No label defined (Q4492676)

This textbook is based on a course on real analysis which the author has given at the University of California, Irvine. The subject of this course is the theory of measure and integration. Its pre-requirement is advanced calculus. The necessary background material can be found in [\textit{R. C. Buck}, ``Advanced calculus'' (1978; Zbl 0385.26002)]. It is particularly suitable for a one-year course at the graduate level. In presentation the material the author follows traditions of the well-known textbooks in pure mathematics, such as [\textit{E. C. Titchmarsh}, ``The theory of functions'' (1975; Zbl 0336.30001)], [\textit{I. P. Natanson}, ``Theory of functions of real variable'' (1955; Zbl 0064.29102)], [\textit{W. Rudin}, ``Real and complex analysis'' (1987; Zbl 0925.00005) and ``Principles of mathematical analysis'' (1976; Zbl 0346.26002)], [\textit{P. Billingsley}, ``Probability and measure'' (1995; Zbl 0822.60002)], [\textit{R. M. Dudley}, ``Real analysis and probability'' (1989; Zbl 0686.60001)]. Every concept is defined precisely and every theorem is presented with a detailed and complete proof. Counter-examples are presented to show that certain conditions in the hypothesis of a theorem cannot be simply dropped. The following description of the chapters will give some impression of the book. Chapter 1 introduces the concepts of measure and measurable function. Section 1 defines a measure as a nonnegative countably additive set function on a \(\sigma\)-algebra of subsets of an arbitrary set. A measurable mapping from a measure space into another one is then defined. Section 2 presents construction of a measure space by means of an outer measure. To have a concrete example of a measure space early on, the Lebesgue measure space on the real line \(\mathbb R\) is introduced in Section 3. Subsequent developments in the rest of Chapter 1 and Chapter 2 are in the setting of a general measure space. (This results from the consideration that in the definition of a measure and an integral with respect to a measure the algebraic and topological structure of the underlying space is irrelevant and indeed unnecessary. Topology of the space on which a measure is defined becomes relevant when one considers the regularity of the measure, that is, approximation of measurable sets by Borel sets.) Section 4 treats measurable functions, in particular algebraic operations on measurable functions and pointwise limits of sequences of measurable functions. Section 5 shows that every measure space can be completed. Section 6 compares two modes of convergence of a sequence of measurable functions: convergence almost everywhere and convergence in measure. The Borel-Cantelli lemma and its applications are presented. A unifying theorem (Theorem 6.5) is introduced from which many other convergence theorems relating the two modes of convergence are derived subsequently. These include Egoroff's theorem on almost uniform convergence and Lebesgue's and Riesz's theorems. Chapter 2 treats integration of functions on an arbitrary measure space. In Section 7 the Lebesgue integral, that is, an integral with respect to a measure, is defined for a bounded real-valued measurable function on a set of finite measure. The Bounded Convergence Theorem on the commutation of integration and limiting process for a uniformly bounded sequence of measurable functions which converges almost everywhere on a set of finite measure is proved here. The proof is based on Egoroff's theorem. On the Lebesgue measure space on \(\mathbb R\), comparison of the Lebesgue integral and the Riemann integral is made. Section 8 contains the fundamental idea of integration with respect to a measure. It is shown here that for every nonnegative extended real-valued measurable function on a measurable set the integral with respect to the measure always exists even though it may not be finite. The Monotone Convergence Theorem for an increasing sequence of nonnegative measurable functions, the most fundamental one of all convergence theorems regarding commutation of integration and convergence of the sequence of integrands, is proved here. Fatou's Lemma concerning the limit inferior of a sequence of nonnegative measurable functions is derived from the Monotone Convergence Theorem. In Section 9 the integral of an extended real-valued measurable function on a measurable set is then defined as the difference of the integrals of the positive and negative parts of the function provided the difference exists in the extended real number system. The generalized monotone convergence theorem for a monotone sequence of extended real-valued measurable functions, generalized Fatou's Lemma for the limit inferior and the limit superior of a sequence of extended real-valued measurable functions, and Lebesgue's Dominated Convergence Theorem are proved here. Fatou's Lemma and Lebesgue's Dominated Convergence Theorem under convergence in measure are included. In Section 10 a signed measure is defined as an extended real-valued countably additive set function on a \(\sigma\)-algebra and then shown to be the difference of two positive measures. In Section 11 the Radon-Nikodým derivative of a signed measure with respect to a positive measure is defined as a function which we integrate with respect to the latter one to obtain the former one. The existence of the Radon-Nikodým derivative is then proved under the assumption that the former is absolutely continuous with respect to the latter and that both are \(\sigma\)-finite. (The fact that the Radon-Nikodým derivative is a derivative not only in name but in fact is the derivative of a measure with respect to another one is shown for Borel measures on the Euclidean space in Section 25.) Chapter 3 treats the interplay between integration and differentiation on the Lebesgue measure space on \(\mathbb R\). Section 12 presents Lebesgue's theorem that every real-valued increasing function on \(\mathbb R\) is differentiable almost everywhere on \(\mathbb R\). The proof is based on a Vitali covering theorem. This is followed by Lebesgue's theorem on the integral of the derivative of a real-valued increasing function on a finite closed interval in \(\mathbb R\). Functions of bounded variation are included here. Section 13 defines absolute continuity of a real-valued function on a finite closed interval in \(\mathbb R\) and then shows that a function is absolutely continuous if and only if it is an indefinite integral of a Lebesgue integrable function. This is followed by Lebesgue's decomposition of a real-valued increasing function as the sum of an absolutely continuous function and a singular function. Such methods of calculating a Riemann integral in calculus as the Fundamental Theorem of Calculus, integration by parts, and change of variable of integration find their counterparts in the Lebesgue integral here. Section 14 treats convex functions and in particular their differentiability and absolute continuity property. Jensen's inequality is included here. Chapter 4 treats the \(L^p\) spaces of measurable functions \(f\) with integrable \(| f|^p\) for \(p\in(0,\infty)\) and the space \(L^{\infty}\) of essentially bounded measurable functions on a general measure space. Here Hölder's inequality and Minkowski's inequality are proved for \(p\in(0,\infty]\). Section 15 introduces the Banach space and its dual. Section 16 treats \(L^p\) spaces for \(p\in[1,\infty]\) as well as for \(p\in(0,1)\). Section 17 treats relation among the \(L^p\) spaces for different values of \(p\). The \(l^p\) spaces of sequences of numbers \((\alpha_n\colon n\in\mathbb N)\) with \(\sum_{n\in\mathbb N} |\alpha_n|^p<\infty\) is treated as a particular case of \(L^p\) spaces in which the underlying measure space is the counting measure space on the set \(\mathbb N\) of natural numbers. The Riesz representation theorem on the \(L^p\) spaces is proved in Section 18. Section 19 treats integration on a locally compact Hausdorff space. Urysohn's Lemma on the existence of a continuous function with compact support and partition of unity, Borel and Radon measures, the Riesz representation theorem on the space of continuous functions with compact support as well as Lusin's theorem on approximation of a measurable function by continuous functions are included here. (The placement of Section 19 in Chapter 4 is somewhat arbitrary.) Chapter 5 treats extension of additive set functions to measures. It starts with extension of an additive set function on an algebra to a measure in Section 20 and completes the theory with extension of an additive set function on a semialgebra to a measure in Section 21. (Semialgebra of sets is an abstraction of the aggregate of left-open and right-closed boxes in the Euclidean space \(\mathbb R^n\). Its importance lies in the fact that the Cartesian product of finitely many algebras and in particular \(\sigma\)-algebras is in general not an algebra, but only a semialgebra.) As an example of extending an additive set function on a semialgebra to a measure, the Lebesgue-Stieltjes measure determined by a real-valued increasing function on \(\mathbb R\) is treated in Section 22. Theorems establishing the equivalence of the absolute continuity and singularity of a Lebesgue-Stieltjes measure with respect to the Lebesgue measure with the absolute continuity and singularity of the increasing function that determines the Lebesgue-Stieltjes measure are proved. As a second example of extending an additive set function on a semi-algebra to a measure, the product measure on the product of finitely many measure spaces is included in Section 23. Tonelli's theorem and Fubini's theorem on the reduction of a multiple integral to iterated integrals are found here. Chapter 6 specializes in integration in the Lebesgue measure space on \(\mathbb R^n\). In Section 24 the Lebesgue measure on \(\mathbb R^n\) is constructed as an extension of the notion of volumes of boxes in \(\mathbb R^n\) to Lebesgue measurable subsets of \(\mathbb R^n\). Then it is shown that the Lebesgue measure space on \(\mathbb R^n\) is the completion of the \(n\)-fold product of the Lebesgue measure space on \(\mathbb R\). Regularity of the Lebesgue measure and in particular approximation of Lebesgue measurable sets by open sets leads to approximation of the integral of a measurable function by that of a continuous function. Translation invariance of the Lebesgue measure and integral and linear transformation of the Lebesgue measure and integral are treated. Section 25 begins with the study of the average function of a locally integrable function. The Hardy-Littlewood maximal theorem and the Lebesgue differentiation theorem are presented. These are followed by differentiation of a set function with respect to the Lebesgue measure, in particular differentiation of a signed Borel measure with respect to the Lebesgue measure, and density of a Lebesgue measurable set with respect to the Lebesgue measure. Section 26 treats the change of variable of integration by differentiable transformations. Chapter 7 is an introduction to Hausdorff measures on \(\mathbb R^n\). Section 27 defines \(s\)-dimensional Hausdorff measures on \(\mathbb R^n\) for \(s\in[0,\infty)\) and the Hausdorff dimension of a subset of \(\mathbb R^n\). Section 28 studies transformations of Hausdorff measures. Section 29 shows that a Hausdorff measure of integral dimension is a constant multiple of the Lebesgue measure of the same dimension.