly recommend book [5]). 4. Characteristic Features of Higher Mathematics. There k distinct boundary between elementary and higher mathema the division being conditional. These are not at all different sciei and the division is mainly accounted for by some historical rea as elementary mathematics and higher mathematics were ere in different historical epochs. But nevertheless we can point some characteristic features of higher mathematics. One of them is the universality, generality, of its methods an example, let us take the problem of finding volumes of so Elementary mathematics gives us different formulas for compu the volumes of a prism, pyramid, cone, cylinder, sphere and s other simple solids. Each formula is obtained on the basis of a cial argument which is rather complicated in certain cases. Bi higher mathematics we have general formulas expressing the vol of any solid, the length of any curve, the area of any surface the like. Take another example. Consider the problem of inves* ting the motion of a material ‘point under the action of given fo: In elementary courses in physics (based on elementary mathema methods) we study only uniform rectilinear motion, uniformly i lerated rectilinear motion, uniformly decelerated rectilinear me and uniform circular motion, and it is rather difficult to investi other types of motion by means of techniques of elementary nu matics. But the methods of higher mathematics make it pos: to investigate any type of motion which can be encounterei practical problems. There is another characteristic feature of higher mathemi (related to the above one). It is the systematic consideratio variable quantities. When investigating various objects and cesses by means of elementary mathematics we usually regard important quantities as velocities, accelerations, densities, ma forces etc. as being invariable, constant (and yet we attain aim only in some simple cases). But if these quantities vary siderably (as is often the case) we cannot regard them as being stant. To solve such problems we usually apply higher mathema There is a branch of higher mathematics (called differential calci which is one of the earliest divisions of mathematics particul intended for solving various problems connected with an inv gation of the dependence of one quantity upon another. The q tities and their interrelations can be of any nature (for insta we can consider the relation between the acceleration, velc and path length of a motion or between the density, mass and i
INTRODUCTION
23
and the like). Therefore differential calculus deeply penetrates into various natural sciences and engineering. . The third characteristic feature of higher mathematics is the close relationship between its various divisions and the systematic unification of the computational, analytical (based on formulas) and geometric methods in contrast to elementary mathematics in which the connection between algebra and geometry is more or less accidental. In higher mathematics, the coordinate method re duces geometric problems to solving algebraic equations, graphs are used for representing relations between variable quantities, analytical methods of integral calculus are applied for computing areas and volumes of geometric figures and so on. Some historical remarks w ill be given in due course in this book. But it is expedient to make some introductory notes here. The most important divisions of higher mathematics which now form the basis of the syllabus for engineers of many specialities were created in the 17th and 18th centuries. They include the coordinate method, differential and integral calculus etc. These divisions are represented in courses of higher mathematics for engineers mostly in the form they appeared after the works of Euler. L. Euler (a Swiss by birth) spent most of his life in Russia and died in Petersburg. Most of his works (473 out of 865) were published in Russia. H is outstanding results in various divisions of mathematics, mechanics, physics and other sciences lie in the foundation of these divisions. Mathematics was created by scientists of many countries. Among Russian mathematicians we should mention N. I. Lobachevsky (17921856), the creator of a nonEuclidean geometry. He also obtai ned some important results in other divisions of mathematics and initiated mathematical studies in Kazan. An intensive development of mathematics in Petersburg began with the works of the prominent mathematician Academician M. V. Ostrogradsky (18011862). The founder of the famous Petersburg mathematical school was the great Russian mathematician and mechanician Academician P. L. Chebyshev (18211894). He obtained many important results in various fields of mathematics and its applications to the theory of mechanisms, cartography etc. After Chebyshev most prominent representatives of the Peters burg mathematical school were Academician A. A. Markov (18561922), a famous mathematician and the creator of the theory of random processes, and Academician A. M. Lyapunov (18571918), the founder of the theory of stability. Since the second half of the 19th century mathematical investi gations have been developing in Kiev, Moscow, Odessa, Kharkov, and other Russian towns. 5. Mathematics in the Soviet Union. In the Soviet Union there are many centres of mathematical research. Among prominent
24
INTRODUCTORY
MATHEMATICS
FOR
EN G IN E E R S
Soviet mathematicians we should mention Academicians A. D. Alek sandrov, P. S. Aleksandrov, N .' N. Bogolyubov, V. M. Glushkov, L. V. Kantorovich, M. V. Keldysh, A. N. Kolmogorov, M. A. Lav rentyev, Yu. V. Linnik, N. I. Muskhelishvili, P. S. Novikov, I. G. Petrovsky, L. S. Pontryagin, V. I. Smirnov, S. L. Sobolev, A. N. Tikhonov, I. N. Vekua, I. M. Vinogradov, and others. Mathematics is being intensively developed both in old centres and in new ones in Baku, Erevan, Gorki, Lvov, Minsk, Novosibirsk, Rostov, Saratov, Sverdlovsk, Tashkent, Tbilisi, Vilnyus, Voronezh, and other towns. The role of mathematics in other sciences, industry and enginee ring has considerably increased. Many mathematicians work out new theoretical problems of other branches of knowledge connected with applications of mathematics. At the same time many physi cists, mechanicians and engineers take part in the development and applications of those divisions of mathematics which are related to their fields of work. As examples of fruitful unification of mathe matics and its applications we can mention the works of the great Russian scientist and one of the founders of modern flight mechanics and hydromechanics N. E. Zhukovsky (18471921), the prominent Russian scientist, mathematician, mechanician and naval architect Academician A. N. Krylov (18631945), the prominent Soviet scientist in the fields of theoretical mechanics, aerodynamics and hydromechanics Academician S. A. Chaplygin (18691942) and others. There is no doubt that development of mathematical education will further increase the role of mathematics in our life and yield fruitful results.
CHAPTER I
Variables and Functions
§ 1. Quantities 1. Concept of a Quantity. It is difficult to give a strict definition of a quantity since the notion is extremely general and universal. Masses, pressures, charges, different kinds of work, lengths and volumes are examples of quantities. It w ill be sufficient for our further aim to regard as a quantity everything that is expressible in certain units and completely characterized by its numerical value. For instance, masses are measured in grams or kilograms and the like. We can say that the area of a circle is a quantity since it is completely characterized by its numerical value (for example, 5, st etc.) if we measure it in certain units, e.g. in square centimetres. The circle itself regarded as a geometric figure is of course not a quan tity because it is characterized by a certain geometric form which cannot be expressed numerically. Many notions which were originally understood only in a quali tative aspect have been recently “advanced” and transferred to the class of quantities (for instance, such notions as effectiveness, infor mation and even likelihood). Every change of this kind is a great event since it enables us to apply quantitative mathematical me thods to investigating the corresponding notions and this usually turns out to be very effective. 2. Dimensions of Quantities. A unit measure which is used for expressing a quantity is called the dimension of the quantity. For in stance, the gram or the kilogram usually serves as the dimension of mass. The dimension of area is the square centimetre or the square metre and so on. A dimension is denoted by square brackets. For instance, if M is a mass and 5 is an area then lilf] = kg (the kilo gram) and 151 = m2 (the square metre) in the international system of units. Usually the units of some quantities are regarded as fundamental units whereas the units of all other quantities are derived units expressed in terms of the fundamental ones. For instance, the units of length (m), of mass (kg) and of time (sec) are the fundamental
26
INTRODUCTORY
MATHEMATICS
FOR
E N G IN EE R S
units in the international system of units (SI), and the unit of velo city (m/sec) or of force (kgm/sec2) is expressed in terms of the fun damental units. We can add together and subtract only quantities of the same dimension, the dimension of a sum being that of the summands. It is permissible to multiply or divide quantities of arbitrary di mensions. The multiplication or division of quantities yields, res pectively, the multiplication or division of their dimensions. We also consider dimensionless (“abstract”) quantities. For instance, the ratio of two quantities of the same dimension is dimensionless. The numerical value of the ratio of a quantity to the chosen unit measure is also dimensionless. For example, the numerical value of the mass of 5 kg is the “dimensionless mass” 5. We can also obtain a dimensionless mass if we take the ratio of the mass to a certain mass which is characteristic of the process in question (such a mass is supposed to be well known, and we choose it as a standard to compare with). Dimensionless length, time etc. are introduced in like manner. In mathematics we usually regard quantities as dimensionless. Finally, a dimensionless quantity is completely characterized by its numerical value, and its “unit measure” is the number 1. 3. Constants and Variables. A quantity entering into an investi gation can take on either different values or only one fixed value. In the first case we call the quantity a variable quantity or, in short, a variable, and in the second case we call the quantity a constant (a constant quantity). Suppose we consider the water in a basin. The water pressure measured at different points of the basin is a variable since it varies and is different at different points. At the same time the water density can be regarded as a constant since it takes on one and the same value (with a sufficient degree of accu racy) at different points. As another example let us consider the process of compressing a given mass of a gas while the temperature is kept constant. Then the pressure and the volume are variables whereas the mass and the temperature are constants. But it should be noted that in a real process the last two quantities inevitably vary a little. Hence, we can schematize the process and conditionally regard the mass and the temperature as constants only in case their real variations are of no importance for our investigation. And in many other cases the constancy of some quantities should be under stood in a conditional sense. We must never forget it since, if we regard a quantity as a constant in a process in which the variations of the quantity, small though they may be, are essential for the investigation, we may arrive at wrong conclusions and our schema tized model will not apply. . It may happen that a quantity which is constant in a certain treatm ent of a phenomenon takes on a different value or even becomes
27
VARIABLES AND FUNCTIONS
a variable under some other (though similar) circumstances. The constant quantities of this kind are called the parameters of the process; they are the characteristics of the process. For example, the mass and the temperature of a gas are the parameters of the process of isothermal compression. When we deal with an electriclight bulb we take into account such parameters as the resistance, the supply voltage the bulb is designed for and the power consump tion. Even in this case there are some other parameters which may also be taken into account (for instance, the sizes of the bulb) but usually we do not regard these parameters as basic ones. Generally, in all cases it is very important to choose the basic, the most signi ficant parameters among various parameters characterizing an object. 4. Number Scale. Slide Rule. Quantities can be represented visually by means of a number scale. For this purpose we usually take a rectilinear axis with a uniform scale. To construct a number scale we choose a straight line and a point on the line which serves as the origin (the origin is usually designated by the letter 0). Wc choose one of the directions on the straight line as the positive direction and take a certain line segment as a unit of length (the positive direction is indicated by an arrow; see Fig. 1). Setting off
 1HUH(11}(ntn[11fn11111nfnro11m(11n111Ht*2
1
0
1
2
3
t
Fig. 1
the unit segment from the origin in both directions and repeating e procedure infinitely we obtain all the points which correspond to the integral values of the quantity. Between the “integer points” there are points representing fractional values, both rational (such as y , —2.03 etc.) and irrational (that is fractional numbers that are not rational, e.g. ^ ± i ,  K and the like). In case we have LauTre? t h r L ^ ntit athe .eSment chosen as the length unit also v X r o f t i m r f d S ^ ^ t lmeT i0n* For examPle ’ the numerical see the d o X s N expressed in seconds; we also there P N ( = Sec)’ 0 1 is represented on such a scale by a point which is ob tained by drawing the line segment of length k log n (where k is a factor of proportionality suitably chosen) in the positive direction from a point A . Positive numbers n < 1 are obtained on the loga rithmic scale by drawing the segment fclog n\ in the negative direction from A because for such n we have log n < 0. A logarithmic scale is, in particular, utilized in the construction of a slide rule. The instrument consists of a ruler and a slide which
VARIABLES AND FUNCTIONS
29
with respect to each other (see Fig. 4) so that two points a and & on the lower scale coincide with the corresponding points a and b on the upper scale. Then we have
k log b — k log a = k log b'
k log a'
since the lengths of the shaded line segments are equal. Now after some simple transformations we obtain (check it up!)
Three of the values a, b, a' and b' being given, we can read the fourth value on the slide rule. This fourth value w ill satisfy relation (1). b If, for example, we put a' = 1 then b = ab' or a = p~. Consequently, to determine the product of two given numbers a and b' we must make the point 1 on the slide coincide with the point a on the ruler
Fig. 4
and then read the value of the product which is indicated on the ruler by the point b' of the slide. (Think how to find the quotient of two given numbers.) It is sometimes convenient to put b' = 10 instead of a' = 1 and to move the slide not to the right but to the left with respect to the ruler. The slide rule was invented in the 17th century. It is widely used now and facilitates the work of many technicians, engineers, physicists etc. Supplementary scales on the slide rule make it possible to perform various additional operations including extracting roots, taking logarithms, raising, solving equations of different types and so forth. There is a number of handbooks on using the slide rule, for example, [36] andj [43] to which we refer the reader. Some curvilinear scales are also of use in certain cases (for example, see Sec. IX .1). But in our course we shall usually use rectilinear axes and uniform scales for representing quantities unless the con trary is explicitly stated. 5. Characteristics of Variables. A variable which takes on all the numerical values or all the values lying between some lim its is called continuous. On the contrary, a variable which assumes •certain “separated” values is called discrete.
30
INTRODUCTORY
MATHEMATICS
FO R
E N G IN EE R S
The set of all numerical values which may be assumed by a variable is called the range of the variable. Now we introduce the notion of an interval which is of use for characterizing ranges of some types of variables. A finite (bounded) interval is the set of all numbers contained between two given numbers a and b. The numbers a and b are called the endpoints of the interval. The endpoints a and b may or may not be included into the interval and this fact should be sometimes indicated. Respectively, in the first case we call the interval closed (i.e. when a t ^ x ^ b and the endpoints are thus included) and denote it as [a, b] and in the second case we say that the interval is open (i.e. a < x < b and the endpoints are excluded) and denote it by (a, 6). Finite intervals are represented by line segments on the number scale. There are also unbounded (infinite) intervals for which a or 6 or both a and b may be infinite. For example, if a variable x assumes all possible values greater than some constant number a the range of the variable is described by the inequalities a < x < oo. This is an example of an infinite interval; it has no finite right endpoint, of course, but in such a case we say, conditionally, that the right endpoint is at infinity. An interval of this kind is also said to have no upper bound since in case a variable may increase unlimitedly we usually interpret the variable as “rising up”. The notion of a lower bound is understood in just like manner. The collection of all real numbers is an interval with neither lower nor upper bound (that is, geometrically, the whole number scale). The range of a continuous variable is an interval or a collection of some number of intervals. For example, if a triangle ABC is defor med in all possible ways the corresponding angle A is a continuous variable whose range is the interval 0 < A < n (in case the nume rical values of the angle are expressed in radians). At the same tim e the area S of the triangle has the interval 0 < S < oo as its range (of course, here we also mean that the numerical values of the area are measured in certain units but we are not going to mention details of this kind in all cases in future). The range of a discrete variable is a set (finite or infinite) of separate real numbers. We can also say, in the geometric sense, that such a range consists of separate points (but not of entire intervals). For example, let an index assume the values 1, 2. . . ., n. Then it is a discrete variable. If a variable changes in a certain process in such a way that its numerical values vary only in one direction, that is they either increase or decrease, it is called monotonlc. The point representing^ a monotonic variable on a number scale moves in one direction. It is inconvenient to consider constant quantities apart from variables and therefore we can regard a constant quantity as a spe cial case of a variable, i.e. a variable which all the time assumes
31
VARIABLES AND FUNCTIONS
only one fixed value (the same idea is used in mechanics when the state of rest is regarded as a special case of motion). The range of a constant consists of only one point. We say that a variable changing in a certain process has an upper hound if all the time it remains smaller than a constant (such a constant is called an upper bound of the variable; it is clear that a variable having an upper bound has in fact an infinitude of upper bounds because every constant greater than a given upper bound of the variable can serve as a new upper bound). We likewise define the notion of a lower bound of a variable. Of course, a variable may not have an upper or a lower bound (or either of them). If a
h
h
h
h
mmrnmkmmxmm — h
0
h
o~h
a
— *
a+h
Fig. 5
variable has both an upper bound and a lower bound it is sim ply called a bounded variable. Variables having upper bounds (lower bounds) are called bounded above (bounded below). When investigating different quantities we often use the notion of the absolute value of a quantity. As is w ell known from elemen tary mathematical courses, the notion is defined in the following way:  a  = a if a ^ 0 and  a \ = —a if a < 0 For instance,  5  = 5,  0  = 0 and \ —5 1 = 5 [that is I —5 I = =  (  5 ) = 5]. Absolute values possess the following simple properties: 1.  a. + b  ^ 1 a ) + I & I The inequality is strict in case a and b have opposite signs and it turns into the equality if otherwise. 2. For any a and b we have I ab  =  a  •  6  and YU? =  a  The significance of the last formula is sometimes underestimated in elementary mathematical courses and this may be the cause of different errors and false conclusions. The quantity  a — b  =  &— a  is equal to the distance between the points a and b lying on the number scale. The inequality I®  < / i (h > 0) defines the interval —h < ® < ft, and the ine quality  x — a  < h defines the interval —h < x — a < h, i.e. ? ~ « < ® < c + h. An interval of the form a — A < x < a + A is called an ^neighbourhood of the point a . The intervals are shaded
w Fig. 5.
32
INTRODUCTORY
MATHEMATICS
FOR
E N G IN EE R S
§ 2. Approximate Values of Quantities 6. The Notion of an Approximate Value. I t is usually impossible to speak about the absolutely precise value of a physical quantity. For example, we can never determine the exact value of the length of a real object. This is so not only because our measurements are imperfect but also because of a complex form of the body which makes it impossible to indicate exactly the points between which the length should be measured. If we recall that the object consists of molecules which are in permanent motion we see that the situa tion becomes still more complicated. Moreover, there is a vast majo rity of cases when the determination of a length with a great accuracy is inexpedient and senseless even when the modern level of measu rement techniques makes such an accuracy attainable. For instance, if we have to design or measure a dwellinghouse it would be obvio usly senseless to determine the sizes of the building with the accu racy to within 0.01 mm. The same can be said about masses, pres sures etc. The numerical values of almost all quantities in physics and engineering (for example, the values of all continuous variables) are therefore approximate. Mathematical operations on approximate values of quantities are called approximate calculations. There exists a special branch of science devoted to approximate calculations and we shall study some of its rules later on. A. N. Krylov (18631945) was one of the initiators of developing approximate calculations in the USSR. His book [28] (the first edition was in 1911) still retains its significance. The appropriate choice of a degree of accuracy for calculations, measurements or for manufacturing machine elements is a very important operation. When making such a choice one should take into account a great many factors, i.e. our requirements, technical means, economy etc. 7. Errors. Let A be the exact value and a an approximate value of a quantity. Then the error, th at is the deviation of the appro ximate value from the exact one, is equal to A — a. It may be posi tive or negative. As a rule, we do not know the error exactly since the exact value A is unknown. Therefore we usually consider the limiting errors a* and a„ which form an interval containing the true error: c*i < A — a < a 2, i.e. a  f  G ^ c A C a + a* Thus the value of the quantity A is estimated from two sides. For instance, the formula of the length L = 9io*.i mm means th at the true value of the length lies between 9—0.1 = 8.9 mm and 9 + 0.2 = = 9.2 mm. It is sometimes inconvenient to consider* two limiting errors and therefore we often indicate the maximum absolute error a , that is
33
VARIABLES AND FUNCTIONS
a value which exceeds the absolute value of the error:  A — a  < a , i.e. —a < A — a < a a —a < A < a + a
or
For example, suppose that the measurement of a length I results in the value 137 cm and that we can guarantee the accuracy of 0.5 cm. This means that we have a = 0.5 cm and 136.5 cm 0) where b is several times smaller than a (and there fore V a 2 — b2 ^ Y a 2 = a) we can transform the expression in the following way: n _
( a  ' \ / g *  ~ b 2) (f l + y fl2 _ 62 ) _______ 62 a + ~]/ a2 — 62
a f ~\fa2—62
The last expression no longer contains the undesirable difference. 10. Multiplication and Division of Approximate Numbers. General Remarks. Let us begin with an example. Suppose it is necessary to determine the area 5 of a rectangle with the sides a = 5.2 cm and b = 43.1 cm. It would be wrong to give the answer S — 5.2 x x 43.1 = 224.12 cm2. In fact, a is contained between 5.1 and 5.3, and b between 43.0 and£43.2. Thus, the area is contained between Si = 5.1 X 43.0 = 219.3 cm2 and S 2 = 5.3 x 43.2 = 228.96 cm2 We see* that alljL e decimal digitsvbeginning with the second one in the above value of S may be incorrect and therefore they may only
VARIABLES AND FUNCTIONS
37
lead to misconceptions. In this case the correct answer we must give is S = 2.2 X 102 cm2. By the way, we note that the calculation of o j and S 2 demonstrates the way that can he followed in estimating the results in other proWThus we see that in multiplying two numbers with two and three correct decimal digits wTe must retain two decimal digits in the answer. The same rule holds for the general case of multiplication of approximate numbers and also for their division: the number of correct decimal digits in the result must be equal to the smallest of the numbers of correct decimal digits in the factors (or in the divident and the divisor in the case of division). The reason for this general rule is that, in the first place, the operations of multi plication and division performed on approximate numbers yield the addition of the corresponding maximum relative errors (this will be shown in Sec. I X .ll) and, in the second place, the number of correct decimal digits and the maximum relative error indicate similar qualities connected with the degree of relative accuracy. In the example of calculating S the maximum relative error of b is considerably smaller than that of a and therefore 5S = 8a + f 6t « 6a, that is S has the same number of correct decimal digits as a. If the factors entering into a product are given with different numbers of correct decimal digits we must round the numbers before multiplying them and retain one reserve decimal digit which is discarded after the operation is performed. In case the factors have the same number of correct decimal digits but there are many factors (for instance, more than four) it is advisable to reduce the number of correct digits in the product by one. As an example, let us take the formula Q = 0.24 P R t which is applied to calculating heat generated by an electric current. In this case the answer cannot have more than two correct decimal digits because the coefficient 0.24 has only two correct digits. There fore there is no sense in taking I , R and t with more than three correct decimal digits (moreover, the third digit is taken only as a reserve digit). If a more accurate value of Q is desirable we must first of all specify the value of the coefficient. It should be noted that absolutely exact factors do not affect the choice of the number of correct decimal digits in a product. For instance, the coefficient 2 entering into the formula L = 2nr of the circumference of a circle is absolutely exact (we can write it as 2.0 or 2.00 etc.) and therefore the accuracy of calculations depends only on the number of correct decimal digits to which n and r are computed. , Let us take an example involving all the above rules. Let D — — 11.3 x 5.4 { 0.381 X 9.1 + 7.43 x 21.1. In order to estim ate
38
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
the magnitude of the summands we calculate them rounding to one correct decimal digit. Thus we get 500. 3.6 and 140. Hence the sum is of the order of several hundreds. The factor 5.4 entering into the first summand (which is the largest one) has only two correct decimal digits and thus the whole result must have two correct digits. Now, according to the rule of a reserve decimal digit, we must calculate to within unity and then round off the result to the nearest ten. Thus we obtain D — 690 + 3 + 157 = 850, i.e. D — = 8.5 X 102. Calculations with unnecessary digits are not only useless but oven misleading because they may give the illusion of an accuracy greater than that we actually have. The choice of a degree of accuracy of approximate quantities for performing mathematical operations on them is made in accordance with a general principle which states that all the degrees of accuracy which we choose must be coherent to each other at every stage of our calculations. This means that none of the degrees must be too great or too low. We shall take an example to illustrate the principle. Suppose we have to calculate the area of a rectangle by the formula S = ab. Let a be measured or calculated to three correct decimal digits. Then we must take b also with three correct digits because the fourth decimal digit of b would be useless whereas if we determined b only with two correct digits the efforts applied to finding the third digit of a would be futile. Therefore when we calculate a product it is convenient to take the factors (at least those factors which are difficult to determine) with the same number of correct decimal digits. Similarly, the summands entering into a sum must be taken with the same number of decimal digits to the right of the decimal point. Here we give an example. Let the expression M — ab + cd be calculated and let it be known that a « 30, b « 6, c « 0.1 and d « 40. Suppose that a is taken with three correct decimal digits. What number of correct digits should be chosen for 6, c and d? It is clear that we must take three correct decimal digits for b according to the accuracy of a. Further, we have ab « 180 and c d » 4. This implies that for calculating M with three correct digits (the accuracy of a makes it impossible to obtain M with more than three correct digits) it is sufficient to determine c and d with only one correct decimal digit. If it is not too difficult the accuracies of b, c and d should be increased by one decimal digit but the extra digit is only a reserve one. When performing practical calculations we often face a problem which is in some sense inverse to the above problem. The degree of accuracy of a desired result is sometimes set beforehand according to some prerequisites and then it is required to determine the necessary
VARIABLES AND FUNCTIONS
39
degrees of accuracy of the quantities involved into the calculations (and the accuracy of the calculations). Some of the quantities may be obtained as a result of an experiment and therefore our discussion also applies to the determination of a desirable precision of an experiment. The solution of the inverse problem is based on the rules of approximate calculations we have studied here. For example, suppose we have to calculate the total surface area of a circular cy linder by the formula S — n (DH + ^ ) . Let it be approxima tely known that D m 20 cm and H m 2 cm. Then S m 700 cm2 (check it up!). Now turning to the inverse problem and reasoning as in the preceding paragraph we see that if, for instance, we want to have the result with three correct decimal digits, i.e. with an accuracy of 1 cm2, then jt and D should he taken with three correct decimal digits and H with two correct digits. Thus, measuring D and H we must attain the accuracy of 1 mm. It is better to calculate with a reserve decimal digit, and jt should also he taken with a re serve digit. But if we wanted to have more accurate values of D and H we would have to perfect our measuring instruments. The rules of determining degrees of accuracy for more complicated formulas will be given in Secs. IV.10 and I X .l l.
§ 3. Functions and Graphs 11. Functional Relation. When investigating a phenomenon or a problem we often deal with several variables which are interrelated so that a change of one of the variables affects the values of the others. Then we say that there is a functional relation between the variables. For example, suppose a mass of a gas is kept under chan ging conditions. Then there is a functional relation between the volume V, the temperature T and the pressure p of the gas because, as is well known from physics, the quantities are interrelated. We also have a functional relation between the area of a circle and its radius, between the distance passed over in a process of motion and the time taken and so forth. Usually it is possible to pick out certain variables from a number of interrelated quantities such that the values of the variables can be taken arbitrarily whereas the values of the other quantities are determined by the values of the variables entering into the first gronp. The variables of the first type are called independent vari ables (or arguments) and the variables of the second type are called dependent variables (or functions). As an example let us consider the relationship between the area S of a circle and the length R of its radius. It is natural to regard R as an independent variable and choose its values arbitrarily; then the area computed by the tormula S = ni?2 is a dependent variable in this functional relation.
40
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
In the example of the mass of gas we could have taken V and T as independent variables. Then the variable p (the pressure) would have been regarded as a dependent variable. A rule (a law) according to which to the values of independent variables there correspond the values of a dependent variable is called a function. Thus, every time there is a law of correspondence between the values of variables we say that there is a functional relation. The concept of a function is one of the most important mathematical notions. By the way, the term “function” is sometimes used in a different sense. As it has been mentioned, independent variables are called arguments and a dependent variable itself is called a function. Such a twofold sense of the term does not, however, lead to any misunderstandings. It should be noted that when we have a functional relation between variables the distinction between the independent variables and the dependent ones is sometimes conditional. For instance, in the example of the mass of gas we could have taken T and p as indepen dent variables and 7 as a function. We can easily construct the scheme of an experiment in which T and p can be varied arbitrarily whereas V depends on T and p. Of course, the choice of variables which are regarded as independent quantities may be important in some cases. The choice should be made in a natural and conve nient way in accordance with the circumstances. Functions may depend on one argument (as in the example of the area of a circle) or on two or more arguments. In the first two chapters of our course we shall consider (almost without exceptions) functions of one independent variable. We must note that when we regard a quantity y as a function of an independent variable x we do not necessarily suppose that there is a meaningful causal relationship between the variables. It is quite sufficient if there exists a rule which attributes a certain value of y to each x even if we do not know this law of correspon dence. For example, the temperature 0 at a point in space can be regarded as a function of time t since it is clear that we always have a certain temperature at the point at each moment t, that is to the values of t there correspond the values of 0, although the variations of 0 cannot be simply accounted for by the changes of t since in reality these variations are determined by some complicated physical laws. 12. Notation. If a variable y is a function of a variable x we usually write y = / (x) (this is read as “y is equal to / of x”) where / is the sign of a function. If we make x assume certain particular (concrete) values the function will assume its particular values. For instance, let y = / (x) be of the form y = x2. Then y = 4 for x = 2, y = 0.36 for x = —0.6 etc. This can be written as / (2) = 4, / (—0.6) = 0.36 and so on, or as y\x=2 = 4, yx==_0.o =
VARIABLES AND FUNCTIONS
41
= 0.36 etc. The vertical lines in the last expressions are the signs of substitution which mean that we substitute the values of x for the argument. ' . , The notation y = / (x) is used when the concrete expression ol a function is too complicated or when we do not know the expression. It is also used for formulating general rules and properties of all functions or of many concrete functions. For example, the formula (fl  f 6)3 = as _j_ 3a2& 30&a _j. &3 which is well known from al gebra is written in letters. Here the letters a and b are not concrete numbers but they can be replaced by any numbers. If we consider several functions simultaneously we can use, besi des /, any other letters: F, cp, ® etc. We can also introduce diffe rent subscripts, superscripts, and other indices: / 2, F® etc. At the same time when we consider different problems we can denote different functions by the same letter /. We remind the reader that we have a similar situation in algebra: a letter, say the letter a, may denote different quantities in different problems, but we must not denote by a different quantities entering into one and the same problem. On the other hand, different quantities may sometimes be connected by one and the same functional rela tion. In such a case we can use one and the same letter / because / designates the law of dependence of one quantity upon another and is irrelevant to the way the quantities are denoted. For example, if y = s 3, z = u6 and v = t3 then we can write y — f (x), and v = f (f). In this case the sign / indicates raising to the third power whereas
(x) = (x2 — 3x) (2x + 1) = 2x3 — 5x2 — 3x; / (9 (x)) = [9 (x)]2 — 39 (x) = (2x + l) 2 — 3 (2x 4 1) = — ^2*2 2 x 2 * cp (f (x)) = 2/ (x) + 1 = 2 (x2 — 3x) + 1 = 2x2 — 6x + 1; / (X } s) = (x f 5)2 — ~ 3 {X j~ s) = x2 f 2x5 ~ 52 — 3x — 35 [this is a function of two variables which can be denoted by © (x, s)] etc. In particular, in the above examples we come across the operation of composing “a function of a function” or, as it is usually said, we deal with a composite function. A composite function is usually obtained in the following way. Let a variable y depend on a variable u and let u, in its turn, depend on a variable x. Thus, y= / {u) and u = tp (x). Then variations in x change u and therefore yalso varies. Hence, y is a function of x of the form y = f (q> (x)). Thus we obtain a composite function. In this case u is an intermediate variable. There may also be several intermediate variables. If we only want to designate that y is a function of x avoiding all the intermediate operations we can write y = y (x). For instance, in the examples in Sec. 11 we could have written S = S (R ) or P = p { V , T) and V = V (T , p). 13. Methods of Representing Functions. If we intend to investi gate a function, that is a dependence of one quantity upon another, the function must be represented in a certain way. There are several methods of representing functions. The analytical method (i.e. representing a function by a formula) is one of the most widely used methods in mathematics. This method describes the mathematical operations which should be performed on the independent variable to obtain the value of the function. The operations are indicated by a formula. For example, the formula y = x2 — 2x says that in order to compute the value of the function y we must raise the corresponding value of the argument to the second power and then subtract the doubled value of the argument from the result. The analytical method is compact (i.e. formulas usually occupy little space), it can be easily reproduced (i.e. it is not difficult to rewrite a formula). Besides, it is the most suitable method for per forming mathematical operations on functions. Here we mean the algebraic operations (addition, multiplication and so on), the ope rations of higher mathematics (differentiation, integration and the like), and others. But the method is not visual enough (this means that when we have a formula it is not always possible to visualize the character of dependence of the function upon its argument).
43
VARIABLES AND FUNCTIONS
The calculation o! particular values of a function ^presented by a formula (in case the values are needed) may be a complicated ope ration In addition, not all functions can be represented by a for^ u~ la, and it may be inconvenient to put down a formula even when exists^ometimes neCessaTy to use several different formulas to represent a function on different parts of the range of its argument. For example, let a material point fall without an initial velocity s=Q
I Position of the 1/ point at moment t u g tz
Position of }the point a t moment t Position o f the platform a t moment t =0 Position of the platform at moment r j vf
Fig. 6
on a platform which is moving downwards uniformly with the velo city v, the distance between the point and the platform at the mo ment t = 0 being equal to h. Then the path s covered by the point is a function of the time t , i.e. s = / (t). According to Fig. 6 the relationship is determined by the formula [£ 1 » = / 0) which guarantee the possibility of obtaining real values. Then the problem of determining the domain of defi nition is reduced to solving these inequalities. If an independent variable is discrete the domain of definition of the corresponding function consists of discrete (separate) points. For instance, if / (x) = x! = 1 y 2 ) . . . x x then x can assume only the values 1, 2, 3, . . . . I n case a discrete argument takes on only integer values, as in the above example, it is usually denoted hy x but by the letters n, m, k and the like whereas the values J W> / (2)» • • • > / ( » ) , . . . are denoted as at, o2, . . ., an __ .
48
INTRODUCTORY MATHEMATICS
FOR
EN G IN EE R S
In such a case we say that there is a sequence; for example, a geometric progression of the form flj —
&2
C LQ ,
« •
=
C LQ
• • •
is an example of a sequence etc. The graph of a function of a discrete argument is not a continuous line but consists of discrete points (see Fig. 11). The range of a dependent variable, th at is the set of all the values assumed by a function as its argument runs over the domain of definition of the function, is called the range of the function. For an■ example, the domain of definition o Qj —3! 5of the function y = x2 is the inter 5val —oo < x < oo and the range a„=n' 4of this function is the interval 30 ^ y 0 if there is an identity of the form / (x f A) = / (*). Such a function behaves in the same way on each of the intervals . . . , a — 2A < z < fl — A, a — A < x < a, a < z < a  f A, a + A < £ < u + 2A, . . . where a is an arbitrary number. Therefore in order to investigate the function it is sufficient to consider its behaviour on one of the
intervals (see Fig. 16). The equality / (z + A) — f (a:) is illustrated in Fig. 16 for one of the values of x. A function y = / (z) is called an even function if it does not change its value when the sign of the argument is changed, that is if / (—x) = / (z). The examples of even functions are y — z 2, y — z 6, y = cos z etc. Fig. 17 shows that the graph of an even function is symmetric with respect to the axis of ordinates. A function / (z) is called odd in case it is multiplied by —1 when the sign of the argument is changed, that is / (—z) = —/ (z). The examples of odd functions are y = z, y — z 5, y — sin z etc. Fig. 18 illustrates the fact that the graph of an odd function is symmetric with respect to the origin of the coordinate system. It should be noted that in the general case a function may be neither even nor odd; for example, this is the case with the functions y — 1 + sin z, y = 1 — z, y = 2X. y = log z etc. 17. Algebraic Glassification of Functions. Functions represented by a single formula (see Sec. 13) are classified depending on the
52
INTRODUCTORY
MATHEMATICS
FOR
EN G IN E E R S
necessary algebraic operations which should he performed on the values of the argument in order to obtain the values of the functions. If only the operations of addition, subtraction and multiplication are used, and also the operation of raising to a positive integral power which is a special case of multiplication, the function is called a polynomial or an entire (integral) rational function; in
forming a polynomial arbitrary constant coefficients can be used. Examples of polynomials are y — a? — 2x + 3, y ~ x2, y = 3, y = V ^ r3, y = axx2 — 2 and the like. On the other hand, the functions y = and y = z3 + 2 x are not polynomials injthe sense of the above definition. Every polynomial is characte rized by its degree which is the highest of the exponents of powers of the independent variable entering into the expression of a poly nomial; for instance, the degrees of the above written polynomials are, respectively, 3, 2, 0, 3, and 2. Rational functions form a wider class of functions: these are the functions which involve the additional operation of division. If a rational function is not an entire function it is called a fractional rational function. An example of such a function is _ x
x—1
y ~ XV 2 —3
ax —
x~2~l
According to the rules of elementary algebra every rational function can be represented as a ratio of two polynomials after all the summands entering into the expression of the function are reduced to a common denominator. There is a still wider class of functions whose analytical expres sions may involve an additional operation of extracting roots. This is the class of algebraic functions. If an algebraic function is not rational it is called irrational. An example of an irrational function is y = x2 — ~ + Y — 1
VARIABLES AND FUNCTIONS
53
Functions which are not algebraic are called transcendental. Examples of transcendental functions are y = sin x, y — x + _l x y = 2Xj y = log ^ otc. We P°ln^ out that the last two functions are transcendental despite the fact that they are some times traditionally considered in elementary courses on algebra. All these definitions are automatically extended to functions of several independent variables. The only new fact is the definition of the degree of a polynomial in several variables: it is defined as the greatest of the sums of the exponents of arguments entering into the monomials which are the summands in the expression of the polynomial. For instance, the function / (x, y) = xty — x h f + i is a poly nomial of the sixth degree in x and y. But if we regard y as fixed the same function w ill be a polynomial of the fourth degree in x. A polynomial of the first degree and a polynomial of the second degree are called, respectively, a linear function and a quadratic function. A polynomial of the third degree is called a cubic function and so on. These terms are applied for any number of independent variables. 18. Elementary Functions. We first enumerate the basic elemen tary functions studied in elementary mathematical courses: y = x 1 (where a is constant) is a power function; y = ax (where a is constant) is an exponential function; y — logo x (where a is constant) is a logarithmic function; y  sin x, y = cos x, y — tan x and y — cot x are trigonome tric functions (circular functions); y — arc sin x, y = arc cos x etc. are inverse trigonometric func tions (inverse circular functions). Elementary functions are all the functions which can be obtained from basic elementary functions by means of algebraic operations (with any numerical coefficients) and the operation of composing a function of a function (see Sec. 12). In view of this definition all the algebraic functions are elementary. But very many transcendental functions are also elementary, for example, the functions y — x + log sin x, y = 210®tan *+sin x etc. (one may come across very complicated expressions of this type). The class of elementary functions includes the greater part of functions treated in general courses of higher mathematics. As an example of a function which is not elementary we can mention p = x! (but we do not give now the general definition of the function involving nonintegral values of the argument. On this question see Sec. XIV.17). Many nonelementary functions are widely used in special branches of mathematics and its applications. Many of these functions have been investigated in detail and therefore the traditional classification into elementary and nonelementary func tions may now be considered out of date.
54
INTRO DUCTO RY
MATHEMATICS
FOR
EN G IN E E R S
19. Transforming Graphs. It often happens that we know the graph of a function and it is necessary to construct the graphs of some other functions which can be expressed in a certain way in terms of the former graph. Here we give several examples of trans forming graphs in this manner. Let the graph of a function y = / (x) be given. It is required to construct the graphs of the functions z = / (x) + a and u — / (x+b) where a and b are some constants. The values of the quantities z and u will be represented on the same axis of ordinates as that for y (see Fig. 19). Then we have z = y f a for any x and therefore
Fig. 19
Fig. 20
the graph of the function z {x) can be obtained from the graph o\ the function y (x) by translating the latter along the i/axis by the distance a in the positive direction of the axis in case a > 0 . Such a translation is depicted in Fig. 19 where each of the vertical line segments has the length a. As for the graph of the function u (x) one may think that it is obtained from the graph of y (x) by transla ting the latter along the xaxis by the distance b in the positive direction in case b > 0 . But this conclusion is wrong; in fact we should displace the graph of y (x) by the amount b in the negative direction (if 6 > 0) in order to obtain the graph of u (x). Indeed, the value u = y is obtained if we take the value of the argument foi u which is smaller by b than the corresponding value of the argument for y since u = / [(x — b) + b] = / (x) = y. Of course, if a < C or b < 0 the corresponding translation should be carried out in the opposite direction. On the other hand, when we say, for example, “to displace upwards by (—3)” we mean, in fact, “to displace down wards by (+ 3 )” etc. Therefore it is permissible to say that we trans late a graph in a certain direction by an amount h no m atter whal the sign of h is.
VARIABLES AND FUNCTIONS
55
The graphs of the functions v — kf (x) and w = f (fez) are construc ted in a similar way (see Fig. 20). The graph of the function v(x) is obtained from the graph of y (x) by the uniform M old expansion of the latter in the direction of the y ~axis so that all the distances from the points of the graph of y (x) to the xaxis should increase k times (in case k > 1). Indeed, the points of the graph of v (x) which have the same abscissas as the points of the graph of y (#) have the ordinates k times the ordinates of y (x). The graph of the function w (x) is obtained from the graph of the function y (x) by the uniform contraction of the latter toward the yaxis with the M old decrease (in case k > 1 ) of all the distances of the points of the graph of
y (x) to the yaxis. Actually, we have w
= / ( &jjr) — / (*) —
= y (x). Of course, what we have said is literally true if k > 1 . In case 0 < k < 1 the expansion is replaced by the contraction and vice versa. But again, when we say, for example, “the jfold expansion” we actually mean “the 3fold contraction” and the like. Therefore we can say that we perform a M old expansion (or contrac tion) without specifying the magnitude of k. In conclusion we re mark that if A: < 0 we should additionally apply the operation of forming the corresponding mirror images of the graphs of v (x) and w (x) (for the function v (x) the mirror image must be taken about the xaxis and for the function w (x) the mirror image must be taken with respect to the yaxis). Now combining the above results we can say that the graph of the function y = kf (mx + &){ a can be obtained from the graph of the function y = / (x) by means of the following transformations (performed in succession): the parallel translation along the xaxis Iwhich yields the graph of y = / (x f 6)1, the contraction [which results in the graph of y = / (mx + &)], the expansion [which gives the graph of y = kf (mx + b)) and one more final translation along the yaxis resulting in the desired graph of y = kf (mx f b) J a. (If necessary, the corresponding mirror images should also be taken.) The same results can be obtained by the corresponding operations on the coordinate axes without changing the graph. For example, instead of displacing the graph to the right we can translate the axes to the left, or, in other words, displace the origin (from which x is reckoned) to the left. Similarly, instead of expanding the graph and increasing the distances from the xaxis k times we can decrease the corresponding unit of length for the yaxis k times. We can perform arithmetical operations on functions represented graphically. For example, Fig. 21 illustrates the graphical addition of two functions: the graphs of / (x) and
b whereas it is threevalued for a < x < b. When separa ting the function into branches it is natural to regard the arc AB ns the graph of the first branch, the arc BC as the graph of the second branch and the arc CD as th at of the third one.
y
Fig. ‘24
Fig. 23
In connection with the question of multiplevalued functions discussed above we can note that for some functions to every value of the independent variable there corresponds an entire interval of values of the function. For instance, the relation between the height of a person and his possible weight is an example of such a functional relation. Functions of this land are usually investigated in the theory of probabilities (see Sec. X V III.16) and they will not occur in other chapters of our course. 21. Inverse Functions. Suppose we are given a function U = / (*) 20. Imp now t ake different values of y and find the c o r re s p o n d in g is "defined *b’ th at us c^ oosc the former dependent variable as „ *__ n _t and retrard the former indeDendent variable as a func
59
VARIABLES AND FUNCTIONS
function. Therefore, if we wanted to denote, as we usually did, the independent variable by x and the dependent variable by y for the inverse function we should simply have to substitute x for y and y for x in (6). Hence, using the new notation we rewrite the relation which defines the inverse function in the form
x =f{y)
.
(7)
Thus, the inverse function turns out to be represented in an im plicit form and therefore (see Sec. 20) it is, generally speaking, multiple valued. We can easily establish a condition which guarantees the
singlevaluedness of an inverse function: this is the monotonicity of the original function. Indeed, this being so, we obtain a certain uniquely defined value x = x (y) for each given value of y (see Fig. 25). Examples. The inverse function of the function y = a? is defined by the equality x = y3, that is y = y / x; the inverse function of y —x“ is the twovalued function y ~ ± V x . Equalities (6) and (7) differ only in the interchange of the notation of the quantities x and y, that is in the interchange of their roles. Therefore we see, as it is shown in Fig. 26, that the graph of the inverse function is obtained as. a mirror image of the graph of the original function about the bisector of the angle between the coor dinate axes (the bisector is represented by the dotted line in Fig. 26). The points M and M' in Fig. 26 both correspond to one and the same equality of the form b = / (a). We remark in conclusion that if the function x (y) is the inverse function of the function y (x) then, conversely, the latter is the inverse function of the former.
60
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
§ 4. Review of Basic Functions Many of the functions which we are going to discuss here are stu died in elementary, mathematical courses. We shall consider them here because of their significance. 22. Linear Function. The general form of a linear function (see the end of Sec. 17) is y = ax + b (8) where a and b are constant coefficients. The graph of a linear function is a straight line (see Fig. 27). The coefficient a is called the slope of the straight line. The greater  a \ (i.e. the greater a in its absolute value), the steeper the slope of the straight line (with regard to the xaxis). If the argument of a func tion changes from a value x 0 to a certain value x it receives an in crement Ax (which is equal to x — x0)*. Then the function recei ves the corresponding increment Ay [which is equal to y — yo = = f (x) — / (x0)L In our case y = f (x) — ax b and therefore y 0 — ax o + b and y = ax f b. Consequently, y — y 0 = a (x — x0), i.e. Ay — a Ax. This implies ^
=a
( if A x ^ O )
(9)
Thus, the ratio of the increment of a linear function to the incre ment of the argument is constant and equal to the slope of the graph. The increment of a linear function is directly proportional to the increment of the argument. In Fig. 27 the case when a > 0 is shown. If a < 0 the straight line is drawn downwards to the right (see Fig. 28). In case a = 0 the straight line is parallel to the xaxis; in this case the function is constant and thus we obtain the graph of a constant. The property of the increment of a linear function forms the basis for the socalled linear interpolation which is used even in elemen tary mathematical courses. The idea of this method is the following. Suppose we know the values of a function y — f (x) (its graph is depicted in Fig. 29 by the dotted line) for x = x0 and for x = x 0 + h: f fro) = I/o, / (*o + h) = yt * The Greek letter A (delta) is used to denote an increment. The symbol Ax should be regarded as an indivisible symbol and by no means as the product of A by x. An “increment” is understood in the algebraic sense, i.e. it can be positive, negative or equal to zero.
VARIABLES AND FUNCTIONS
61
but the intermediate values of the function for the values of x lying between i = and x = x 0 + h are unknown. Then we approxi mately replace the given function by a linear function which assu mes the same values for x = x 0 and for x — x 0 + h, that is we replace the arc u AE by the straight line segment AE. The simila rity of the triangles ABC and ADE then implies y — yp ___ Ui— yo
x —xq
h
i.e. y = y 0+   Y l   ( x ~ x °)
Such a replacement is possible in case the function / (a:) slightly differs from the linear function on the interval between x 0 and f h. The interpolation method is widely used, in particular,
for tables with a sufficiently small step when the successive values of the function differ slightly from each other. More precise methods of interpolation will be discussed in Secs. V.68. The linear extra polation (see Sec. 13) is performed in like manner. Formula (9) and Fig. 27 imply that a = tan q>, i.e. the slope of ■a straight line is equal to the tangent of the angle of inclination of the line to the axis of abscissas. If the quantities x and y have certain dimensions the slope also has a dimension. Formula (8) shows that [hi = [y] and [ax] = [y] which implies [a] = (the dimensions of coefficients entering . I*17] into other formulas can be determined similarly). The geometrical meaning of the slope can be easily interpreted in the general case: if lx units of length for the icaxis correspond to the unit measure of the quantity x and lv units of length for the yaxis correspond to
62
INTRODUCTORY
MATHEMATICS
FOR
E N G IN E E R S
the unit measure of the quantity y (lx and ly are the socalled scale factors) then the sides A B and BC of the triangle ABC in Fig. 27 are of lengths lx Ax and lv Ay, respectively. Consequently, ly Ay fay I tan(p = £A F and a = A7 = 1 7 tan
. 23. Quadratic Function. The general form of a quadratic func tion is y — ax2 f bx + c From elementary mathematical courses it is known that the graph of a quadratic function is a parabola. In the simplest case when a = 1, 6 = 0 and c = 0, i.e. y = x2, the graph has the form depicted in Fig. 30. Then the function is even and the yaxis is the symmetry axis of the graph (the axis of the parabola). The intersection point
of a parabola with its axis is called the vertex of the parabola. The vertex in Fig. 30 is placed at the origin of the coordinate system. In the general case when a 0, b and c are arbitrary numbers the parabola is obtained from the parabola depicted in Fig. 30 by the operations of uniform expansion and parallel translation. To determine the position of the vertex we can apply the socalled me thod of completing a square which we shall demonstrate here by considering a concrete numerical example. Let the quadratic func tion y = 2.r2 — 3x + 1 be given*. Then we perform the following * In practice we usually have quadratic functions (trinomials) whose coeffi cients are approximate numbers (in contrast to the above trinomial with exact coefficients). For instance, we can take the trinomial y 2.17a;2 — 3.21a; + 0.84 and the like. But if we investigate the case with exact coefficients we can easily pass to a more complicated case. The comment also refers to further exam ples of this type.
VARIABLES AND FUNCTIONS
63:
simple transformations: j, = 2 ( x»  4 x + 4 ) = 2 [ ( x —  ) 2 ( 4 ) 2+5] = 2 ( i — I ) ' — S’
Consequently (see Sec. 19) we obtain the soughtfor graph ^from the parabola depicted in Fig. 30 by translating the parabola unit. of length to the right, expanding it along the direction of the yaxis with the twofold increase of the distances from the points of the graph to the xaxis and, finally, by transla ting!unit downwards. The graph thus obtained is Q
shown in Fig. 31. To construct a more accurate graph we cant additionally take several values of x and determine the cor responding values of y which enables us to construct the corres ponding points of the graph. For instance, we have y — 1 for x — 0. y = 0 for x — 1 and y =■ 3 for x = 2; the corresponding points are indicated on the graph. The vertex of the constructed parabola is situated at the point M with the coordinates x = g and y = — i. The parabola is “narrower” than the one depicted in Fig. 30 (with the same unit of length). Generally, the greater  a , the narrower the parabola. If a 1 then the greater n , the greater the values of the function. In Fig. 33 we see the graphs for n = 1, 2, 3 and 4. While constructing the parts of the graphs of y = xn for the values x < 0 ■one should lake into account that the function i/ = re71 is even for even n and odd for odd n. In particular, let us consider in detail the graph of the function y ~ a r (the cubic parabola). The graph is convex upwards (or concave downwards) for x < 0, that is it lies under the tangent drawn at any of its points. For x > 0 the graph
is convex downwards (or concave upwards). If we pass from left to Tight through the origin of the coordinate system the direction of convexity changes to the opposite one. The tangent to the graph at the origin coincides with the ;raxis but at the point of tangency 0 but for x < 0 as well because, for negative numbers, we can extract real roots with odd indices of radicals. In particular, let us take the graph of the function y — x* (the semicubical parabola) depicted in Fig. 35. The graph first approaches
Fig. 35
the origin of the coordinates (for example, when we pass from left to right) and then departs from the origin. At the origin this curve has the socalled spinode, or cusp. Later on we shall investigate some other curves having cusps. Finally, let us take the case of a negative n {n = — m 0; we leave to the reader the construction of the parts of the graphs corresponding to x < 0. All these graphs stretch along the coordinate axes and approach them unlimitedly as x or y approaches infinity. Gene rally, when a curve and a straight line have a mutual disposition of such a kind the straight line is called the asymptote of the curve. Hence, each of the above graphs has two asymptotes which are the coordinate axes. 50141
66
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
One must not think th at a curve cannot intersect its asymptote in all other cases. For example, investigating the socalled damped oscillations we obtain a graph of the form shown in Fig. 37. Here the zaxis is the asymptote of the graph which intersects it infinitely many times. 25. LinearFractional Function. A linearfractional function has the general form ax + 6
0 = cx\~d
In the simplest case when a = d — 0 we obtain, denoting
(ii)
= k,
the expression y = ~ which describes the inverse proportional relation. The corresponding graph, as is well known from elementary
mathematical courses, is called a hyperbola. The graph is depicted in Fig. 38 for the two cases k > 0 and k 0 (a ^ 1). The graphs of logarithmic functions are shown for different bases in Fig. 40. They have neither symmetry axes nor centres of symmetry but have an asymptote which is the z/axis. All the logarithmic functions are proportional to each other since taking logarithms to the base b of both sides of the equality a log&x = = x we get log;, x = loga a:*logj, a = /cloga x
{k = logb a = j — j )
(13)
Therefore we can obtain all the graphs depicted in Fig. 40 by expanding or contracting one of them along the direction of the yaxis with a uniform increase or, respectively, with a uniform dec rease of the distances of the points of the graph from the xaxis. Now let us consider the angles of inter section of the graphs with the xaxis. According to the general definition of the angle between inter secting curves as the angle between the tangents to the curves at the point of their intersection, we mean here the angles formed by the tan gents to the graphs with the xaxis. When the graph is expanded or contracted in the way described above the tangent rotates about the*point of intersection. We see that the tangent has a very slant inclination (to the xaxis) for very large values of a and a very steep inclination for values of a close to 1. For a certain value of a the angle of intersection of the graph of logarithmic function (12) with the xaxis is equal to 45°. This value of a is denoted by the let ter e. It plays an important role in mathematics as we shall see later. We see in Fig. 40 th at the angle of intersection is greater than 45° for a = 2 and smaller than 45° for a — 4; hence, the number e lies between the limits 2 and 4. More accurate calculations w7hich will be described in Sec. IV.16 show that e = 2.71828 with an accuracy of 10"5. The notation e for this number was introduced by Euler. Logarithms to the base e are called natural logarithms [Napierian logarithms after the Scottish mathematician J. Napier (15501617)].
69
VARIABLES AND FUNCTIONS
They are denoted as In x = log, x. The graph ol the natural loga rithm is shown in Fig. 41. A logarithm to any other base can be expressed in terms of the natural logarithms in accordance with formula (13): log«x =  I ^
'
(i4 )
Hence, the formulas for passing from the common (decimal) loga rithms to the natural ones and vice versa are log x = 0.4S43 In x and In x — 2.303 log x where the values of the proportionality factors are accurate to four decimal places. Besides the natural logarithms we also use the common logarithms (in numerical calculations) and the logarithms to the hase 2^ (in information theory and some other branches of modern mathematics).
27. Exponential Function. An exponential function is a function of the form y = a* (15) The function is defined for all x, and we always consider the values ® > 0 (because raising a < 0 to a fractional power may result in an imaginary number). Equality (15) can be obtained from formu la (12) if we solve it for x (which yields x = av) and then inter change x and y. Consequently (see Sec. 21) the exponential function and the logarithmic function are the inverse functions with respect to each other. Therefore the graphs of exponential functions which are depicted in Fig. 42 for different bases a are obtained as the mirror images of
70
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
the corresponding graphs (shown in Fig. 40) with respect to the bi sector of the angle between the coordinate axes. If a > 1 the expo nential function is an increasing function, and the greater a, the greater the rate of its increase. In case 0 < a < 1 the exponential function is a decreasing function. We often deal with the exponential function with a = e. In this case there is special notation: y = ex == exp #. Any exponential function with an arbitrary base a can be reduced to the base e; indeed, the definition of a logarithm implies that a ein a an(i therefore ax — (eIn a)x = ekx where k = In a. 28. Hyperbolic Functions. The hyperbolic sine, cosine and tan gent are, respectively, the functions ex — €~x e*—ex ex + ex sinh x sinh x cosh x tanh# ex^ex 2 2 cosh x At first these terms sound strange but their genuine sense (for exam ple, the connection between sin x and sinh x or between the hyperbolic functions and a hyperbola) will be explained only when we get to Secs. V III.4 and XIV.8. Let us now establish some formulas connecting these functions. Squaring the first two equalities we get e2*_2i 0, and tanh x « —1 for large  x I and x < 0.
We sometimes consider the inverse hyperbolic functions which are denoted as sinh1 x, cosh1 x and tanh1 x, respectively. Figs. 43 and 44 show that the first and the third functions are singlevalued (compare with Fig. 25) whereas the second one is twovalued. All
72
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
these functions can be expressed in terms of logarithms. In fact, let; for example, y = sinh"1#. Then, by the definition of an inverse function, we have . , eV~ eV x = smh y = — ^— i.e. ev _ ev  2x = 0,
e2I/  2xev  1 = 0
which implies ev — x ± V x 2 f 1. The lefthand side being positive, the righthand side should also be positive. Therefore we can take only in front of the radical. Now taking logarithms we obtain y = sinh1 x = In (x f J^x2 f 1)
(16)
29. Trigonometric Functions. The function y = sin:r with period 2ji is well known from courses on trigonometry. Its graph (the sinusoid, the sine curve) is represented in Fig. 45. The function
is odd, has no points of discontinuity and is bounded (its values lie between the limits —1 and + 1 ). We have cos x = sin ^ + y ) and therefore the graph of the function cos x is the same sinusoid but translated ~ units of length to the left; this graph is also shown in Fig. 45. In many applications we encounter a sinusoidal, “harmo nic11 relation of the form y = M sin (cot + a)
(17)
where the independent variable t is interpreted as time, the con stant M > 0 is called an amplitude and cd > 0 is called a frequency (circular, angular frequency). The sum co£ + a is called a phase and the constant a is an initial phase which is obtained from the phase by substituting t = 0 for t. We can easily investigate in what way parameters Jkf, co and a affect the form and the disposition of the sinusoid (compare with Sec. 19). The amplitude M increases the range of the sinusoid and brings it from —M to M, the frequency
73
VARIABLES AND FUNCTIONS
to changes the period 2n into T — — and tlie presence of the initial phase' a displaces the sinusoid to the left by the distance ©f fa = © \\ t t ^\ and therefore the value C — is 1 G> / O
[since
added
to
the argument]. The graph thus obtained is represented in Fig. 46. A function of form (17) is obtained, in particular, when we trans form the expression A cos cat f B sin cot. The righthand side of (17) can be rewritten in the form M sin a • cos cot + f M cos a • sin cot and thus in order to obtain the equality
A cos cot B sin ©t == = M sin (at + a) (18) we must have A — M sin a and B = M cos a. From this it is easy to find M and a: M = S A z \B2 and tana =
the quarter in
Fig. 46 which a should be taken is defined by the signs of sin a and cos a, i.e. by the signs of A and B . In case the independent variable is interpreted not as tim e but as a geometrical coordinate the sinusoidal relation is usually written m the form y = M sin (kx  f a) instead of (17). In this case k
is called a wavenumber and X = ^2 is a wavelength. The function y — tan x has the period jt since tan (x } ji) == S3 tan a:. It has the points of discontinuity at x = ~ £ 71
2
n » • • •
(this can be written in the general form as x
£ —
2
f
k it
vhere k — 0. ± 1 . ± 2 , . . .). Indeed, at these points cos x = 0 and therefore tan x — ± o o . The graph of the function (the tangent curve) is represented in Fig. 47; it consists of an infinitude of similar components and has infinitely many asymptotes. The graph of the Junction y = cot x is also shown in Fig. 47. We have V
cot x = — tan (x — —j
74
INTRODUCTORY
MATHEMATICS
FOR
ENG INEERS
and therefore the curve has the same form but its disposition is changed in the corresponding way. The function y — Arc sin x is the inverse function of y = sin x and therefore its graph (see Fig. 48) is the mirror image of the graph of y — sin x about the bisector of the first quadrant angle. This function is multiplevalued (more precisely, infinitevalued) and
therefore (see Secs. 2021) one usually considers its principal branch (the principal value of the arc sine) which is shown in Fig. 48 in heavy line; this branch is denoted as y = arc sin x,
—  ^ a r c sin
and is a singlevalued function." Other branches of the function have no special names. The functions y = Arc cos x and y = Arc tan x can be investiga ted in a similar way and we leave this to the reader. We should note in conclusion that we shall always deal with dimensionless (abstract) values of arc sin x . For example, we have
2arc sin 1 2 * = 2 1' 57 2 .9 7 * Similarly, the values of the function y = sin x are taken for di mensionless values of x . We mean here th at the sine of a number x * The last two equalities are approximate. If one intends to stress this
n
fact one writes 22 « 2 1*57. We are not going to mention stipulations of this hind in the future.
VARIABLES AND FUNCTIONS
75
is the sine of the angle of x radians. For instance, sin 1 = sin 57°1S' = = 0.8415. 30. Empirical Formulas. We have already mentioned (see Sec. 13) that an experiment often results in a function y = f (x) which we are interested in and represents the function in a tabular form (2). In such a case the problem of selecting an appropriate empirical formula for the function may arise. We usually begin with repre senting the values of the function on the graph paper or some other appropriate paper. Then we select a certain form of the formula we are going to use. If the form is not implied by general considerations we usually choose one of the functions described in Secs. 2229 or a simple combination of such func tions (a sum of power functions or of exponential functions and the like). In order to select such a for mula in the best way one must know the graphs of these functions well. When selecting a function we must try to achieve the resemblance bet ween the characteristic peculiarities of a soughtfor function q> (x) and of the function / (x) under consi deration. For example, if the phy sical meaning of the function indi cates that / (x) is even and / (0) = 0 then the function q> (x) should also have these properties and so on. It sometimes turns out that we cannot find a single formula for the whole interval of x. Then it is necessary to divide the interval into several parts and select, for each of the parts, its own appropriate formula. After the form of the formula has been chosen it is necessary to determine the values of parameters entering into the formula. For example, suppose that after plotting the points we obtain the drawing shown in Fig. 49. If we have certain reasons to suspect that the experiment or the calculations of the values of the function ®9~“ contain essential errors we must simply discard the points which fall out of the general form of the relationship described by the data represented in our drawing. For instance, the point P m rig. 49 is a point of this kind. By the way. such points may some times indicate that certain important factors were not taken into account and then, of course, we must pay much attention to them. ine remaining points in Fig. 49 resemble a linear relation of the lorm y = ax + b. In order to determine the parameters a and b let us draw' a straight line such that the experimental points should lie as close as possible to the line. This can easily be done by means o a transparent ruler. We apply the ruler to the drawing and then
INTRODUCTORY
76
MATHEMATICS
FOR
EN G IN EE R S
approximately find the soughtfor position of the ruler. For example, the straight line drawn in Fig. 49 yields b = 0.50 and a = = = 0.58, i.e. y = 0.58x + 0.50. The selection of a linear relation described above is comparatively simple. Therefore when choosing some other kind of functional relation one often tries to introduce new variables so that there should be a linear relation between the new variables and then to determine the parameters entering into the linear relation. We can apply this method only if there are no more than two such parameters since a linear function contains two parameters. For example, let an experiment yield the following table of values: X
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
y
0.00
0.01
0.03
0.08
0.17
0.29
0.45
0.66
0.91
1.22
1.57
We leave to the reader to represent the experimental points on the graph paper. The disposition of the points thus constructed resembles the graph of a power function of the form y = axa. In
Fig. 50
order to determine the parameters a and a we take logarithms of both sides of the equality and denote log y — Y , log x = X and log a = A . Then we arrive at the equality Y = a X + A and thus
77
VARIABLES AND FUNCTIONS
we see that there is a linear relation between the new variables. By means of a table of logarithms we compile the table of the values of the new variables: X 1 .0 0 .7 0 —0.52 0 .4 0 0 .3 0 —0.22 —0.15 0 .1 0 Y 2
—1.5 —1.1 —0.77 0 .5 4 —0.35 —0.18 —0.041
0 .0 5
0.00
0.086 0.196
The points thus obtained are lying close enough to the straight line drawn in Fig. 50. In drawing the straight line we should pay more attention to the last three points whose positions are determined with greater accuracy. Our construction yields the values A = 0.196 and a = 2.44, that is a = 1.57. Hence we finally obtain y = 1.57x2 44. Some further rules and examples see in [511.
CHAPTER II
P lane A nalytic G eom etry
Analytic geometry is a branch of mathematics in which geometri cal problems are investigated on the basis of the coordinate method by means of algebraic techniques. § 7. Plane Coordinates 1. Cartesian Coordinates, Cartesian coordinates are known from elementary mathematical courses and we have already used them (see Chapter I). Cartesian coordinates are called after R. Descartes. R. Descartes and P. Fermat (16011665) are the founders of the coordinate method. Several points are depi cted in Cartesian coordina tes in Fig. 51. It should be noted that we take mutual ly perpendicular coordinate axes here and that the unit of length is the same for both axes. The origin of the coordinate system is placed at the point of intersection of the axes from which the distances along the axes are reckoned. (As it was mentioned in Sec. 1.14, when we construct graphs, we can sometimes take different scales for the axes and change the position of the point the coordinates are recko ned from.) Each point in the coordinate plane has certain uniquely determined coordinates and, conversely, to each ordered pair of coordina tes x and y there corresponds a certain uniquely determined point of the plane. This basic property makes it possible to consider the coordinates of points instead of the points themselves.
PLANE ANALYTIC GEOMETRY
The coordinate axes break the plane into the quarters (quadrants) which are numbered in the way shown in Fig. 52. Each of these quadrants is characterized by its specific combination of the signs of abscissas and ordinates; this is also shown in Fig. 52. 2. Some Simple Problems Concerning Cartesian Coordinates. (1) The distance between two given points. Let the points Mi (x„ z/j) and M2 (x 2, z/2) be given (i.e. their coordinates are known). It is required to determine the distance d — M XM 2 (see Fig. 53). The
y m~,+)
/(+,+)
0 m(~)
m+,~) Fig. 52
formula for the distance is implied by Pythagoras’ theorem applied to the rectangular triangle M tM 2P. Thus we have f+ PMl, i.e. d2 = (x2 — x,)2 + (y2 — z/,)2, or d = K (x, — xt)2  f (y2 — y,)2
(1)
This formula and all the following formulas hold for any two points Mi and M2 placed in an arbitrary manner in the coordinate plane. (2) Division of a line segment in a given ratio. Let some points Mi (xj, iji) and M 2 (x2, y2) be given. It is required to find a point M (x, y) lying on the segment M tM 2 such that the ratio of division should be equal to X where % is a given number (see Fig. 54). The solution of the problem follows from the similarity of the tri angles MiPM and MQM2 which implies 1 i.e. MQ
MMn
1
x2—x ~ ^ an(^ x — xi = ^ 2 — kr From the last relations we deduce ___*1 + ^*2 X ~ l+ 3 t ’
_ 2/i“t~^#2 l+ X
(2)
(the expression for y is obtained in like manner). In particular, in the case X = 1, that is when the segment is halved, we have
x = Xt+2* 2 '
..
y
V i + ’Jz 2
INTRODUCTORY
80
MATHEMATICS
FOR
EN G INEERS
(3) Transformation of coordinates without changing the scale. Suppose we have an “old” coordinate system x , y. Now let a “new” coordinate system xf, y' be introduced. It is required to establish the relation
ship between the old coordinates and the new ones. We shall con sider the following three cases. I. Let the new coordinates be the result of a translation of the original ones. Suppose the new origin of the coordinate system has
y ' b) and let M (;x, y) be an arbitrary (moving, current) point of the circle. Then the basic property characterizing a circle can be written as A M = R where R is the radius of the circle. Now, applying formula (1) for
PLANE ANALYTIC GEOMETBY
83
the distance between two points we obtain
Y(xaY + { y  b f =R or, squaring, we derive
(* 
a?
+
(y
 b)2 =
W
This relation is the one which is satisfied by th e coordinates of all the points of the given circle and a t the same tim e only the points belonging to the circle m ay satisfy this relation. Thus, th is relation is an equation of the circle and a , b and R entering into the equa tion are some fixed numbers (i.e. param eters which characterize the position and the size of the circle) whereas x and y are the current coordinates of a variable point of the circle. Now, contrary to the previous example, let an equation be ori ginally given, for example, the equation x~ + J/2 — 3x + A y — 1 = 0 (7) Transforming the equation and completing the squares we ob tain ( *  l ) 2  ( f ) i!+ < S + 2 ) a  2 ,  l = 0,
( x   ! ) !+te + 2 )* f = 0 Consequently the given equation is the equation of a circle w ith centre at the point (1.5, —2) and of radius = 2.69. If two curves w ith the equations F t (x , y ) = 0 and F z (x, y ) = 0 are given we can pose the problem of finding the point of intersection of the curves. The soughtfor point of intersection m ust belong to both lines simultaneously and therefore its coordinates x and y must satisfy the equations of both lines. Hence, in order to determine the coordinates we must solve the following system of equations: F i (x, y) = o  F 2 (*, y ) = 0 I
W
Such a system of equations can have a number of distinct solutions and this number corresponds to the number of possible points of intersection. Of course, a solution is understood as a pair of certain values x and y satisfying (8 ). For example, let it be necessary to determine the point of inter section of circle (7) w ith the straight line y = x + b where b is a constant. To do this we should solve the system of equations s 2 + y 2 — 3® rf Ay — 1 = 0 1 y = x + b j
84
INTRODUCTORY
MATHEMATICS FOR ENGINEERS
Expressing y in terms o£ x from the second equation, substituting the expression thus obtained for y into the first equation, removing brackets and solving the quadratic equation in x we obtain after some transformations — 1 — 26 + "l/9 — 286 — 462 4 *
1 + 26 + V 9 —286—462 4“ ;

— l _ 2b — 1 /9 — 286—462 4 t
^2 =
— 1 + 26— V 9 —286462 4
Let us find for what values of ft both points of intersection coincide. This will happen in case the expression under the radical sign vani shes which implies bu2 = d~_V 58 , i.e. bt = 0.31 and ft2 = = —7.31. The straight line y = x + 6, as it is shown in Fig. 62, is tangent to the circle for these values of ft. If ftj l!
104
INTRODUCTORY
MATHEMATICS
FOR
ENG INEERS
It is easy to verify that if we take equation (29) with e = 1 and pass from the polar coordinates to the Cartesian coordinates accor ding to formulas (5) we obtain the equation of a parabola. In fact, P = l+^coTy and P + P cos
* oo or xn — oo. The compre hensive statement of the notion “increases infinitely" is analogous to the one given in Sec. 1 for the notion of “infinitely approaches zero" but, of course, here we should consider inequalities of the form  x  > Ar. This means that from a certain moment on the variable must satisfy the inequality [ x  > 1 and from some other (later) moment on it must satisfy the inequality  x  > 1 0 . Further, there must exist a certain moment from which on we shall have  x  > 1 0 0 and so on. Now let us discuss some simple properties of infinitely large variables. A variable which is the inverse of an infinitely large variable is an infinitesimal and, conversely, the inverse of an infinitesimal vari* able is an infinitely large variable. These properties can be condi tionally expressed as oo = 0 and ~
LIMIT. CONTINUITY
113
We shall use this notation but one must understand it correctly. For example, the first of the properties means that if the variable x entering into the equality = a increases unlimitedly then in the same process the variable a approaches zero (or, as in Sec. 1, if a: is a “practical” infinitely large variable then a is “practically” infinitesimal). All formulas containing the symbol of infinity oo should be understood in a similar way. For instance, the formula tan 1 3 = too is a conditional and abbreviated form of expressing the fact that when the variable
increases unlimitedly in its absolute value, that is x is infinitely large and so on. This enables us to operate on the symbol oo as if it were a usual number in many cases but, of course, oo is not a concrete number but only a symbol indicating infinitely large variables which are different in different circum stances. The sum of an infinitely large variable and a bounded variable is infinitely large since the first summand “gains over”. The sum of two infinitely large variables of a similar sign is also infinitely large. In contradistinction to it the sum of two infinitely large variables of opposite signs may not be infinitely large since these infinitely large variables may “compensate” mutually. These facts are written as oo ( oo = oo; the expression oo — oo denotes an indetermi nate form. This shows that it is impossible to operate on the symbol oo as on a usual number in all cases; not always oo — oo = 0 since 00 — oo is an abbreviated and conditional way of denoting a diffe rence of the form X — Y where X and Y are infinitely large variab les. The behaviour of these infinities may vary in different cases and therefore it is impossible to have a good judgment on the be haviour of their difference unless an additional investigation has been carried out. We shall discuss indeterminate forms of various types in detail in our course later on. The product of two infinitely large variables is an infinitely large variable. Moreover, the product of an infinitely large variable by a variable which is larger than a positive constant in its absolute value is an infinitely large variable. At the same time, the ratio of two infinitely large quantities is an indeterminate form like the ratio of two infinitesimals.
§ 2. Limits 4. Definition. It is said that a variable x approaches (tends to) a finite limit a in some process if a is constant and x approaches a un limitedly in this process. Then we write x* a or lim x = a 80141
INTRODUCTORY
114
MATHEMATICS
FOR
EN G INEERS
Thus, the lim it of a variable in case it exists is a constant value* According to the above definition infinitesimals are variables that approach zero, that is having zero as a limit. But infinitely large variables, of course, have no finite limit. To say ilx approaches a unlimitedly'’ is to say “the difference between x and a approaches zero unlimitedly”, that is x — a = a is an infinitesimal. The last equality may be rewritten as x = a+ a where a is the infinitesimal. If a variable x approaches its lim it a but always remains smaller than a, that is approaches a from the region of smaller values, then we conditionally write x a — 0 or lim x = a — 0 (this is a conditional way of denoting the lim it since if we understand the / x
a a
x
a
x
x
Fig. 92
expression a — 0 as a real difference then a — 0 = a). If a: in its process of approaching a always remains larger than a then we write x a + 0. Finally, x may tend to a in such a way that it could take on values larger than a and values smaller than a all the time (such a process resembles damped a* ■ ■ ■ oscillations). All the cases described _ here are depicted in Fig. 92. x p tc f ^ Now we can sum up our discussion on the types of variables. A variable x Fig. 93 may be of one of the following types in a certain process: (1) x is bounded and has a limit; in a special case when the limit is equal to zero x is an infinitesimal variable. To distinguish between these cases a bounded variable is sometimes called finite only if it is not an infinitesimal; for instance, it is possible to speak about an infinitesimal mass and about a finite mass etc. (2) x is bounded but has no limit; as an example we may consider the deviation of a pendulum from its equilibrium position in the case of undamped oscillations. Variables of this type are called oscillating (see Fig. 93). In Fig. 93 we see the point a which possesses the following pro perty: the variable x approaches the point a infinitely many times in the process of its change and takes on the values that are arbi trarily close to a but at the same time x does not remain near a all the time. In this case a is called a limit point of the variable x .
J
115
LIMIT. CONTINUITY
There are the greatest value q and the least value p among these limit points, q and p are denoted, respectively, as lim x and lim x and are called the limit superior and the limit inferior of the variable x. But in this case x does not have a “unique” lim it which we discussed in the beginning of this section. Therefore we should remark that the everyday notion of a “lim it” (in the sense of some kind of a “border”, “frontier”) differs from the mathematical one. A bounded variable x always has the lim it superior and the lim it inferior and lim x ^ lim x. The “unique” lim it, that is the lim it in the sense of our previous definition, exists if and only if lim x = = lim x. (3) x is unbounded and besides infinitely large. In this case we write lim x = ± o o , and x is said to have an infinite limit.

6
«27
X
Fig. 94
(4) ar is unbounded but not infinitely large. The deviation of?an oscillating body m the case of a resonance may serve as an example Here. Such a variable oscillates and from time to time travels “to ward infinity” further and further but at the same time it permanent ly returns to regions lying near the initial point (see Fig 94) 5. Properties of Limits. 6 '* i . If a variable has a limit then this limit is unique, i.e. there
s
z p
j z s s t f *
~l and y+ oo. Suppose, for example, that x —> 1 + 0 , that is x > 1 . Then the expression xy “has a cer tain tendency to approach 1” (since 1v — 1) and at the same time it “wants to tend to infinity” (since x > 1 and x*° = oo because if we raise a constant number larger than unity to an infinitely increasing power we shall arrive at an infini tely large variable). Therefore, these two tendencies “act” upon the expression in the opposite directions and hence the result may be different in different prob lems depending on which of these tendencies “wins”. For example, in case (13) the lim it turned out to be equal to e whereas the imme diate substitution of h = 0 yields 1°°. In this case we see that the tendencies are equal in a certain sense; they are in a state of “balan ce”. Similar conclusions may be derived in connection with other indeterminate forms. 2. Solving Inequalities. Let a function / {x) be considered on an interval (a, b) (in particular, the interval may be the whole araxis) and let it be necessary to solve the inequality /(* )> 0 (17) that is to determine all the values of x for which it holds. In the geo metrical sense this means that we must find such regions on the rraxis where the graph of the function y = f ( x ) lies over the rcaxis (such regions are shaded in Fig. 107). But it must be stressed that in this problem we regard the function f {x) as known whereas its graph may be unknown.
To solve inequality (17) let us mark all t he zeros of Hie function f (that is the points where / vanishes) on the interval (a, 6) and also all the points of discontinuity (there are three zeros and one point of discontinuity in Fig. 107). The interval is divided into several parts by these points (five parts in Fig. 107). Since we have taken into account all the points of discontinuity of the function it is continuous inside each of these parts. Besides the function does not vanish inside the parts because we have reckoned all its zeros. Thus, the function y — f (x) retains its sign inside each of the parts (see
3
~w 00— 1 ~Z
is
— 0. x  —3 The numerator equals zero when x = 1 and therefore it is divisible by x — 1. This implies that the expression ..
lJ~
*3+3*2—4
*2—3
(* — 1) ( * 2  f 4 * + 4 )
—
a:2—3
{x—
1) ( * + 2 p
x2—3
must be positive for the values of x which we are interested in. Hence, the function is defined over the whole xaxis except x = ±1/"3 and has two zeros (x = 1 and x = —2) and two points of discontin uity (x = ± ] / 3 ) . These points break the xaxis into five parts (see Fig. 108). Now we choose a point in each of the intervals, substitute these values into the last fraction and determine the signs of the fraction inside the intervals (the numerical values themselves do not matter and only their signs are essential). Thus we receive the table —3
—1.9
0
1.1
2
—
—
+
—
+
Thus, the solution of the inequality is a totality consisting of two intervals; —V 3 < i < 1
and
l / l l < x < OO
CHAPTER IV
D erivatives, D ifferentials, In vestigation of the Behaviour of Functions
§ 1. Derivative 1. Some Problems Leading to the Concept of a Derivative. We come to the notion of a derivative, one of the most important notions in mathematics, when investigating the rate of change of a function. For example, let us turn to the notion of the velocity (rate) at a given instant of a rectilinear motion of a material point. A mate rial point is understood in physics as a material body such that it is permissible to neglect its geometrical sizes while investigating the state of the body under some concrete conditions. In different circumstances a particle of a substance, or an airplane, or a heavenly 0 +V—
y
S
A B HH
~S
AS
Fig. 109
body etc. may sometimes be regarded as a point. Let a material point move along the saxis from left to right. In the general case the motion may be nonuniform, that is the velocity of the motion may be variable. The law of motion is expressed mathematically as a dependence of the coordinate s on time t: s = f (t). Since the velocity is variable the ratio of the distance passed over to the time taken represents the average velocity only. As for the “true” velocity, that is the velocity at a given instant, it can be obtained by means of the following procedure. Let the moving point occupy the position A (see Fig. 109) at an instant t. Suppose during the period of time At (see Sec. 1.22 on this notation) the moving point transits to the position B , the distance As being passed over. Then s = f {t), s + As = / (t + At) '
As
i.e. As = f (t + At) — f (t). Hence, the ratio vov = (which is the distance passed over per unit of time taken) is the average
d e r iv a t iv e s ,
d iffe r e n tia ls ,
b e h a v io u r
o f fu n c t io n s
135
velocity of motion during the time period At from < to t + A*. Now the instantaneous velocity of motion at time t is obtained as me
limit of the average velocity in the process of decreasing Vie interval At unlimitedly, that is V jn st
— h m Va v '■ =lim ~ A0 At At*0
= iim Af+ o
Af
(1)
It is also said that the instantaneous velocity (that is the velocity at a given instant, the true velocity) is the average velocity during an infinitesimal interval of time (“element” of time) or that the instantaneous velocity is the ratio of an infinitesimal distance to an infinitesimal time interval. Both definitions briefly express the mea ning of the general definition (1). The rate of a physical process is not in all cases represented by the Fig. 110 distance passed over related to the unit of the time taken. Let us con sider, for example, the process of filling a vessel. In this^ case the dependence V — f (t) of the volume already filled on time t expresses the law of the process of filling. The average rate of filling during the interval of time from t to t f At is represented by the ratio AT __ /( l + AQ—/(*) At
At
whereas the limit Winst =
,.
,.
AtvO
At»0
lim w aB= lim
AT
.. /(t+At) —/(0 lim At At*0
( 2)
serves as the instantaneous rate, i.e. the rate of filling at time tt Thus we have arrived at an expression similar to (1). But we can understand the velocity, the rate, even in a wider sense relating the change of a quantity not to the unit of time but to the unit of some other quantity. For example, let us consider the notion of the linear density of a material line, that is of a body such that, under given concrete conditions, it is permissible to take into account only its size in onedimensional extent (the longitudinal size) neglecting the crosssection sizes. At the same time we do not neglect its mass. If this line (“thread”) is homogeneous its linear density is equal to the ratio of its mass to its length. In case the thread is nonhomogeneous its linear density is different at different points. Let us reckon the distance from one of the ends of the thread (see Fig. 110) and let the mass of the part of the thread corresponding to the distance s be equal to M ~ f (s). If now some additional
[INTRODUCTORY MATHEMATICS FOR
136
ENGINEERS
distance As is passed the ratio n — Aitf — /(s+As)—/(s) Pao~" As ~~
As
represents the average linear density of the thread corresponding to the part A B . The lim it AAf lim /(«4As) —/(*) p = lim pav= lim As (3) \ As As~+0 A5+0 now gives the linear density of the thread at a point (namely, at the point A). We may say that p is the rate (velocity) of change, of the mass of the thread, i.e. the change of the mass per unit of distance passed. 2. Definition of Derivative. From the mathematical point of view expressions (1), (2) and (3) are quite similar. This enables ns to state the following definition. Let a function y = / (a:) be given. Then the rate of its change related to the unit of change of the argu ment x is equal to Ay y '= lim Ax Ax*0
lim Ax*0
/ ( x  f Ax) — / (x) Ax
This rate (velocity) is called the derivative of the variable (func tion) y with respect to the variable (argument) x ; in other words, the derivative is the limit of the ratio of the increment of the function to the increment of the argument taken in the process when the incre ment of the argument approaches zero. Since this rate has, in general, different values for different values of x the derivative itself is a new function of x. This new function is designated as y' = f (x). Hence, in the examples of Sec. 1 the velocity of motion is equal to the derivative of the distance passed with respect to the time, i.e. v = st (the subscript t in the expression st indicates that the derivative is taken with respect to the variable t) etc. For example, let us compute the derivative of the function y = ax2. Increasing the argument by an increment Ax we receive the new value of the argument x + Ax and the new value of the function y + Ay = a (x + Ax)2 since x j Ax should be substituted for x into the expression of the function. Thus, Ay = a (x + Ax)2 — ax2 = 2ax Ax + a (Ax)2 This implies 2ax Ax + a (Ax)2 = Ax &X+ 0
lim (2ax + aAx) = 2ax AxvO Note that in the latter passage to the lim it only Ax varied as A x> 0 whereas x was considered to be constant. The result thus obtained can be written in the form (ax2)' = 2ax. We leave it to the reader to verify th at (ax3)' = 3ax2, (ax)' = a and the like. We particularly note here that x' = 1. y'
lim —  = lim
Ax*0 a x
d e r iv a t iv e s ,
d if f e r e n t ia l s ,
b e h a v io u r
OF FUNCTIONS
137
3. Geometrical Meaning of Derivative. Let us consider the grapb
oi a function / (a) (see Fig. 111). We see that ■— = = t a n pr i.e. the ratio is equal to the slope of the secant mm. If A** 0 then the secant turns round the point M and tends to the position of the
tangent ll in the lim it process since the tangent occupies the limiting position of the secant when the points of intersection merge. (This obvi ous property ■which we have already used is in fact nothing but the definition of a tangent.) Therefore
y’0= lim £7 = lim tan p = tan a
[y'0= /' (*)]
(4)
A**0 a x
that is the geometrical meaning of the derivative of a function is that it is equal to the slope of the tangent. By formula (11.21) it is easy now to put down the equation of the tangent ll:
y — yo = Vo — x0) (5> where *„ and y0 are the coordinates of the point of tangency, x and y are the moving coordinates of the point on the tangent straight line. Similarly, the equation of the normal to the curve, that is of the line perpendicular to the tangent at the point of tangency, has the form y — y0 = ~ ( x — x0) (see problem 5 in Sec. II.9). In Sec. 1.26 we said that the angle between two curves at the point of their intersection was defined as the angle between the tangents to the curves at that point, and therefore we are able now to deter mine the angle by means of formula (11.23) since we know how to>
138
INTRODUCTORY
MATHEMATICS
FOR
EN G INEERS
•determine the tangents. Note that the angle may turn out to be zero in case these curves are tangent to each other, i.e. when their tangents coincide. When the graph of a function y = / (z) is given the geometrical meaning of the derivative makes it possible to indicate the slope of the tangent to the graph and this enables us to draw immediately •a sketch of the graph of the derivative (see Fig. 112). For more accu rate “graphical computation of the derivative” it is necessary to draw tangents to the given graph and measure their slopes. It turns •out that it is practically simpler to draw normals to the graph by means of a shiny (metallic) ruler and to measure their slopes with
Tespect to the yaxis which is just the same. One of these procedures is shown in Fig. 112. We apply the ruler perpendicularly to the plane of the graph to one of its points, e.g. to the point M , and turn the ruler in such a way that the reflection of the graph in the Tuler should prolong the graph without a break at M. In this posi tion the ruler will lie exactly along the normal to the graph at M . Then we draw a straight line passing through the point (0; 1) and parallel to the normal thus constructed. In this way we get the line segment OP which is then transferred to the position qq. After a number of such procedures are carried out we obtain a rather accurate graph of the derivative. While discussing the geometrical meaning of the derivative [formula (4)] we supposed that both variables x and y were dimen sionless and that the scale was the same for both axes. But this is not always the case in practical problems. It follows from formu la (1.10) that in the general case we must write = ta n a . *y Thus in the general case the derivative is also equal to the slope of the tangent. Note that if the derivative y' approaches infinity for some value of x (for x = Xi in Fig. 113) then the tangent at the corresponding
DERIVATIVES, DIFFERENTIALS, BEHAVIOUR OF FUNCTIONS
139
point of the graph has the slope equal to infinity, th a t is the tangent line is parallel to the yaxis. If the derivative has a jum p discon tinuity at a point then the tangent turns jumpwise, i.e. the graph is broken at the point (see the point x = x 2 in Fig. 113). In case the function approaches infinity the derivative m ay also tu rn into infinity (see the point x = x 3 in Fig. 113). 4. Basic Properties of Derivatives. 1. The derivative of a constant equals zero. (The property is ob viously interpreted as the fact th a t the velocity of a body in a sta te of rest is equal to zero.) The formal proof of the property looks as follows: if y — C — const then A y —C — C —0,
^ = 0 ,
y ' = lim  ^ r  = lim 0 = 0
2. The derivative of a sum is equal to the sum of the derivatives of the summands. Indeed, if y (x) = u (x) + v (a;) then y (x f Ax) = — u (x + Ax) + v (a: + Ax) and by = y (a: + Ax) — y (x) = [u (x + Ax) + v (x f Ax)] — — [u ( x )  f v (a:)] = [u (a:) { A u + v (x) + Ad] — — [w (x) + v (x)] = Au + Ad that is the increment of a sum is equal to the sum of the increments•! (w + v) — Au J Ad. Hence, it follows th a t
A
y =
lim A.r»0
lim An + Ay = lim ( Au A.r^0 Ax Ax*0 ' Ax
Av ^ __ Ax J
t , t lim ~At> K I~ U + v
= lim ArcM)
(in the deduction we have used the fact that the lim it of a sum is equal to the sum of the limits) which is the required proof. This property can be rewritten in a different way as (w +
d) '
= u' +
d'
We have taken a sum of two summands. It is clear that the same l ™ e, fo r.a.ntt arbitrary number of summands. Similarly, the increthp . w S .. ° ere_n c e *s equal to the difference of the increments and vatiyes1VatlVe °f 3 difference is e;  < ^ ) 'end + wof ' Sec.+ (5)ox + (see the 2). =
INTRODUCTORY
140
MATHEMATICS FOR
EN G INEERS
3. A constant factor can be taken outside the derivative sign, i.e. (Cu)f = Cu' (where C = const). Virtually, if y = Cu then Ay = y (x r Ax) — y (x) — Cu (x + Ax) — Cu (x) = = C lu (x + Ax) — u (x)] — C Au in other words, if a function is multiplied by a constant its increment is multiplied by the same constant: A {Cu) = C Au. Hence. y = lim
= lim C^ U —C lim
Ar*0 A x
Ax*0
l~ = Cu'
Ax*0 A x
A x
4. Formula for the Derivative of a Product. Let y = uv. Then Ay = {u + Au) {v j Av) — uv = (Au) v f u Av j Au Av which implies lim y’ = Ax+0
= lim “
Ax>0
(Au)
lim
u Av
Ax+0
lim Ax*0
Au Au A
,
Au Au Ax ,
/
— Ax = u z; 4 uv lim Ax" — Ax
— lim 4Ax r  y  r lim w Ax*0
V
Ax
Ax*0
Ax*0
Thus, (uz;)' = u'v f uz/ For example, [(3x2 + 5x) (4x2  6)]' = (3X2 + ox)9 (4x2 — 6) + + (3s2 + 5x) (4x2  6)' = (6x + 5) (4x2 — 6) + + (3x2 t ox) 8x = 48X5 + 60x2 — 36x — 30
(6)
From formula (6) we can easily deduce the formula for the deri vative of a product of several factors. For example, (uvw)' = [(uz;) w]' = (uv)r w p {uv) w' = = {u'v r uv') w + uvw' = u vw 4 uv'w  f uvwr The formula for the derivative of a product of an arbitrary number of factors looks quite similar. We note that property 3 can be easily deduced from formula (6) by putting v = C. 5. Formula for the Derivative of a Quotient. Let y = — . Then Ay
upAu ujAu
u v
(Au) v — u^Au d (ufAu)
From this, representing Av which enters into the denominator in the form 4Ax—Ax, we obtain y yf = lim A Ax Ax+0
lim Ax*0
Au Ax'
Av Ax
u'v— uv' v (uj u'*0)
DERIVATIVES, DIFFERENTIAI«S, BEHAVIOUR OP FUNCTIONS
141
Thus,
()= T
P>
For example, /
5x2
\ '
(5x2)' (3 x 2 + 4 )  ( 5 * 2 ) ( 3 * 2 + 4 ) '
\ 3x2+4 ) ~
(3x2+4)2
10x(3x2+ 4 ) —5x26x _ —
(3x2+ 4 ) 2
40x
— (3 x 2 + 4 )2
6. The Derivative of a Composite Function. Let y = / (u), u = = q> (x) and let y be regarded as a composite function of x. If x receives an increment Ax then the intermediate variable u receives an increment Au and therefore y receives an increment Ay too. We have Ay _ Ay Au * Ax Au Ax *' Now let Ax —v 0. Then •—
Ax
= 0. Therefore 
ux and hence Au = 4Ax ^ Ax
u * 0 =
y'u. Passing to the lim it in formula (8) we
obtain J/x ~ yrthx
(9)
The last formula may be rewritten as 1/ (
) and v = v (x) then y x = yU'Uvv'x. For instance, let y — (x2 — 5x j 3)3. Then we can denote y = uP where u = x 5x j 3, and by formula (9) we get
y’x = y'uux = (u3);(x2 — 5x + 3); = 3 u2 (2x — 5) = = 3 (x2  5x  f 3)2 (2x  5) which is, of course, simpler than removing the brackets! In practical computations there is no need to write down all this in such a detailed foHoura F°r mStanCe’ the former calculations can be put down as I(x2 — 5x + 3)3]' = 3 (x2 — 5x + 3)2 (x2 — 5x  f 3)' = = 3 (x2 — 5x + 3)2 (2x — 5) u
# o  But
^
«
142
INTRODUCTORY
MATHEMATICS
FOR
EN G INEERS
We use here formula (10), of course. After some practice one can write the result immediately without intermediate transformations. For this purpose it is advisable to remember the formulas for the derivatives in the form (u2)' = 2u u \ {u?)r = 3u2u! and the like (the derivatives are taken with respect to x). 7. The Derivative of an Inverse Function. Suppose the equality y = y (x) defines the inverse relation x = x (y) (see Sec. 1.21) for which we can determine the derivative x ’y. Then it is easy to compute the derivative of the original function y (x). Indeed, — •= which implies, as Ax v 0 and Ay
0,
Ay ( 11)
For example, let y = y ^ x which yields x = yz. Then 1 _ 1 _ 1 = 1 xv ~ W v ~3y2 3?/T2 8. The Derivative of an Implicit Function. If a function is deter mined in an implicit form F (x, y) = 0 (see Sec. 1.20) then to com pute the derivative y'x one should simply equate the derivatives of the lefthand side and of the righthand side of the latter relation taking into account that y isja function of x which turns the relation into an identity. Generally, it is permissible to equate the deriva tives of both sides of an equality if and only if the equality is an identity (but not an equation!). For instance, let us take 5 + S 1
(« )
Then
( # ) ; + ( £ ) > < « . < « > Computing the derivative of the second summand we have used property 6: { ^ ) x = ^ { y 'l)vyx = ^ 2^ y y '■ Thus, (13) implies
5, Derivatives of Basic Elementary Functions. 1. The Derivative of Sine. Let y = sin x . If the argument changes and becomes equal to x } Ax then the function becomes equal to sin (x + Ax). This implies Ay = sin ( x A x ) — sin x = 2 sin 4^* cos (2 + ^ ) 5
DERIVATIVES* D IFF E R E N T IA L S, B E H A V IO U R
.
..
to s in
= lim ■
OF FUNC TIO NS
145
[ 2 ‘* > x  c o s( * +  T  ) ] A*
lim cos ^v— vo '
" *
Isee formula (III.11)]. Hence, (sin x)' — cos x
s=l»cosa;
(15)
2. We leave it to the reader to verify in an analogous way that (cos x)' = —sin x (16) 3. The derivative of tangent is calculated hy formula (7): . . , _ / sin * \ ' (sin x ) ’ cos a—sin x (cos x ) ' ' • Vcosar/ COS2 a: cos x«cos x —sin x (—sin x ) 1 cos2 cos2 a: co s2 a; a: 1 4. Similarly, we can verify that (cot a:)' = sin2* * 5. The Derivative of Arc Sine. Take y = arc sin x. Then x — = sin y and, hy formula (11), j#, ___ 1 1 _ 1 ________ 1 1 y* ~ x'v ~ (siny)’v ~ cosy ~ ^ l / l —sin2y ~ Y T ^ We have written f in front of the radical sign because the values of arc sin x, as is well known, are taken in the interval — ^ arc sin x ^ y which corresponds to nonnegative values of c o s y ! > 0 . Thus, (arc sin x)' = ■ 1 l / l —*2 6. We verify similarly that , (arc cos x)’ =
(17) y i —*2 using the inequality 0 ^ arc cos a: ^ it. The resemblance between the last two results is explained by the formula arc sin x + arc cos x = ~
(18)
which can be deduced in the following wayi if we denote sin a = x then cos ^
=
x
INTRODUCTORY
144
MATHEMATICS
FOR
EN G IN EE R S
and these formulas yield a = arc sin x and ~ — a = arc cos x . The addition of the last results implies (18). 7. The Derivative of Arc Tangent. Let y = arc tan x. Then x = = tan y. Using the formulas for the derivatives of an inverse func tion and of the tangent we obtain y*~
x'v ~~
1 cos2 y
~~ CQS y ~
1 + tan2 y ~~ 1 + x2
Thus, (arc tan x y = — ^ 8. The Derivative of a Logarithmic Function. Take y = lna:. Then putting h =  ^  i n formula (III.12) we see that y ’ = lim
= lim
AacvO
lim 1_d ( 1+ t ) = _1_
Aar>0
Aar+0
x
x
x
Therefore (In
=
Applying formula (1.14) and taking into account that In a = const we receive fl0®**)' = ( t e t ) ,
(l nx) ' = THT7
9. The Derivative of an Exponential Function. If y = ax then x = loga y and
, 1 1 , X1 yx — ——= —:— = y In a = a In a xv —— y In a
Hence, (ax)' ~ a x In a In particular (e*)' = ex. 10. The Derivative of a Power Function. According to the formula of the derivative of a composite function we have
(xny = I(elD*)n]' = (en ln *)’ = en In xn ^ =■■xnn— — nx"~l Thus (xnY = nx*1
D E R IV A T IV E S ,
D IF F E R E N T IA L S ,
B E H A V IO U R
OF
F U N C T IO N S
\
145
This formula holds for any n, both integral and nonintegral. (V * )' = (x2)' = x2 =■' ( * “) “ & ^ A = and the like. ar i l . The Derivatives of Hyperbolic Functions. We have
For example,
(sinh x)’ —
^ c°s h x
C~ g 2(
Similarly, (cosh x)' = sinh x; /sinh x \ ' (sinh *)' cosh a:—sinh x (cosh *)' (tank x) = { — ^ )  ^  x _cosh^j—sinli2.T _ 1. cosh2# cosli2 a: ’
(sinh1 x)r = [\n{x + Y z2+ l)] ' = = _____ 1 X + V ^ + i
 ( r + ] / ~ ^ + r y = *( i + = V ^ ' X +1A 2+1 V ^ 2 V * 2 + l /
1 jA , 2x \ z f l/x 2 + l \ ’r 2 V x 2 4 lj
1 1 /^ + T + a r a r + y ja + l \/x 2 + l
1 ~\/x* + l \
These formulas also demonstrate a rather close analogy between trigonometric and hyperbolic functions. 12. The above formulas (comprising the table of basic differentia tion* formulas) should be learnt by heart since they w ill be per manently used in what follows. With the help of the formulas it is possible to compute the derivative of any elementary function by using the rules of Sec. 4. For example, ( V x 2tan 5*)' = { ^ x ) 2tan B* + ¥ x (2,an 5xy
But
and, by the formula of the derivative of a composite function, we obtain (2tan57 = 2tan5xIn 2 (tan 5x)' = 2tan f I n 2 ' l l .  (5x)' = *
= 2
tan 5a
duiJSLisrg? i ' J o * ! ? 100141
In 2
1 cos2 5x
',erivative oi
COS* u X
•5 ’
a imcUo° is
146
INTRODUCTORY
MATHEMATICS
FOR
EN G IN EE R S
Taking the common factor outside the brackets we derive Wy x 2,an 5X)7 ‘' = ^3 x± \(1 f 15 In 2 cos2 4 ^o x)) After some practice calculations of this type can be carried out much faster without intermediate transformations. 13. In some cases it is useful to take logarithms before calcula ting a derivative. For example, let it be necessary to find the deri vative (a;5in:c)'. Then we write y = x sinx; ln y = s i n x l n x and (In y)x = (sin x In x)r. Therefore, 1 u, = c o s x ln , x  r. sm . x — 1 — y y
x
To calculate the lefthand side we have applied the formula of the derivative of a composite function. Finally, from the last relation, y r = (xsln XY — x sinx ^co szln zjsin rr y j This method is sometimes used when it is necessary to find the derivative of the product of several factors since after taking the logarithm the product turns into a sum, and, generally speaking, it is easier to find the derivative of the sum than that of the product. 6. Determining Tangent in Polar Coordinates. The problem of determining the tangent to a curve which is represented by its equa tion in Cartesian coordinates was solved in Sec. 3. Now let a curve be given by its equation p = / ((p) in polar coordinates. To determine the tangent we could transform the equation into Cartesian coordi nates but it is simpler to solve the problem directly in polar coor dinates. Let the position of the tangent be determined by the angle 0 (see Fig. 114). Let us give
dy=siJs3 (Aar)2=S2
of the tangent. Hence, the replacement of the increment of a function by its differential is equivalent to the replacement of the graph of the function by the segment of the tangent drawn through the point A . This replacement is justified in case Ax is small enough. To investigate the connection between the differential and the increment we take into account th at * y ', i.e. = 4a Ax a.tj0 * Ax * 1 where a 0 as Ax 0. This yields Ay = y' Ax + a Ax = dy + P (23) where p = a Ax is an infinitesimal variable of higher order than Ax (it is represented by the segment CD in Fig. 117). The equality
DERIVATIVES, DIFFERENTIALS, BEHAVIOUR OF FUNCTIONS
151
123) can be formulated as the differential is the principal linear part of the increment of a function. It is called here the principal part since the difference between the differential and the increment is the infinitesimal p of higher order and it is called the linear part since it is directly proportional to Ax (compare with Sec. t.LL). then dy and dx = Ax are infinitesimals of the same order and therefore 6 in formula (23) is an infinitesimal of higher order than dy, that js dy and Ay are equivalent infinitesimals (see bee. H Let us take an example to illustrate the error which occurs if the increment of a function is replaced by its differential. Let y — x~ and let the argument first assume the value x = 1 and then receive the increment Ax. We have Ay = (1 + Ax)2 — l 2 = 2 Ax + Ax2; dy = y' Ax = 2  1 Ax = 2 Ax Therefore, Ay and dy differ from each other by the infinitesimal (Ax)2 of the second order (see Fig. 118). In particular, for A i=
0.1
Ay=
0.21;
dy =
0.2;
for Ax =
0.01
Ay=
0.0201;
dy=
0.02;
for A x=—0.001 Ay = —0.001999;
dy= — 0.002;
the error is 5 per cent; the error is 0,5 per cent; the error is 0.05 per cent etc.
It is obvious here that the relative error generated by the repla cement of Ay by dy decreases rapidly as  Ax  decreases. A function which has the differential is called differentiable. In other words, a differentiable function is a function such that its small increment has the principal linear part, i.e. a function which may be approximately replaced by a linear function on every small interval of change of the argument (such a replacement is the socalled process of linearizing). A differentiable function must have a finite derivative, and the function itself must be continuous for the considered values of the argument since (23) shows that Ay is infinitesimal when Ax is infinitesimal. At the same time a con tinuous function may turn out to be nondifferentiable at some points. For instance, the function shown in Fig. 113 is nondifferen tiable not only at the point of discontinuity x = x 3 but also at the points x = x l and x = x 2 where it is continuous. B. Bolzano, a famous^ Czech mathematician (17811848), and, independently of him, K. Weierstrass (18151897), a prominent German mathemati cian, discovered (the former in 1830 and the latter in 1860; Bolzano’s result was not published) the existence of continuous functions which are nondifferentiable for all values of the argument. Such functions had been considered a mathematical trick for a long time but it
152
INTRODUCTORY
MATHEMATICS
FOR
ENG INEERS
turned out that they were of essential importance for describing processes of the type of a Brownian motion. We shall not pay atten tion to the possibility of existence of such “monsters” in our intro ductory course. 9* Properties of Differential. The differential of a function is obtained by multiplying its derivative by the differential of the argument and therefore each property of the derivative (see Sec. 4) obviously implies the corresponding property of the differential. For example, multiplying both parts of the equality (u + v)f = = u' + v' by dx we receive (u + v)f dx = ur dx + vf dx or, which is the same, d [u + v) = du + dv (i.e. the differential of a sum is equal to the sum of the differentials). Similarly, we deduce the formula d (uv) = [du) v u dv (24) and the like. We shall see in Sec. IX .12 that these formulas also hold for the case of an arbitrary number of independent variables. The implication of the formula for the derivative of a composite function is of special importance. Let y — f (x) and let x first be an independent variable. Then each of the formulas (21) and (22) can be used for calculating dy since in this case Ax = dx. Now let x depend on a third variable, for example, x = x (t). Then A x ^ dx but it turns out that nevertheless formula (22) remains true [whereas formula (21) does not hold, in general]. Virtually, dy = y \d t = z/ixj dt = y'x dx which is what we set out to prove. Therefore it is natural to use formula (22) [and not (21)] for calculating the differential since this formula remains true (invariant) in all cases. Now we shall apply this invariance property* to computing the derivative of a function represented parametrically (see Sec. II.6). Let x = x (t) and y = y (/) [t is a parameter). Then dx — x dt and dy — y dt (the dot usually denotes the derivative with respect to a parameter) which implies dy y_ (25) dx
All the linearizing a general, increments
x
properties of differentials are used, in particular, for relations between variables, that is for passing from nonlinear, relation to the linear relation between the of the variables. Such a linearization is possible in case
* This property is usually called the invariance of the form of the diffe rential.— Tr.
DERIVATIVES, DIFFERENTIALS, BEHAVIOUR OF FUNCTIONS
453
th e c h a n g e s o f t h e v a r i a b l e s a r e s m a l l , a n d i t i s b a s e d o n d r o p p i n g in fin ite s im a ls o f h ig h e r o r d e r . T h u s, fo r in s t a n c e , e q u a t io n r e la tio n
b e tw e en
th e
( 1 1 .3 0 ) c h a r a c t e r iz e s t h e
n o n lin e a r
M (x, y)
b e lo n g in g
c o o r d in a te s
of
a
p o in t
to a c u r v e o f t h e s e c o n d o r d e r . B u t n o w l e t t h e p o s it i o n o f t h e p o i n t
M
M 0 (x0,
ch a n g e n e a r a fix e d p o in t
a: —
x0 =
£
sm tl
y
yo
—
( 1 1 .3 0 ) a n d r e p l a c i n g to
th e
lin e a r iz e d
=
th e
T
be
Po)> t h a t i s l e t t h e i n c r e m e n t s
s m a ll.
d iffe r e n tia ls
D iffe r e n tia tin g
by
th e
e q u a tio n
in c r e m e n ts w e
2By