Calculus 1 to 4 (2004–2006)
Axel Schuler¨
January 3, 2007 2 Contents
1 Real and Complex Numbers 11 Basics ...... 11 Notations ...... 11 SumsandProducts ...... 12 MathematicalInduction...... 12 BinomialCoefficients...... 13 1.1 RealNumbers...... 15 1.1.1 OrderedSets ...... 15 1.1.2 Fields...... 17 1.1.3 OrderedFields ...... 19 1.1.4 Embedding of natural numbers into the real numbers ...... 20
1.1.5 The completeness of Ê ...... 21 1.1.6 TheAbsoluteValue...... 22 1.1.7 SupremumandInfimumrevisited ...... 23 1.1.8 Powersofrealnumbers...... 24 1.1.9 Logarithms ...... 26 1.2 Complexnumbers...... 29 1.2.1 TheComplexPlaneandthePolarform ...... 31 1.2.2 RootsofComplexNumbers ...... 33 1.3 Inequalities ...... 34 1.3.1 MonotonyofthePowerand ExponentialFunctions ...... 34 1.3.2 TheArithmetic-Geometricmeaninequality ...... 34 1.3.3 TheCauchy–SchwarzInequality ...... 35 1.4 AppendixA...... 36
2 Sequences and Series 43 2.1 ConvergentSequences ...... 43 2.1.1 Algebraicoperationswithsequences...... 46 2.1.2 Somespecialsequences ...... 49 2.1.3 MonotonicSequences ...... 50 2.1.4 Subsequences...... 51 2.2 CauchySequences ...... 55 2.3 Series ...... 57
3 4 CONTENTS
2.3.1 PropertiesofConvergentSeries ...... 57 2.3.2 OperationswithConvergentSeries...... 59 2.3.3 SeriesofNonnegativeNumbers ...... 59 2.3.4 TheNumber e ...... 61 2.3.5 TheRootandtheRatioTests...... 63 2.3.6 AbsoluteConvergence ...... 65 2.3.7 DecimalExpansionofRealNumbers ...... 66 2.3.8 ComplexSequencesandSeries ...... 67 2.3.9 PowerSeries ...... 68 2.3.10 Rearrangements...... 69 2.3.11 ProductsofSeries ...... 72
3 Functions and Continuity 75 3.1 LimitsofaFunction...... 76 3.1.1 One-sided Limits, Infinite Limits, and Limits at Infinity...... 77 3.2 ContinuousFunctions...... 80 3.2.1 TheIntermediateValueTheorem...... 81 3.2.2 Continuous Functions on Bounded and Closed Intervals—The Theorem about Maximum and Minimum 3.3 UniformContinuity...... 83 3.4 MonotonicFunctions ...... 85 3.5 Exponential, Trigonometric, and Hyperbolic Functions and their Inverses . . . 86 3.5.1 ExponentialandLogarithmFunctions ...... 86 3.5.2 TrigonometricFunctionsandtheirInverses ...... 89 3.5.3 HyperbolicFunctionsandtheirInverses ...... 94 3.6 AppendixB...... 95 3.6.1 MonotonicFunctionshaveOne-SidedLimits ...... 95 3.6.2 Proofs for sin x and cos x inequalities ...... 96 3.6.3 Estimates for π ...... 97
4 Differentiation 101 4.1 TheDerivativeofaFunction ...... 101 4.2 TheDerivativesofElementaryFunctions ...... 107 4.2.1 DerivativesofHigherOrder ...... 108 4.3 LocalExtremaandtheMeanValueTheorem ...... 108 4.3.1 LocalExtremaandConvexity ...... 111 4.4 L’Hospital’sRule ...... 112 4.5 Taylor’sTheorem ...... 113 4.5.1 ExamplesofTaylorSeries ...... 115 4.6 AppendixC...... 117
5 Integration 119 5.1 TheRiemann–StieltjesIntegral ...... 119 5.1.1 PropertiesoftheIntegral ...... 126 CONTENTS 5
5.2 IntegrationandDifferentiation ...... 132 5.2.1 TableofAntiderivatives ...... 134 5.2.2 IntegrationRules ...... 135 5.2.3 IntegrationofRationalFunctions ...... 138 5.2.4 PartialFractionDecomposition ...... 140 5.2.5 Other Classes of Elementary Integrable Functions ...... 141 5.3 ImproperIntegrals ...... 143 5.3.1 Integralsonunboundedintervals ...... 143 5.3.2 IntegralsofUnboundedFunctions ...... 146 5.3.3 TheGammafunction...... 147 5.4 IntegrationofVector-ValuedFunctions ...... 148 5.5 Inequalities ...... 150 5.6 AppendixD...... 151 5.6.1 MoreontheGammaFunction ...... 152
6 Sequences of Functions and Basic Topology 157 6.1 DiscussionoftheMainProblem ...... 157 6.2 UniformConvergence...... 158 6.2.1 DefinitionsandExample ...... 158 6.2.2 UniformConvergenceandContinuity ...... 162 6.2.3 UniformConvergenceandIntegration ...... 163 6.2.4 UniformConvergenceandDifferentiation ...... 166 6.3 FourierSeries ...... 168 6.3.1 AnInnerProductonthePeriodicFunctions ...... 171 6.4 BasicTopology ...... 177 6.4.1 Finite,Countable,andUncountableSets ...... 177 6.4.2 MetricSpacesandNormedSpaces...... 178 6.4.3 OpenandClosedSets ...... 180 6.4.4 LimitsandContinuity ...... 182 6.4.5 ComletenessandCompactness...... 185 k 6.4.6 Continuous Functions in Ê ...... 187 6.5 AppendixE ...... 188
7 Calculus of Functions of Several Variables 193 7.1 PartialDerivatives...... 194 7.1.1 HigherPartialDerivatives ...... 197 7.1.2 TheLaplacian...... 199 7.2 TotalDifferentiation...... 199 7.2.1 BasicTheorems...... 202 7.3 Taylor’sFormula ...... 206 7.3.1 DirectionalDerivatives ...... 206 7.3.2 Taylor’sFormula ...... 208 7.4 ExtremaofFunctionsofSeveralVariables ...... 211 6 CONTENTS
7.5 TheInverseMappingTheorem ...... 216 7.6 TheImplicitFunctionTheorem...... 219 7.7 LagrangeMultiplierRule ...... 223 7.8 IntegralsdependingonParameters ...... 225 7.8.1 Continuity of I(y) ...... 225 7.8.2 DifferentiationofIntegrals ...... 225 7.8.3 ImproperIntegralswithParameters ...... 227 7.9 Appendix ...... 230
8 Curves and Line Integrals 231 8.1 RectifiableCurves...... 231 k 8.1.1 Curvesin Ê ...... 231 8.1.2 RectifiableCurves ...... 233 8.2 LineIntegrals ...... 236 8.2.1 PathIndependence ...... 239
9 Integration of Functions of Several Variables 245 9.1 BasicDefinition...... 245 9.1.1 PropertiesoftheRiemannIntegral ...... 247 9.2 IntegrableFunctions ...... 248 9.2.1 IntegrationoverMoreGeneralSets ...... 249 9.2.2 Fubini’sTheoremandIteratedIntegrals ...... 250 9.3 ChangeofVariable ...... 253 9.4 Appendix ...... 257
10 Surface Integrals 259 3 10.1 Surfaces in Ê ...... 259 10.1.1 TheAreaofaSurface ...... 261 10.2 ScalarSurfaceIntegrals...... 262 10.2.1 Other Forms for dS ...... 262 10.2.2 PhysicalApplication ...... 264 10.3 SurfaceIntegrals ...... 264 10.3.1 Orientation ...... 264 10.4 Gauß’DivergenceTheorem...... 268 10.5 Stokes’Theorem ...... 272 10.5.1 Green’sTheorem ...... 272 10.5.2 Stokes’Theorem ...... 274 10.5.3 Vector Potential and the Inverse Problem of Vector Analysis ...... 276
n
11 Differential Forms on Ê 279 n 11.1 The Exterior Algebra Λ(Ê ) ...... 279
11.1.1 The Dual Vector Space V ∗ ...... 279 11.1.2 The Pull-Back of k-forms ...... 284 n 11.1.3 Orientation of Ê ...... 285 CONTENTS 7
11.2 DifferentialForms...... 285 11.2.1 Definition...... 285 11.2.2 Differentiation ...... 286 11.2.3 Pull-Back...... 288 11.2.4 ClosedandExactForms ...... 291 11.3 Stokes’Theorem ...... 293 11.3.1 Singular Cubes, Singular Chains, and the Boundary Operator . . . . . 293 11.3.2 Integration ...... 295 11.3.3 Stokes’Theorem ...... 296 11.3.4 SpecialCases ...... 298 11.3.5 Applications ...... 299
12 Measure Theory and Integration 305 12.1 MeasureTheory...... 305 12.1.1 Algebras, σ-algebras,andBorelSets...... 306 12.1.2 AdditiveFunctionsandMeasures ...... 308 12.1.3 ExtensionofCountablyAdditiveFunctions ...... 313 n 12.1.4 The Lebesgue Measure on Ê ...... 314 12.2 MeasurableFunctions...... 316 12.3 TheLebesgueIntegral ...... 318 12.3.1 SimpleFunctions...... 318 12.3.2 PositiveMeasurableFunctions ...... 319 12.4 SomeTheoremsonLebesgueIntegrals...... 322 12.4.1 TheRolePlayedbyMeasureZeroSets ...... 322 12.4.2 The space Lp(X,µ) ...... 324 12.4.3 TheMonotoneConvergenceTheorem ...... 325 12.4.4 TheDominatedConvergenceTheorem ...... 326 12.4.5 Application of Lebesgue’s Theorem to Parametric Integrals...... 327 12.4.6 TheRiemannandtheLebesgueIntegrals ...... 329 12.4.7 Appendix:Fubini’sTheorem...... 329
13 Hilbert Space 331 13.1 TheGeometryoftheHilbertSpace...... 331 13.1.1 UnitarySpaces ...... 331 13.1.2 NormandInnerproduct ...... 334 13.1.3 TwoTheoremsofF.Riesz ...... 335 13.1.4 OrthogonalSetsandFourierExpansion ...... 339 13.1.5 Appendix ...... 343 13.2 BoundedLinearOperatorsinHilbertSpaces ...... 344 13.2.1 BoundedLinearOperators ...... 344 13.2.2 TheAdjointOperator...... 347 13.2.3 ClassesofBoundedLinearOperators ...... 349 13.2.4 OrthogonalProjections ...... 351 8 CONTENTS
13.2.5 SpectrumandResolvent ...... 353 13.2.6 TheSpectrumofSelf-AdjointOperators ...... 357
14 Complex Analysis 363 14.1 HolomorphicFunctions...... 363 14.1.1 ComplexDifferentiation ...... 363 14.1.2 PowerSeries ...... 365 14.1.3 Cauchy–RiemannEquations ...... 366 14.2 Cauchy’sIntegralFormula ...... 369 14.2.1 Integration ...... 369 14.2.2 Cauchy’sTheorem ...... 371 14.2.3 Cauchy’sIntegralFormula ...... 373 14.2.4 ApplicationsoftheCoefficientFormula ...... 377 14.2.5 PowerSeries ...... 380 14.3 LocalPropertiesofHolomorphicFunctions ...... 383 14.4 Singularities...... 385 14.4.1 ClassificationofSingularities ...... 386 14.4.2 LaurentSeries ...... 387 14.5Residues...... 392 14.5.1 CalculatingResidues ...... 394 14.6 RealIntegrals ...... 395 14.6.1 RationalFunctionsinSineandCosine ...... 395 14.6.2 Integrals of the form ∞ f(x) dx ...... 396 −∞ 15 Partial Differential Equations I —R an Introduction 401 15.1 ClassificationofPDE ...... 401 15.1.1 Introduction...... 401 15.1.2 Examples ...... 402 15.2 FirstOrderPDE—TheMethodofCharacteristics ...... 405 15.3 Classification of Semi-Linear Second-Order PDEs ...... 408 15.3.1 QuadraticForms ...... 408 15.3.2 Elliptic,ParabolicandHyperbolic ...... 408 15.3.3 ChangeofCoordinates ...... 409 15.3.4 Characteristics ...... 411 15.3.5 TheVibratingString ...... 414
16 Distributions 417 16.1 Introduction — Test Functions and Distributions ...... 417 16.1.1 Motivation ...... 417 n 16.1.2 Test Functions D(Ê ) and D(Ω) ...... 418 n 16.2 The Distributions D′(Ê ) ...... 422 16.2.1 RegularDistributions...... 422 16.2.2 OtherExamplesofDistributions ...... 424 CONTENTS 9
16.2.3 ConvergenceandLimitsofDistributions ...... 425 P 1 16.2.4 The distribution x ...... 426 16.2.5 OperationwithDistributions ...... 427 16.3 TensorProductandConvolutionProduct ...... 433 16.3.1 TheSupportofaDistribution ...... 433 16.3.2 TensorProducts...... 433 16.3.3 ConvolutionProduct ...... 434 16.3.4 LinearChangeofVariables...... 437 16.3.5 FundamentalSolutions ...... 438
n n Ê 16.4 Fourier Transformation in S (Ê ) and S ′( ) ...... 439 n 16.4.1 The Space S (Ê ) ...... 440 n 16.4.2 The Space S ′(Ê ) ...... 446 n 16.4.3 Fourier Transformation in S ′(Ê ) ...... 447 16.5 Appendix—MoreaboutConvolutions ...... 450
17 PDE II — The Equations of Mathematical Physics 453 17.1 FundamentalSolutions ...... 453 17.1.1 TheLaplaceEquation ...... 453 17.1.2 TheHeatEquation ...... 455 17.1.3 TheWaveEquation...... 456 17.2 TheCauchyProblem ...... 459 17.2.1 MotivationoftheMethod ...... 459 17.2.2 TheWaveEquation...... 460 17.2.3 TheHeatEquation ...... 464 17.2.4 PhysicalInterpretationoftheResults ...... 466 17.3 FourierMethodforBoundaryValueProblems ...... 468 17.3.1 InitialBoundaryValueProblems ...... 469 17.3.2 EigenvalueProblemsfortheLaplaceEquation ...... 473 17.4 Boundary Value Problems for the Laplace and the Poisson Equations...... 477 17.4.1 FormulationofBoundaryValueProblems ...... 477 17.4.2 BasicPropertiesofHarmonicFunctions ...... 478 17.5Appendix ...... 485 17.5.1 Existence of Solutions to the Boundary Value Problems...... 485 17.5.2 Extremal Properties of Harmonic Functions and the Dirichlet Principle 490 17.5.3 NumericalMethods...... 494 10 CONTENTS Chapter 1
Real and Complex Numbers
Basics
Notations
Ê Real numbers
C Complex numbers
É Rational numbers
Æ = 1, 2,... positive integers (natural numbers) { }
Z Integers
Z É Ê C Ê É Z We know that Æ . We write , and for the non-negative ⊆ ⊆ ⊆ ⊆ + + + real, rational, and integer numbers x 0, respectively. The notions A B and A B are ≥ ⊂ ⊆ equivalent. If we want to point out that B is strictly bigger than A we write A ( B. We use the following symbols
:= defining equation y, implication, “if ..., then ...” ⇒ “if and only if”, equivalence ⇐⇒ for all ∀ there exists ∃ Let a < b fixed real numbers. We denote the intervals as follows
[a, b] := x Ê a x b closed interval { ∈ | ≤ ≤ }
(a, b) := x Ê a [a, b) := x Ê a x < b half-open interval { ∈ | ≤ } (a, b] := x Ê a < x b half-open interval { ∈ | ≤ } [a, ) := x Ê a x closed half-line ∞ { ∈ | ≤ } (a, ) := x Ê a < x open half-line ∞ { ∈ | } ( , b] := x Ê x b closed half-line −∞ { ∈ | ≤ } ( , b) := x Ê x < b open half-line −∞ { ∈ | } 11 12 1 Real and Complex Numbers (a) Sums and Products Let us recall the meaning of the sum sign and the product sign . Suppose m n are ≤ integers, and a , k = m,...,n are real numbers. Then we set k P Q n n a := a + a + + a , a := a a a . k m m+1 ··· n k m m+1 ··· n kX=m kY=m In case m = n the sum and the product consist of one summand and one factor only, respec- tively. In case n < m it is customary to set n n ak := 0, (empty sum) ak := 1 (empty product). kX=m kY=m The following rules are obvious: If m n p and d Z are integers then ≤ ≤ ∈ n p p n n+d ak + ak = ak, ak = ak d (index shift). − kX=m k=Xn+1 kX=m kX=m k=Xm+d n We have for a Ê, a =(n m + 1)a. ∈ − kX=m (b) Mathematical Induction Mathematical induction is a powerful method to prove theorems about natural numbers. Theorem 1.1 (Principle of Mathematical Induction) Let n Z be an integer. To prove a 0 ∈ statement A(n) for all integers n n it is sufficient to show: ≥ 0 (I) A(n0) is true. (II) For any n n : If A(n) is true, so is A(n + 1) (Induction step). ≥ 0 It is easy to see how the principle works: First, A(n0) is true. Apply (II) to n = n0 we obtain that A(n0 + 1) is true. Successive application of (II) yields A(n0 + 2), A(n0 + 3) are true and so on. n Example 1.1 (a) For all nonnegative integers n we have (2k 1) = n2. − k=1 Proof. We use induction over n. In case n =0 we have anX empty sum on the left hand side (lhs) and 02 =0 on the right hand side (rhs). Hence, the statement is true for n =0. Suppose it is true for some fixed n. We shall prove it for n +1. By the definition of the sum and by induction hypothesis, n (2k 1) = n2, we have k=1 − n+1 n P (2k 1) = (2k 1)+2(n + 1) 1 = n2 +2n +1=(n + 1)2. − − − ind. hyp. Xk=1 Xk=1 This proves the claim for n +1. 13 (b) For all positive integers n 8 we have 2n > 3n2. ≥ Proof. In case n =8 we have 2n =28 = 256 > 192 = 3 64=3 82 =3n2; · · and the statement is true in this case. Suppose it is true for some fixed n 8, i.e. 2n > 3n2 (induction hypothesis). We will show ≥ that the statement is true for n +1, i.e. 2n+1 > 3(n +1)2 (induction assertion). Note that n 8 ≥ implies n 1 7 > 2 = (n 1)2 > 4 > 2 = n2 2n 1 > 0 − ≥ ⇒ − ⇒ − − = 3(n2 2n 1) > 0 = 3n2 6n 3 > 0 +3n2 +6n +3 ⇒ − − ⇒ − − | = 6n2 > 3n2 +6n +3 = 2 3n2 > 3(n2 +2n + 1) ⇒ ⇒ · = 2 3n2 > 3(n + 1)2. (1.1) ⇒ · By induction assumption, 2n+1 = 2 2n > 2 3n2. This together with (1.1) yields · · 2n+1 > 3(n + 1)2. Thus, we have shown the induction assertion. Hence the statement is true for all positive integers n 8. ≥ For a positive integer n Æ we set ∈ n n! := k, read: “n factorial,” 0!=1!=1. kY=1 (c) Binomial Coefficients For non-negative integers n, k Z we define ∈ + n k n i +1 n(n 1) (n k + 1) := − = − ··· − . k i k(k 1) 2 1 i=1 Y − ··· · n The numbers k (read: “n choose k”) are called binomial coefficients since they appear in the binomial theorem, see Proposition1.4 below. It just follows from the definition that n =0 for k > n, k n n! n = = for 0 k n. k k!(n k)! n k ≤ ≤ − − Lemma 1.2 For 0 k n we have: ≤ ≤ n +1 n n = + . k +1 k k +1 14 1 Real and Complex Numbers Proof. For k = n the formula is obvious. For 0 k n 1 we have ≤ ≤ − n n n! n! + = + k k +1 k!(n k)! (k + 1)!(n k 1)! − − − (k + 1)n!+(n k)n! (n + 1)! n +1 = − = = . (k + 1)!(n k)! (k + 1)!(n k)! k +1 − − We say that X is an n-set if X has exactly n elements. We write Card X = n (from “cardinal- ity”) to denote the number of elements in X. n Lemma 1.3 The number of k-subsets of an n-set is k . n The Lemma in particular shows that k is always an integer (which is not obvious by its defi- nition). n n n Proof. We denote the number of k-subsets of an n set Xn by Ck . It is clear that C0 = Cn = 1 since ∅ is the only 0-subset of Xn and Xn itself is the only n-subset of Xn. We use induction 1 1 1 1 over n. The case n = 1 is obvious since C0 = C1 = 0 = 1 = 1. Suppose that the claim is true for some fixed n. We will show the statement for the (n + 1)-set X = 1,...,n +1 and { } all k with 1 k n. The family of (k +1)-subsets of X splits into two disjoint classes. In the ≤ ≤ first class A1 every subset contains n +1; in the second class A2, not. To form a subset in A1 one has to choose another k elements out of 1,...,n . By induction assumption the number { } A n n A is Card 1 = Ck = k . To form a subset in 2 one has to choose k +1 elements out of 1,...,n . By induction assumption this number is Card A = Cn = n . By Lemma1.2 { } 2 k+1 k+1 we obtain n n n +1 Cn+1 = Card A + Card A = + = k+1 1 2 k k +1 k +1 which proves the induction assertion. Proposition 1.4 (Binomial Theorem) Let x, y R and n Æ. Then we have ∈ ∈ n n n n k k (x + y) = x − y . k Xk=0 Proof. We give a direct proof. Using the distributive law we find that each of the 2n summands n n k k of product (x + y) has the form x − y for some k = 0,...,n. We number the n factors n as (x + y) = f1 f2 fn, f1 = f2 = = fn = x + y. Let us count how often the n k k · ··· ··· summand x − y appears. We have to choose k factors y out of the n factors f1,...,fn. The remaining n k factors must be x. This gives a 1-1-correspondence between the k-subsets − n k k of f1,...,fn and the different summands of the form x − y . Hence, by Lemma1.3 their { n } n number is Ck = k . This proves the proposition. 1.1 Real Numbers 15 1.1 Real Numbers In this lecture course we assume the system of real numbers to be given. Recall that the set of m É Z integers is Z = 0, 1, 2,... while the fractions of integers = m, n , n = 0 { ± ± } { n | ∈ 6 } form the set of rational numbers. A satisfactory discussion of the main concepts of analysis such as convergence, continuity, differentiation and integration must be based on an accurately defined number concept. An existence proof for the real numbers is given in [Rud76, Appendix to Chapter 1]. The author É explicitly constructs the real numbers Ê starting from the rational numbers . The aim of the following two sections is to formulate the axioms which are sufficient to derive all properties and theorems of the real number system. The rational numbers are inadequate for many purposes, both as a field and an ordered set. For instance, there is no rational x with x2 =2. This leads to the introduction of irrational numbers which are often written as infinite decimal expansions and are considered to be “approximated” by the corresponding finite decimals. Thus the sequence 1, 1.4, 1.41, 1.414, 1.4142, ... “tends to √2.” But unless the irrational number √2 has been clearly defined, the question must arise: What is it that this sequence “tends to”? This sort of question can be answered as soon as the so-called “real number system” is con- structed. Example 1.2 As shown in the exercise class, there is no rational number x with x2 =2. Set 2 2 É A = x É x < 2 and B = x x > 2 . { ∈ + | } { ∈ + | } Then A B = É and A B = ∅. One can show that in the rational number system, A ∪ + ∩ has no largest element and B has no smallest element, for details see Appendix A or Rudin’s book [Rud76, Example 1.1, page 2]. This example shows that the system of rational numbers has certain gaps in spite of the fact that between any two rationals there is another: If r 1.1.1 Ordered Sets Definition 1.1 (a) Let S be a set. An order (or total order) on S is a relation, denoted by <, with the following properties. Let x, y, z S. ∈ (i) One and only one of the following statements is true. x (ii) x In this case S is called an ordered set. (b) Suppose (S,<) is an ordered set, and E S. If there exists a β S such that x β for ⊆ ∈ ≤ all x E, we say that E is bounded above, and call β an upper bound of E. Lower bounds are ∈ defined in the same way with in place of . ≥ ≤ If E is both bounded above and below, we say that E is bounded. The statement x Definition 1.2 Suppose S is an ordered set, E S, an E is bounded above. Suppose there ⊆ exists an α S such that ∈ (i) α is an upper bound of E. (ii) If β is an upper bound of E then α β. ≤ Then α is called the supremum of E (or least upper bound) of E. We write α = sup E. An equivalent formulation of (ii) is the following: (ii)′ If β < α then β is not an upper bound of E. The infimum (or greatest lower bound)ofaset E which is bounded below is defined in the same manner: The statement α = inf E means that α is a lower bound of E and for all lower bounds β of E we have β α. ≤ Example 1.4 (a) If α = sup E exists, then α may ormay not belong to E. For instance consider [0, 1) and [0, 1]. Then 1 = sup[0, 1) = sup[0, 1], however 1 [0, 1) but 1 [0, 1]. We will show that sup[0, 1] = 1. Obviously, 1 is an upper 6∈ ∈ bound of [0, 1]. Suppose that β < 1, then β is not an upper bound of [0, 1] since β 1. Hence 6≥ 1 = sup[0, 1]. We show will show that sup[0, 1)=1. Obviously, 1 is an upper bound of this interval. Suppose that β < 1. Then β < β+1 < 1. Since β+1 [0, 1), β is not an upper bound. Consequently, 2 2 ∈ 1 = sup[0, 1). É (b) Consider the sets A and B of Example1.2 as subsets of the ordered set É. Since A B = ∪ + (there is no rational number with x2 =2) the upper bounds of A are exactly the elements of B. 1.1 Real Numbers 17 Indeed, if a A and b B then a2 < 2 < b2. Taking the square root we have a < b. Since B ∈ ∈ contains no smallest member, A has no supremum in É+. Similarly, B is bounded below by any element of A. Since A has no largest member, B has no infimum in É. Remarks 1.1 (a) It is clear from (ii) and the trichotomy of < that there is at most one such α. Indeed, suppose α′ also satisfies (i) and (ii), by (ii) we have α α′ and α′ α; hence α = α′. ≤ ≤ (b) If sup E exists and belongs to E, we call it the maximum of E and denote it by max E. Hence, max E = sup E and max E E. Similarly, if the infimum of E exists and belongs to ∈ E we call it the minimum and denote it by min E; min E = inf E, min E E. ∈ bounded subset of É an upper bound sup max [0, 1] 2 1 1 [0, 1) 2 1 — A 2 — — (c) Suppose that α is an upper bound of E and α E then α = max E, that is, property (ii) in ∈ Definition1.2 is automatically satisfied. Similarly, if β E is a lower bound, then β = min E. ∈ (d) If E is a finite set it has always a maximum and a minimum. 1.1.2 Fields Definition 1.3 A field is a set F with two operations, called addition and multiplication which satisfy the following so-called “field axioms” (A), (M), and (D): (A) Axioms for addition (A1) If x F and y F then their sum x + y is in F . ∈ ∈ (A2) Addition is commutative: x + y = y + x for all x, y F . ∈ (A3) Addition is associative: (x + y)+ z = x +(y + z) for all x, y, z F . ∈ (A4) F contains an element 0 such that 0+ x = x for all x F . ∈ (A5) To every x F there exists an element x F such that x +( x)=0. ∈ − ∈ − (M) Axioms for multiplication (M1) If x F and y F then their product xy is in F . ∈ ∈ (M2) Multiplication is commutative: xy = yx for all x, y F . ∈ (M3) Multiplication is associative: (xy)z = x(yz) for all x, y, z F . ∈ (M4) F contains an element 1 such that 1x = x for all x F . ∈ (M5) If x F and x =0 then there exists an element 1/x F such that x (1/x)=1. ∈ 6 ∈ · (D) The distributive law x(y + z)= xy + xz holds for all x, y, z F . ∈ 18 1 Real and Complex Numbers Remarks 1.2 (a) One usually writes x x y, , x + y + z, xyz, x2, x3, 2x, . . . − y in place of 1 x +( y), x , (x + y)+ z, (xy)z, x x, x x x, 2x, . . . − · y · · · (b) The field axioms clearly hold in É if addition and multiplication have their customary mean- Z Z ing. Thus É is a field. The integers form not a field since 2 has no multiplicative inverse ∈ (axiom (M5) is not fulfilled). (c) The smallest field is F = 0, 1 consisting of the neutral element 0 for addition and the neu- 2 { } + 0 1 tral element 1 for multiplication. Multiplication and addition are defined as follows 0 0 1 1 1 0 0 1 · 0 0 0 . It is easy to check the field axioms (A), (M), and (D) directly. 1 0 1 (d) (A1) to (A5) and (M1) to (M5) mean that both (F, +) and (F 0 , ) are commutative (or \{ } · abelian) groups, respectively. Proposition 1.5 The axioms of addition imply the following statements. (a) If x + y = x + z then y = z (Cancellation law). (b) If x + y = x then y =0 (The element 0 is unique). (c) If x + y =0 the y = x (The inverse x is unique). − − (d) ( x)= x. − − Proof. If x + y = x + z, the axioms (A) give y = 0+ y = ( x + x)+ y = x +(x + y) = x +(x + z) A 4 A 5 − A 3 − assump. − = ( x + x)+ z = 0+ z = z. A 3 − A 5 A 4 This proves (a). Take z = 0 in (a) to obtain (b). Take z = x in (a) to obtain (c). Since − x + x =0, (c) with x in place of x and x in place of y, gives (d). − − Proposition 1.6 The axioms for multiplication imply the following statements. (a) If x =0 and xy = xz then y = z (Cancellation law). 6 (b) If x =0 and xy = x then y =1 (The element 1 is unique). 6 (c) If x =0 and xy =1 then y =1/x (The inverse 1/x is unique). 6 (d) If x =0 then 1/(1/x)= x. 6 The proof is so similar to that of Proposition1.5 that we omit it. 1.1 Real Numbers 19 Proposition 1.7 The field axioms imply the following statements, for any x, y, z F ∈ (a) 0x =0. (b) If xy =0 then x =0 or y =0. (c) ( x)y = (xy)= x( y). − − − (d) ( x)( y)= xy. − − Proof. 0x +0x = (0+0)x =0x. Hence 1.5(b) implies that 0x =0, and (a) holds. Suppose to the contrary that both x =0 and y =0 then (a) gives 6 6 1 1 1 1 1= xy = 0=0, y ·x y ·x a contradiction. Thus (b) holds. The first equality in (c) comes from ( x)y + xy =( x + x)y =0y =0, − − combined with 1.5(b); the other half of (c) is proved in the same way. Finally, ( x)( y)= [x( y)] = [ xy]= xy − − − − − − by (c) and 1.5 (d). 1.1.3 Ordered Fields In analysis dealing with equations is as important as dealing with inequalities. Calculations with inequalities are based on the ordering axioms. It turns out that all can be reduced to the notion of positivity. In F there are distinguished positive elements (x> 0) such that the following axioms are valid. Definition 1.4 An ordered field is a field F which is also an ordered set, such that for all x, y, z F ∈ (O) Axioms for ordered fields (O1) x> 0 and y> 0 implies x + y > 0, (O2) x> 0 and y> 0 implies xy > 0. If x> 0 we call x positive; if x< 0, x is negative. Ê For example É and are ordered fields, if x>y is defined to mean that x y is positive. − Proposition 1.8 The following statements are true in every ordered field F . (a) If x Proof. (a) By assumption (a + y) (a + x)= y x> 0. Hence a + x < a + y. − − (b) By assumption and by (a) we have x + x′ Proposition 1.9 The following statements are true in every ordered field. (a) If x> 0 then x< 0, and if x< 0 then x> 0. − − (b) If x> 0 and y < z then xy < xz. (c) If x< 0 and y < z then xy > xz. (d) If x =0 then x2 > 0. In particular, 1 > 0. 6 (e) If 0 Proof. (a) If x > 0 then 0 = x + x > x +0 = x, so that x < 0. If x < 0 then − − − − 0= x + x< x +0= x so that x> 0. This proves (a). − − − − (b) Since z>y, we have z y > 0, hence x(z y) > 0 by axiom (O2), and therefore − − xz = x(z y)+ xy > 0+ xy = xy. − Prp. 1.8 (c) By (a), (b) and Proposition 1.7 (c) [x(z y)]=( x)(z y) > 0, − − − − so that x(z y) < 0, hence xz < xy. − (d) If x > 0 axiom 1.4(ii) gives x2 > 0. If x < 0 then x > 0, hence ( x)2 > 0 But − − x2 =( x)2 by Proposition1.7(d). Since 12 =1, 1 > 0. − (e) If y > 0 and v 0 then yv 0. But y (1/y)=1 > 0. Hence 1/y > 0, likewise 1/x > 0. ≤ ≤ · If we multiply x Remarks 1.3 (a) The finite field F = 0, 1 , see Remarks 1.2, is not an ordered field since 2 { } 1+1=0 which contradicts 1 > 0. 2 (b) The field of complex numbers C (see below) is not an ordered field since i = 1 contradicts − Proposition1.9 (a), (d). 1.1.4 Embedding of natural numbers into the real numbers Let F be an ordered field. We want to recover the integers inside F . In order to distinguish 0 and 1 in F from the integers 0 and 1 we temporarily write 0F and 1F . For a positive integer n Æ, n 2 we define ∈ ≥ n := 1 +1 + +1 (n times). F F F ··· F Lemma 1.10 We have n > 0 for all n Æ. F F ∈ 1.1 Real Numbers 21 Proof. We use induction over n. By Proposition1.9(d) the statement is true for n =1. Suppose it is true for a fixed n, i.e. nF > 0F . Moreover 1F > 0F . Using axiom (O2) we obtain (n + 1)1F = nF +1F > 0. From Lemma1.10 it follows that m = n implies n = m . Indeed, let n be greater than m, 6 F 6 F say n = m + k for some k Æ, then n = m + k . Since k > 0 it follows from 1.8(a) that ∈ F F F F n > m . In particular, n = m . Hence, the mapping F F F 6 F Æ F, n n → 7→ F is a one-to-one correspondence (injective). In this way the positive integers are embedded into the real numbers. Addition and multiplication of natural numbers and of its embeddings are the same: nF + mF =(n + m)F , nF mF =(nm)F . From now on we identify a natural number with the associated real number. We write n for nF . Definition 1.5 (The Archimedean Axiom) An ordered field F is called Archimedean if for all x, y F with x> 0 and y> 0 there exists n Æ such that nx>y. ∈ ∈ An equivalent formulation is: The subset Æ F of positive integers is not bounded above. ⊂ Choose x = 1 in the above definition, then for any y F there in an n Æ such that n>y; ∈ ∈ hence Æ is not bounded above. Æ Suppose Æ is not bounded and x> 0,y > 0 are given. Then y/x is not an upper bound for , that is there is some n Æ with n > y/x or nx>y. ∈ 1.1.5 The completeness of Ê Using the axioms so far we are not yet able to prove the existence of irrational numbers. We need the completeness axiom. Definition 1.6 (Order Completeness) An ordered set S is said to be order complete if for every non-empty bounded subset E S has a supremum sup E in S. ⊂ (C) Completeness Axiom The real numbers are order complete, i. e. every bounded subset E Ê has a supremum. ⊂ The set É of rational numbers is not order complete since, for example, the bounded set 2 É A = x É x < 2 has no supremum in . Later we will define √2 := sup A. { ∈ + | } The existence of √2 in Ê is furnished by the completeness axiom (C). Axiom (C) implies that every bounded subset E Ê has an infimum. This is an easy conse- ⊂ quence of Homework 1.4 (a). We will see that an order complete field is always Archimedean. Proposition 1.11 (a) Ê is Archimedean. É (b) If x, y Ê, and x 22 1 Real and Complex Numbers Ê Part (b) may be stated by saying that É is dense in . Proof. (a) Let x,y > 0 be real numbers which do not fulfill the Archimedean property. That is, if A := nx n Æ , then y would be an upper bound of A. Then (C) furnishes that A has a { | ∈ } supremum α = sup A. Since x> 0, α x < α and α x is not an upper bound of A. Hence − − α x < mx for some m Æ. But then α< (m +1)x, which is impossible, since α is an upper − ∈ bound of A. (b) See [Rud76, Theorem 29]. Ê É Remarks 1.4 (a) If x, y É with x − 1 1 Æ Ex class: (b) We shall show that inf n Æ = 0. Since n > 0 for all n , > 0 by n | ∈ ∈ n Proposition1.9 (e) and 0 is a lower bound. Suppose α > 0. Since Ê is Archimedean, we find m Æ such that 1 < mα or, equivalently 1/m < α. Hence, α is not a lower bound for E ∈ which proves the claim. (c) Axiom (C) is equivalent to the Archimedean property together with the topological com- pleteness (“Every Cauchy sequence in Ê is convergent,” see Proposition 2.18). (d) Axiom (C) is equivalent to the axiom of nested intervals, see Proposition2.11 below: Let I := [a , b ] a sequence of closed nested intervals, that is (I I I ) n n n 1 ⊇ 2 ⊇ 3 ⊇··· such that for all ε> 0 there exists n such that 0 b a <ε for all n n . 0 ≤ n − n ≥ 0 Then there exists a unique real number a Ê which is a member of all intervals, ∈ i.e. a = I . n Æ n { } ∈ T 1.1.6 The Absolute Value For x Ê one defines ∈ x, if x 0, x := ≥ | | ( x, if x< 0. − Lemma 1.12 For a, x, y Ê we have ∈ (a) x 0 and x =0 if and only if x =0. Further x = x . | | ≥ | | | − | | | (b) x x , x = max x, x , and x a (x a and x a). ± ≤| | | | x{ − x} | | ≤ ⇐⇒ ≤ − ≤ | | (c) xy = x y and y = y if y =0. | | | | | | | | 6 (d) x + y x + y (triangle inequality). | |≤| | | | (e) x y x + y . || |−| ||≤| | Proof. (a) By Proposition 1.9 (a), x < 0 implies x = x > 0. Also, x > 0 implies x > 0. | | − | | Putting both together we obtain, x = 0 implies x > 0 and thus x = 0 implies x = 0. 6 | | | | Moreover 0 =0. This shows the first part. | | The statement x = x follows from (b) and ( x)= x. | | | − | − − (b) Suppose first that x 0. Then x 0 x and we have max x, x = x = x . If x< 0 ≥ ≥ ≥ − { − } | | then x> 0 > x and − 1.1 Real Numbers 23 max x, x = x = x . This proves max x, x = x . Since the maximum is an {− } − | | { − } | | upper bound, x x and x x. Suppose now a is an upper bound of x, x . Then | | ≥ | | ≥ − { − } x = max x, x a. On the other hand, max x, x a implies that a is an upper bound | | { − } ≤ { − } ≤ of x, x since max is. { − } One proves the first part of (c) by verifying the four cases (i) x, y 0, (ii) x 0, y < 0, (iii) ≥ ≥ x< 0, y 0, and (iv) x,y < 0 separately. (i) is clear. In case (ii) we have by Proposition 1.9 (a) ≥ and (b) that xy 0, and Proposition1.7(c) ≤ x y = x( y)= (xy)= xy . | || | − − | | The cases (iii) and (iv) are similar. To the second part. Since x = x y we have by the first part of (c), x = x y . The claim follows by multipli- y · | | y | | cation with 1 . y (d) By (b) we| | have x x and y y . It follows from Proposition1.8(b) that ± ≤| | ± ≤| | (x + y) x + y . ± ≤| | | | By the second part of (b) with a = x + y , we obtain x + y x + y . | | | | | |≤| | | | (e) Inserting u := x + y and v := y into u + v u + v one obtains − | |≤| | | | x x + y + y = x + y + y . | |≤| | | − | | | | | Adding y on both sides one obtains x y x + y . Changing the role of x and y −| | | |−| |≤| | in the last inequality yields ( x y ) x + y . The claim follows again by (b) with − | |−| | ≤ | | a = x + y . | | 1.1.7 Supremum and Infimum revisited The following equivalent definition for the supremum of sets of real numbers is often used in the sequel. Note that x β x M ≤ ∀ ∈ = sup M β. ⇒ ≤ Similarly, α x for all x M implies α inf M. ≤ ∈ ≤ Remarks 1.5 (a) Suppose that E Ê. Then α is the supremum of E if and only if ⊂ (1) α is an upper bound for E, (2) For all ε> 0 there exists x E with α ε < x. ∈ − Using the Archimedean axiom (2) can be replaced by 1 (2’) For all n Æ there exists x E such that α < x. ∈ ∈ − n 24 1 Real and Complex Numbers Ê (b) Let M Ê and N nonempty subsets which are bounded above. ⊂ ⊂ Then M + N := m + n m M, n N is bounded above and { | ∈ ∈ } sup(M + N) = sup M + sup N. Ê (c) Let M Ê and N nonempty subsets which are bounded above. ⊂ + ⊂ + Then MN := mn m M, n N is bounded above and { | ∈ ∈ } sup(MN) = sup M sup N. 1.1.8 Powers of real numbers n We shall prove the existence of nth roots of positive reals. We already know x , n Z. Itis n n 1 1 n 1 ∈ recursively defined by x := x − x, x := x, n Æ and x := n for n< 0. · ∈ x− Proposition 1.13 (Bernoulli’s inequality) Let x 1 and n Æ. Then we have ≥ − ∈ (1 + x)n 1+ nx. ≥ Equality holds if and only if x =0 or n =1. Proof. We use induction over n. In the cases n = 1 and x = 0 we have equality. The strict inequality (with an > sign in place of the sign) holds for n =2 and x =0 since (1 + x)2 = ≥ 0 6 1+2x + x2 > 1+2x. Suppose the strict inequality is true for some fixed n 2 and x = 0. ≥ 6 Since 1+ x 0 by Proposition1.9(b) multiplication of the induction assumption by this factor ≥ yields (1 + x)n+1 (1 + nx)(1 + x)=1+(n + 1)x + nx2 > 1+(n + 1)x. ≥ This proves the strict assertion for n +1. We have equality only if n =1 or x =0. Æ Lemma 1.14 (a) For x, y Ê with x,y > 0 and n we have ∈ ∈ x ⇐⇒ Æ (b) For x, y Ê and n we have ∈ + ∈ n 1 n n n 1 nx − (y x) y x ny − (y x). (1.2) − ≤ − ≤ − We have equality if and only if n =1 or x = y. Proof. (a) Observe that n n n n k k 1 y x =(y x) y − x − = c(y x) − − − Xk=1 1.1 Real Numbers 25 n n k k 1 with c := k=1 y − x − > 0 since x,y > 0. The claim follows. (b) We have P n n n n 1 n k k 1 n 1 y x nx − (y x)=(y x) y − x − x − − − − − − k=1 Xn k 1 n k n k =(y x) x − y − x − 0 − − ≥ k=1 X n k n k since by (a) y x and y − x − have the same sign. The proof of the second inequality is − − quite analogous. Proposition 1.15 For every real x> 0 and every positive integer n Æ there is one and only ∈ one y> 0 such that yn = x. 1 This number y is written √n x or x n , and it is called “the nth root of x”. n n Proof. The uniqueness is clear since by Lemma1.14(a) 0 E := t Ê t < x . { ∈ + | } Observe that E = ∅ since 0 E. We show that E is bounded above. By Bernoulli’s inequality 6 ∈ and since 0 < x < nx we have t E tn Lemma⇒1.14 Ê Hence, 1+ x is an upper bound for E. By the order completeness of Ê there exists y such ∈ that y = sup E. We have to show that yn = x. For, we will show that each of the inequalities yn > x and yn < x leads to a contradiction. Assume yn < x and consider (y + h)n with “small” h (0 n n n 1 n 1 0 (y + h) y n (y + h) − (y + h y) < h n(y + 1) − . ≤ − ≤ − n 1 n Choosing h small enough that h n(y + 1) − < x y we may continue − (y + h)n yn x yn. − ≤ − Consequently, (y + h)n < x and therefore y + h E. Since y + h>y, this contradicts the fact ∈ that y is an upper bound of E. Assume yn > x and consider (y h)n with “small” h (0 n n n 1 n 1 0 y (y h) ny − (y y + h) < hny − . ≤ − − ≤ − n 1 n Choosing h small enough that h ny − Consequently, x < (y h)n and therefore tn 1/n 1/n 1/n Remarks 1.6 (a) If a and b are positive real numbers and n Æ then (ab) = a b . ∈ Proof. Put α = a1/n and β = b1/n. Then ab = αnβn = (αβ)n, since multiplica- tion is commutative. The uniqueness assertion of Proposition1.15 shows therefore that (ab)1/n = αβ = a1/n b1/n. (b) Fix b> 0. If m, n, p, q Z and n> 0, q > 0, and r = m/n = p/q. Then we have ∈ (bm)1/n =(bp)1/q. (1.3) Hence it makes sense to define br =(bm)1/n. (c) Fix b> 1. If x Ê define ∈ x p b = sup b p É,p 0 is a positive real number. − Example 1.3 (a) The intervals [a, b], (a, b], [a, b), (a, b), ( , b), and ( , b] are bounded −∞ −∞ above by b and all numbers greater than b. 1 1 1 (b) E := n Æ = 1, , ,... is bounded above by any α 1. It is bounded below { n | ∈ } { 2 3 } ≥ by 0.