Wendell Fleming Functions of Several Variables 2nd Edition With 96 Illustrations Springer-Verlag New York Heidelberg BerlinWendell Fleming Brown University Department of Mathematics Providence, Rhode Island 02912 US.A. Editorial Board J. H. Ewing F. W. Gehring Department of Mathematics Department of Mathematics Indiana University University of Michigan Bloomington, IN 47401 Ann Arbor, MI 48109 USS.A. US.A. AMS Subject Classifications: 26-01, 28-01, 58-01 Library of Congress Cataloging in Publication Data Fleming, Wendell Helms, 1928- Functions of several variables. (Undergraduate texts in mathematics) Bibliography: p. Includes index. 1. Functions of several real variables. I. Title QA331.F63 1977. -515’.84 = 76-40029 © 1965 by Wendell Fleming © 1977 by Springer-Verlag New York Inc. P. R. Halmos Department of Mathematics Santa Clara University Santa Clara, CA 95053 US.A. All rights reserved. No part of this book may be translated or reproduced in any form without written permission from Springer-Verlag, 175 Fifth Avenue, New York, New York 10010, U.S.A. Printed and bound by R. R. Donnelley & Sons, Harrisonburg, VA. Printed in the United States of America. 98765 43 (Corrected third printing, 1987) ISBN 0-387-90206-6 Springer-Verlag New York Heidelberg Berlin ISBN 3-540-90206-6 Springer-Verlag Berlin Heidelberg New York-To FloPreface The purpose of this book is to give a systematic development of differential and integral calculus for functions of several variables. The traditional topics from advanced calculus are included: maxima and minima, chain rule, implicit function theorem, multiple integrals, divergence and Stokes’s theorems, and so on. However, the treatment differs in several important respects from the traditional one. Vector notation is used throughout, and the distinction is maintained between n-dimensional euclidean space E” and its dual. The elements of the Lebesgue theory of integrals are given. In place of the traditional vector analysis in E°, we introduce exterior algebra and the calculus of exterior differential forms. The formulas of vector analysis then become special cases of formulas about differential forms and integrals over manifolds lying in E”. The book is suitable for a one-year course at the advanced undergraduate level. By omitting certain chapters, a one semester course can be based on it. For instance, if the students already have a good knowledge of partial differentiation and the elementary topology of £", then substantial parts of Chapters 4, 5, 7, and 8 can be covered in a semester. Some knowledge of linear algebra is presumed. However, results from linear algebra are reviewed as needed (in some cases without proof). A number of changes have been made in the first edition. Many of these were suggested by classroom experience. A new Chapter 2 on elementary topology has been added. Additional physical applications—to thermo- dynamics and classical mechanics—have been added in Chapters 6 and 8. Different proofs, perhaps easier for the beginner, have been given for two main theorems (the Inverse Function Theorem and the Divergence Theorem.) : viiPreface The author is indebted to many colleagues and students at Brown Uni- versity for their valuable suggestions. Particular thanks are due Hildegarde Kneisel, Scott Shenker, and Joseph Silverman for their excellent help in preparing this edition. Wendell H. Fleming Providence, Rhode Island June, 1976 viiiChapter 1 Euclidean spaces 1.1 1.2 1.3 1.4 ales The real number system Euclidean E” Elementary geometry of E” Basic topological notions in E” Convex sets Chapter 2 Elementary topology of E” 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 Functions Limits and continuity of transformations Sequences in E” Bolzano—Weierstrass theorem Relative neighborhoods, continuous transformations Topological spaces Connectedness Compactness Metric spaces 2.10 Spaces of continuous functions *2.11 Noneuclidean norms on E” Chapter 3 Differentiation of real-valued functions 3.1 3.2 Directional and partial derivatives Linear functions Contents 10 14 19 28 28 31 37 43 47 50 56 60 62 67 70 76 76 79 ixContents 3.3. Differentiable functions 3.4 Functions of class C” 3.5 Relative extrema *3.6 Convex and concave functions Chapter 4 Vector-valued functions of several variables 4.1 Linear transformations 4.2 Affine transformations 4.3. Differentiable transformations 4.4 Composition 4.5 The inverse function theorem 4.6 The implicit function theorem 4.7 Manifolds 4.8 The multiplier rule Chapter 5 Integration 5.1 Intervals 5.2 Measure 5.3 Integrals over E” 5.4 Integrals over bounded sets 5.5 Iterated integrals 5.6 Integrals of continuous functions 5.7 Change of measure under affine transformations 5.8 Transformation of integrals 5.9 Coordinate systems in E” 5.10 Measurable sets and functions; further properties 5.11 Integrals: general definition, convergence theorems 5.12 Differentiation under the integral sign 5.13 L?-spaces Chapter 6 Curves and line integrals 6.1 Derivatives 6.2 Curves in E” 6.3 Differential 1-forms 6.4 Line integrals *6.5 Gradient method *6.6 Integrating factors; thermal systems Chapter 7 Exterior algebra and differential calculus 7.1 Covectors and differential forms of degree 2 7.2 Alternating multilinear functions x 82 89 99 107 119 119 125 128 134 140 147 153 161 167 168 170 181 186 190 200 206 209 216 222 227 237 240 245 245 247 253 258 265 268 275 276 2837.3, Multicovectors 7.4 Differential forms 7.5 Multivectors 7.6 Induced linear transformations 7.7 Transformation law for differential forms 7.8 The adjoint and codifferential *7.9 Special results for n = 3 *7.10 Integrating factors (continued) Chapter 8 Integration on manifolds 8.1 Regular transformations 8.2 Coordinate systems on manifolds 8.3. Measure and integration on manifolds 8.4 The divergence theorem *8.5 Fluid flow 8.6 Orientations 8.7 Integrals of r-forms 8.8 Stokes’s formula 8.9 Regular transformations on submanifolds 8.10 Closed and exact differential forms 8.11 Motion of a particle 8.12 Motion of several particles Appendix | Axioms for a vector space Appendix 2 Mean value theorem; Taylor’s theorem Appendix 3 Review of Riemann integration Appendix 4 Monotone functions References Answers to problems Index Contents 287 291 295 306 309 311 316 318 321 322 329 334 340 350 353 356 362 367 369 375 380 383 385 386 388 389 391 405 xiEuclidean spaces 4 | This book is concerned with the differential and integral calculus of functions of several variables. For this purpose one needs first to know some basic properties of euclidean space of arbitrary finite dimension n. We begin this chapter with a brief review of the real numbers and the elements of vector algebra and geometry of such spaces. Later in the chapter the concepts of neighborhood, open set, and closed set are introduced. These constitute the basis for studying what are called topological properties of n-dimen- sional space. Format The word “Theorem” has been reserved for what the author considers the most important results. Results of lesser depth or interest are labeled “Proposition.” The symbol C indicates the end of the proof of a theorem or proposition. Occasionally part of a proof is left to the reader as a homework exercise. The sections marked with an asterisk (*) may be omitted without disrupting the organization. References are given at the end of the book. We presume that the reader is acquainted with the most elementary aspects of set theory. The symbols 6&€ U9, —, stand, respectively, for is an element of, is not an element of, union, intersection, difference, and inclusion. Sets ordinarily are denoted by capital italicized letters. A set is described either by listing its elements or by some property characterizing them. Thus {2, 5, 7} is the set whose elements are the three numbers 2, 5, and 7. If S is a set and m a property pertaining to elements of S, then {p eS: 7} denotes the set of all p€S with property x. For example, if Z = {1, 2,...} is the set of natural numbers, then S = {xe Z:x = 2y—-1 : 11 Euclidean spaces for some y € Z} is the set of odd natural numbers. The set {x € Z: x? = 3} is the empty set. The set {x € Z: x(x — 1) = x? — x} isallof Z. When the set S in question is clear from the context, we abbreviate by writing simply {p: 7}. 1.1 The real number system While calculus has been motivated in large part by problems from geometry and physics, its foundations rest upon the idea of number. Therefore a thorough treatment of calculus should begin with a study of the real numbers. The real number system satisfies axioms about arithmetic and order, which express properties of numbers with which everyone is familiar from elemen- tary mathematics. We list these properties as Axioms I and II. Axiom I (a) Any two real numbers have a sum x + y and a product xy, which are also real numbers. Moreover, Commutative law x+y=ytx, xy = yx, Associative law xt(ytz)=(x+ y) +2, x(yz) = (xy)z, Distributive law X(y + 2) = xy + xz for every x, y, and z. There are two (distinct) real numbers 0 and 1, which are identity elements under addition and multiplication, respectively: Ss x+0=x, xl=x for every x. Every real number x has an inverse — x with respect to addition, and if x # 0, an inverse x~! with respect to multiplication: (c LY x + (—x) =0, xx t= 1, Axiom I]. There is a relation < between real numbers such that: (a) For every pair of numbers x and y, exactly one of the following alternatives holds: x < y,x = y,y c, then b is also an upper bound for S. Axiom III. Any set S of real numbers that has an upper bound has a least upper bound. The least upper bound for S is denoted by sup S. If S has no upper bound, then we set sup S = +00. A number d is a lower bound for S if d < x for every x ES. If S has a lower bound, then (Problem 2) S has a greatest lower bound. It is denoted by inf S. If S has no lower bound, then we set inf S = — oo. ExampLe |. Let S = {1, 2, 3,...}, the set of positive integers. Then sup S = +oo and inf S = 1. EXAMPLE 2. Let a and b be real numbers with a < b. The sets [a, b] = {x:a a}, (a, 00) = {x:x > a} are called closed and open, respectively, and have a as greatest lower bound. The corresponding intervals (— 00, b], (— 00, b) have b as least upper bound. Let S be a set that has an upper bound. Example 2 shows that the number sup S need not belong to S. If sup S does happen to be an element of S, then it is the largest element of S and we write “max S” instead of “sup S.” Similarly, if S is bounded below and inf S is an element of S, then we write for it “min S.”1 Euclidean spaces EXamMpLe 3. Let S = {x: x? <2 and x is a rational number}. Then 2) = sup S and —,/2 = inf S. Since \/2 is not a rational number, this example shows that the least upper bound axiom would no longer hold if we replaced the real number system by the rational number system. Exampte 4. Let S = {sin x:x €[—2, 2]}. Then —1 = min S, 1 = max S. The real number system also satisfies the archimedean property. This means that for every ¢ > 0, x > 0 there exists a positive integer m such that x < me. To prove it, suppose to the contrary that for some pair e, x of positive numbers, me < x for every m = 1,2,.... Then x is an upper bound for the set S = {e, 2, 3¢,...}. Let c = sup S. Then (m+ l)e or Ix-y? <|xPly/?. The last inequality is equivalent to Cauchy’s inequality. 0 From the proof we see that equality in Cauchy’s inequality is equivalent to the fact that |x + toy| = 0, that is, that x + toy = 0. Thus, if y #0, |x y| = |x|ly| if.and only if x is a scalar multiple of y. If x - y = |x||y|, then x is a nonnegative scalar multiple of y (and conversely). 61.2 Euclidean E” PrRooF OF (1.2). We write, as before, (x+y)((X+ty)=x°xt2xryt+yry. From Cauchy’s inequality, Ix + yl? <|x/? + 2ixllyl + lyl?, or Ix + yl? < (|x| + lyl). This is equivalent to the triangle inequality. O If y 4 0, equality holds in (1.2) if and only if x is a nonnegative scalar multiple of y. Using the fact that |cx| = |c||x|, one can easily prove by induction on m the following extension of the triangle inequality: Ms (1.3) c m 2X; < Lle'llx;1 j= j=1 for every choice of scalars c!,...,c” and of vectors x,,...,X,-. We recall (Appendix A.1) that }); c’x; is called a linear combination of x1, ..., Xm- The euclidean distance between x and y is |x — y|. If x, y, and z are vectors, then x —Z=(x—y) + (y — 2). Applying (1.2) to x — y and y — z, we have (1.4) Ix—z| <|x—yl+ly-2l, which justifies the name “triangle inequality” (see Figure 1.1). x—Z ly—2| Ix y| y Figure 1.1 If x and y are nonzero vectors, the angle 0 between x and y is defined by the formula * ‘y (1.5) cos @ = ly r 00 if x¥0 . Using Problem 2, show that (w + cx)*(y + dz) =wey + cx*y + dw°z + cdx°z. . Show that )%., [x!| < \/n|x|, for any x = (x!,..., x"), [Hint: First suppose that x! > 0. Use Equation (1.1) with y' = 1.] . Show that 2|x|? + 2ly|? =|x + y|? + |x —y|?. What does this say about parallelograms (see Figure 1.3)? Figure 1.3 . Show that |x + y||x — y| < |x|? + ly? with equality if and only if x-y =0. What does this say about parallelograms? + Prove (1.3), using (1.2) and induction on m. . Letn = 4, and v, = 43e, + 4e3), v, = $(4e, — 3e4), v3 = (/2/10)(—4e, + 3e, + 3e; + 4e,). Show that v,, ¥2, ¥3 are mutually orthogonal unit vectors. Find a unit vector vy such that v,, V2. V3, V4 form an orthonormal basis for E*.1 9. 10. ll. Euclidean spaces Let {v,,...,¥,} be an orthonormal basis for E”, and let c= fain = Yrv,0 have analogs in E” for any dimension n. Let us begin with the concept of line in E”. Definition. Let x,, x, € E” with x; # x,. The line through x, and x, is xy {x:x = fx, + (1 — dx,,¢ any scalar}. If we set z = x, — X2, then this can be rewritten as {x:xX =X, + Z, t any scalar}. In the plane E? the vector equation x = x, + tz becomes X= X24 tx, — x2) Y= V2 + 1 — Ya), which, in elementary analytic geometry, are called “ parametric equations” of the line through (x,, y;) and (x2, yo). The line segment joining x, and x, is {xix = fx, + (1 — 4x,,1r€ [0, 1]}, where [a, b] denotes the set of real numbers t such that a < t < b (Section 1.1). For example, if t = 4, then x is the midpoint of the line segment joining and x, (Figure 1.4). The points corresponding to t = 4, 3 trisect the line segment. 101.3. Elementary geometry of E" Figure 1.4 Hyperplanes, half-spaces Given two points x,, x, € E", consider the set P = {x: (x, — X2)*(X — X2) = O}. Such a set P is called a hyperplane through x,. Geometrically, one can think of P as follows. Let / denote the line through x, and x,, and /' the line through x, and x. We call / and I’ perpendicular if (x; — x2)*(x — x2) = 0. Thus P is made up of all lines through x, perpendicular to |. If we set z = x, — x, and c = Z* x, then (x; — X,)° (KX —X,)=Z°x- Cc. Thus the definition of hyperplane can be conveniently formulated as follows: Definition. A hyperplane in E" is a set of the form {x:z* x = c}, where z # Oand c are given. A hyperplane is a point for n = 1, a line for n = 2,a plane for n = 3. For any nonzero scalar b, {x:z° x = c} = {x:(bz)* x = bc}. Thus the vector z and scalar c defining a hyperplane are determined only up to a scalar multiple. If c 4 0, we may, for instance, always take c = 1. A hyperplane P = {x:z*x =} is parallel to P, = {x:z°x =c,} for any c, # c. If cy = 0, then P, contains 0, and P, is a vector subspace of E” of dimension n — 1. This last statement follows from Problem 5. ExampLe |. Find the hyperplane P in E* which contains the four points e,,€; + 2e,,€, + 3e3,e; + 4e,. Every xe P must satisfy the equation z*x =c, where z and c must be found. Taking in turn x =e,, x = e, + 2e,,..., we obtain c=z'e, =2!, c=2Z-+(e, + 2e,) = z! + 227, c= 2° (e, + 3e3) = 27 + 329, c= 2° (3 + 4e4) = 2° + 424.1 Euclidean spaces From these equations, the components z!, ..., z* of the vector z satisfy 1 2 3_¢ 4_o¢ zZ=¢, 2° =0, z=5, z=. 3 6 Taking for convenience c = 6, we have P = {x:6x! + 2x? + x* = 6}. Definition. Let z 4 0. A closed half-space is a set of the form {x:z+x > c}, and an open half-space is a set of the form {x:z+ x > c} (see Figure 1.5). P= {x:z-x =c¢} Figure 1.5 A set H of the form {x:z-x < c},z # 0, is a closed half-space, since H is also {x:(—z)*x = —c}. The same remark applies to open half-spaces. A hyperplane P = {x:z* x = c} divides E” into two half-spaces. More precisely, E" — P is the union of the open half-spaces {x:z°x>c} and {x:z°x 0, the set {y:|y — x| = 6} is called the (n — 1)-sphere with center x and radius 6. For n = 1, it consists of two points x + 6; for n = 2, it is a circle, and for n = 3 a sphere. The set {y:|y — x| < 5} is the open spherical n-ball with center x and radius 6. The corresponding set {y: ly — x| < 5} is called a closed spherical n-ball. The terms “open” and “closed” are justified in Section 1.4. It turns out that an open half-space or spherical n-ball is an open set, and a closed half-space or spherical n-ball is a closed set. Convex sets This concept is defined as follows. Definition. Let K < E". Then K is a convex set if the line segment joining any two points of K is contained in K (Figure 1.6). 121.3. Elementary geometry of E” a> x Convex Not convex Figure 1.6 E" itself is a convex set. The empty set and sets with just one point trivially satisfy the definition; hence they are convex. The reader should be able to think of several kinds of geometric objects such as lines, planes, spherical balls, regular solids, and so on, which appear to be convex sets. However, geometric intuition is not always a reliable guide, especially in four or more dimensions. In any case, intuition is no substitute for a proof that the set in question is actually convex. To show that a set K is convex directly from the definition, we must verify that for every x,, x, € K and re [0, 1], the point x = tx, + (1 — t)x, also belongs to K. In the definition, we assumed that x, # X. But ifx, = xp, it is trivial that x € K, since x = x, = Xp. ExampLe 2. Any closed half-space is a convex set. Let H = {x:z*x > c}, z#0. Let x,,x,¢H and x =tx, + (1 —1)x,, where te[0,1]. Then z*x,; =>c and z* x, >c. Since t > 0, tz* x, >ftc; and since 1 —¢t > 0, (1 — tz* x, > (1 — tc. Consequently, Z°x=tz-x, + (1 —tz+x, =>te + —te=c. This shows that x € H. Therefore H is a convex set. Similarly, any hyperplane is a convex set (Problem 12) and any open half-space is a convex set. EXAMPLE 3. Let U be an open spherical n-ball, namely, U = {x:|x — xo| < 5}, for some X, and 6 > 0. To show that U is a convex set, we proceed as in Example 2. Let x,,x, ¢ U and x = tx, + (1 — t)x, where re [0, 1]. Then IX; — Xo] <4, [x2 — Xol < 4, X — Xq = U(X, — Xo) + (1 — 1(X2 — Xo), |X — Xo] S t1X; — Xo] + (1 — t)IX2 — XQ] < 6. Hence x € U. The convex subsets of E” have many remarkable properties. There is an extensive mathematical literature devoted to them [6] [9] [13]. In Section 1.5 we give a brief introduction to the theory. The main result (Theorem 1.1) is the characterization of closed convex sets as intersections of closed half- spaces.1 Euclidean spaces PROBLEMS 1. Let n = 3. Find the plane that contains the three points e,, e,, and e, — 3e,. Sketch its intersection with the first octant in E°. 2. (a) Find the hyperplane in E* containing the four points 0,e, + e2,e, — e, +2e5, 3e, — ep. (b) Find the value of t for which t(e, — e,) + (1 — te, is in this hyperplane. 3. Let | denote the line in E* through e, — e, and —e, + e, + 2e,4. Find the hyper- plane P through e,; — e; to which / is perpendicular. 4. Let W = {(x, y, z):2x + 3y — z = 0}. Show that Y is a 2-dimensional vector subspace of E>, and find a basis for ¥.(¥ is a vector subspace of E” if x, y ¢ ¥ imply x +yeVv andcxe ¥ for any scalar c.) 5. Let ¥ = {x:z+x = 0},wherez # Ois given. Show that ¥ is an (n -- 1)-dimensional vector subspace of £”, and find a basis for W. 6. Show that {x:|x — x,| = |x — x2|}, where x, and x, are given points in E", is a hyperplane. 7. Show that {x:|x — x,| = c|x — x,|}, where x, and x, are given points in E" and 0 0 is called the radius of U. We also call U the 6-neighborhood of xg. Let A be any subset of E”. Definition. A point x is called interior to A if there is some neighborhood U of x such that U c A. If some neighborhood of x is contained in the complement A‘ = E" — A, then x is exterior to A. If every neighborhood of x contains at least one point of A and at least one point of A‘, then x is a frontier point of A. An interior point of A necessarily is a point of A, and an exterior point must be a point of A‘. However, a frontier point may belong either to A or to A‘. Figure 1.7 ExampLe |. Let U be the 6-neighborhood of xo (Figure 1.7). Let us show that every point of U is interior to U. Given x € U, let r = 6 — |x — Xo| and let V be the r-neighborhood of x. If ye V, then y — xp = (y — x) + (x — Xq) and by the triangle inequality ly —Xol sly — x] + [x — Xol, ly —Xxol dis exterior to U. If |x — x9| = 6, x is a frontier point. EXAMPLE 2. Let n = 1. Then neighborhoods are open intervals (x9 — 6, Xo + 6). Let A be the set of all rational numbers. Any open interval contains both rational and irrational numbers. Hence every point of E! is a frontier point of A. Definition. The interior of a set A is the set of all points interior to A. It is denoted by int A. The set of all frontier points of A is the boundary (or frontier) of A, and is denoted by fr A. The set A u fr A is the closure of A and is denoted by cl A. , 151 Euclidean spaces In Example 1, int U = U and fr U is the (n — 1)-dimensional sphere of radius 6. In Example 2, int A is the empty set, and fr A = cl A = E!” EXxamPLe 3. Let A = E". Then int E” = cl E” = E" and fr E” is the empty set. EXAMPLE 4. Let A be a subset of E', which has an upper bound. Then sup Ae frA, where sup A denotes the least upper bound (Section 1.1). Similarly, inf A € fr A if A has a lower bound. The closure of a set A consists of those points not exterior to A. Thus (cl A)’ = int (A‘). It is always true that int A c A. If these two sets are the same, then A is called an open set. Definition. A set A is open if every point of A is interior to A. From the above examples any neighborhood is an open set and E” is an open set. The empty set furnishes another example of an open set. The interior of any set is an open set (Problem 4). We encounter many more open sets, starting with the homework problems at the end of the section. The collection of all open sets define what is called the topology of E". In Chapter 2 we see that such properties as connectedness of a set and continuity of a function can be expressed in terms of open sets. These are examples of topological properties. Let us next show that the union of open sets is open, and that the inter- section of finitely many open sets is open. We begin by considering two sets A and B. Proposition 1.1. If A and B are open sets, then A U B and A Bare open. Proor. To show that A u B is open, we show that every point of A U B is interior to AU B. Let xe A U B. We must find a neighborhood U of x such that U c AU B. By the definition of the union of two sets, either x € Aorxe B.Ifx e€ A, then there is a neighborhood U of x such that U c A. Since A < AU B,U cA UB. Similarly, if x € B there is a neighborhood of x contained in A u B. This proves that A U B is open. If x € AB, then x has neighborhoods U,, U, such that U, < A and U, c B. Let U3; = U, V Uj. Then U; is a neighborhood of x and U3; ¢ A © B. Therefore A - B is open. O In order to discuss unions and intersections of more than two sets, let us introduce some set-theoretic notation. By indexed collection of sets let us mean a function with domain some nonempty set ¥ (called an index set) whose values are subsets of some set S. In the present instance § = E”. 161.4 Basic topological notions in E” Let A, denote the value of the function at p € %. Moreover, let U A, = {pe S:peéA, for some pe ¥}, wes () A, = {pe S:peéA, for every pe F}. ues These sets are, respectively, the union and intersection of the indexed collec- tion. If ¥ is a finite set, then the indexed collection is called finite. For instance, if % = {1,2,...,m}, we write the collection as A,,A2,...,A,. If 4 = {1, 2,...}, then the indexed collection is an infinite sequence of sets and is written A,, A2,..., or [A,,], m = 1, 2,... In that case the union is written A, U A, U-++ or ()Z_, A,, with similar notations for the intersection. Proposition 1.2. The union of any indexed collection of open sets is open. The intersection of any finite indexed collection of open sets is open. The proof is almost the same as for Proposition 1.1. Examp_e 5. Let n=1 and A= {x:0 0. (b) {x:|x — Xo] = 6},6 > 0. (c) {(x, y):0—-I}. (d) {(r cos 0, r sin 0):0 c} is a closed set, using (a), and that the hyperplane {x:z°*x = c} is a closed set. 181.5 Convex sets 6. Show that: (a) fr A = fr(A‘). (b) cl A = cl(cl A). (c) fr A = cl A ncl(A‘). (d) int A = (cl(A9)*. 7. Show by giving examples that the following are in general false: (a) int(cl A) = int A. (b) fr(fr A) = fr A. 8. Let A be open and B closed. Show that A — B is open, and that B — A is closed. 9. Show that: (a) int(A q B) = (ant A) 7 (int B). (b) cl(A U B) = (cl A) U (cl B). (Hint: Part (a).] 10. Show that: (a) int(A U B) > (int A) uv (int B). {b) cl(A 7 B) < (cl A) A (cl B). Give examples in which = does not hold. *1.5 Convex sets In this section we give a brief introduction to the theory of convex sets. Our treatment of convexity continues in Section 3.6 with a discussion of convex and concave functions. Proposition 1.4. If K,,...,K,, are convex sets, then their intersection K,---OK,, is convex. Proor. Let x,, x, be any two points of K, N--- A Ky, X; # X2. Let | denote the line segment joining x, and x,. For eachj = 1,...,m,x,,X. € Kj. Since K; is convex | c K; for each j = 1,...,m. Thus! cK, 0--- 0 Ky. In particular, Proposition 1.4 applies if each K; is a half-space. A set that is the intersection of a finite number of closed half-spaces is called a convex polytope. Since a half-space is a convex set, any convex polytope is a convex set. ExampLe 1. Let T be a triangle in the plane E?. Then T is the intersection of three half-planes, bounded by the lines through the sides of T. A convex polytope is the set of all points x that satisfy a given finite system of linear inequalities of the form z/-x > ci, j = 1,...,m. The theory of linear programming is concerned with the problem of maximizing or minimizing a linear function subject to such a system of linear inequalities. It has various interesting economic and engineering applications [10, 13]. In Section 3.6 it is shown that the maximum and minimum values of a linear function must occur at “extreme points” of K, at least if K is compact. 191 Euclidean spaces In the proof of Proposition 1.4 we did not really use the fact that the num- ber of sets K ; is finite. Therefore we have: Proposition 1.5. The intersection of any collection of convex sets is a convex set. O In particular, the intersection of any collection of half-spaces is convex. The intersection of any collection of closed sets is a closed set. Hence, if each of the half-spaces is closed, the intersection is a closed, convex set. An im- portant fact about closed convex sets is that, excluding trivial cases, the converse holds. The converse can be stated in a slightly sharper form, in that only half-spaces bounded by supporting hyperplanes need be used (Theorem 1.1). In order to do this we first state the following. Definition. Let K be a closed convex set. Assume that K is neither the empty set nor E”. A hyperplane P is called supporting for K if PO K is not empty and K is contained in one of the two closed half-spaces bounded by P. If P is supporting for K, the set P ~ K is convex by Proposition 1.4, and contains only boundary points of K (Problem 5). Given any boundary point of K, there is at least one supporting hyperplane containing it. This can be deduced from Theorem 1.1 and results to be proved in Section 2.4 (Problem 7). If K has interior points and the boundary fr K is “sufficiently smooth,” then given y e fr K, there is just one supporting hyperplane containing y. It is the tangent hyperplane to fr K at y, and can be found by the methods of calculus. This will be explained in Section 4.7. If, for example, T is a triangle in E”, then each vertex is contained in an infinite number of supporting lines to T. Each other boundary point is con- tained in a single supporting line, the line through the edge of T containing it (Figure 1.8). 7 7 4 Supporting line Figure 1.8 EXAMPLE 2. Let B = {x:|x| < 1}, the closed unit n-ball. Then fr B = {x:|x] = 1}. 201.5 Convex sets Given y € fr B, let Hy = {x:y-x < 1}, Py = {x:y+x = 1}, so that P, is the hyperplane bounding H,. By Cauchy’s inequality and the fact that |y| = 1, y-x <|yl |x| = |x|. Equality holds if and only if x is a positive scalar multiple of y. Hence B < Hy and B 2 Py consists only of y. The supporting hyperplane to B at y is Py (Figure 1.9). YS OQ. “< L, . Figure 1.9 Again let K be any nonempty, closed convex set that is a proper subset of EK # E"). Let #, denote the collection of all closed half-spaces H such that K c H and the hyperplane P bounding H is supporting for K. For instance, the collection #3 in the above example consists of the various half- spaces H, for all possible choices of y € fr B. The notation NH He #x stands for the intersection of all half-spaces H € #. Theorem 1.1.K = () H. He.#K PROOF. For convenience let us set K,= ()\ H. Hew#K Since K c H for each H € #x, K < K,. Let us show that K = K,. Suppose it is not. Then there exists some x, € K, — K. Since K is a closed set, there is a point xy € K nearest x,, that is, |x — x,| => |Xo — x,| for every xéK (Figure 1.10). This fact about closed sets is proved in Section 2.4. 211 Euclidean spaces fe ; 4 xy za Figure 1.10 LSS é ts Consider the closed half-space Ho = {x:(Xq — X,)*(X — Xo) > Of. Then x, ¢ Ho since (Xo — X41) * (XK, — Xo) = — [Xo — X4/? <0. To show that K < Hp, let x be any point of K. Since K is convex, tx + (1 — t)xo € K for every t € |0, 1]. Then [tx + (1 = 1x9 — X17 = 1X) — x1”, or (Xo — X41) + (x — Xo)? = [Xo — x17, [Xo — xy |? + 2t(xq — x4) *(K — Xo) + t?]K — Xo|? > [xq — x, [?. Subtracting |x) — x, |? from both sides and dividing by r, we get for0 <1 <1 Xo — X1)*(K — Xo) + t|X — Xo|? = O. Letting t+ 0* we find that 2(x9 — x,)*(x — x9) > 0, which shows that x € Hy. Thus K c Ho. The boundary of Ho is Po = {X:(Xq — X1)* (XK — Xo) = 0}, and Xo € Py. Hence Py - K is not empty, and therefore P, is supporting for K. This shows that Hy € Hx. Consequently, K, ¢ Hy. But x,;€K,, x; Ho, a contradiction. Therefore K = K,. oO Convex combinations The definition of convex set is expressed in terms of pairs of points. It can also be given in terms of convex combinations of any finite number m of points. Let x,,...,X,, be distinct points (x; # x, ifj # k). Definition. A point x is a convex combination of X,,...,X if there exist scalars t',...,f" such that x= Y rx, l= ¥e, v>0 forj=1,...,m. j=l 221.5 Convex sets (x1, ¥1) (u, v) (x2, Yo) (xo. Yo) Figure 1.11 To say that x is a convex combination of two points of S is merely to say that x lies on some line segment with endpoints in S. For instance, if S is the circle with equation x? + y? =a’, then every point in the circular disk {(x, y): x? + y? < a?} bounded by S is a convex combination of two points of S. On the other hand, if S consists of three noncollinear points (x9, Yo), (x1, ¥1); (X2, Y2), then each boundary point of the triangle with these points as vertices is a convex combination of two points of S, but the interior points are not. However, each interior point (x, y) is a convex combination of (x2, y2) and some point (u, v) on the edge opposite (x2, y2) (Figure 1.11). Since (u, v) is a convex combination of (x9, yo) and (x,, y,), we can write (x, y) as a convex combination of the three points (x9, Vo), (X1, ¥1), (Xz, 2) as follows. Writing x = (x, y), x; = (x;, y,), there exist s, te [0, 1] such that x = t[sxg + (1 — 5)x,] + (1 — dx. = 19x + t'x, + 1?x2, where t° = ts,t! = t(1 — s),t? = 1 — tarenonnegativeandr® + t! + 1? = 1. Proposition 1.6. A set K is convex if and only if every convex combination of points of K is a point of K. Proor. Let K be convex. Let us prove by induction on m that if x is any convex combination of x,,...,x,,¢K, then xe K. The case m = 2 is the definition of convexity. Assuming the result true for the integer m > 2, let x be a convex combination of points X,,..., Xm41 Of K, mt+1 m+. | x= Yrx, 1= Yer, #20 forj=l,...,mt. j=l j=l Ifr"*! = 1, then = Oforj < mand x = x,,,; isin K. If r"*! < 1, let J m+ (Pleo eS ff (OTe al eerie y= DY six;,. j=l Then y is a convex combination of x;,...,X,,. By the induction hypothesis, yéK. But x=tyt+ (1 —8)Xn41, and t € (0, 1]. Hence xe K. 231 Euclidean spaces Conversely, assume that every convex combination of points of K is a point of K. In particular, this is true for convex combinations of any pair X,, X2 of points of K. Hence K is convex. oO Let x9, X,,...,X,,r n+ 1, />0 for j = 1,...,m, and x,,...,x,,€ 5S. Let us show that x is a convex combination of m — 1 points of S. Since m — 1 > n, there exist c',...,c™~! not all 0 such that OMX, Xp) fe FCT X= — Xm) = 0. Let c™ = —(c! + +--+ c™7'). Then m . m . ¥ cix; = 0, Yc =0. j=l j=l Let si=th—aci forj =1,...,m, where « is a positive number chosen so that s/ > 0 for each j = 1,...,m and s* = 0 for some k. Explicitly, 1 c! c™ — = max4-;,.--,—?- a t ” x= Vox, 1= Ys, j#k j#h Then and consequently x is a convex combination of the m — 1 points x,,...,X,-}, Xp tasss ss Xme Either m — 1 =n + 1, or else the same argument shows that x is a convex combination of m — 2 points of S. Continuing, we find that x must be a convex combination of n + 1 or fewer points of S. oO If S is the set of vertices of an n-simplex T and each of the barycentric coordinates of x is positive, then x is not a convex combination of fewer than n + 1 points of S. Hence the number n + 1 is the best possible in Proposition 1.7. A slightly better result is possible if S is a connected set. A set S is called disconnected if there exist open sets D and E such that S CDUE,DQS and E 4S are both nonempty, but Dm E 4S is empty. Connected means not disconnected. For further discussion of connectedness, see Section 2.7. Proposition 1.8. If S is a connected subset of E" and x is a convex combination of points of S, then x is a convex combination of n or fewer points of S. Proor. Suppose that x* is a point which is a convex combinatiun of n + 1 points X9, X,,..., X, of a connected set S, but not of fewer than n + 1 points of S. The differences x; — X9,...,X, — Xo form a linearly independent set; if not, the reasoning used to prove the proposition above shows that x* is a convex combination of n of the points x9, X;,...,X,. Therefore Xo, ' 251 Euclidean spaces Figure 1.13 X,,...,X, are the vertices ofan n-simplex T, and all the barycentric coordinates of x* are positive. Let Ty be the face of T opposite xy, and Ko = {x:x* = tx + (1 — ty, where y € Ty and t € (0, 1]}. K, is a convex polytope, and its boundary fr Kg consists of portions of the hyperplanes which contain x* and the (n — 2)-dimensional faces of Ty (we leave the verification of this to the reader). If fr Kg intersects S, then x* is a convex combination of fewer than n + 1 points of S, contrary to hypothesis (Figure 1.13). Hence S 4 fr Ky is empty. The interior int Ky and the comple- ment Ko = E" — Kg are open sets, their union contains S, and their inter- section is empty. But x, € int Ky and x; € K6 for i = 1,...,n. This implies that S is disconnected, which is a contradiction. By slightly refining the proof, an even stronger result is obtained. Suppose that S=S, U---US,, where k O}. (d) {((x, y):0 3. Let x be a convex combination of x,,..., X,, and let x; be a convex combination Of Yi1,--+5 Yim,» J = 1,...,m. Show that x is a convex combination of z,,...,Z,, which are the distinct elements of the set {y,:k = 1,...,mj,j = 1,...,m}. Let S be any subset of E”. The set S of all convex combinations of points of S is the convex hull of S. (a) Using Problem 8, show that S is convex. (b) Using Proposition 1.6, show that if K is convex and S < K, then § ¢ K. Thus the convex hull is the smallest convex set containing S. Given xo and 5 >0, let C = {x:|x' — xh] <6, i=1,...,n}, an n-cube with center Xq and side length 26. The vertices of C are those x with |x‘ — x)| = 6 for i = 1,...,n. Show that C is the convex hull of its set of vertices. [Hint : Use induction onn.] Let K be a closed subset of E” such that both K and its complement E" — K are nonempty convex sets. Prove that K is a half-space. Let A and B be convex subsets of E”. The join of A and B is the set of all x such that x lies on a line segment with one endpoint in A and the other in B. Show that the join of A and B is a convex set.Elementary topology of E” In this chapter we explore various ramifications of the topological notions introduced in Section 1.4. We begin with some basic properties of functions, limits, and continuity. Then we turn in Sections 2.3 and 2.4 to some properties of E" that depend essentially on the least upper bound axiom for real numbers. Among these properties are the convergence of Cauchy sequences (Theorem 2.2) and the Bolzano-Weierstrass theorem. In Section 2.6 a very general concept is introduced, that of topological space. In particular, any set S < E" becomes a topological space with the relative topology which S inherits from E”. The concepts of connected set and compact set are intro- duced in Sections 2.7 and 2.8. It turns out that these properties are preserved by continuous functions (Theorems 2.8 and 2.10). At the end of the chapter, the concept of metric space is introduced, and some interesting special cases are considered. Uniform convergence of a sequence of functions is studied as convergence in a certain metric on a space of bounded functions. 2.1 Functions One should think of a function fas assigning to each element p of some set S an element f(p) of another set T. The element f(p) is called the value of f at p. However, this is not a satisfactory definition of “function” because of the ambiguity of the word “assigns.” We give a more careful definition, in terms of cartesian product sets. Cartesian product sets If S and Tare sets, then the cartesian product set S x Tis formed by taking all ordered pairs (p, q) where p € S and q € T. For example, if S = {1, 2,...,n} and T = {1,2,...,m}, then the elements of S x Tare pairs (i,j) of positive 282.1 Functions integers with 1 1, are called vector valued and are indicated by boldface letters (say, f or g). A vector valued function f from a subset D of some euclidean E" into E™ is also called a transformation from D into E”. By merely writing “transformation” in place of “vector-valued function,” we have of course introduced no new mathe- matical idea. However, the word “transformation” is supposed to have a geometric flavor which aids intuition. Some authors say “mapping” instead of “transformation.” The differential calculus of transformations is developed in Chapter 4. If f and g are functions with the same domain S and values in a vector space ¥, then the sum f + g is defined by (f + 9)(p) = f(p) + 9(P) for every peéS. In particular, it makes sense to speak of sums of real valued functions or of transformations. If fhas values in Y and ¢ is real valued, then of is the Y-valued function given by ($f )(p) = OP) Ff (P) 292 Elementary topology of E” for every péS. If ¢ is a constant function, $(p) = c for every peé S, then we write cf instead of $f. Restriction of a function Often one is interested only in the values of a function f for elements of some subset A of its domain. The restriction of f to A is the function with domain A and the same values as f there. It is denoted by f | A. Thus SIA = ((p, f(p)): pe A}. For instance, if a real-valued function fis integrated over an interval I < E', then it is only f |J which is important. The values of f outside I do not affect the integral. Images, inverse images Let f be a function from a set S into a set T. The image under f of a set A c S is the set f(A) = {f(p): p€ A}. It is a subset of T, and in fact the restriction fA is a function from A onto f(A). The inverse image of a set B < T is the set f~'(B) = {p: f(p)€ B}. It is a subset of S. Example |. Let f(x) = x?. Then f([—2,2]) = [0,4], f(E’) = [0, 00), f7MOL 3D = [-/3, -1] v [1 3]. The function f is not univalent since f(-x) = f (>). EXAMPLE 2. Let f be a function from a set S into a set T. Show that (*) Ac f~ (f(A) for any A c S. Consider any pé A. Then f(p) € f(A) by definition of f(A). Take B= f(A) in the definition of inverse image set above. Then pe f~'(f(A)). Since this is true for each p € A, we get (*). EXAMPLE 3. Show that if fis univalent in Example 2, then (**) A= f~'(f(A)). It suffices to show that A > f~'(f(A)), since the opposite inclusion is (*). Consider any pe f~'(f(A)). Then f(p) € f(A). Therefore f(p) = f(p’) for some p’€ A. Since fis univalent, p = p’. Thus pé A as required. PROBLEMS 1. (a) Let f(x) = cos x. Find f(E’), f([—1/4, 2/2]), f~ \({0, 1). (b) Let g = f |[0, x]. Find g([0, 7), g~ ‘([0, 1]). Is g univalent? Nn . The equations s = (x? + y?)!/?,r = x — ydefine a transformation f from E? into E?, such that f(x, y) = (s, t). Let A = {(x, y):x? + y? < a?}, where a > 0 is given. (a) Find f(A). (b) Find f~1(A). 302.2 Limits and continuity of transformations 3. Let f be a function from S into T. Show that, for any B c T: (a) B> f(f~*(B)). (b) B= f(f~ \(B)), if f is onto B. 4. Let f be a function from S into T. Show that, for any subsets A and B of S: (a) f(A v B) = f(A) U f(B). (b) {(A 0 B) < f(A) 9 f(B). (c) If f is univalent, then f(A 4 B) = f(A) 7 f(B). 5. Let f be a function from S into T: Show that, for any subsets C and D of T: (a) f- (CUD) =f" (Chu f(D). (b) f- (COD) =f" (C)nf- (D). (c) f- (DY = Lf (DY. 6. Let S and T be sets, and let (p,q) = p for all p € S, q € T. The function x projects S x T onto S. Let R < S x T bea relation. Show that R is a function if and only if n|R is univalent and onto S. 7. Let 1 yo as x > Xo are used to mean that yo is the limit of f at xo. 312 Elementary topology of E” © f If we let ¢ and 6 denote the radii of V and U respectively, then the defi- nition may be rephrased as follows: f(x) > yo as x > Xq if for every ¢ > 0 there exists 6 > 0 such that | f(x) — yo| < € whenever 0 < |x — xo| < 6. The number 6 depends, of course, on ¢ and may also depend on xy. Given ¢ and Xo, there is a largest possible 5. However, there is ordinarily no reason to try to calculate it. Let us first show that limits behave properly with respect to sums and products. Let f and g have the same domain and values in the same euclidean E™. Proposition 2.1. If yo = lim,..,, f(x) and Z) = lim,..,, 9(x), then (1) Yo + Z = lim, ..,,[ f(x) + g(x)]. (2) cyo = lim,.,, ¢f(x), for any scalar'c. (3) Yor Zo = lim, +,, f(x) - g(x). Proor. Let V be any neighborhood of yg + Zo and let ¢ be its radius. Let V,, V2 be the neighborhoods of radius ¢/2 of yo, Zo, respectively. If ye V, and ze V,, then (y + Z) — (Yo + Zo) = (Y — Yo) + (Z — Zo). By the triangle inequality é é€ ly + 2) — (Yo + Zo) Sly — Yol + 1Z— Il <5 +5 =8 Hence y + ze V. By hypothesis there exist punctured neighborhoods U,, U, of x9 such that f(U,) c V, and g(U,) c V,. Let U = U, NU), which is also a punctured neighborhood of xo. If xeU, then f(x)eV,, g(x)e V2. Consequently, f(x) + g(x)€V, which shows that (f + g)(U) < V. This proves (1). The proof of (2) is left to the reader (Problem 2). To prove (3) let Vo be the neighborhood of yy of radius 1, and Uy be a punctured neighborhood of x such that f(Uo) < Vo. Let C = max{lyol + 1,1Zol}- 322.2 Limits and continuity of transformations If ye VY, then y = yg + (y — Yo). By the triangle inequality lyl < lyol + ly — yol < l¥ol + 1, and hence |y| < C. Now f(x) B(x) — Yo-Zo = f(x)- [e(x) — Zo] + 2 -[A(x) — yo. From the triangle inequality and Cauchy’s inequality, (*) f(x) + B(x) — Yo * Zol < If(x)I1g(x) — Zol + 1ZolIf(x) — yol- Given ¢ > 0, let V,, V. be the neighborhoods of radius ¢/2C of yo, Zo, respec- tively, and let V, = WO V,. If e < 2C, then V', = V,. By hypothesis there are punctured neighborhoods U,, U2 of Xp such that f(U,) < V';,g(U2) < Vp. Let U = U, \ U3. For every x € U, f(x) € Vo and hence | f(x)| < C. From (*), & € | f(x) B(x) — Yo+Zol < Coet Cre=e for every x € U. This proves (3). O Definition. A transformation f is called bounded on a set A if there exists C such that |f(x)| < C for every x € A. In the course of the proof we showed that if f has a limit at xo, then f is bounded on some punctured neighborhood U) of Xo. Limits of components Since f has values in E”, f(x) = (f108),.-.5 3) where f! is a real valued function called the ith component of f (with respect to the standard basis of E”). For instance, in Problem 2, Section 2.1,f '(x, y) = (x? + y?)/? and f(x, y) =x — y. The next proposition shows that con- vergence of f(x) to yo is equivalent to convergence of each component f'(x) to the corresponding component of yo. Proposition 2.2. yo = lim,..,, f(x) if and only if yi) = lim,.,, f'(x) for each i=1,...,m. , The proof is left to the reader (Problem 10). Limits along lines Proposition 2.3. If yo = lim,.,, f(x), then for any v 4 0, Yo = lim f(xq + tv). 170 332 Elementary topology of E” Proor. Let V be any neighborhood of yo. There exists 6 > 0 such that f(x) € V whenever 0 < |x — xo| < 6. If 0 < |t| < 4/|vI, then I(Xo + tv) — xo] = ItlI¥] < 6. Hence f(xg + tv) € V for every r in the punctured 6/|v|-neighborhood of 0. O The points xp + tv lie on the line through xg and x9 + v. Roughly speaking, Proposition 2.3 states that if f has a limit yo at xo, then yo is also the limit as x9 is approached along any line containing xy. When f fails to have a limit at xo this fact can often be discovered by testing f along various lines. Example |. Let f(x, y) = x?/(x? + y?), (x, y) # (0,0), and let x9 = (0,0). Taking v = e, = (1,0), f(t,0) = 1 for every t ¥ 0. Hence f(t,0) + last > 0. Similarly, taking v = e, = (0, 1), f(0, 1) = 0 for every t # 0 and f(0, 1) > Oas t > 0. Since these limits are different, f has no limit at (0, 0). As we show below, if x9 = (x9, yo) ¥ (0,0), then f(x, y) > f(x9, yo) as (x, y) > (x9, Yo). Thus (0, 0) is the only point where there is no limit. The converse to Proposition 2.3 is false, as shown in Example 2. EXAMPLE 2. Let f(x, y) = (y? — x)?/(y* + x?), (x, y) # (0,0), and again let Xo = (0,0). Consider any v = (h, k) ¥ (0,0). Then (tk? — h)? kt +h? which tends to 1 as t > 0. However, every punctured neighborhood of (0, 0) contains part of the parabola y? = x, and f(y?, y) = 0. Hence fdoes not havea limit at (0, 0). f(th, tk) = Continuity Let us now suppose that x, is an interior point of the domain D of the trans- formation f. Definition. A transformation f is continuous at Xo if f(xo) = limy.,, f(x). When f is continuous at x9, punctured neighborhoods may be replaced by neighborhoods. The definition may be restated: f is continuous at x9 if for every neighborhood V of f(x) there is a neighborhood U of xo such that f(U) < V. ExampLe 3. Let f(x, y) = (xy)'/?. Show directly from the definition that f is continuous at (0, 0). Denote the radii of the neighborhoods U and V by 6 and é, respectively. We must show that given ¢ > 0 there exists 6 > 0 such that | f(x,y) — f(0,0)| < ¢ whenever x? + y? < 67. Note that (0,0) = 0. 342.2 Limits and continuity of transformations From the inequality x? + 2xy + y? > 0, we get Ixyl 0, we choose 6 = ./2e3. Then x? + y? < 6? implies | f(x, y)| < € as required. One does not ordinarily have to verify continuity of a transformation directly from the definition (as in Example 3). Instead, one builds up a stock of transformations known to be continuous and shows that other trans- formations can be expressed in terms of these by such operations as addition, multiplication and composition of functions. Let us consider for the moment real valued functions f, g. Proposition 2.4. Ler f and g be continuous at Xy9. Then f + g and fg are con- tinuous at xq. If g(x) # 0, then fg~' is continuous at xq. This is a consequence of Proposition 2.1 and Problem 8. The continuity of a transformation f = (f',..., f™) can be verified from continuity of its components f': Proposition 2.5. f is continuous at Xo if and only if f' is continuous at Xo for eachi = 1,...,m. This is a consequence of Proposition 2.2. ExampLe 4. A constant function, f(x) = c for all x, is continuous. In the definition, the neighborhood U can be chosen arbitrarily. ExampLe 5. Let I(x) = x, the identity transformation. Then I is everywhere continuous (take U = V above). Therefore the components of I are every- where continuous. In this book these components are called the standard cartesian coordinate functions, and are denoted by X',...,X". For each x = (x!,...,x"), Xx) = x! ExampPLe 6. Any polynomial in n variables is everywhere continuous. This is proved by induction on the degree of the polynomial using the continuity of the coordinate functions X‘ and of constant functions. EXAMPLE 7. A rational function f(x) = P(x)/Q(x), where P and Q are poly- nomials, is continuous at each point where Q(x) # 0 by Proposition 2.4. For instance, in Examples | and 2, fis continuous at each (x, y) # (0, 0). 352 Elementary topology of E" It is shown in Theorem 2.7 that the composite of two continuous trans- formations is continuous. In Section 4-4 it is shown that any differentiable transformation is continuous. For a real valued function f of one variable, differentiability of f at xo is equivalent to the existence of the derivative S' (Xo). Limits at oo Let us call a set of the form {x :|x| > b} a punctured neighborhood of oo. The definition of “limit at oo” then reads: yo = lim),)... f(x) if for every neighbor- hood V of yo there exists a punctured neighborhood U of oo such that f(U) < V. When f is real valued we say that lim,.,, f(x) = +00 if for every C > 0 there is a punctured neighborhood U of xg such that f(x) > C whenever x € U. The definition of “lim,..,, f(x) = — 00” is similar. PROBLEMS 1. Find the limit at xp if it exists. (a) f(x, y) = xy/(x? + y?), Xo =e, + e2. (b) f(x, y) = xy/(x? + y?), Xo = (0, 0). (c) f(x) = (1 — cos x)/x?, x9 = 0. [Hint: lim, 9 (sin x)/x = 1.] (d) f(x) = |x — 2]e, + |x + 2]e2, xo = 3. (e) f(x, y) = ye, + (xy)?/[ixy)? + (x — y)?Je2, Xo = (0, 0). At which points is each of these functions continuous? 2. Prove (2) of Proposition 2.1. 3. Show that if'yo = limy..,, f(x), then |yo| = lim,..,, |f(x)|. Prove that the converse holds if yp = 0. 4. Let f(x) = |x|*%, where a > 0. Show that f is continuous at x9 = 0 directly from the definition of continuous function. 5. Let f(x, y) = x cos(y”')if y # 0, and f(x, 0) = 0. At which points is f continuous? 6. Find the limit if it exists. x* + yt xy? ) lim ——. (x.3)+(0.0) 7 + y* (a lim : ) (x.y+(0.0) X7 + y? () lim (x* X1)(X* Xp) Ixl= 0 , where x, and x, are given vectors not 0. (d) tim % =! Ix]+00 1X — X2| 7. (a) Let g(x) = |f(x)|* wherea > 0. Suppose that f is continuous at Xo and f(xo) = 0. Show that g is continuous at Xo. (b) Use (a) and Problem 3 to give another proof that the function in Example 3 is continuous at (0, 0). 362.3 Sequences in E” 8. Let yo = limyx. f(x), Zo = lim, x9 g(x). Show that if z) # 0 then li f(®) _ Yo im ——=—, x-x9 AX) Zo 9. Show that lim,..,, f(x) = +00 if and only if lim,..,,[f(x)]"' = 0 and f(x) > 0 for every x in some punctured neighborhood of xg. 10. Prove Proposition 2.2 in two different ways. [Hints : For one proof use the definition of limit directly and the inequality, for any vector h = (h',..., h"), [hi] < |b] < /m({ht] + ++ + Lh"). Take h = f(x) — yo. For the other proof, write f = f'e, +--- + f"e, and note that f(x): e; = f'(x).] 2.3 Sequences in E” An infinite sequence is a function whose domain is the set of positive integers. For brevity, we use the term “sequence” to mean infinite sequence. In this section let us consider sequences with values in E”. It is customary to denote by x,, the value of the function at the integer m = 1,2,..., and to call x,, the mth term of the sequence. The sequence itself is denoted by x,,x2,..., or for brevity by [x,,]. It must not be confused with the set {x,,x.,...} whose ele- ments are the terms of the sequence. This set may be finite or infinite. For instance if x,, =(—1)" then the sequence is —1, 1, —1,..., and the set {x1,X2,..-} has only two elements —1 and 1. Definition. Suppose that for every ¢ > 0 there exists a positive integer N such that |x,, — Xo| < € for every m > N. Then xg is the limit of the sequence [x,,]. The notations “x9 = lim, Xm” and “Xx, > Xo aS m— oo” are used to mean that X, is the limit of the sequence [x,,]. A sequence is called convergent if it has a limit, otherwise divergent. The integer N in the definition depends of course on ¢. Given ¢ there is a smallest possible choice for N. However, for purposes of the theory of limits it is of no interest to calculate it. What matters is the fact that some N exists. Proposition 2.6. Let xo = lim, Xm>Yo = liMmso Ym» Then: (a) Xo + Yo = limy, o(Xm + Ym): (b) cxXo = lim,,— 0 CX»,for any scalar c. (C) Xo* Yo = Vimy, Xm *Ym- Let xi, denote the ith component of the vector X,»,.2 Elementary topology of E” Proposition 2.7. Xo = lim, Xm if and only if xi, = lim... xi, for each i=1,...,n. We leave the proof of Propositions 2.6 and 2.7 to the reader (Problem 6). The proofs of the corresponding Propositions 2.1 and 2.2 in Section 2.2 can be adapted for this purpose. In this section we prove three important results (Theorems 2.1-2.3) that depend on the least upper bound property of the real numbers (Axiom III, Section 1.1). The first of these results is about convergence of monotone sequences of real numbers. A sequence [x,,] of real numbers is called monotone if either x, < x2. < x3<-:: or x, >x, =x; 2---. In the first instance the sequence is non- decreasing, in the second nonincreasing. A sequence [x,,] is bounded if there is a number C such that |x,,| < C for every m = 1,2,.... Theorem 2.1. Every bounded monotone sequence of real numbers has a limit. Proor. Let [x,,] be nondecreasing and bounded. Let x9 = sup{x,,x2...}. Given ¢ > 0 there exists an N such that x9 — xy < ¢. Otherwise x,, < Xo — & for every m = 1,2,...,and xo — € would be a smaller upper bound than the least upper bound x9. Since the sequence is nondecreasing, xy < Xm < Xo for every m > N. Hence |x,, — Xo| = Xo — Xm < € for every m > N. This shows that x,, > Xo aSm — 00. If the sequence is nonincreasing and bounded, let xo = inf{x,,x2,...}. In the same way x,, > Xo asm —> ©. EXAMPLE 1. Let 0 < a < 1. The sequence a, a’, a°, ... of its powers is de- creasing, and is bounded below by 0. The limit of the sequence is b = inf{a, a?,a°,...}. Since b < a™*!, a~'b < a™ for m = 1,2,.... Therefore a~'bis a lower bound for {a, a”, a>, ...}. Since b is the greatest lower bound, a~'b < b. However, a~'b > b since 0 < a < 1; and hence a~*b = b. This implies that b = 0. EXAMPLE 2. Let ym = )i'=o(1/k!). The sequence [y,,] is nondecreasing. Since (k))7' <27** fork >1 1 1 1 <1l4+14+e4+-4--4+—-5<3. Ym S 1+ t5tgt + amt <3 Thus [y,,] is bounded. This sequence has a limit yo. ExamPLe 3. Let x,, = (1 + 1/m)". From the binomial formula fan 1 mm —- Wt _m! i rm 2! m? "mim 382.3 Sequences in E” which can be rewritten as 1 1 1 1 2 meat t1+z(I eH! “(i =) et B12 -2)...(1- 24), m! m m m When x,, is replaced by x,,4;, the product multiplying (k!)~! becomes larger for k = 2, 3,...,m, and an additional term appears. Thus x,, < Xm44- Moreover, Xm < Vm < Yo < 3, with y,, as in Example 2. Thus the sequence [x,,] has a limit x9, and x9 < yo. In fact, x9 = yo (Problem 11). This number is the base for natural logarithms xy = e = 2.718 ---. Let us turn to the second main result of the present section. The definition of limit has the sometimes inconvenient feature that it refers not only to the sequence [x,,], but to the limit xy. This limit x9 is not known in advance, although one can sometimes guess what x, must be. The following is a re- lated concept due to Cauchy. It is stated purely in terms of the sequence, without reference to a limit. Definition. If for every ¢ > 0 there exists a positive integer N such that |x: — X,| < € for every |, m > N, then [x,,] is a Cauchy sequence. We may agree that / > m in this definition without loss of generality. ExamPLe 4. Let x,, = Yr. (cos k)/k?. Let us show that [x,] is a Cauchy sequence. If / > m, then I cos k Xr Xm = x “Ke kh=m+1 ! 1 Ix, — Xml < > 7 homer kK Since k~? < t~? fork — 1 0, choose N > ¢~!. Then [x1 — Xml 0 there exists N such that |x,, — Xo| < ¢/2 for every m > N. Now X1 — Xm = (X; — Xo) + (Xo — Xm)- If l,m => N, then by the triangle inequality é & IX; — Xml S 1X) — Xol + [Xo — Xml <5 +5 =e 2 2 Therefore [x,,] is a Cauchy sequence. The proof of the converse is more difficult. Let us first show that every Cauchy sequence is bounded. If [x,,] is Cauchy, then taking ¢ = | in the definition there is an N such that |x, — x,,| < 1 for every l,m => N. In par- ticular, let / = N and let C = max{|x,|,...,|Xw—1],1Xw/ + 1}. Then by the triangle inequality and the fact that x,, = xy + (Xm — Xy), [Xml << [Xwl + [Xm — Xwl< [xv] + 1 ifm > N. Therefore |x,,| < C for every m = 1,2,.... Next, let [x,,] be a Cauchy sequence of real numbers, and C be a number such that |x,,| < C for every m. For each m = 1,2,... let Vn = if {X ns Xin tas + Since {Xm4isXm+29---} S {Xm>Xm+is-+-$, we must have ym < Vm44. Since Xm < C for every m, y, < C for every m. The sequence [y,,] is nondecreasing and bounded. By Theorem 2.1 the sequence [y,,] has a limit yy. Let us show that x,,— Yo aS m— oo. Given ¢ > 0 there exists N such that for every m= N,|xy — Xml < &/2, and hence € € Xn — 5S Xm < XW + 5. Ifm = N, then xy — é/2 is a lower bound and xy + &/2 an upper bound for {Xm>Xm+19++-}- Therefore when m > N € € xy — 5 S Vm S Yo S XW +5; & é [Xm — Yo! S1Xm— xwl + lxw — Yo <5 +5 = % This proves that x,, > Yo aS m — 00. Finally, if [x,,] is a Cauchy sequence in E”, then for each i = 1,...,n the components form a Cauchy sequence [xi,] of real numbers. This follows from the definition of Cauchy sequence and the inequality |x} — xi,| < |x; — Xm|- Each of the sequences [x!,] has a limit y},i = 1,...,n. By Proposi- tion 2.7, x,, > Yo aS m— 00. 402.3. Sequences in E” For the next theorem we need the concept of diameter of a set. A nonempty set A is called bounded if there exists C > 0 such that |x| < C for all xe A. The diameter of a bounded set A is diam A = sup{|x — y|:x, ye A}. Theorem 2.3 (Cantor). Let [A,,] be a sequence of closed sets such that A, > A, >--: and 0 =lim,,.,, diam A,,. Then Ay 0 A, O°: contains a single point. ProoF. For each m = 1,2,... let x,, be a point of A,,. Let us show that [x,,] isa Cauchy sequence. Given ¢ > 0, there exists N such that diam Ay < . If l,m > N, then x;, x,,€ Ay since A; © Ay, Am © Ay. Therefore |x; —X,| < diam Ay I since A,, € A. Since A, is closed, X9 € A; (Problem 5). Since this is true for each 1, xg €A, AV AQ O-°°. IfxeA, OA, ----, then xeA,, and 0 < |x — xo| < diam A,,, m = 1,2,... Since diam A,, > 0 as m > 00,|x — Xo| = Oand x = Xo. O Exampte 5. Letn = 1, A,, = {m,m + 1,m + 2,...}. Then each A,, is closed, unbounded, and A, > A, >---. The intersection A, A A, --: is empty. If in Theorem 2.3 it is not assumed that the diameter of A,, tends to 0, then it is still true that A, 0 A, 1--- is not empty, provided some set A,, is bounded. This is proved in Section 2.4. Note. Theorems 2.1-2.3 depend in an essential way on the least upper bound Axiom III. Let us suppose, on the other hand, that instead of Axiomt III we took the following. Axiom III, (a) E! has the archimedean property. (b) Every Cauchy sequence in E' has a limit. Then Axioms III and III, are equivalent in the presence of Axioms I and II, Section 1.1. This means that if either III or III, is taken as an axiom, then the other can be proved as a theorem. We have shown that III implies III,. To show that III, implies III, we first note that if III, is taken as an axiom, then Theorem 2.3 remains valid (with the same proof). It then suffices to deduce the least upper bound property from the archimedean property and Theorem 2.3. To prove this, let S be any subset of E! that is bounded above, and let c be some upper bound for S. Let us define a sequence of closed intervals : 42 Elementary topology of E” Ss — ~ Y + + a \ J Cc Lo Figure 2.2 I,, 1,,... aS follows: Let a be some point of S, and I, = [a,c]. Divide I, at the midpoint (a + c)/2 into two congruent closed intervals. If (a + c)/2 is an upper bound for S, let I, be the left-hand interval, otherwise let I, be the right-hand interval. In general, suppose m > 1 and I,, has been defined. If the midpoint of J,, is an upper bound for S, let J, , be the left half of I,,, otherwise the right half. The archimedean property implies that for any x > 0, the sequence [m7 'x] tends to0 asm > oo. Since0 < 2°" < m™!, the sequence [2~”x] also tends to 0. Let x = 2(¢ — a). Now 1, > 1, > +++ and the length of I, is 2-"x. By Theorem 2.3, 1, © 1, © --- contains a single point xo. By the construction, x9 = sup S (see Figure 2.2). Infinite series Formally, an infinite series is an expression written )& ; x,0rx, + X2 + °°: To be more precise, with any sequence [x,] is associated another sequence [s,,], where s, = x, +--+ +-x,, is called the mth partial sum. This pair of sequences defines an infinite series. If the sequence of partial sums has a limit s, then the series is convergent and s is its sum. This is denoted by s = x, + x, +---. If the sequence of partial sums has no limit, then the series is divergent. If s=x, +x, +-:-, t=yy+y2+---, then s+t=(x, +y,)+ (x2 + y2) +--+ and cs = (cx,) + (cx2) +--+ for any scalar c. This follows from the definition and Proposition 2.6. Some further elementary properties are given in Problems 7(c) and 8. PROBLEMS In Problems 1 and 2 you may use the results of Problems 9 and 10. 1. Find the limit if it exists. (a) X_ = (2" — 2>-™)(3" + 37"). (b) X_ = sin(mn/2). (c) X» = sin mn. (d) Xm = (m + Im — 1)". (Hint: (1 + 1/m)" > e asm >.) (©) Xm = ((m? + 1m? — 1)". N . Find the limit if it exists, using Proposition 2. (a) (Xm Ym) = (1 + m1 — 2m), 11 + m)). (b) (Xs Ym) = (27, 1 + m). (C) (ms Ym) = (1 — 27", (m? + 3")/m!). 4210. 11. 2.4 Bolzano-Weierstrass theorem . Show that a sequence [x,,] has at most one limit x). [Hint: If yy were another limit, let € = |Xq — yo|/2.] |. (a) Let A bea closed set. Show that if x,, € A form = 1, 2,...and Xo = lim, Xm then x9 € A. (b) If A is not closed, show that there exists a sequence [x,,], with x,,¢ A for m = 1, 2,..., converging to a limit Xp ¢ A. . Show that if x,, € A for every m > | and Xo = lim,,... X,,, then X9 € cl A. . (a) Prove Proposition 2.6. (b) Prove Proposition 2.7. « (Comparison tests.) Show that: (a) If 0 < Xm < Ym for every m > land y,, + 0 asm — oo, then x,, - 0 asm — oo. (b) If [xn], Lm] are nondecreasing sequences such that x,, < y,,foreachm = 1,2,... and y,— y as m— oo, then [x,,] has a limit x < y. (c) If O < xm < Ym for every m = 1,2,... and t = y; + y2 +---, then the series xX, + x2 +++: converges with sums 0, then (a) lim,,+. a" = 1. (b) lim,,.., a"/m! = 0. (c) Litmyy+0c(Xm)'/" = 1 provided limy,¢ Xm = a (Hints: For part (a) reduce to the case 0 < a < 1. By Example |, if b < 1 then a < b” for only finitely many m. For part (b), compare with the sequence [c/m] for suitable c and suitable | in Problem 7(a).] Let xo = limpnx Xm> Yo = limn~ Ym, and assume that y,, # 0 form = 0,1, 2,.... Show that Xo/yo = lity) Xm/Ym- [Hint: By (c) of Proposition 2.6 it suffices to show that yo! = lim,,.4 Yn) J Show that x9 = yo in Example 3. 2.4 Bolzano—Weierstrass theorem Suppose that an infinite number of points lie in a box. It is intuitively reason- able that they cannot remain scattered but must accumulate at some points of the box. The purpose of the present section is to put this idea on a precise basis. We begin with definitions of the concepts of isolated point and accumu- lation point of a set A c E”. Definition. A point x is an isolated point of A if there exists a neighborhood U of Xo such that A MU = {xo}. Definition. A point x9 is an accumulation point of A if every neighborhood of Xo contains an infinite number of points of A. , 432 Elementary topology of E” In the definition of accumulation point we do not require that x9 € A. However, the next proposition shows that accumulation points are in the closure of A. Proposition 2.8. A point Xq is an accumulation point of A if and only if xo € cl A and Xq is not an isolated point of A. Proor. Suppose that x €cl A and X, is not an isolated point of A. Let U, be any neighborhood of x9. From the definitions of closure and isolated point there exists x; € AM U,, xX, # Xo. Let us show that Am U, contains infinitely many points. Suppose not. Then Am U, contains finitely many points x,,...,x, different from xg. Choose a smaller neighborhood U of radiusd < min{|x,, — Xo|,m = 1,...,/}. Either A M7 Uisempty or A NU = {xo}. The first possibility contradicts the assumption that x9 €cl A, and the second the assumption that xo is not isolated. Thus x9 is an accumulation point of A. The converse follows immediately from the definitions. oO EXamPLe |. Every point of the closed interval [a, b] is an accumulation point of the open interval (a, b). ExampLe 2. Let A = {1,4,4,...}. Every point of A is an isolated point of A. The set A has the single accumulation point 0. ExampLe 3. Let A = {all rational numbers}. Then every point of E! is an accumulation point of A. The next proposition characterizes accumulation points in terms of con- vergent sequences. Proposition 2.9. A point xq is an accumulation point of A if and only if there exists a sequence [X,,] with limit xo, such that x,,€A and Xm, # Xo for m= 1,2,.... The proof is left to the reader (Problem 2). We come now to the main results of the present section. A set A is called infinite if the number of its elements is not finite. A set I < E" of the form ; ; a l= fave —xol I, > 13 > --- such that AQ], is infinite for each | = 1,2,... and diam I, > 0 as | > oo. By Theorem 2.3, I, O1,¢.--- has a single point xo. If U is any neighborhood of xo, then I, < U for large enough |. Since AN 1, ¢ AN U, A QU is an infinite set. Therefore x9 is an accumulation point of A. O The Bolzano- Weierstrass theorem has the following useful consequences. Corollary 1. Let S < E" be closed and bounded. Then every infinite set A < S has at least one accumulation point Xo € S. Proor. Since S$ is a bounded set and A c S, A is a bounded set. By the Bolzano- Weierstrass theorem, A has an accumulation point x9. By Proposi- tion 2.8, x9 €cl A. Since S is closed, cl A < S. Thus xg €S. Note. The converse to Corollary 1 is true. To prove it consider first an unbounded set S. Foreachm = 1, 2,...there exists x,, € S such that |x,,| > m. The set A = {x,,X,,...} is infinite and has no accumulation point. If S is not closed, then there exists a point xy € fr S — S. Form = 1, 2,...there exists Xm €S such that |x,, — Xo| < 1/m. The set A = {x,,X2,...} is infinite and has the single accumulation point xy. But xo ¢S. Corollary 2. Let A,,A2,... be nonempty bounded, closed subsets of E" such that A, > A, > +++. Then (\2=, Am is not empty. Proor. For m = 1,2,... choose some point x,,€A,,. Let A = {x,,X2,...}. If A is a finite set, then some x = x,, is repeated infinitely often in the sequence [x,,]-Since A, > A, >--:,x€A,,foreverym = 1,2,....Thusxe (2-1 An- , 452 Elementary topology of E” Let us consider the case when A is an infinite set. Then A < A, since each A, © Ay. Since A, is bounded, A is bounded. Let xo be an accumulation point of A. As in the proof of Corollary 1, we have xq € A, since A, is a closed set and A c A,. For each m = 1,2,..., Xo is also an accumulation point of the set {x,,,Xm41,---}. Since this set is contained in A,, and A,, is closed, we get in the same way Xq € A,,. Since this is true for each m, Xp €()\@-, An. O Corollary 3 shows the existence of a point in any closed set A nearest a given point Xo ¢ A. Corollary 3. Let A be a closed, nonempty subset of E" and Xo ¢ A. Then there exists x, € A such that |x — Xo| => |X, — Xol forall xeé A. Proor.Let S, = {x:|x — Xo| yp as! > 00, yo € S. Show that a nonempty set S ¢ E” is sequentially compact if and only if S is closed and bounded. 7. Let y be any frontier point of a closed convex set K. Show that K has a supporting hyperplane P that contains y. [Hint: Let {y,,} be a sequence of points exterior to K such that y,, tends to y asm — oo. Let x,, be a point of K nearest to y,, and Ym — Xm Un = 7° lYm — Xml Then |u,,| = 1 and x,, tends to y as m — oo. By the proof of Theorem 1.1 there is a supporting hyperplane of the form {x:u,,°(x — x,,) = 0}. Let u be an accumulation point of the bounded set {u,,u,,...} and P = {x:u+(x — y) = 0}.] 2.5 Relative neighborhoods, continuous transformations For the definition in Section 2.2 of a transformation continuous at a point Xo, it is assumed that xq is interior to the domain D. However, we often wish to discuss continuity at points which are not interior to the domain. Moreover, even if the domain D is an open subset of E", we may be interested only in the restriction of the transformation to some set S c D. We easily circumvent this apparent difficulty by introducing the idea of relative neighborhood. Definition. Let S be a nonempty subset of E”. A relative neighborhood of a point x € S is any set U such that U = S m W, where W is a neighborhood of x in E” (Figure 2.5). , 472 Elementary topology of E” U=SAW Ss Figure 2.5 In considering the relative neighborhood U = Sq W one simply ignores all points of W not in S. Let us now consider a transformation f, whose domain is a set S c E". Definition. The transformation f is continuous at Xo if, for every neighborhood V of f(x), there exists a relative neighborhood U of xo such that f(U) c V. If f is continuous at every point xX, € S, then we call f continuous on S. An equivalent form of the definition of continuity at xo is: for every & > 0 there exists 6 > 0 such that |f(x) — f(x )| < ¢ whenever x eS and |x — X9| < 6 (cf. Section 2.2.) ExampLe |. Let S = [a,b]. The relative neighborhoods are described in Problem 3. Let f be real valued. For a < xo < b, continuity at x9 means the same as that in Section 2.2. Continuity at a means the following: given é > 0 there exists 6 > 0 such that | f(x) — f(a)| < ¢ whenever a Yo aS m — oo. Choose Zn €S with y,, = £(z,,). AS before, the set B = {z,,z,,...} has an accumu- lation point Zo € S. Since yo ¢ f(S), Yo # f(z). Let V be a neighborhood of f(z) of radius $|f(Z9) — yo|. Since f is continuous at Zo, there exists a relative neighborhood U of zp with f(U) c V. In particular, y,, = f(z,,) is in V for infinitely many m for which z,, € U. This contradicts the fact that y,, > yo as m— oo. Thus f(S) is closed. If we specialize Theorem 2.4 to real valued functions, we get a theorem about the existence of maxima and minima. Theorem 2.5. Let f be a real valued function continuous on a closed, bounded set S. Then there exist x,,X€S such that J (x1) < f(x) < f(x2) for all xéS. Proor. We recall that if T ¢ E' is bounded and closed, then y, = inf T and y2 = sup T are points of T (Example 4, Section 1.4). Let T = f(S). By Theorem 2.4, T is closed and bounded. Take x; such that y, = f(x;),i = 1,2. oO The function f is said to have a minimum on S at x, and a maximum on S at x2. PROBLEMS I. In each case show that f has a minimum on S, but no maximum on S. Which assumption in Theorem 2.5 is violated? (a) S=(0, 1], f(x) = x71. Ix| (b) S = E", f(x) = T+ Ix" 2. Given Xo, let f(x) = |x — Xo|. Show that f has a minimum on any closed, nonempty set A c E”. (This gives another proof of Corollary 3, Section 2.4.) 3. Let S = [azb]. (a) Show that the relative neighborhoods of a are the half open intervals [a, c) with a0 and f(x) =0 if x <0. Let S; = [0, 0), S=E'. Show that f |S, is continuous, but f is not continuous at each point of S,. 6. Let f be a transformation with domain S. Show that f is continuous at Xo if and only if f(xo) = lim,,.,, £(x,,) for every sequence [x,,] such that x,,¢S for m= 1,2,... and x,, Xo aS m — 00. 7. Let f be continuous on E”. Suppose, moreover, that f(x) > 0 for all x 4 0, and that S (cx) = of (x) for any x and c > 0. Show that there exist a > 0 and b > 0 such that a|x| < f(x) < b|x|. [Hint: First consider {x : |x| = 1}.] 8. (Uniform continuity.) A transformation f is uniformly continuous on S < E” if given é > 0 there exists 6 > 0 (depending only on ¢) such that |f(x) — f(y)| < ¢ for every x, y€ S with |x — y| < 6. Show that if S is closed and bounded then every f con- tinuous on S is uniformly continuous on S. [Hinr: If not, then there exists « > 0 and for m = 1, 2,...,Xms Ym € S such that |f(x,,) — f(y,,)| = € and |x,, — y,,| < 1/m. Let Xo be an accumulation point of {x,,x,...}. Show that the continuity of f at Xo is contradicted.] 2.6 Topological spaces In order to proceed further with the study of subsets of E” and continuous transformations, it is convenient to introduce a very general concept—that of topological space. In this section S denotes a set, not necessarily a subset of E", and p denotes a point of S. The notion of topological space occurs in practically all branches of mathematics. There are several equivalent definitions; of these, we give the one in terms of neighborhoods. Definition. Let S be a nonempty set. For every pe S let %, be a collection of subsets of S called neighborhoods of p such that: (1) Every point p has at least one neighborhood. (2) Every neighborhood of p contains p. (3) If U, and U, are neighborhoods of p, then there is a neighborhood U; of psuch that U; c U; N U2. (4) If U is a neighborhood of p and q€ U, then there is a neighborhood V of q such that V c U, Then S is a topological space. More precisely, the topological space is S together with the collections %, of neighborhoods. However, it is common practice to omit explicit reference to the collections of neighborhoods when no ambiguity can arise. For our purposes, the following two examples of topological spaces are of primary importance. 502.6 Topological spaces ExaMPLE |. Let S = E”, and as in Section 1.4, let %, be the collection of all open spherical n-balls with center x. Clearly, Axioms (1) and (2) of the definition above are satisfied, and in (3) we may take U3; = U,; NU}. Axiom (4) is verified in Example 1, Section 1.4. Thus E” is a topological space. EXAMPLE 2. Let S c E". Let neighborhoods of x € S be all relative neighbor- hoods U = S$ m W, where W is a neighborhood of x in E”, as in Section 2.5. The topology on S defined by the collections of relative neighborhoods is the relative topology. It is discussed further later in the section. Roughly speaking, the relative topology is the one obtained by simply ignoring the comple- mentary set SC = E” — S. Further examples of topological spaces appear in the homework problems. The metric spaces (Section 2.9) furnish an important class of topological spaces. In any topological space S, the basic notions of interior, frontier, and closure are defined just as in Section 1.4 for the topological space E". For instance, p is interior to a set A < S if some neighborhood of p is contained in A. Definition. Let S be a topological space. A set A < S is open if every point péA is interior to A. A set A < S is closed if S — A is open. Axiom (4) (in the preceding definition) guarantees that any neighborhood is an open set. Propositions 1.1-1.3 in Section 1.4 about unions and inter- sections of open sets (or closed sets) remain true in any topological space. The proofs are almost the same as before. It often happens that two different collections of neighborhoods %,, %, lead to the same collection of open subsets of S. In E” we need not have started with spherical neighborhoods. For instance, the neighborhoods obtained from any noneuclidean norm on E" lead to the same open sets as in Section 1.4. (cf. Section 2.11). The open sets, and not the particular kinds of neighborhoods from which they were obtained, determine all of the topological properties of S. Thus we say that the collections %,, W, define the same topology on S if they lead to the same collection of open sets. EXAMPLE 3. Let S = E". Let %, consist of all n-cubes U' = fy: ly — x! <5, =1,...n} for any 5 > 0. The reader should verify that the collections W;, of “neighbor- hoods” in this sense define the same topology on E” as the usual topology of E" in Example | (Problem 1). The key fact is that each U’ € %, is contained in some U € &%,, and vice versa. 512 Elementary topology of E” Continuous functions The concept of continuous function from one topological space into another is a natural extension of that already considered for transformations between euclidean spaces. Definition. Let fdenote a function from a topological space S into a topological space T. The function f is continuous at po if for every neighborhood V of f (po) there exists a neighborhood U of pg such that f(U) < V. IfS c E",T = E”,and S is given the relative topology, then this definition of continuity agrees with that in Section 2.5. Proposition 2.4’. Let f and g be real valued functions, continuous at po. Then f + gand fg are continuous at po. If (po) # 9, then fg ' is continuous at po. Proposition 2.5’. A function f, from a topological space S into E™, is continuous at Po and only if each component f' is continuous at py,i = 1,...,m. We leave the proof of Propositions 2.4’ and 2.5’ to the reader (Problems 11 and 12). Definition. If fis continuous at every p € S, we call f continuous on S. The following is a convenient characterization of continuous functions, in terms of open subsets of S and T. Theorem 2.6. f is continuous on S if and only if the inverse image f~'(B) of any open set B is open. . Proor. Let f be continuous on S and B c T be open. Let p be any point of f~'(B) and V be a neighborhood of f(p) such that V c B. Since f is con- tinuous, there is a neighborhood U of psuch thatf(U) <¢ V.ThenU c f~*(B), which shows that f ~ '(B) is open. Conversely, let f ~ ‘(B) be open for each open set B. Let p be any point of S, and V be any neighborhood of f(p). Since V is open, f ~ '(V) is open and con- tains p. Let U bea neighborhood of p such that U c f~'(V). Then f(U) < V, which shows that f is continuous at p. Since this is true for every pe S, f is continuous on S. Since the complement of an open set is closed and f~'(B°) = [f~'(B)]*, we have the following. Corollary 1. f is continuous on S if and only if the inverse image f ~ '(B) of any closed set B is closed. Let us next specialize to real valued functions f and special choices for B. 522.6 Topological spaces Corollary 2. If f is real valued and continuous on S, then {p:f(p) > c} is an open set, and {p:f(p) = c}, {p:f(p) = c} are closed sets. Proor. In this case T = E!. The semiinfinite interval B = (c, 00) is open, and {p:f(p) > c} =f *(B). The other statements are obtained by taking B = [c, oc), B = {c}, which are closed sets. By applying Corollary 2 to —fand —c, the set {p:f(p) < c} is open and {p:f(p) < c} is closed. Corollary 2 often provides a convenient way to show that a given set is open or closed. Examp_e 4. Let A = {(x,y):y? 0}; hence A is open. Similarly, the set {(x, y): y? < x? + 1} and the hyperbola {(x, y): y? = x? + 1} are closed sets. Composites Let f be a function from S into T, and g from R into S. The composite fe g is defined by (fo g(r) =f[g(r)]_ for every re R. Theorem 2.7. If g is continuous at ro and f is continuous at po = g(ro), then f°g is continuous at ro. Proor. Let V be any neighborhood of f(po) (see Figure 2.6). There is a neigh- borhood U of po such that f(U) < V. Moreover, there is a neighborhood W of ro such that g(W) < U. Then (f° g)(W) = f[g(W)] < f(U) < V. This shows that f° g is continuous at ro. Oo Exampte 5. Let D c E" be open. Given xg and ¥, let g(t) = x9 + tv for every scalar t. Then g is continuous from E! into E". By Proposition 2.6, A = {t:Xq + tve D} is open. Let f be continuous on D and let (t) = f(x + tv) forte A. Then ¢ = f° (g|A), and by Theorem 2.7, ¢ is continuous on A. This result is similar to Proposition 2.3. (fF: g(W) J Vv le Figure 2.6 532 Elementary topology of E" The following consequence of Theorem 2.7 is used in Chapter 5. EXamPLeE 6. Let 1 < s D, where D is an open subset of So. Similarly, A is closed in the relative topology if and only if A= SD where D is a closed subset of So. Proor.Suppose that A = $1 D, where D is an open subset of Sy. Any péA has a neighborhood W in S, such that W c D. Then U = SQAWisa relative neighborhood of p such that U c A. Hence A is relatively open. Conversely, let A be relatively open. For each p€é A, let U, be some relative neighborhood of p with U, c A. Then U, = Sq W,, where W, is a neigh- borhood of p in Sg. Let D = ener W,. Then D is an open subset of Sp, and A=SQOD. The statements about closed sets are then obtained by considering the sets S — Aand Sy — D. 542.6 Topological spaces EXamPLe 7. Let S = [a,b], Sy = E', as in Example 1, Section 2.5. For a-x>0O}. (b) {(x, y)ix4 + y* = 1}. (c) {x:yo*xX < |x|}, Yo a given vector. 3. Let Sy = E'. Find whether A is open relative to S, closed relative to S, or neither. (a) S= {x:a y*}. (b) S = {(x, y):x? + y? = 4}, A = {(x, y) Six? < y?}. 5. Let S be an open subset of Sy. Show that the relatively open sets are just those open subsets of Sy contained in S. , 552 Elementary topology of E” Problems 6 through 9 give some examples of topological spaces. 6. (Indiscrete spaces.) Let S be any set, and let every p € S have exactly one “neighbor- hood,” namely, S itself; that is, each %, consists of the set S only. (a) Verify Axioms (1) through (4). (b) Show that the only open sets are S and the empty set. (c) Show that any real valued function continuous on S is constant. 7. (Discrete spaces.) Let S be a topological space such that the set {p} with the one element p is a “neighborhood” of p; that is, {p} € , for each pe S. (a) Show that every subset of S is open. (b) Show that every function with domain S is continuous. (c) Suppose that S < E” with the relative topology. Show that S is a discrete space if and only if every x € S is isolated. 8. Let S=E' with the following nonstandard topology. In this topology the “6-neighborhood” of xo is the interval [xo, Xo + 4). (a) Verify Axioms (1) through (4). (b) Show that [0, 00) is both open and closed in this topology. (c) Let f be a real valued function from E! with this topology into E! with the standard topology. Show that continuity of f at xo is equivalent to f(xo) = lim,..¢ f(x), where superscript * denotes righthand limit. 9. Consider the following nonstandard topology on the plane E?: In this topology the “d-neighborhood” of (xo, yo) is the set {(x, y):x9 —d 0}, H_ = {(x, y):x < O}. Let us next characterize completely the connected subsets of E’. Definition. A nonempty set J < E! is an interval if for every x, ye J,x < y, the set [x, y] is contained in J (Figure 2.7). J x Zz y [x,y] Figure 2.7 In Example 2, Section 1.1, four types of finite interval and four types of semi-infinite interval were considered. Each of these is an interval according to the definition just given. Moreover, E' is an interval, and any set {x} with asingle point is an interval. These 10 types of intervals are the only possibilities (Problem 5). Proposition 2.11. 4 set S < E' is connected if and only if S is an interval. ProoF. If S is not an interval, then there exist x, ye S,x < y,andz¢S such that x 0 such that [x,, x, + 6,) ¢ A. Similarly, there exists, > Osuch that(x, — 6,,x,] ¢ B.LettingB, = {xe B;x > x} and y = inf B,, we have x, < y < x2. Since J is an interval, ye J. If ye A, then some interval (y — 6, y + 6) is contained in A and y + 6 is a lower bound for B,, contrary to the fact that y is the greatest lower bound. Similarly, if y € B, then some interval (y — 6, y + 6) is contained in B,, and y is nota lower bound. This is a contradiction. Oo Let f be a function from Sy into a topological space T. For S < So, wecall f continuous on S if the restriction f |S is continuous at each p € S when S is given the relative topology. This agrees with the terminology “continuous on S” in Section 2.5 when Sy = E" and in Section 2.6. 57