PROBLEMS IN Calculus and Analysis ALBERT A. BLANK Courant Institute of Mathematical Sciences New York University John Wiley & Sons New York London SydneyCopyright © 1966 by Richard Courant All Rights Reserved This book or any part thereof must not be reproduced in any form without the written permission of the publisher. Library of Congress Catalog Card Number: 65-27663 _, Printed in the United States of AmericaPreface This book is intended as a companion to Introduction to Calculus and Analysis by Courant and John. At the time this textbook was undertaken, it became clear that the inclusion of more exercises and problems plus the expansion of solutions would require a book too bulky for convenient use. For that reason the Examples of the original text were divided into two classes: the Problems which are kept in the text for their special interest or difficulty, and the Exercises appearing here which are of more routine character and are intended primarily to improve skill through practice. Hints and solutions for both Exercises and Problems are in- cluded here. For cl: and for ease of reference, the Problems are reprinted from the original text. The amplification of material is intended to meet the demands for more practice exercises and for a fuller and clearer exposition of the solutions of the more difficult problems. Some additional problems have been provided. The core of this book is an outgrowth of the Examples from the original text. I am grateful to Richard Courant and Fritz John, who read much of this material and gave me the benefit of their valuable counsel. I am also deeply indebted to Alan Solomon, who contributed to every aspect of its production. Brigitte Hildebrandt kindly read the solutions of the problems of the first five chapters and eliminated some errors and vague- ness. Although these kind and conscientious friends have done much to help correct mistakes, errors will inevitably remain. I shall appreciate being told of them and any other suggestions for the improvement of this work. ALBERT A. BLANK vNote to the Reader In the statements of the Exercises and Problems and in their solutions, it is assumed that the functions involved are continuous and have all the required derivatives unless there is an indication to the contrary. The solutions presented range from complete explanation to the merest hint. In some cases no answer is given. If, upon careful check, you are unable to verify a solution, investigate the possibility that a mistake was made in the formulation.Contents Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 Introduction 1-81 The Fundamental Ideas of Integral and Differential Calculus 82-114 Techniques of the Calculus 115-173 Applications in Physics and Geometry 174-202 > ~ Taylor’s Theorem 203-216 Numerical Methods 217-225 Infinite Sums and Products 226-252 Trigonometric Series 253-261 Differential Equations for the Simplest Types of Vibration 262-2641 Introduction EXERCISES SECTION 1.la, page 2 \ @ For any fixed integer q > 1, prove that the set of points x = p/q*, p,s ranging over all positive integers, is dense on the number line. 1 (6) Show that if p is required to range only over a finite interval, p < M for some fixed M, the set of all « is not dense on any interval. (c) Prove that if we demand only that p +a,X +a, (an #0), if written in lowest terms as p/q, that the numerator p is a factor of a) and the denominator q is a factor of a,. (This criterion permits us to obtain all rational real roots and hence to demonstrate the irrationality of any other teal roots.) @) Prove the irrationality of V2 + ¥2 and V3 + 72. (Hint: For each of these numbers obtain a polynominal which has the given number at a root. Then apply the result of part a.)2 Sec. 1.1b: Exercises Answers to Exercises SECTION I.la, page 2 1. (a) Show for any s that it is possible to find a point x closer to any given P than 1/48. (6) Show that if zero is outside any interval, then that interval contains only a finite number of points of the set. Solutions and Hints to Problems SECTION I. 1a, page 2 1.(a) Assume that a + 2 and az are rational and obtain a contradiction. (6) If a and b are rational, a < 6, then a + (b — a)/ V2 is irrational and lies between the two. 2. (a) See proof of irrationality for V2 in the text (p. 5). (6) If Vn = pig, where p and q are relatively prime and q > 1, then ng? = p*; that is, q® is a factor of p?. But g has no prime factor in common with p; hence q® has no prime factor in common with p®. It follows that q = | and therefore that n = p*. 3. (a) If pig is a rational root of the given polynomial, written in lowest terms, then yp" + dy yp""q + Gn sp" 9G? + °° + apg" + agg! Since all the terms but the last on the left side have the factor p, it follows that P divides ayq” as well; but p and g" have no common factor and so p divides @. Similarly, we conclude that g divides a,p" and hence is a factor of ay. (6) Show that these numbers satisfy, respectively, the equations 0. a8 — 624 — 42? + 1222 — 2dr — 4 =0, and 2 — 924 — 429 + 272% — 36x — 23 = 0, and apply the preceding result. EXERCISES SECTION 1.1b, page 7 1.Prove that the Postulate of Nested Intervals on p. 8, cannot be true for the “punctured” real line x < a, x > a, excluding the point 2 = a.Ch. 1: Introduction 3 Answers to Exercises SECTION 1.1b, page 7 1. Consider the nested set of intervals a — I/n <2 Sa + I/nforn= 1223; EXERCISES SECTION 1.1c, page 9 WH. We can estimate V2 to one decimal place as follows. First, 12 =1 <2 and 2?=4>2; therefore 1 < V2 <2. Next, LI? =1.21 <2, ..., 1.4? = 1.96 < 2, and 1.5% = 2.25 > 2; therefore 1.4 < V2 <15. (a) Continue this process one step further. (6) Calculate V7 to two decimal places by the same method. (c) Calculate the real root of 23 + x + 1 accurately to two decimal places. 4 £2. (@) Give the representations in the base 3 of 12; of 4. (6) Give the binary representation of 156; the representation to the base 4 of the same number. (c) Express the following decimal numbers in the base 12: (i) 10,000, Gi) 20,736, (iii) 4, (iv) 24, (v) 3. 3, (@) Let p be any integer, p > 1. Let the representation of an integer N to the base p be N = dydy 1° > dydo, and to the base p*, N = eei-4*** e160: Show that: 9 = dq + pd; + p*d,, &1 = ds + pd, + pds, ep = de + pd; + p'dy.... (6) Let N = 100111 to the base 2. Find the representation of N to the base 8 (the octal representation). PROBLEMS SECTION 1.1c, page 9 U1, Let [x] denote the integer part of x; that is, [z] is the integer satisfying a-1l<[e] y for two real numbers in terms of their decimal representations (see Supplement, p. 92). *3. Prove if p and q are integers, q > 0, that the expansion of pig as a decimal either terminates (all the digits following the last place are zeros) or is periodic; that is, from a certain point on the decimal expansion consists of the sequential repetition of a given string of digits. For example, } = 0.25 is terminating, 4, = 0.090909 --- is periodic, ‘The length of the repeated string is called the period of the decimal; for ,} the period is 2. In general, how large may the period of p/q be? Answers to Exercises SECTION l.Ic, page 9 1. (a) 1.41 < V2 < 1.42. ©) 2.64 < V7 < 2.65. (c) —0.68. 2. (a) 110; 0.00202. (b) 10011100; 2130. (©) (i) 5954, (ii) 10,000, (iii) 0.2, (iv) 0.023, (v) 0.2497. 3. (a) The number d, yd, -*** do is the remainder in the division of N by p’. Apply this observation repeatedly. () 47. Solutions and Hints to Problems SECTION I.1c, page 9 ¢ a0 toto as and Yn, = 2%, + ion We have from 1, Set t, = cy + the definition of cn a 10%(e — 2) — 1 < cy S$ 10%@ — 24). First we establish that c, is a digit for m > 1. We set « =2 — cy and e, = 10%@ — 2) — c for k > 1. From (1) we conclude that 0 0 we have cn 10 (a) Tair = In + Zt and from ¢, <9 cee! 1 (3b) Yar =%n + Ton + Ton Ste + Poa = Ye From (2), (3a), and (35) we conclude that the closed intervals [x,, y,] con- stitute a nested sequence. Since x belongs to all the intervals of the sequence and the length y, — x, = Jorscan be made arbitrarily small, it follows that x is uniquely determined by this representation. Let « = cy + cyC2°** Cy 9999..., where we may assume that either n = 0 or Cc, #9. The decimal Y = Cy ° C402 *** (Cy + 1) 000-+* also represents the number since ones gang, SL a 1 yerezaqtiy To" + Joni +i + + ie 1) fie ail 29 Tor + ioe! — io 1 ae raed It follows that jz — yl < Since the difference between x and y 1 on is smaller than any positive quantity, it follows that « = y. However, if the digits were chosen by the prescribed method, clearly for the nth digit we obtain from (1) 10° — x,) = cy and the infinite string of 9’s could not occur. Let = and y be given by the decimal representations T= Ay.aydg..., ¥ = by bib... 56 Sec. J.le: Solutions and Hints to Problems we say that x 0), it follows that the decimal expansion can be written in the form Cg Cale" Car traCapts2 "Cars where the s —¢ digits cy, 11'** Coss Tepeat periodically. For proof we show first that R, = R, We have R tet = 108 a = Begs = 10H a — AIOE, + ey y41) = 10R, = Bey spre Tose But 10#x aueses) = (10R,/8], whence Ry. = 10R, — BOR, /6]-Ch. 1: Introduction 7 It follows that R,,, can be expressed in terms of R, alone. Hence, since Ry = Re, Rear = Rests Resa = Ress -.-+ Finally, since s — t < B — 1, we have obtained an upper bound on the period. This is a best upper bound in the sense that it is actually attained (for example, for p/q = 1/7). EXERCISES SECTION I.le, page 12 1. Show for each of the following statements whether it is true or false: (a) V2 > 1.41, (6) V2 > 1.414214, () ¥4 > 1.59, (d) 15 < V240 < 16, (@) 2V2 >1 + V3, (f) 10(7.1)(7.2) — 491 > Ww (g) 100V7 — (16 <8, (h) (42)? > (12)%. 2. On the number axis depict those parts which satisfy the following inequalities. (@) |z| <2, (6) |e - 11 <1, (© le +11 <1, (d) x >5, (e) |z| > 5, (is«s3, (g) @ -2| <1, (4) @ -1) >0, @ 2@* — 1) >0, Qj) le - 1) V3x -2 (m) +221. 3. Prove (a) |x - 1] + |e - 2) 21, (6) |x — 1] + |x — 2) +12 — 3] 22. When does equality hold? 4.(a) If x > 0, b > 0, and a # 5, prove that (a + )/(b + 2) lies between 1 and a/b. (6) More generally, if a/b and c/d are distinct fractions with b > 0 and d > 0, prove that (a + c)/(b + d) lies between a/b and c/d. 5. If (a,b) and (c, d) are neighborhoods of «, 8 respectively, prove: (a) (a +c, 6 +) is a neighborhood of x + B, (6) (2a, 78) is a neighborhood of 2x for all positive numbers 2, (c) (a — d, 6 — c) is a neighborhood of « — 8, (d) (ac, bd) is a neighborhood of «8, if ¢ > 0. (e) If c > 0, then (a/d, b/c) is a neighborhood of «/8.8 Sec. 1.1c: Problems 6. Prove that for a given number a > 0 and any x > 0, - atl vi te For what value of x is equality attained? Use this inequality for x = ¢, §, $, 3, $ to obtain an upper bound for V2. 7. Prove that the set of all rational numbers of the form p*/q* for p, ¢ integers is dense in the set of all positive real numbers. PROBLEMS SECTION I.le, page 12 1. Using signs of inequality alone (not using signs of absolute value), specify the values of x which satisfy the following relations. Discuss all cases. (a) |x —a| < |x — 5). (6) |z —al <2 —b. © | -al a,x > a (sometimes denoted by (—~, ~)] and (—, a], (—©, a), (a, ©), [a, 2), respectively) (see also footnote, p. 22). *(a) Prove that the cases of intervals specified above exhaust all possibilities for connected subsets of the number axis. (6) Determine the intervals in which the following inequalities are satisfied. @ 22-3242 <0. (ii) (@@ — aw — b\(@ —c) > 0, fora V2/2. J © Prove ifa < < 6, then |x| < |a| + |b). {| Iz +-| <6. aCh. 1: Introduction 9 3. Derive the inequalities 1 @ «+22, forx >0, 1 () # +> < -2, fore <0, © 1 + + > 2,000 #0. 4, The harmonic mean & of two positive numbers a, 6 is defined by yal ;) e72\a ta) Prove that the harmonic mean does not exceed the geometric mean; that is, that & < Vab. When are the two means equal? 5. Derive the following inequalities: (a) 2 + ay + y? 20, #(b) a2 4 nly 4 2-22 4. 4 ym DO, *(c) xt — 323 + de? —32 +1230. When does equality hold? *6. What is the geometrical interpretation of Cauchy's inequality for n=2,3? 7. Show that the equality sign holds in Cauchy’s inequality if and only if the a, are proportional to the b,: that is, ca, + db, =0 for all » where ¢ and d do not depend on and are not both zero. 8. (a) |x — ay] + |x — al + le — ag] > ay — ay, for ay < ay < ay. For what values of x does equality hold? *(®) Find the largest value of y for which for all x lz — a) +e — al +--+ le -al >y, where a < a, <--* ab + be + ca. (b) (a + BN + Mc + a) > Babe. (c) a®b? + bc? + ca? > abe(a + b +c). 10. Assume that the numbers 2, 9, x3, and aj, (i,k = 1,2, 3) are all positive, and in addition, ay, < M and x, + 2,2 + 2,2 <1. Prove that Ay 2)? + Ay Xy + °° + Agata? < 3M. *11, Prove the following inequality and give its geometrical interpretation forn <3. + VOFE TFBS:10 Sec, 1.de: Answers to Exercises 12. Prove, and interpret geometrically for n < 3, ea a oe S Vai +o a2 + VEE + +b +e + VEE He. 13. Show that the geometric mean of » positive numbers is not greater than the arithmetic mean; that is, if a; > 0 (i =1,..., 7), then 1 Vayay SG tay to tay). (Hint: Suppose a, < a, < +++ 0, the stated inequality is equivalent to (Vaz — 18 20. Equality holds for z = 1/Va. Use x = $ to obtain V2 < 1.417. 7. For a given g choose p so that p? < xq* <(p + 1)*. Then pete 2Ve 4! F qf Consequently, for fixed « the difference between x and p%/q* can be made arbitrarily small by choosing q large enough. P O (a+) 2 Fora > 6, <2, (6) Not possible for a < b. (a+b) Fora >b2 ><> (c) Not possible for b < 0. Forb >0,@>6, Va—b <|2| < Va +b. For b >0, —b 0, a < —6, it is impossible. 2. (a) Let S be a connected set of the number axis. If there exists any number m such that x > m for all xin S, then S has a unique “lower end point” a such that for all x in S, x > a and such that every neighborhood of a contains points of S. We determine a by giving a nested sequence of intervals which contains it. Let fy be any point of S and take ay =m. We divide [2, Bo] into equal halves. We take as [2, f,] the leftmost of the halves which contains points of S. Similarly, we define [21 8,41] as the leftmost of the halves of [«,, 8,] which contains points of S. Any neighborhood of the point a defined by this nested sequence contains points of S. Furthermore, since « > a, for all points of S, it follows that « > a. Similarly, if there exists a number M such that z < M for all x in S, then has a unique “upper end point” b such that for all in S, x < 6 and such that every neighborhood of 6 contains points of S. The separate cases correspond to the existence of both end points (bounded intervals), one end point (rays), no end point (whole line), end points in S (closed), end points not in S (open). @) @ (1,2). (ii) (a, b) and (c, ~). Gili) (— ©, 3). (iv) (— 2, —lal), (lal, ©).12 Sec. 1.Je: Solutions and Hints to Problems Include end point « = a in the appropriate interval when a # 0. (v) 3 — V8,3 + V8), (-3 — VB, —3 + V8). (vi) (—, 1). (il) Qr + 2nz, $2 + 2nn), 1 =0, +1, 42,.... (c) We have lal < lal +|b] and [| < lal +101, whence — lal — |b] 2ab, B® + c® > 2bc, c + a? > 2ca. (6) Multiply a + 6 > 2Vab,b +. >2Vbe,¢ +a >2Vea. (c) Add the three inequalities of the type a°b® + bc? > 2b¥ac. 10. Observe that + yyy? + AygtyZy + +> + Aagrs? < M(a,? + xq? + 23%) + 2M(wyxq + xyrq + 2 yr) and apply Cauchy’s inequality to the numbers (2, , 73) and (2, #3, 24). 11. Square both sides and use Cauchy’s inequality. For the triangle with vertices 0 = (0, 0, ..., 0), A = (a, dy, -- +, @n), B = (by, ba, -.., by), the sum of two sides OA and OB of the triangle is greater than or equal to the third side. 12. Note first that [ay + by)? + (cy + ba)® 4° °° + Gy + bn)" < Vat 4a) + VBE FBS follows by squaring and apply Cauchy’s inequality. We then have [ay + by + ey)? + +++ + Gn + bn + Cn) = {Iq +b) +P +-°* + [Gq + bn) + enP}4 S Ula, + by? 00+ + Gy + BI + VEER Heat . Let P be the perimeter of the ellipse, A its area, e its eccentricity, and (F, 0), (—F, 0) the foci, Which of the following statements are true? (a) Aisa function of p. (6) A is a function of p and g. (©) pis a function of A. (d) e is a function of A. (e) ¢ is a function of p, g. (f) Pisa function of p, g. (g) Pisa function of A. (A) Aisa function of P. (i) p is a function of A and g. Answers to Exercises SECTION 1.2a, page 18 1. Note that a? + 6? = 1/(x* + y*). A circle in the («, y) plane has the equation (x — 2)? + (y — yo)" = 7°, or a2 y® = nx, + yyy +r? — a Uo. Divide by z? + y? to obtain 1 = 2ary + 2byy + (7? — xg — yo? Ma® + 5%). Pr): (a — a)* + (6 — bo) = p* This is the equation of a circle (unless 22 + yo?Ch. 1: Introduction 15 where =Yo P= a? — yo" and by = If the original circle passes through the origin (r? = + y,%), the equation represents a straight line. For a straight line in the (x, y) plane given in the form Ax + By +C=0, we obtain on dividing by 2? + y* Aa + Bb + + C(a® + b*) =0, which, if C #0, is the equation of a circle through the origin with center ay = —A/2C, by = —B/2C. If the original line passes through the origin (C = 0), then its image is the line in the a,b plane which passes through the origin and has the same direction. By= az-a 3. 0), (0. EXERCISES SECTION 1.2b, page 21 1. Given f(z) = 22, F(z) = 1/2, g(x) = 2%, Ge) = 1/(1 — 2). Find @ f), Fa), g@), GG); ©) SEQ), FQ), FQ), gFQ), 1 — FG), 1 — GF@); f@ +h) —f@ Fle +h) — FQ) ge +h —g) a They a pete 2. Which of the following relations define y as a function of x? In those cases which do, give the domain and range of the function. (a) |z —y| =0, (6) |z? — y*| =0, © z-y=1 @e+y=1 @y=-vi—2, (Ne=vi-¥#. (g) az? — y? = b, a,b >0, (i) & + y® = 1, x,y real numbers. 3. Let f(x) = ax + 6, F(z) = Ax + B. Find f(F(2)), F(f(2)), fU/F(/A)), SE@)).16 Sec. 1.2c: Exercises Let H(x) = (1/F(x)) + 1. Find H(0), H(1), H(2), f(A(2)), fA@)» S(H@)). What is the domain and range of f, F, and H? Answers to Exercises SECTION 1.2b, page 21 1. (a) 0, 1, 4, —4. OL.spnr -1 (CO) 2 aeth’ 2z +h. 2.(a) All x, (e) -1 <2 <1. EXERCISES SECTION 1.2c, page 24 1, Plot the graph of y =2* from calculated values. Without further calculation obtain the graph of y = 2. 2. Sketch the graphs of the following functions where defined and state whether the functions are (1) monotonic or not, (2) even or odd. Which two of these functions are identical? (@) y=2, (©) y = 2* on [0, 1], (c) y=zon[—l, 1), (4) y =|2| on [—1, 1), (©) y = Vz on [-1, 1], (g) y =I lz] - 1. @y= QMy =z +l, (k) y =l21 +12 — I, © y =| + 4% +2) on [—4, 3], (m) y = [2], where [2] denotes the integer part of z @ y =«-[), (0) y = Vx — [x], (p—) y =a t+ Va — lel, @y —1| +|z + 1| -2o0n(-5 <2 <5), Wy =le-1 22) +e +1, y= 2@?-1), md 1 Ou= Zar’ Wy=Ch. 1: Introduction 17 3. Sketch the graph of y =(x —a)(x — b\(x —c), where a z, () y > lel, (c) Iyl > lel, @y>t, () |e -2) <1, QNe+ytlyl- (m) 0 1. Answers to Exercises SECTION 1.2c, page 24 2. (a) Even, (6) monotonic, (c) monotonic odd, (d) and (e) same function, even, (g) even, (4) monotonic odd, (i) odd, (p) monotonic, (4) even, (r) even, (s) odd, (t) odd; (u) odd. 4. (a) Odd, (6) even, (d) odd, (e) odd, (g) odd. 5. Sums of even functions are even; sums of odd functions are odd. The product of an odd function and an even function is odd; of two even functions even; of two odd functions, even. 6. After $ sec, the motion of the body is not described.18 Sec. 1.2d: Exercises EXERCISES SECTION 1.2d, page 31 1. (a) Let f(x) = 62. Find a 8 which may depend on &, so small that \f@) —f(@i Ois arbitrary. Do the same for (6) f@) =2* — 2x, -1<2x 3f@ > 0. 2. Hint: Any open interval about 2, contains a centered interval. 3. Hint: Every interval contains irrational points. 4.(a) Hint: Every interval contains both rational and irrational points. (6) For any rational number x = p/g, every interval about x contains irrational points &, so that | f(«) — f(§)| = 1/g. For an irrational value of x there exist only finitely many rational points £ such that f(z) > and lz — &| < 1; choose 4 as the lesser of 1 and the distance of the closest such point. 5, For any natural number 7, f@ whence f@-) +f0) (a — 2) + 2f) = nf). Furthermore, since f(0) = f(0) + f(0), we have f(0) = 0; hence from £O) =f) +f(-n) a f(-m = fn) = —nfld.22 Sec. 1.2e: Exercises Finally, for every rational number r = p/q P (2) =s(9-2) = =o. Consequently, fis a linear function on the rational numbers SQ) = ar, where a = f(1). If f is continuous, it follows that f(x) = ax for all 2, since the difference |f@) — ar| can be made arbitrarily small by taking r close enough to 2. 6. (a) |u™ — EM] = fae — Ext + Exe + i a Choose 6 < 1, so that |x| < |é| + 13 fom — Er] < of(l8] + "7 $+ [41 + +} < d{(lé] + 7 + (el + 4 ++}, or la” — &"| < dn(ls| + 1)". It is sufficient to choose « 6=———__ gl +” where « <1. 6) [f@) — SO! = lane" = Ean"? — 4) + | S lanl [2™ — &"| +++ + lal le — €l. From the preceding results we conclude that If) —fI < Adfn(\E] + 1)? + Ge — IE] + + °°}, where 4 is the largest of the |a,|, ¥ = 1,..., 2, Consequently, \f@) —fO| < Ad{n(le| + 7 + allel + 7 +--+}. It is sufficient to take « ° 1. @) f@) =24+2e4+2,c21.Ch. 1: Introduction 23 1 ©f@)=r+ 5221. asi, yc 2. In each of the following cases demonstrate that f(«) is monotonic on the whole real axis. In which cases does f(x) have a real root, that is, a number £ such that /(&) = 0? (@) fi) = Os =1+ Fy. D420 © s@ = TS. PROBLEMS SECTION 1.2e, page 44 1. Prove that if f(x) is monotonic on [a, 5] and satisfies the intermediate value property, then f(z) is continuous. Can you draw the same conclusion if fis not monotonic? 2. (a) Show that 2” is monotonic for x > 0. As a consequence, show for a > O that 2" =a has a unique positive solution 7a. (6) Let f(z) be a polynomial S(®) = ay2™ + Gye”? +++ + aye +a, (Gn 0). Show (i) if 2 is odd, then f(z) has at least one real root, (ii) if a, and a) have opposite signs, then f(z) has at least one positive root, and, in addition, if nis even, n # 0, then f(z) has a negative root as well. *3. (a) Prove that there exists a line in each direction which bisects any given triangle, that is, divides the triangle into two parts of equal area. (6) For any pair of triangles prove that there exists a line which bisects them simultaneously. Answers to Exercises SECTION 1.2e, page 44 1, The inverses are given by ot; (@) (1 ee (d) -14+Ve-1,221; x+v2e2—4 a, asl OZ, 223 Neale as.24 Sec. 1.2e: Solutions and Hints to Problems 2. Assume x 0. Root at x = 0. (6) Since f(x) — 1 is an odd function, it is sufficient to establish the result forx >0. Fory >z 20, y x y-2 [0 -f© =p, 4277s a0 ep 2" root. fo Pitty _ B42 _ Gaeta oy tv ty T+ 1+e ~ + y)0 +24) a Solutions and Hints to Problems SECTION 1.2e, page 44 1. If f is not continuous at é, then for some « and every 6 > 0 there exist points x such that | f(x) — f(é)| > « and |x — &| < 4, It follows that every neighborhood of £ contains infinitely many such values, and so for the two sets of intervals (€ — a, &) or (¢, + a) every interval of at least one of the two sets contains such values of x. Suppose f(x) is increasing and every interval (, § +a) contains a point 7 such that |/(7) — f(é)| >. Then SE +a) =f) >fO™ + > fl. The value f(¢) + « cannot be taken on, in contradiction to the assumption of the intermediate value property. No, for consider f(x) = sin (1/2) for x # 0, f(0) =0. 2. (a) If z > y > 0, then 2? > y? > 0, and upon repeated multiplication, a" >y">0. If 0 1, then a is intermediate between 1 and a*. In either case, the existence of a positive root is assured. Since x” is monotone, there cannot be two distinct positive roots. (6) (i) If nis odd, then we may set SL (%) = aq gt" + dapat" * + Aang > 4° ++ Gana Gon-3 Banca 5 Sans, On Pana = an (1 +Ch. 1: Introduction 25 If |2| >A +1, where A is the largest of the numbers |@,/ayn—l, ¥ = 0,1,..., 2 — 2, then ray Gan—2 Fen—3 Fy” Fay. |S and we conclude that f(x) has the same sign as a,_2?"-!, Thus for positive and negative values of x of large magnitude, f(x) takes opposite signs. The existence of a root follows by the intermediate value property. (ii) If a, and ap have opposite signs, then for large positive values of 2, Sf(@) has the sign of a, and since /(0) = ap, it follows by the intermediate value property that f(x) has a positive root. If n is even, then f(x) has the sign of a, for negative and absolutely large values of x, and it follows similarly that f(x) has a negative root. 3. (a) It is known from elementary geometry that every line through the intersection point of medians has this property. From the intermediate value theorem, it also follows that in each direction there is a line which bisects a triangle (or any other area). Consider the lines L(x) of a parallel family, where = is the distance measured along a common perpendicular. Let f(x) be the difference between the areas of the parts of the triangle on opposite sides of L(x). Since there exist lines on either side of the triangle which do not intersect it, it follows that f(x) takes on both negative and positive values; hence, by the intermediate value property, f(z) = 0 for some value of x. (b) For each direction 6, choose the line L(8) which divides the first triangle into two equal parts. Let L(0) be oriented positively in the direction 6. Let &(6) = 2(8) — A(8), where «(6) and £(8) are those parts of the area of the second triangle which lie on the right and left of L(6). If g(@) > 0, then &(n + 6) = B(@) — 2(6) < and by the intermediate value property it follows that g(0) = 0 for some value of 6. EXERCISES SECTION 1.3a, page 47 1. Draw the graphs of the rational functions: 1 1 1 @y=ay7° Onis as @yarte, ae +] al 1 @y=a7> ®v=a7° Ny=sy7-26 Sec. 1.3b: Solutions and Hints to Problems EXERCISES SECTION 1.3b, p. 49 1. Draw the graphs of the following algebraic functions: , 1 1 @y=@ +0 Ov-a@rp Ov-@ypF @y=@ +04, Oy=@+)%, Ny =@ - 1% PROBLEMS SECTION 1.3b, page 49 1. (a) Prove that Vz is not a rational function. (Hint: Examine the possibility of representing Vz as a rational function for x = y®. Use the fact that a nonzero polynomial can have at most finitely many roots.) (6) Prove Vz is not a rational function. Solutions and Hints to Problems SECTION 1.3b, page 49 1. If a tac + + bo + byt + +t aye” Vrs # +b_n™? then, for « = +2, where » is any natural number, P(») = ay — boy + ay? — by? ++ 0. Thus the polynomial P(») has infinitely many distinct roots. We conclude that P(») is identically zero, or that ay = by =a, = by =*** =0; hence that Vx cannot be represented as a rational function.Ch. 1: Introduction 27 EXERCISES SECTION 1.3c, p. 49 1. Find the radian measures of the angles: (a) 10°, (6) 25%, (©) 93°, (a) 173°, (© 453°, (f) 169° 2. Draw the graphs of the functions (@) y =sinx-cos=, (6) y =sin( + 7), ie z (c) y =tanx-sinz, -7<7<5- PROBLEMS SECTION 1.3c, page 49 1. (@) Show that a straight line may intersect the graph of a polynomial higher than first degree in at most finitely many points. (6) Obtain the same result for general rational functions. (c) Verify that the trigonometric functions are not rational. Answers to Exercises SECTION 1.3c, page 49 1. (a) 0.17453, (b) 0.43633, (c) 1.62316, (d) 3.01942, (e) 7.90634, (f) 13.42158. Solutions and Hints to Problems SECTION 1.3, page 49 1.(a) The graphs y=ar+ 6B and y=a,r" +a, 42") +++* tay intersect at points for which ax + B = a,2" + ay 42" ++ ++ +a,28 Sec. 13d: Exercises that is, a (ay — B) + (a, = a)2 + ayx® + agz? + +++ + 4,2" =0, If the number of intersections is infinite, then the polynomial of (1) must be identically zero, and ay = B, ay = %, dy = 0,6, aq = 0, contradicting the assumption n > 1. () It P(x) Gy + aye ++++ +4,2" oF + B= OG) = 5 tbe tt bg for infinitely many values ~, then by the argument used above, boB = ao, bot + 8B = ay,.-5 byt + bessB = des (where, for convenience, we set a, =0 for k > and b, =0 for k > m). Thus Pla) = ay + aye + +++ + aq” = boB + (box + biB)e + (by + byB)x® +--+ = b(B + ax) + b(B + axle +--- = (6 + a2) Q(z). It follows that [except for zeros of Q(z)] P(x)/ Q(x) is precisely the same linear polynomial. (c) Because of its periodicity, the horizontal line y = 1 has infinitely many intersections with the graph of any trigonometrical function. EXERCISES SECTION 1.3d, p. 51 1, Assuming the exponential function y = a® to be defined as indicated on p. 52, draw the graphs of: (a) y =(0.5)7, x = logy.sy3 (+) y =2%, x = logy.Ch. 1: Introduction 29 EXERCISES SECTION 1.3e, page 52 1, Find the domain and range of each of the following functions, and draw their graphs: (a) y =sin (2), () y =(sin2)*, sin x cos x, (d@) y = (sin 2), (e) y =(1 + cos2)4, (fy = cos [(1 — 2%)4], () y = sin 2%), @ y =28In2, () y =sin [sinz + 1). 2. Show that the function y = 2z* — 3x + 1 is monotonic increasing for « > 3. Draw the graph of this function and of its inverse for x > 3. 3. (a) Let z—n2, 2 <0, so-( 2 >0, and A(u) = sin u. Prove that the symbolic product Ag is continuous for all z, although g is not. (b) Show that if g is a continuous function whose range contains a neigh- borhood of a discontinuity of 4, then Ag is discontinuous. (©) Using the continuity of the identity mapping J, produce a pair of discontinuous functions g, 4 whose symbolic product Ag is continuous. Answers to Exercises SECTION 1.3e, page 52 1.(a) Alla, -1 & > ¥, then the difference in the corresponding function values is @ — H[Ax + &) — 3] >0.30 Sec. 1.5: Exercises 3. (a) sin g(x) = —sin2, for all x. () For some « > 0, every neighborhood of a point of discontinuity « contains points v such that | f(v) — f(w)| > «. (c) Take g as the function of part a, utm ug—n, hu) ={u—7, u>a, 1; -r 1 and for all positive integers 7, nett © Gay zitetet st QD, nok Om; =k DL IME $ IME eE, 6. With the aid of the binomial theorem, prove that the number 20" is irrational for all integers n > 2.Ch. 1: Introduction 31 7. Using the binomial theorem, prove that the decimal number 2 = Mey yy = My + 10m +++ +10, Moy .. +4 Me <10 is divisible by 11 if and only if ng — m +, +++ + (—1)'n, is divisible by 11. PROBLEMS SECTION 1.5, page 57 1, Prove the following properties of the binomial coefficients. wore()e)er ele) +l) or-(()+6)-() rer) -n n n n n (© (7) +2(3) +3(3) +e +n(") =n(2""), (Hint: Represent the binomial coefficients in terms of factorials.) @! -2(;) +2-3(3) $e $d (") = n(n — 2, 1 1 (n lj[n 1 n ant 1 Ol+s\y #57 ee ced Nal eee dS ne (n\* nP (2n : . “(3) + (") toe +( ) = (a: (Hint: Consider the coefficient n of a" in (1 +2)".) *(g) Sn = (<) =3 (}) +5(3) 5 (3) x or (" 4n(nty® “Qn + DI" ; 2n +2 (nin Prove nad Sp = Sux) 2. Prove (1 + 2)" >1 +x, forz > —1. 3. Prove by induction that 1 +2 +--- + =4n(7 + 1). *4, Prove by induction the following: 1 =—(m + 1)q" +g" a -¢? Lag! l—q ° 5. Prove for all natural numbers x greater than | that n is either a prime or can be expressed as a product of primes. (Hint: Let A,_, be the assertion for all integers k with k < n that k is either prime or a product of primes.) (@) 1 +29 43g? +075 tnght = OA+Qi +e) +g") =32 Sec. 1.5: Answers to Exercises *6. Consider the sequence of fractions 137 Pn Degg where Past = Pn + 2gn and gaia = Pn + One (a) Prove for all 1 that p,/q, is in lowest terms. (6) Show that the absolute difference between pn/qn and V2 can be made arbitrarily small. Prove also that the error of approximation to V2 alternates in sign. 7. Let a, b, a, and 6, be integers such that (a + bV2)" =a, +b, V2, where a is the integer closest to V2. Prove that a, is the integer closest to by V2. *8. Let a, and b, be defined by = 3, ayy = 3%, and by =9, ayy For each value of , determine the minimum value m such that a,, > by. 9. If 1 is a natural number, show that (+ v5" -(1 = v5" 2v5 is a natural number. 10. Determine the maximum number of pieces into which a plane may be cut by straight lines. Show that the maximum occurs when no two of the lines are parallel and no three meet in a common point, and determine the number of pieces when concurrences and parallelisms are permitted. 11. Prove for each natural number n that there exists a natural number k such that (V2 - 1)" = Vk - VEAL. 12. Prove Cauchy’s inequality inductively. Answers to Exercises SECTION 1.5, page 57 land 2. Take the formula for the sum to # terms; add the (m + 1)th term to obtain the formula for the sum to # + 1 terms. n(n + 1m + 2) 3.@ 5Ch. 1; Introduction 33 4. True for n = 1. If true for n = k, then 28+ > 2(2k) Bk +k) = Ak + 1). 5.(a) True for n = 1. If true for », then 2 kth (+e Pie ae 2" Seat Seer since the middle expression consists of the first two terms of the binomial expansion of (n + I)*/k(k + 1). (6) The result is certainly true for n = 1. If the result holds for a given n, it will hold for m + 1, provided that 11k 4 ye = pta+i a. Lee Uy This inequality is equivalent to 1 mk <(nt vn + i: n Wer 4 1s(4)'[ +a] and, for the kth powers, to (chp a] lbealestt so that the last inequality is valid. 6. If 21/" is rational, then set aun (1 +2), q. where p/q is given in lowest terms. Take the nth power to obtain hence, to 2g" =p" + (ema te tg" Since q is a factor on the left, it must be a factor on the right; this implies that q divides p" and contradicts the fact that g and p have no common factors. 7. n = Mg + 10m + 100m + +++ + 10*n, =n + (LL — 1m, + °° + (Ll -— In, = [my — my +00 + (—1iu] + Lm,34 Sec. 1.5: Solutions and Hints to Problems Solutions and Hints to Problems SECTION 1.5, page 57 1. (a) Expand (1 + 1)". (6) Expand (1 — 1)". « (1) +2) +~+4()) a re ea Wi ea (a) n-1 = n2-, by part a. n (k — I)k(a!: nt! (0 Ue -0K(,) = gmat Ea n-2 = nn — (1-3). n nt 1 n+l (0 User aula) -&F DIG mh! =cnilt +): (f) The coefficient of 2” in (1 +2)" = (1 + 2)"(1 +2)" is C2) =) «Nar por #0 ~(fe heh (g) Use the relation int? . (=1)FQn — 2k +2)(n +1 +3" &y Qn + 32k +1) \ k ) z eS 1 = Foye Om) ale ) oo Qn+3)Qk + \ k ned (IE (an $1 ne “("*) -3 al k )- Praha k 1 = Sua — Fg tl — = Snare 2. For n = 1, the statement is trivially true. 1 + ka, then If, forn =k, (+a) > (1 + kay +2) 21t+k + De. (+a >Ch. 1: Introduction 35 3. True for n = 1. If true for n =k, then Ltt Fk tk +i) =k +) 4+ RK +0 =(k + DGk +1) = Mk + Ik +2). 4.(a) True for n = 1. If true for n = k, then ($29 +205 +g) +k + Dg 1 —(k + 1)g* + kg? To +(k + 1g _ 1 =k + Dg + kg? +k + DG — 2g + gi?) % ag _ 1 =k + 20g +k + gh? a= : which proves the result form =k + 1. (6) True for n = 1. If true for n = k, then (1 +9)---G +g) +92") d= 0 +g") I-q Lag 9 atte q 5. Let A, be the statement, “If 1 1, en _ ya] 2by + 1. This is certainly true for # = 1, If for m =k, ayy, > 2b + 1, then iggy = 301 > HL > 3-32 23-9 > Sar > Derr + besa > We +1 From this result if follows that no smaller value of m will suffice, and since a, < 6, the proof is complete. nd — v5)n 9. Set F, = CickeV DR US VS): Then Fy + Faia =Fnsg. The anv5 numbers F,, are known as the Fibonacci numbers. 10. Let 4, be the maximum number of pieces into which n lines cut the plane. Proceeding along the (n + 1)th line we cut each region containing the path into two parts and enter a new region if and only if we cross one of the preceding lines. We see then that 4,,; = 2, + (n + 1) provided that the (n + 1)th line is not parallel to any preceding line or passes through a point where any two preceding lines intersect. By induction it follows that dn = Hn? +0 + 2). For each pair of parallel lines 4, is reduced by 1, since one crossing is eliminated. If there are k families of parallel lines with p,, Pa... » Px lines in the respective families, then 4, is reduced by Sip, -.Ch. 1: Introduction 37 Similarly, if a line crosses an existing intersection of two lines the number of regions crossed is reduced by 1. If we assume j families of concurrent lines with cy, a, ..., ¢; lines in the respective families, then 4, is reduced by SH, — Vc, — 2. Thus, in general, the number of regions into which S the plane is divided by 7 lines is & 24 dor e429 — § ANDY _ YF yee, — 6, - 2) = vel 11. Define p, and q, as in Problem 6. We have (V2 = 1)" = (=n — gn V2) = (DV pt? — V 292). Pret = Pn + 29m Gner = Pn + Ins Prva ~ 2Gnaa = 2gn? — Pr 1, it follows inductively that 2qn® — pa? = (-1)"7; hence, for n even, p,? = 29,2 + 1, and (V2 - 1)" = V2q,2 +1 — V2q,3, whereas for n odd, 24,2 = p,? + 1 and (V2 - 1)" = Vp, +1 — Vp,2. 12. The relation is obvious for n = 1. Suppose it is true for k: (Se) <(Zer)(So} Then, for k +1, (Sul (Eo tut im im From we have Since 29,° — p,* Sesh) + 2anubaer 0b + abba ( 2 be + ahha -(3)(3) where we have used the well-known inequality 24B < A? + B®, or (4 — BY) 20.38 Sec. 1.6: Exercises EXERCISES SECTION 1,6, page 60 1, Find the limits of the following expressions as n -> 0: ant? 2n +3” 6n® + 2n +1 m+ 6n' +2n +1 w+ re (a) (b) © n m+n, mont] w+ntl’ an” + ayn?) + +a ban? + by Ps + + by’ 144494+::: +7? (e) (6, # 0), n 2. For each of the following expressions the limit is given as n > «0. In each case find an N such that for n > N the difference between the expression and its limit is (a) less than ;'o, (b) less than 1/1,000, (c) less than 1/1,000,000. ite ES (@) lim = = 0. n _ 1-2 en Tee 2+n—-1 1 Gi) im Se 3 cea ne () lim = =0. no 2 (vi) lim 7? nse (vii) lim (- zy =0. noCh. 1: Introduction 39 3. Find the limits of the following as n -» 0: sin (I/n) at cosn (©) sin (;) cos”, a (@) sin(w- 2), 2n — 5 © yya—sm 4. Using the binomial theorem (p. 59), prove that for any fixed # > 0, and any fixed integer k, lim Tp 7° 5. Prove that for any positive integer k, lim (Vny’™ = 1, 6. Prove that for any positive integer k, lim (Vn +1 — Vn¥) =0. nso 7. Prove the following relations: (a) lim (Vn? + 2n — Vn® +n) =}, (6) lim (V2n® + 3n +1 — Vik +5) = @, (©) lim (Vin? + 2kn $1 — Vn® +5) =k, for all k > 0. no 8. For any given positive integer k, consider the difference Dn = Vagyn® + a, ++ -- Fayn + dy = Vbyn™® + by nt 4 +++ + bin + by; prove that if a, b, for any i > k, D, tends to infinity as n + «0, whereas if a; = b; for i > k, D,, converges to (a, — b,)/2V ayy, as n> 0. | PROBLEMS SECTION 1.6, page 60 a 1. Prove that lim (Vn +1 — Vn)(Vn +) =}. 2. Prove that lim (Wn +1 — ¥n) =0.40 Sec. 1.6: Problems 3. Let a, = 10"/n!. (a) To what limit does a, converge? (b) Is the sequence monotonic? (c) Is it monotonic from a certain n onward? (d) Give an estimate of the difference between a, and the limit. (e) From what value of n onward is this difference less than 1/100? n! 4. Prove that lim — = ey ae ki 5. (a) Prove that im (F ta tee +3) =}. (b) Prove that li I i L 0. (Hint: Co ve Se ape ‘ e that tim (oe +5 + qt + op ). (Hint: Compare the sum with its largest term.) 1 1 (c) Prove that lim (+ watt) =o, no\Vn Vat] v2n. 1 1 *(d) Prove that lim as + 4H H+ zs) nae\Vne +1 Vn +2 Vit +n 6. Prove that every periodic decimal represents a rational number. (Com- pare Section 1.1c, Problem 3.) 1300 7. Prove that lim or exists and determine its value. 8. Prove that ita and b < aare positive, the sequence 7a" + b” converges toa. Similarly, for any k fixed positive numbers a), as, ..., a, prove that Va," +a.” + - >> + a,” converges and find its limit. 9, Prove that the sequence V2, V2V2 ,/2V2V?, ..., converges. Find its limit. (n) 10. If »(m) is the number of prime factors of , prove that lim ae = 11. Prove that if lim a, = &, then lim o, = &, where a, is the arithmetic ne mean (a, + a, +---a,)/n. 12. Find ’ 1 1 1 (@ tim (5 2+7at° +t) : 1 1 1 (ai: EH Me+) ~k k41 :) 1 1 @) tin (35 sy Cae sree + ares): 13. If ay + a, + °°» +a, =0, prove that sates eae a +++: +a,Vn +p) =0. (Hint: Take Vn out as a factor.)Ch. 1: Introduction 41 14. Prove that lim aeat +n=1. *15. Let a, be a given sequence such that the sequenceb, = pan + 9an415 where |p| q >0, show that a, need not converge. 16. Prove the relation lim = A nal! for any nonnegative integer k. (Hint: Use induction with respect to k and use the relation Spee —@ — yey oats, ast expanding (i — 1)*# in powers of i.) Answers to Exercises SECTION 1.6, page 60 1. (a) 3, (6) 0, (c) 6, (d) 2, (€) aglby, (f)4- 2.(i) 50, 5000, 5000000. Gi) 31, 3001, 3000001. Gi) 4, 334, 333334, (iv) 3, 32, 1000. (v) 10, 19, 30. (vi) 22, 66, 132. 3. (a) 0, (6) 0, (c) 0, (d) 1, (e) 2. 4. Fork > 0, r=> "" he z fH. Mee 3(")> (aa) Sin ty 0 =e ee i (n= gente &+0! Consequently, for }n > k nt Kt Dine M+ DC asi SG@—- QA Ret ne where C is constant. Obvious for k < 0.42 Sec. 1.6: Solutions and Hints to Problems 5. Set 1 +h =(n)"", Employing the binomial theorem, we have ik +A > . ) R. 6. Multiply numerator and denominator by Vn¥ +1 — Va¥. 7 and 8. In each case proceed as in Exercise 6. n Thus A? < 2/k(nk — 1). Solutions and Hints to Problems SECTION 1.6, page 60 1. ay =(Vn +1 — Vina +h va+h Vnt1+Va 1 1+ 2n = fit-+1 a For any « > 0 we have l 7. 2 an = Vn 1 — Vn 1 G+ +n + 1) +08 1 3a (3<)%. <Ch. 1: Introduction 43 3. (a) 0, (6) no, (c) for n > 10, (d) set k = 1019/19!. For > 19 we have a, 101 100 101! 100 where the constant k is jae NOLIN Le a a So The limit is 0.44 Sec. 1.6: Solutions and Hints to Problems Ba< va" +b" Sat. Similarly, if a is the largest of the numbers a;,... , aj, then a N. Consequently, 1 lon — §] => {lar — €1 + lag — 8] +°+* + 1am — EI} 1 S {NG $16) + layer — +0 + lan — EI} N en — N) SO eh N s74 +s) te. <2, provided n > N(A + in 1 na +1) 11 Niel 12, (a) a, “5 5 fo yea 1 1 “Taatraat ti@ spat 14 Wat) (vd -(5-5)| + [33-a) -(G~a 2\EBL if wl 1 1 Pa Ne eee, ntl n+2 Tflinsde oth 1 1 a) "alia -aei aa) (2 tne, 1 4 “42h 4DCh. 1: Introduction 45 13. c, =a)VntaVn+1 +--+ +a,Vnt+p = Vales + 0 [istaes tay i+) = Vilas — a + 0( fi that) etal +2 -1)}. Consequently, an | ae / lal < Yall 1+--1 +| 142-4) n n where A is the largest of the numbers |a;|, |a2|, ..., |aj|. From the estimate V1 +a <1 + for > 0 (see the solution to Problem 1), we have eH peer? lenl < vaa(; = teo48) non n Aplp +1) 2Vvn * 14, Set a, =""Vn® Fan =1 +h. We then have Qn +1 r+n=(1 +A > ( 3 Je. The result follows from CG + ‘) _ Qn + 1)2n)Q2n = 1) 3 6 15. Set b = limb,, Putting b, = 6 + ¢, r = plg, we have ee Cbd O41 = fen b+e = = Fan, bts Qe = — Tansy abt in Ot), a, gq q 6 1 my CSA 5 (ene — req) + Paqy b ane = Gl srt bees + (=rP] 1 5 + lenge = renga too + (=O ten] + (— han.46 Sec. 1.6: Solutions and Hints to Problems Note that 6 6 1-(-9F 2a = 24. xa =? [a-rtrt ete +(e] go ier S22 -(-1- +4 Since we can ensure |«,| < « for » > N(e), we obtain 6 |b] = [rik nek 4g pag tel Inl® tet =r If |p| > 4 then a, = (—p/q)" diverges, but av prt modo P, 2 q “(Ha ( dP NG ; 16. For k =1, Sie 1on@+1)_ 1 teat 3 Suppose the relation is true for / s k. Then i allj p>o, rn? —3n? + 4n —1 “We 3 tn” nsin (I/n) + n® 2n—-1 e n®-cos (1/n) + nl/™ 1+ 5. Using the convergence of bounded monotone sequences, verify that infinite decimal fractions are convergent representations for the real numbers. 6. Use Cauchy’s convergence criterion to show that the following sequences converge: (a) (©) (d) (@) a, = (6) a, =Ch. 1: Introduction 49 Dep el ttl 1 Omaltaptat yatta Epil 1 (-1" Mage LTS t te Sree tad . Oam=lt+ G+ =,» for any integer p > 2. (Him Use mathematical induction on k to show that 1 1 Onek — On sr f= - Gor (-—y, nao kes 44 osm: > a 2 © Sa@-2, Ndi =k, x5 mS 3 () Dk. Xo 8. Verify the following formulas: (@) Sx =(n—m +1), OS +b) =Sa + 3b ‘ n, whereas b, = 0 forj <0, j >m. 9.(a) Prove that the sequence {a,} defined by 2-a,2 ans1 = ty t+, in with ay any number greater than 0, converges to V2. (6) More generally, show that the sequence {a,} defined as k-a,2 ta, Ont = An + with ay > 0, converges to Vk, for k any positive number.50 Sec. 1.7: Problems 10. Prove that for any rational number x > 0, e” = lim (1 + ) n 11. Prove that for a, = (1 + 1/n*)", we have lim a, = 1 fork > 1, whereas no a, ~ © asn— co whenk <1. PROBLEMS SECTION 1.7, page 70 *1. Let a, and 6, be any two positive numbers, and let a, 6 derive the sharper inequality n\* m 0, and lim = L, then lim Wa, = L. n+ In n+ 11. Use Problem 10 to evaluate the limits of the following sequences: @ivn, P+, |} ae 12. Use Problem 11c to show nl =n"e"a,, where a, is a number whose ath root tends to 1. (See p. 504, formula (14).) 13. (a) Evaluate 1 1 teat tiweD 1 133: (Hint: Compare Section 1.6, Problem 12a.) 2 (6) From the result above, prove that > peconverges. mm52 Sec. 1.7: Answers to Exercises 14, Let p and q be arbitrary natural numbers. Evaluate 1 i 3 + pk +p +q)" x 1 oe (o) lim 2 pe ap TD 15. Evaluate vail 1 1 Orr Sat te pe ee) wos : Oi kk + DK +3)* (©) Evaluate the limit of each of the above expressions as n -+ 00. *(d) Let ay, ay, ..., m be nonnegative integers with a, a, converges, show that im kay = 0. a ee 17. If a, is monotone decreasing with limit 0 and by = ay — 2az41 + Que 2 > 0 for all k, then show > kby = ay. ma ANSWERS TO EXERCISES SECTION 1.7, page 70 1.(a) Bounded, (6) convergent, (c) convergent, (d) monotonic, conver- gent, (e) bounded, (/) monotonic, convergent, (g) monotonic, (/) bounded, -10, for n>1. (©) Gnya — Gn = 6n +1. ! > 0. (a + 1m + 2) (©) nya — An = Oe writ 1 (©) asi — an = tiaeD: 3. (@) “Ht = 2Q-Unintd) < 1, ‘ a, ri Pe ©) ay = anya = sin 5 — sin aeost(Z 7 \.i(n 7 = 2008555 + an +3)" 2\2n In +2 20, es NEU 1 (Al eee ‘ since 5 7 In tnd <57lg ta <3 for a>. 1 (.) a. = —=_=.. "Vatl+Vn 4. (a) 0, (6) 0, (c) 3, @) &, (e) 1. 5. The infinite decimal fraction ; © = CgCyC2Cg °° is the limit of the sequence o Gigs Sheree, 10° * Jo + ioe" hence is increasing since c, > 0 for k > 0. On the other hand, since ¢ <9 for k > 0, Cor Co + on ata St ete “shige 9 9 tote + 9 So +i9 t 108 To" N. 1 1 1 1 (@) lay, — aul Sot gy << forN >=. 1 (0) aq = 1 +=. Now use part a. 1 1 1 Gai Oe SGD GDS GEOL 1 1 1 -aatastacet +arp ers 1 1 1 1 x < ai 43+qteo te) (a +1) 2a 2 = SGepi 0, then en,1 > 0, and if ay > 0, then e, and all succeeding terms are positive. Since a, > ey, We have a eo" en z n= Peg," ann S 3 Thus the sequence a, converges to V2 for any choice of a) > 0. (6) Proceed as in part a. 10. Let x = p/g for p,gintegers. Clearly, e = lim (1 + 1/n)", if westipulate ne that m ranges over integer multiples mq of g, m= 1, 2,..., or 1 \a e= lim (1 +z) . m0 mq, Now 1 \nra/a ef =e?!1 = lim (1 +z) mq m0 1\"" = lim (1 +2) mae\ mg! Let r = mp. Then as m —» «, we have r > «0, and m = r/p. Thus a(t 5) altos) @ =lim{1+—) =lim{1+-). "9! 40 r 11, Fork <1, 1 sy zt s +a)21+g-.56 Sec. 1.7: Solutions and Hints to Problems 1\" 1< l+a n For k > 1, Since n/n* —- 0, the result follows. Solutions and Hints to Problems SECTION 1.7, page 70 1, From the inequality between the geometric and arithmetic means we have for all n, a, < by. Consequently, a an + By Ong = Vn = Om Png =—Z— S by and the sequences a, and b, are bounded increasing and bounded decreasing, respectively. Furthermore, an + by Data — nya = — Vanb, = VanVb, + aq = Van(V by — Van) It follows that b, — a, bast — Onur Sy and hence that a, and 5, have the same limit.Ch. 1: Introduction 57 2. We have ani; = V2 + aq. Furthermore, if a, <2 then aq,, <2 and the sequence is bounded; hence the sequence is monotone increasing. Finally, if ¢, =2 — aq, then Onur = VE, where ¢, <2 —a, <2 — V2 <1. It follows that Consequently, €ny, < 4/2". 1 1 1 Sidi Oe 23 tad A ee 1 Stn tn It follows that the sequence is monotone decreasing. Furthermore, a, < (n+ Din =1 + In, and eee ee ee 2n 1 1 on+1 1 Pag Rag San ae 4. b, =a, — 1/n; consequently, 5, has the same limit as a,. Furthermore, 1 1 1 butt — be = 32 toed aed 1 1 1 os fe EO 7 2@ 4 *3@4+h n+l so that 6, is monotone increasing. 5, We have by veo?! by the Cauchy condition: The sequences «, and f,, both converge 1 1 1 Ionee — nls Baez — Bul SGanit@d+o! TEST GEE 1 +m We have anB, = esp even at x 1)e-# a aS = 22,601 wie! oun (Co $30 te ip pao (k — Hy)! ole where we have set » = k — yt. For k > 0, k (1) ino (k — )! Thus «,8, = 1 + E and we shall show that E tends to zero with n: 1 ns, 5 2 am re kenyl pontCh. 1: Introduction 59 From the inequality m! > 2"-1 we have ee 4 lis 2 2 x Ran pant eoka=n Lt24---40 nn + 1) Son: It follows that lim «8, = elim B, = 1. Ang, [+ 2)" (nn +1" (ral fad -[ ae) 5) (Cae n n+2 (n+1% +1 > [1-atal(GH)e @+p 7" (From the inequality (1 + 4)" > 1 + nh for h > —1, which is easily proved by induction. See Problems 1.5, No. 2.) Similarly, bn _ (ae + Pn + baa La +a) \ne 8.(a) 1 n+l atl 2 tia@+D n+2 w+ 4n? + 4n +1 +4 +40 7 (+1 He e Oo (i xs, for n>2. 9.(a) On observing that a, of Problem 8a increases monotonically to the limit e, we have OT > aya? Oy : (ty & ( Sa -@—pi~ whence n! > nt/e"-! = e(nje)". a60 Sec. 1.7: Solutions and Hints to Problems Similarly, since 6, is monotonically decreasing, PN Biba by 2\2/3)2 n\n fa - (7) (4) “@—D!’ n\n nt Toa >! and Ste _ 6! x 1098 __ 720 x 1088 TT <7 x Gy STx GO — 1 792 x 10% 792 ar Sm 1, 6. 10. If L #0, choose « > 0, so that « < L. Then form > N(«) me |< ay whence (L = €ay < ny, <(L + Gn and consequently, (L — "ay < an <(L + Nay. It follows that ofc <1 use| w-ole et ——— = . 2 REED (6) From the preceding result lim S, = 3. no The sequence T;, “35 z is monotone increasing. Furthermore, for k > 1 we have 2) p> HAD 3 whence Tn $35, S% (since S,, is monotonically increasing). Since T, is increasing and bounded, it follows that T,, converges. 2 1 14. Sy @) Set Su = 2 ES pe tp FD 1 1 1 1 From ee =~ pts) we have 1 1 1 1 i 1 spats t ptq n+p+l ntpt2 1 ee) 1 1 1 © im 56 = prot tp 4g 15. (a) Same as Problems 1.6, No. 126. 1 1 1 © GREE FD ~ MEF) EFDE FS’62 Sec. 1.7: Solutions and Hints to Problems Consequently, by Problem 14a z 1 2 (145 +h-o5 : : EE + DE +3) 3 n+l n42 43 1 (3 1 1 1 NETS eee. aes (c) Limit in Problem 15a is 4; in 156, 54. (d) Reduce to sums of the form in 14a by a sequence of steps for which the first is 1 1 a\k +a, k + ak + a) 1 1 1 1 ay =e +ayk +43) (kK + ak aale $a) +m)” 16. Since a, is monotone, we see that the terms have a constant sign for sufficiently large n. Suppose, for example, that 0 < dnj1 < aq for sufficiently large n. Then by the Cauchy criterion, for sufficiently large m, Amst + Omy2 + 7°" + Amie < & independently of &. From Ansa + Ames + °° + Ome 2 kOmex we have Kamp < & Similarly, for sufficiently large k Manin < & Adding, we observe that (m + Kam, can be made arbitrarily small by taking m and & large enough. 17. By induction "SF nby =a, — kay + (k — Dagyn ON ay — ayy — May — ae42)- Now we use Problem 16. Alternatively, since b, > 0, we have Oy — Any 2 Acyy — Mey It follows that Oe — Gam =D (@ea4s — Mee) jaa 2 M4 m—1 — Ge4m)-Ch. 1: Introduction 63 Entering this result in kame Mby = Ay — Agim — (k +m — IN(Qe4 m1 — Ae+m)s n=l we obtain bared ktm-1 2 Mbn — ay] S anim + ——— (ae — ym) Sara DE & mata ‘k +3m—1 S 2a,\—— : For any y choose k and m so that m +k —2—=vand S4+10, © f@) = vz. 7.(a) Let f(z) = 6x in the interval 0 0. Do the same for () f@) 2x, -1<2 0. _e+2 3 _ 10. (a) Prove that lim ——— =>. Find a 6 such that for |1 —2| <6 the patti 2 x 2 is, in absolute value, less than «, < > 0. Hl 3 difference between 3 and Do the same for (6) lim V1 +23, 22 tim 2 (©) ‘im Seat 11. Prove that vi¢e-1 1 ime 20 z 2 (© lim Vz +4 (Vz +1 — V2) =}. ew 12, Prove that the elementary trigonometric functions sin x, cos x, are continuous. PROBLEMS SECTION 1.8, page 82 1. (a) Prove that lim (cos 2x)” exists for each value of x and is equal to 1 or 0 according to whether x is an integer or not.66 Sec. 1.8: Answers to Exercises (6) Prove that lim [lim (cos n! 7x)?™] exists for each value of x and is equal to 1 or 0 according to whether = is rational or irrational. (©) Discuss the continuity of these limit functions. 2. Let f(x) be continuous for 0 + lal. Solutions and Hints to Problems SECTION 1.8, page 82 1.(q) If x is an integer, then y = (cos wx)® = 1 and y™ +1. If x is not an integer, then0 we have z, = (cos n! wx)? = 1 and lim w,, = 1, where w, lim z,”™. If x is irrational, then no mo a! x is irrational and by (a) for fixed 1, w, = lim z,” = 0, hence lim w, = 0. (c) The function of 1a is continuous except for integral values of x, The function of 15 is nowhere continuous. 2. If f(x) assumed any value y # }, then f(z) would take on all values between $ and x and would therefore take on irrational values. (See Problems 1.la, No. 16.) PROBLEMS SECTION 1.S.1, page 89 1. Let r =p/g, s = mn be arbitrary rational numbers where p, g, m, n are integers and q, n are positive. In terms of the integers p, g, m, n, define @rts, Wr—s (rs, =, Or 0, then xz < yz. 4. Prove that the following principles are equivalent in the sense that any one can be derived as a consequence of any other. (a) Every nested sequence of intervals with real end points contains a real number. (6) Every bounded monotone sequence converges. (c) Every bounded infinite sequence has at least one accumulation or limit point. (d) Every Cauchy sequence converges. (e) Every bounded set of real numbers has an infimum and a supremum. Solutions and Hints to Problems SECTION 1.$.1, page 89 L@rts= Pot Or-sa Pt Ora, in @i=2. sy mg (e) r N so large that lay = an'| < ay — 2; it follows that ay — in’ Say — 2, whence a,’ > x + (a, — ay) > x. Thus, if x is an element of the first class for [a,, 5,], it is an element of the first class for [an’, by’).70 Sec. 1.8.1: Solutions and Hints to Problems If x > by for some N, then choose m > N so large that , « nan <5, See lan — an’ <5, « an <> bn ;> where « =x — by. From the last two of these inequalities we have af 2 a’ = ql <5 and, thence, from the first, by’ — bn Se or bn’ <% + (bp — by) Se. Thus, if x is a member of the third class for [a,, by], it is a member of the third class for [4y’, by’I. Conversely interchanging the roles of a and a’, b and b’, we see that the first and third classes are the same for the two nested sets of intervals. It follows that the second classes are also the same. (®) If a, N so large that By — Ay < ay — by. Using this inequality, we find bn + Bn S by + Bn O and a, > 0. We then define TY = {[An%ms bnBn]}, — (@n > 0, tq > 0). In general, we define the product of any two real numbers in terms of the product of nonzero numbers by ay = |x| -|yl sgna-sgny, where _[ 4 =>0 SEN ade. fe and ay =Oifeither ory is zero. Here —z is defined by —2 = {[—Bp, —a,J}. In proof, for positive x and y, we first observe that since 0 0, From a, <2 Sb,<0 and O 0. We consider the case x 2, then, since b, — 2 < by — @, = (by — ap)/2”, there would be some value 6, closer to x than s, such that 5 > by > 2, contradicting the fact that b, is an upper bound for S. Further- more, x is the supremum of S since every interval [a,, b,] and therefore every interval [a,, «] contains points of S. A similar proof establishes the existence of an infimum. Thus (a) implies (e). (ii) Next consider a bounded monotone sequence ay, SAY dy, < any, N a—a, »,}. The supremum b, of this set is less than 6. If it is a limit point of the sequence, we are through; otherwise choose », as the largest index such that a,, = by and repeat the process. In this way we either obtain a b, which is a limit point of the original sequence or a new sequence by, bs, .. . which is bounded and strictly decreasing. Since by (6) the sequence {b,} must converge, and since each by is a term in {a,}, it follows that {a,} has a limit point. Thus from (6) and (e) we derive (c). But since (6) and (e) are equivalent, either one alone implies (c). (v) Let @, be a Cauchy sequence. For some sufficiently large n we have lan — Gnyxl < € for all k. Thus lang! < lanl + = My, and, on setting lay] + lag] + +++ + lana] = My we have la, < My + Mp for all »; that is, the sequence is bounded. It follows from (c) that the sequence has a limit point «. For any « we can find an N so that if n,m > N « lan = aml <5 + Furthermore, since « is a limit point of the sequence, we can fix m so that ‘ lam — al < 3° It follows that la, - al N. Thus (c) implies (d). (vi) Now consider any nested sequence [4p, Bq]. Since a <6; for all i and j, we have 0 Sdn = bape < bn — An It follows that 6, is a Cauchy sequence and by (d) has a limit x. We have b, > xforalln, for if by <, then for n > N we would have b, < by < cor lb, —2| 22 — by > 0,74 Sec. 1.8.1; Solutions and Hints to Problems and « could not be the limit of 6,. Furthermore, a, <« for all n, for if ay > x, then b, > ay > x and b, —® Bay —2, and again x could not be the limit of 6,. Thus (d) implies (a). Letting an arrow denote the direction of implication, we indicate the train of arguments below. Clearly, assuming any one of the principles, we can deduce all the others. MISCELLANEOUS PROBLEMS, Chapter 1 1. If Wy, Wo, «++, Wn > 0, prove that the weighted average Wye, + Woy + ** + Ween w+ Wa + We lies between the greatest and the least of the z's. 2. Prove — 1 1 1 ad Wn +1 —I 0 arty" (x ty\" a" Sy. Interpret this result geometrically in terms of the graph of x”, 4. Ifa, >a, > °°: Da, and b, > by D++- D by, prove a,b; > (3 a)( 6,) . 5. (a) Show that the sequence a, a, ag, . . . can be written as the sequence of partial sums of the series 1, i, Ug... WHETE Uy = Gy — Gq for n > 1 and 1, =a. (b) Write the sequence a, = 7° as the sequence of partial sums of a series.Ch. 1: Introduction 75 (c) From the result obtain a formula for the nth partial sum of the series 1L+4494---+nb +---, (d) From the formula for 1? + 2? + --- + n®, find a formula for P43 4St*$--- + Qn $1 6. A sequence is called an arithmetic progression of the first order if the differences of successive terms are constant. It is called an arithmetic pro- gression of the second order if the differences of successive terms form an arithmetic progression of the first order; and, in general, it is called an arithmetic progression of order k if the differences of successive terms form an arithmetic progression of order (k — 1). The numbers 4, 6, 13, 27, 50, 84 are the first six terms of an arithmetic progression. What is its least possible order? What is the eighth term of the progression of smallest order with these initial terms? 7. Prove that the nth term of an arithmetic progression of the second order can be written in the form an? + bn + ¢, where a, b, c are independent of n. *8. Prove that the nth term of an arithmetic progression of order k can be written in the form an* + bn*-! +--- + pn +, where a, b,..., p,q are independent of n. Find the nth term of the progression of smallest order in Problem 6. 9. Find a formula for the nth term of the arithmetic progressions of smallest order for which the following are the initial terms: (a) 1, 2, 4,7, 11, 16,.... (6) -7, -10, —9, 1, 25, 68, .... *10. Show that the sum of the first n terms of an arithmetic progression of order k is Sy + ay Spa + -* > + aS + ayn, where S, represents thesum of the first n»th powers and the a, are independent of n, Use this result to evaluate the sums for the arithmetic progressions of Problem 9. 11, By summing oy 410% +2) +k 41) —-@-IO H+): © +H from » = 1 to » =n, show that $10 4042-4 =D vel (n +k +1) k+2 12, Evaluate 1° + 2% + --- + n® by using the relation Fay +10 +2) -3 + 4%.76 Solutions and Hints to Problems 13. Show that the function x #0 S@) = (18 0, x=0 is continuous but not Hélder-continuous. (Hint: Show Hélder continuity with exponent « fails at the origin by considering the values # = 1/2"/2.) 14. Let a be a monotone decreasing Sequence of nonnegative numbers. Show that > a, converges if and only it 2%ayy does. vo 15. Investigate for convergence and determine the limit when possible, (a) nle — [ne] (6) anlansa, Where a, = 0, a, = 1, and dyye = ayy1 + ay. Solutions and Hints to Problems MISCELLANEOUS PROBLEMS, Chapter 1 1, Suppose x Sa Wt D4) ovata, Vvatl whence nah ech * AVvn+2-N< eaCh. 1: Introduction 77 On the other hand, 1 2Vam +1) +1 2V@4+hF +1 2Va+ ——< = Vn+1 vn+1 vn+1 2 2 ees vnt+1 2Vvn +1; hence $1 cavagi <2Vn +1 fa VE | and the proof is complete. 3. For n = 1, the inequality is obviously true. If the inequality is true for n =k, then ety) [at py) (2 + 2 SAT 2 aH og yPHL yk + ya aa 4 Observe that (y — 2)(y* — 2*) > O and 2+) 4 y+ > wy! + yx", Applying this result to the second term on the left above, we obtain the desired con- | clusion. The midpoint of the chord joining two points of the graph lies on or above { the graph. | 4, Since (a; — a,)(b; — 6,) > 0, we have ~ 05d YG - ab; - 4) SDD (ab; + a;5,) — (ib, + :a))) | 5.(@) Sy = uy + ug too $y = ay + (Gy — ay) + (gy — Gg) + °° + Gn — Gna) = ay + (ay + a) + (—ay + ag) +> + (Gna + Gn) = (ay = ay) + (aq = 2) +2 + Gnd = Gna) + One (0) uy =n = (n = 1) = 3n? — 3 $1, S Gxt - 3k +1). a © >= if +36 - »}. ma a we78 Solutions and Hints to Problems Applying the formula for the sum of an arithmetic progression, we obtain Soe -y= ween mi and hence a (n + 1)(2n +1) B= eee 4 are sai si baceaunae @) Sek - 1 = 4S 4S 4S ey mee aig = 3(n + I)(n + 2)Qn + 3) — An + In +2) + $1) = An + 1)(2n + 1)(2n + 3). 6. Take the differences until one term or a sequence of constants is reached: 4 6 13 27 50 84 131 193 2 7 14 23 34 47 62 5 7 9 11 13 15 2 2 2 2 2 To calculate the next terms, add back the successive differences. Answer: 3rd order. 193. 7. Let the terms of the arithmetic progression be denoted by an. Set bn = ns — Ons Cn = basa — On = Cy. The second differences c, are con- stant. We have then by = Oy ey Hep te tent =bh +(n — Icy, whence Sh a, =a, +> & a (n= In = 2). in = fn = 2e, =a, + (1 = Ny +, which is quadratic in 7. 8. Let a, be the arithmetic progression of kth order and define the sequence of differences dy! = any, — dn, dy? =d},, — d,},..., da® = dk} — dit = constant. We prove -1 n-1 n-1 a anat("] Jan +( A Jat ao +( k ) as from which the result follows, since x) _ na =e ar th) r rt is a polynomial of degree r in n.Ch. 1: Introduction 79 The proof is inductive. We observe that (1) holds for 2 = 1. If the result is true for n Gr* + ayy +--+ +a) vet” vei = aySp + Oa Sea t+ + ayn. To obtain the sums of the progressions of Problem 9, we take S, from Problem Sc and Sj from Problem 12 below. n(n? + 5) n(n — 5)(Sn® + 11n + 26) 6 * 24 . Answers:80 Solutions and Hints to Problems 11. Setting 4, = (» — 1) + 1I)@ + 2)---(@ +4), we obtain DY Gyr — 4) = Wag = ng nn 4 2 nek ED. But ta Wy =O FD OF HIO +k 4+) -0-DI =v +1)---() +k +2), from which the result follows. 12. From the preceding problem a nr + +2) nin +) “ 4 3 2 _ n(n +1)? ae. 13. Set x, = 1/2", Then 1 \f@n) — fl] = jog") But L Lit, — OF =>. Since 2" > [n(n — 1)]/2 for all x, we have L 2L a B SiG (2L/«) + 1); hence [f@n) —f@| > L le, — Ol and Hélder-continuity is not possible. 14, If 3 2¥a,, converges, then mo Sn = Da, = ay + (aq + a) + (ay + °° +.) a to tage to ak) H gk + Ages, Hi + Any where 2B on <2kH, Consequently, since a, is monotone decreasing, Sy Sa, + ay + day ++ >> + Page x =>2,. v=o Thus S,, is monotone increasing and bounded; hence convergent.Ch. 1: Introduction 81 Conversely, if > a, converges, then vet Sgn = ay + ay + (Gy + ay) +> ++ + (Gyn-1,, + Gyn) a, DP ay tay te +2 lays 1 2 d Vv IMs ayy. 0 Thus > 2a,» is bounded and monotone; hence convergent. v0 15. (a) Observe that nle—[nlel =n > 1 1 terest’ 1 The limit is 0. (6) For n > 4 we have 4m 2 Ang 2 34n- The proof is by induction. The inequality is verified for n = 4, If the inequality holds for n = k, then from ay,» = a4, + a), we have 8 2 Bane S Gis + Beir Saxe S Aegr + Fey S Sans from which the inequality holds for n =k +1. From the inequality we have for n > 4 ni)? S Gan Bane) < Ondnsar so that fa ton 3 nia Anse 5 and the sequence is both monotone decreasing and bounded below. The limit is (5 — 1)/2. (Compare Section 1.5, Problem 7.)2 The Fundamental Ideas of Integral and Differential Calculus EXERCISES SECTION 2.1, page 120 Approximate the area under the following curves from above and below by using circumscribed and inscribed rectangles. Obtain the area to within one decimal place accuracy and show that your estimate of error is correct. 1. f@) =2 +1 on (0, 2]. 2. fle) = 2? on [0, 2]. 3. f@) = on [1, 2]. 4. f(@) = Vz on (0, 1). PROBLEMS SECTION 2.1, page 120 1. Let f bea positive monotone function defined on [a, 5], whereO 80. 2. Proceed as in Problem 1. S — s = 8/n. Taken > 160. 3. For a subdivision into equal parts, Le 1 $= 21ee=Dh’ 18 1 “n+ kin” S —s =1/2n. Taken > 10. 4, Subdivide using the points z, = k2/n?, Then “E0)G-"S") 12 =5 RH -k ae! ) 2 (k-1\ (Kk -1P eee ie a] 1s =p Zee - k) — Qk — 1). Tus S— 5 = 55 3 @k =) = Taken > 20° Solutions and Hints to Problems SECTION 2.1, page 120 b Bo 1. The numbers 4 = f f@) de and B =f $y) dy are represented by areas indicated in the figure on p. 84. It is geometrically obvious that A + B = bf — az from which the result is obtained (see also p. 275).84 Sec. 2.2: Problems PROBLEMS SECTION 2.2, page 128 1, Prove for natural number p that > 1 ; ? dz = —— (grt — grt [: de = 5 Ph — ah) using a subdivision of [a, b] into cells of equal length. Employ the techniques in Chapter 1, Miscellaneous Problems 5 to 12, to evaluate the approximating sums Fy, ° 2. Derive the formula for [ x* dx, a, b > 0, when « is rational and a negative, say « = —r/s, where r and s are natural numbers. (Hint: Set qs = +, where g = Vbja.) 3. By the method used to find the integral of sin x, derive the formula o f cosa dx = sinb — sina. a 4. Make a general statement about S(@) dz when f(x) is (a) an odd function and (6) an even function. 12 12 5. Calculate f sin x dz and cos x dz, Explain on geometrical 0 0 grounds why these should be the same. Furthermore, explain why at2n bar f sing dz = cos x dr ° a for all values of a and b.Ch. 2: The Fundamental Ideas of Integral and Differential Calculus 85 6. (a) Evaluate J, -[ xi/n dx, What is lim I,? Interpret geometrically. 0 n0 (6) Do the same for /, -{ 2" de. 0 7. Evaluate li lit + +z) nao VAN VB Val" Solutions and Hints to Problems SECTION 2.2, page 128 1. Taking an equal subdivision of the interval into cells we form the sum not 3D (a + kAyh, 0 where h = (6 —a)/n. Using the binomial theorem, we have [a +k + Ih = [a + kA a (’ ‘ite + Kh)? + is 2 ')wre + kit po =(p + h(a + kh? + wl (? . ‘Nia + kh 4. ‘ =(p + Dia + kh)? + PT. For / <1, it is easily proved that 7), is bounded independently of k, reap +1 ITd <> ( \e la + kh i= <3’ a ) la + kara r=0 S@+kh +1" S Ca] + |b] + 1)?** (cf. Problems 1.1e, No. 2c).86 Sec. 2.2: Solutions and Hints to Problems It follows that (a +khPh = [a + (k + 1)h}PH — (a + kh — PT, pti whence " (a + niet — grt + kiyph — ST germ ptt ed <# SIT & < nb (\a| + |b] + 4. Using mh = b — a and passing to the limit as n tends to infinity, we obtain the desired result. 2. In the notation of Section 2.2, the integral is lim F,, where q-1 Fy = (6841 = at*ge x and the problem is to evaluate fon Se lim oa ft We have g = 7',q = 77%, (r # 1), whence -1 -1 ltrte tet TS f Lr bee be? ba is rayon oh (trt +77) ie aee PL Pe eae)? s In either case we obtain for the limit as 7 tends to 1, —— = . e=F +a 3. In the summation formula of Section 2.2e, we apply the identities wrod Ama maslensCh. 3: The Fundamental Ideas of Integral and Differential Calculus 87 to obtain S heos(a + kh) => sin (« +kh +3) een Rea __A hom 2n +1 7 =— |= a+5t5 cos(a+—Z—h +5 2sin5 ee . an+1, : h ==, ats —sin{a +5 2sin = at {= (6 +5) =sin (« +4)| 2sin5 nv) which yields the result on taking the limit as tends to infinity. 4. Subdivide the interval into 2 equal parts of length A = a/n. We then have if fe) de = tim 5 1fkh) + f(-kiONA. Si ae (a) The integral is zero. ® i fle)de = [709 de, B k aft F/2 sf sins de = f cosa dz = 1. lo fo The integrals correspond to the areas of congruent regions. Since cos« = sin (x + 7/2), the region “under the curve” y = cos between x = b and x = b + 2m is congruent to the region under the curve y =sinz between x = b + 2/2 and b + 37/2. Moreover, the area under the curve y = sin x is the same over every interval of length 27 because of the periodicity property sin (w +27) =sinz. The area over the interval [a, a + 27] can be decomposed into areas over the intervals [a, 2n7] and [2nx, a + 27], where 1 is the unique integer for which a/27 0. at (6) In ©, a>l limI,={0, -Il (a) Je + 1)% de, (6) fisn ax de, (c) if cos ax dx, where « is an arbitrary integer.Ch, 2: The Fundamental Ideas of Integral and Differential Calculus 89 7. Use the formulas obtained on p. 135 along with the identities sin? z = 4 — 400s 2x, cos’ x = } + } cos 2x to prove that e b- sin 2b — sin 2 [leoste de =°5* eee Ly b- in 2b — sin 2« [ana sin 2b — sin 2x is 2 4 a 8. By use of Exercises 1.5, No. 2b, evaluate f 2° dz, using division into equal subintervals. a 1 9. Evaluate fi (1 — )" dx (where n is an integer), 0 10. In each of the following cases give the mean of /(«) on the interval [a, 6] and a value & for which it is achieved. @) f@)=4e-1, a=1,b=3. ®f@)=2+1, a=0,b=1. Of@=@+D, a=0, (a) f(@) =2 —32, a=-1,6 Of@=Vz, a=0,b=1. Nf@) =2". 11. Let f(z) and g(z) be continuous on [a,b]. Prove if f(z) > g(@) and definite inequality f(£) > g(&) holds for at least one point § in the interval, then definite inequality holds for the integrals > > f f@) de > fi (2) de, PROBLEMS SECTION 2.3, page 136 *1. Cauchy's inequality for integrals. Prove that for all continuous functions f(z), g(2) . > 5 2 f U@P dz f [g@P dz > ( I S@g@) de) $ *2, Prove that if f(z) is continuous and S@) i fat, then f(z) is identically zero.90 Sec. 2.3: Answers to Exercises *3, Let f(x) be Lipschitz-continuous on [0, 1]; that is, If@ -f@)| g(z) + « for « = 31f(&) — g(@)]. Then for h(x) = f(x) — g(x) ° 3 Pete ph fine) de -[ +f +f Hex) de a a ns Jess 20425 +0>0, (if ¢ is an end point the proof must be modified slightly.) Solutions and Hints to Problems SECTION 2.3, page 136 1. Observe that for all f, ° f Uf@ + g@P dr 20 and consequently, [ rca)? dx +2 if ” padgla) de +f i r (a)? de > 0. Set A = "fede, B -f pte) de, c= { "gay de, and follow the proof of Cauchy’s inequality in Section 1.1e. Alternatively, express the integrals as limits of sums using equal subdivisions of (a, 6] and apply Cauchy’s inequality to these sums. 2. By the mean value theorem of integral calculus, S@) = xf) = rx, f(x) = erry f(a) ..., where 2, lies between 0 and 2, x, between 0 and 2, 25 between 0 and 2», etc, Let M be an upper bound for |f(z)| on [—1, 1]. We have || <|z| and consequently, IF@) < lzl" |f@n)l < lel". Since lim x" = 0, we have f(x) = 0 for all x in [—1, 1], (at the end points, by continuity). Applying the same argument to flt¢a= f"r0 dt and f(—1—2) = h for |x| < 1 we can extend the result to [—2, 2], and, inductively, to [—n, n] for all n. ie S(t) dt aa92 Sec. 2.4: Exercises -|[roe-23rGi) ane (i) | “Ete But for x in [(k — 1)/n, k/n]) k k ra -1(7)| 0. (4) Show that the line x + y = 0 and the curve y = log x have exactly one intersection. PROBLEMS SECTION 2.5, page 145 1. Prove loge < Jet @ —1. 1 ol +u (6) Show for x > 0 that 2 @—F 1 for all positive x. 1 (®) Since + log x is monotone increasing, there can be at most one intersection where x + log x has a root. That x + log z has a root is clear by the intermediate value theorem from Exercise 1; + log} <0, but «x +logz > Oforz >1.Ch. 2: The Fundamental Ideas of Integral and Differential Calculus 95 Solutions and Hints to Problems SECTION 2.5, page 145 , (([Lal<('a dae Le | 2.(a) log (1 +2) ={ zat 1 oon ee =lim 24m neo het = du “Jol tu’ where for the common representation of the two integrals as the limit of a sum we use a subdivision of the interval into m equal parts of length h = 2/n. (6) Integrating the inequality 1 V=esae 0), we obtain the desired result. (©) For 0 for all x (cf. Exercise 2.5, No. 3a). (6) Show that the line z + y = 0 and curve y = e* have exactly one interaction.96 Sec. 2.6: Answers to Exercises 3. Prove forO <# 0 that an 2 an gzyntl a ltety t+ Se slteda te +o too. (Hint: Obtain upper and lower estimates for I e“ du and integrate repeatedly. ) ° Obtain estimates of the same type for e* when x < 0. Answers to Exercises SECTION 2.6, page 149 1. (abt = erlozad = gatloga + logo) — grloga+aloge = e2log aga log b =a 2. (a) Since log x < 2, it follows from the monotonicity of the exponential function that x = el#* < ex, (b) The intersections occur at points logy = —y and by Exercises 2.5, No. 3b only one such value of y exists. 3. From Section 2.6c, nlog (1 + 2/n) =2Ja 0, f() =1, and for x = 0 there is no derivative. (3) Slope of chord is zero, but for x #0, | f(x)| = § |zI-4 and f’@) not defined. () (1 = £4; 216 = -$,4. 3. If f(z) had two roots, then by Rolle’s theorem for some value & between them, f'(é) = 0. 4.(a) If the roots are given by 2 <2, <-:+ <2, then by Rolle’s theorem there exist values &, where 2, < & sin « for positive x and x < tan ¢ for x in (0. 5). 2. If f(@) is continuous and differentiable for a 0 for § 0. Consequently, f is nondecreasing and is strictly increasing in the intervalCh. 2: The Fundamental Ideas of Integral and Differential Calculus 105 (0,27). It follows that f(z) > f(0) =0. For g(z) =tanz — x we have &(@) =14+tan?x-1>0, (0 g(0) = 0. 2. Fora —4 we have on integrating from 1 to 1, 0<4—4. For the total displacement we have rT % a froaw F-. If Ft = F- we call the common value the Darboux integral of f. Prove that the Darboux integral of f is actually the ordinary Riemann integral; furthermore, show that the Riemann integral exists if and only if the upper and lower Darboux integrals exist and are equal. 2. Let f(x) be a monotone function defined on [a, 6].Ch. 2: The Fundamental Ideas of Integral and Differential Calculus 107 (a) Show that the difference between the upper and lower sums for a subdivision into n equal cells is given exactly by 2-7 =O - S16 — an, and explain this result geometrically. (4) Use the result of (a) to prove that the Darboux integral exists. (c) Estimate = — o in terms of f(a), f(6) and the span of the subdivision if the cells of the subdivision may be unequal. (d) Usually f(z), if not monotone, can be written as the sum of monotone functions, f(z) = 4(z) + y(e), where $ is nonincreasing and y is non- decreasing. Estimate the difference between the upper and lower sums in that case. 3. Show that if f(x) has a continuous derivative in the closed interval [a, 5], then f(z) can be written as the sum of monotone functions as in Problem 2d. Solutions and Hints to Problems SUPPLEMENT TO CHAPTER 2, page 192 1. (a) Let x, = & < & <+-++ < & =, be those points of the refine- ment of the original subdivision which lie in the cell [z;_, 2;] and let «, and A, be the infimum and supremum, respectively, of f in the subcell [£, 1, §]- If m, and M; are the infimum and supremum of fin the original cell [z;_1, 7], then they are lower and upper bounds for fin the subcell [£,_1, &;] and we have 3 a; — $a) 2 s mE; — §51) =1 ms = mx; — xj). Similarly, ‘ > AME; — 85-2) S Mie; — 25-4). Fat Summing from i = | to n, we obtain the desired results. (6) Let 2, and o; denote upper and lower sums, respectively, for the subdivision S;. Let S, and S, be any subdivisions whatever and let S; be the joint subdivision consisting of all the subdividing points of S, and S, taken together. S; is a refinement of both S, and S,; consequently, from the result of a, Xi 2 Ya2 > (c) First we prove for any positive « that there exists a subdivision S for which the upper sum & and lower sum o satisfy D-o f(a)(v; — 21) and B= S/(BXe; — 214), we then have e i « « x-A Kg (B= 953) Thus « -o0 £O —f@l(b - a) 7 D-ox F* > F~ > a, we have F+ — F~ < ¢, which yields the desired result. (c) © — o S|f(b) — f(@| A, where A is the span of the subdivision. (4) = — 9 S {(b) — w(a)] + [4(a) — $(6)]} 4. 3. f@) =f@ + ; i “ro +ironae ar +3 f yo -Ifolde.Ch. 2: The Fundamental Ideas of Integral and Differential Calculus 111 PROBLEMS MISCELLANEOUS PROBLEMS, Chapter 2 1, Prove that : @ fe -Ddr =, ® (orf Gt = nde = is Bs Qn+1!" 2. Prove for the binomial coefficient () that n 1 =a (2) = [« + » fa — 2) *as| +3. If f(z) possesses a derivative f’(x) (not necessarily continuous) at each point x ofa <= < 6, and iff'(z) assumes the values m and Mit also assumes every value “ between m and M. 4. Iff"(@) > 0 for all values of x in a 0 for all values of x in u 0, then ( * f ruaon a>s (; [io a) , fe directly and write down the corresponding integration 8. (a) Differe formulae: (i) =, (b) Evaluate 1 7 2a nw im = Cae one tim Z(1 + sect Z + sect + + sec’ ). 9. Let f(z) have first and second derivatives for all real values of x. Prove that if f(z) is everywhere positive and concave, then f(z) is constant.112 Solutions and Hints to Problems Solutions and Hints to Problems MISCELLANEOUS PROBLEMS, Chapter 2, page 199 1. (6) Since the integrand is an even function (See Problems 2.2, No. 46), the integral equals aon! = yrae =2(- wf £ (fortes 0 k=0 * Sen “Gis 2G em where at the last step we use the result of Problems 1.5, No. lg. 2, Set n =j +k, and ; 1 p -(7f'Joreenf a1 — 2) de. 0 Clearly, 2 = 1. We prove that if 4, = 1, then 4,,, = 1, so that 4, = 1 for allj, We have a ays Jgreen ‘Seo! ) a oro or J | cul) afte pesmi @) | k Jute sod at We rewrite this formula in the form Q aol ecwltt) r=0 r jorti k+rtl? On the other hand, replacing j by j + 1 in (1), we obtain _(itk tlle j+l\jtk+2 e)) Ana = ( g ) ZC. \eeeat”Ch. 2: The Fundamental Ideas of Integral and Differential Calculus 113 Using (2) and (3) we form the difference aac a= (ae M ‘)[Seo(/t ’) +(-0| 70 j+k+1 ae a - 1)? =0. It follows that 4;,1 = 4). 3. Let «and f be points of [a, 6), where f'(x;) = m, f(x) = M. Consider the continuous function oz) th fe) fora fixed h > 0. If m 0 so that f@ +h) —fe hy —mi|0 hy “\¢@i-h, for hy <0, it follows that there exist points , f in [a, 6] such that H(a) 0. Furthermore, since h(x) = h(x.) = 0, by Rolle’s theorem h’(é) = Ofor some £ in (x, 22). Consequently, h’(x) < Oforz, 0 for € g(x) for all values of z. Enter z = u(s) in this inequality, and integrate (note that the result is independent of the sign of a). von Silla fone 4 2 4 acting ss 9. If f(z) is not constant, then there exist points x,, x, such that f(x) < Sf (@2). Consider the case x, <,. By the mean value theorem, for some & between 2, and 2», we have f@) -f@D , 4 (@ = Om Since f“(z) <0, it follows that f’(z) >f'(@) > 0, and for # f'@q). The curve y = f(z) lies on or below the tangent line y = f'(x,e — 2) + 44 which takes on negative values for z sufficiently to the left of 2, namely ecm, * fe For such values of 2, f(z) must be negative, contrary to our assumption.3 Techniques of the Calculus EXERCISES SECTION 3.1, page 201 1, Find the numerical values of all the derivatives of 25 — 24 at x = 1. 2. What is the numerical value of the eleventh derivative of 3172° — 202x? + 76 at x = 1347 3. Differentiate the following functions and write down the corresponding integral formulas: (a) az +b, (6) 25cz", az +b (©) a + 2bx + cx’, Om ax? + 2br +e 1 1 © alt,’ Ota ies’ GB — VBx4 + 48 + VBxt + 4) ®) z+ 16 a 4. Differentiate the following functions and write down the corresponding integral formulas. 1 1+tana’ sinz + cosz sinz —cosx’ (@) 2sinxcos x, ) (©) xtanz, @) sinz O->.- 5. Differentiate 1 1 sec x = —— and cosecz =——., cos x sin 6. Recalling that sec x = 1/cosx, cosec x = 1/sin x, find the derivatives indicated in the following: a a (a) me sec x, (6) wa sec x tan 2, a dt 7 © Facosee 2, (4) Fytanzsinz. us116 Sec. 3.1: Problems 7. Differentiate a@sinz — x5 cosx a tana Evaluate: 8. foe + b) dz. Ee foes + 2bx +c) dz. 10. fox + Ta® + 5x4 + 3x? + 1) de. 1 1 Lt anf (58 tle 1 v2 fle +h) # 6 in ([akas la sin® 2, F Sty ss) 14, J» +7sine +5 - an) de, 15. fre « tanx dr, 16. Find the nth derivative of: (a) sin z sin 2z, (8) cos ma: sin ke. PROBLEMS SECTION 3.1, page 201 1. Let P(x) = ay + a,x + gx? +--+ + 4,2". (@) Calculate the polynomial F(z) from the equation F@) — F(@) = P@). *(6) Calculate F(x) from the equation oF (@) + CF (2) + cgF"(x) = P@). 2. Find the limit as n > © of the absolute value of the nth derivative of 1/x at the point x = 2. 3. Prove if f("(x) = 0, for all z, then f is a polynomial of degree at most n — 1, and conversely.Ch. 3: Techniques of the Calculus 117 4. Determine the form of a rational function r for which ar(x) zoo Mt) 5. Prove by induction that the mth derivative of a product may be found according to the following rule (Leibnitz’s rule): a dg (n\dfd™g | (n\d®fd"*g an) =f Gen * (ie a * Jaca n \d™"fdg df de dc n-1 Here (") ain (3) = =) ore. denote binomial coefficients. 2 2! (a = Ayx™ = ne $1 not iri} = 6. Prove that oe GoD Answers to Exercises SECTION 3.1, page 201 2. £40) = 1, FO) = 8, PU) = 36, FC) = 96, $40) = 12, sed) = 0, fd) = 2.0. 3.(a) a, (0) 175ex8, (ce) A+ a tak .(a) a, , a. OCT 22%(ap — ab) + Delay — ac) + by — fe) © (ax + Ope +P 4x(1 + x4) DP Ta + ae) -1 4.(a) 20822, (6) —2 sinz cosx ® (ram © -Gt yo: d 5. secu = secw tan, dz d > cosecr = —cosec x cot x. dx118 Sec. 3.1: Answers to Exercises a d 6. (a) Fa seca = 5 seca tanz = 2 seca — seca, () From 6a, 3 Beeoztans =F secs — sec x) =A (sect xtanz — sec x tan x) = 24 sec5 x — 20 sect x + sec x. © Cee 8 Oe, BOE: @ Sanz sins =F, cecstanz + sinz) Now applying part 5, we obtain 24 sec’ x — 20 sec? x + sec — cos x. 32% = a cosx — +——. sin” sin? 7. —xsinx + cosx + 32*sina + 2° cosxz — ax* 8. zr + be. 3 9, & + bet + cx, 10. 2% +27 +25 +a5 +2, 1 t lj 13.a sinz — b cot 2. 3 5 14. Be — Tes — 75 — Stanz, 15. secz. (- p= » {cos x — 3°" cos 3x}, forn = 2m, 16. (a) (-)" ; 7 (84 sin 32 sina}, form = 2m +1. 6) fon + bY sin (m + oe — (m — sin (mm — Ke), forn = 21; co [mn + 4 cos (m + Kz — (mm — ki cos (m — kz], for n=21+1.Ch. 3: Techniques of the Calculus 119 Solutions and Hints to Problems SECTION 3.1, page 201 1, Use the method of undetermined coefficients, equating the coefficients of like powers on the left and right. If F() = > 6, 2», then f=} (a) by = ay +k + Daye +k + MK + Dayse $00 + + [Kk + Ik + 2)--- na, ga St alt = Vg, etc. 2. By induction, d* (1) _ nl(-1)" ae\z) ~ 3 lim n!/2"41 = c (cf. Problems 1.6, No. 3). 3. Proof by Induction. The proposition is true for n = 1. Suppose it is true for n =k. If f(2) is a polynomial of degree at most k, oe f'@) isa polynomial of degree at most k — 1. We then have f@*+0(2) = fr (@) = Conversely, if f(z) = 0 for all x, then f(z) = C, where ei is constant. Consequently, for g(x) = f(z) — Cx*/k! we have g(*(z) = 0 for all x. We conclude that g(z) is a polynomial of degree at most k — 1 and hence f(z) is of degree at most k GE" + Ay ye” 1+ Byt™ + By z™ 7 + +a Foy’ Where dy #0, bm #0. 4, Let r(x) = 21) _ (= Magy t™*™ ++ r(2) Gnbyz™ ho Answer: numerator and denominator have the same degree.120 Sec. 3.2: Exercises 5. Proof by Induction. The theorem is proved for n suppose it is true for =k. Then . SupPt qui drat gan lg) = $ sea] dad &(k = E(Aroege @) & (k 5 :) [. SO @ge—tD@) + SVE)gO-v) @| =f@g"*@e) 4 s{(‘) “a ( k rove (et 10(x) eakkeh Wat 16 + FM @g(@) ka (ke +1 = = ( i ) FOI), where we have used the addition property of the binomial coefficients (See Section 1.5) at the last step. EXERCISES SECTION 3.2, page 206 1. Ify = 2°/4, y = 16 corresponds to x = 8. Find dy/dx for z = 8; solve y = x"/4 for x and find dz/dy for y = 16, and show that the values of these derivatives are consistent with the rule for inverse functions. 2. Prove that (@ are sin « + arc sin B = arc sin(aV1 — & + BV1 — 2). (b) arc sin a + arc sin 8 = arc cos (V1 — a V1 — # — a). (c) arc tan « + arctan # = arc tan av. Differentiate the expressions in Exercises 3-16 and write down the corre- sponding integral formula. Va = |. Vx cos? x. 3135 4, Vxcos? 2, 14 vz VrCh. 3: Techniques of the Calculus 121 1 +arctana 7. (arc sin 2) (arc cos z). . Tacos are sin x ) cane" 10. Sarceot = + =. 11. 2(log x — 1). 12. log (ae). 13, e* log z. 14. e*(acosz + bsinz). gs eet 15, 2%e*. 6. ST 17. Find the nth derivative of: (a) xe, (©) (log x), (©) & cos 2z, @ (1 +2)%. PROBLEMS SECTION 3.2, page 206 1. Let y = e*(asinz + bcosz). Show that y” can be expressed as a linear combination of y and y’, that is, yo =py +o, where p and q are constants, Express all higher derivatives as linear com- binations of y’ and y. *2. Find the nth derivative of arc sin x at « = 0, and then of (arc sin z)* atz =0. Answers to Exercises SECTION 3.2, page 206 Aol © dz|dy" 2. Verify identity of the derivatives with respect to 6 and equality when B=0. 1 5. —____.. vz(l — V2)?122 Sec, 3.2: Solutions and Hints to Problems (1 — tanz) + 32(1 + tan? x) 3a4(1 — tana? 7. (arc cos x — arc sinz)/ V1 — 22, 2 Tae ceca” 9 1 are sin x ‘ warctanz (1 +22arctanz)?” 5 1 10. — 1+a2 VI —2*are cosa)? 11, log. 2) 12. - +7. 1 13. e(: + log) : 14, e[(a + 6) cos x — (a — b)sin x]. 15. 2" Vea + 2). 4 ere 17. (a) e[a"x? + 3na™!a? + 3n(n — Ia8x + n(n — 1m — 2a"), =n = I/"e A=)Mn = 1)! (3 1 ). 16. @) - —logz = 2,5 7's oe (a5) (n=1/2) n +(" >, (- D( ie) sin2s| = 5"/¢ cos (28 + na) where tan « = 2. @e. 3 (*) (") Gene, Solutions and Hints to Problems SECTION 3.2, page 206 = 9 =2, yi) =(—4yry, yet = (—4yry’, ym = 2(-4") x Fy, yor) = 2-4 — 2D.Ch. 3: Techniques of the Calculus 123 2. Let y = arc sinz. Then d'y a 1 ) ca ® a de NV a) al = : Apply Leibnitz’s rule to this last expression. avy ane 1 ee z=(q = 7) me =3- * at zt =3°@ 2) goal T= | ice le=0 and continue the process CH! ing gece 2a = 4 +n ae"|,9 ~ (Q» —1)-(n — 2) - 4) @—2 +2) z * C—aAyerye] |, For n even, ary aan for n odd, n = 2m + 1, dry OY _ 42.32.52... (2m — 1) Gon TEBE SP Om — Apply Leibnitz’s rule to obtain mei a 2.92.52... (2k —1))]- [12-32-52 = (2m — 2k —3)3)(_ 2" EyUt Se Ok DATE 38- 58+ Qm — 2k ol and amt Saari (aresinz)?] = 0. dtm. emo EXERCISES SECTION 3.3, page 217 Differentiate the following functions. 1. @ +1, 2. Gx + 5)% 1 . (x? — 328 — 23)8, } 3. (@@ — 3. y 4, cS124 Sec. 3.3: Problems 5. 6. (ax +b)” (nan integer). 7 1 ja + be +e z+V@—1) "N bt + me tn 9. (VL — ay, 10, sin? 11, sin (22), 12. V1 +sinta, 13, atsin 4 14:'ten a l-« 1S. sin (2? + 3x +2). 16, arc sin (3 + 2°), 17. arc sin (cos z). 18, sin (arccos V1 — 2°) 19, 2V2 — V2, 20. [sin@e + 7] ¥8, 21. [are sin (a cos x + b)}*. 22, etan' s+ loging, 23. (@ + 2)4(1 — 2)4(@? + 1)%, 24. (@) 289, (6) (Cos x)tanz, PROBLEMS SECTION 3.3, page 217 1, Find the second derivative of flg(h(z)}]. 2. Differentiate the following function: log,,.) “(z), [that is, the logarithm of u(x) to the base v(x); v(x) > 0). 3. What conditions must the coefficients «, f, a, b, c satisfy in order that ax +B V (az? + 2bx + 0) shall everywhere have a finite derivative that is never zero? 4. Show that [d"(e**/?)]/dx" = u,(z)e*”/, where u,(z) is a polynomial of degree n. Establish the recurrence relation Ung, = lly + Up’. *5, By applying Leibnitz’s rule to d@ 2, — /2) 2/2 qe) = ae", obtain the recurrence relation Unga = By + My.Ch. 3: Techniques of the Calculus 125 *6. By combining the recurrence relations of Problems 4 and 5, obtain the differential equation Un” + Ln’ — Ny = 0 satisfied by u,(). 7. Find the polynomial solution U(x) = 2" + aye"? 4 ++ + ay of the differential aw Un” + 2Up’ — nu, = 0. *8. If P,(z) = "at — 1)", prove the relations nl P i a P eel oy +2 ,, n+2 ©) Pais =e py Pe + eT Pt (0) Pay = 2Pp’ + (0 + Pa Pr. d (co) ne = 1)Pn'] — n(n + 1)Pp = 0. 9. Find the polynomial solution nt x » = arene + a Vps+s tay) of the differential equation d leet — DPn'] — n(n + )P, = 0. 10. Determine the polynomial P,(x) = binomial theorem. zz = ~ (a? = 1)" by using the “11. Let 2960) = (P ) ea —2)-*,n =0,1,2,...,p. Show that 1 = 3 A000)126 Sec. 3.3: Answers to Exercises Answers to Exercises SECTION 3.3, page 217 1. 3@ + 1% 2. 63x +5). 3. 1524326 — 6x? — 12% — 328 - 1)8, geet 2x “d +a7 ar 1 Vat — A(x + V2? = 1)" x%(am — bl) + 2x(an — cl) + (bn — em) 2V (ax? + bx + c)(iz® + mz + n>} * 9. —S(1 — 2). 10. sin 2x. 11. 2x cos (2%), sin 2 cos x aye. al. 12. ————.. 13. 2( sing ~ Leos). V(l + sin? x) aon a 2 6. an(ax + by". Jom 8. 14. (1 — 2)" cos® 15. (2 + 3) cos (2® + 32 +2). 16. ————————.._17. - 1. 18.1 vai-3 +27} sae : Ln aks : 1». 2 (Vita V9, 20, VScos@ +7) (sine + >), a4 OF. tare sin ( +b) .——— - - {arc sin (a cos tag V{l — (az + b)} (- tana tan® x +logsin 2 Bane 4 ate) dnt, 23. 4(x + 29 (a? + :1)% VO = 23) - (@ + 2)@? + 1% 2 3¥0 — 2)? + Aina? + 1 H@ + 2)°V0 = 2%), 24. (a) wane(S + log cos) : .. x log cos *) : fant a pe (b) (cos x)" -( tant + SayCh. 3: Techniques of the Calculus 127 Solutions and Hints to Problems SECTION 3.3, page 217 1. f’Igth@)ilg*th(a) h(a) +f Ig th@)i]g” (h(x) h(x) Hf Tg Wa)’ a) h'(e) « _logu we) v(x) log u (x) 2. Use loge 4 = ioe * uta) log 2) ~ v2) log ofa) * 3. Neither the numerator nor the denominator of the derivative may vanish for any value of x, The derivative is (ab — Bale + ac — 6B (ax® + 2be Foe” The conditions are either ab — fa =0, ac ¥ bf, a > 0 and ac — B® >0 ora =0,b =0,c >0% #0. 4. Verify directly that up =1, m4 = and m = (2 +1). Ifm(z) isa polynomial of degree k, then dee") i (4)z* Gee = Gy Mealadet’2*] = [rug @) + uy’@)e2* = Up ya(aeC=*, Thus u,,; is a polynomial of degree k + 1 and satisfies the stated recurrence relation. 5. By Leibnitz’s rule dnl am ane a #12 & geit) (ety +n (2). dx” dx” 6. From the two recurrence relations we have u,’ = mul,_,. Differentiate either recurrence relation and employ this relation to obtain the differential equation. 7. For the coefficient a, of 2"-” in uy, the equation up” + cup’ — nu, = 0 yields (n =n — » — 1a, — @ + 2)a,49 = 0. Thus mt = 1) ag 4 Mam IM = 2" = 3) ae ¢ pipes My = 2 + a ; amt pee,128 Sec. 3.4: Exercises 8. Apply Leibnitz’s rule to nto d ana (a) wae =p = gaa I@* - 1)-@ - 1, ane an 0) Fa — "= ~ (20 + Ix -(@? — 1"). (c) Equate the two expressions for P*,,, in (a) and (b). 9. For the coefficient a, of x" the differential equation yields _ a= Mn » 8 = OF DQn 9-1)” whence on! n= 1) ag , ma = Wa — 20 — 3) P(e) = sonle 3@n—)* +2-4-@n = NQa = ze] [nl2) (2n — 2k)! - mi (i) Goat 2%)! 10. Same as solution of Problem 9. 11. By the binomial theorem, Sans (x) =[z + (1-2) =1. Also, differentiating (+2) = 3 (2) n=0 lesen S,((Q}r= Multiplying by z* and putting a = 1 — 2, we have Jeers a8 nan k times, we have EXERCISES SECTION 3.4, page 223 1, A quantity of radium weighs 1 gm at time ¢ = 0. At time ¢ = 10 (years) it has diminished to 0.997 gm. After what time will it have diminished to 0.5 gm?Ch. 3: Techniques of the Calculus 129 2. Under favorable conditions, bacterial cultures grow at a rate in pro- portion to their size. Assuming the size of the culture increases by one-third in an hour, find the instantaneous rate of change of the culture. Estimate the time it would take to increase the size of the culture one-thousandfold. 3. Solve the following differential equations: (@) y = ay — A), (b) y’ — wy = B, © y — xy = few, (d) y’ — wy = fer, y xa, PROBLEMS SECTION 3.4, page 223 1, The function f(z) satisfies the equation fe +9 =fOfH. (a) If f(z) is differentiable, either f(z) = 0 or f(z) = &**. *(b) If f(@) is continuous, either f(x) = 0 or f(z) = e*. 2. If a differentiable function f(z) satisfies the equation flew) = fe) + flv, then f(z) = «log z. 3. Prove that if f() is continuous and fee) -["r0 dt, then f(z) is identically zero. Answers to Exercises SECTION 3.4, page 223 1. —10 log 2/log 0.997 = 2.31 x 10* years. 2. For a culture of size » a get ett Gm ae a = log5. Twenty-seven hours. 3.(a)y = B+ce*; ()y = 2 + cer, «#0; y=fr+c,a=0; (©) y = Bret + ce; (dd) y = oF + cer™ yr130 Sec. 3.4: Solutions and Hints to Problems Solutions and Hints to Problems SECTION 3.4, page 223 1. (a) We have d afer» pea ‘@) =f@f'@. Consequently, F(z) = Cer, where C = (0). In order to satisfy the given functional relation for x = 0, y = 0, we have C = C®, Consequently, either C = 0 or C = 1. (&) From f(z) = f(z/2)%, it follows that f(z) > 0 for all x. If f(a) =0 for any 2, then f(z) = f(x — x) f(x») = 0 implies that /(2) is identically zero. Otherwise /(x) is everywhere positive as we assume henceforth. Since (0) = f(0)®, we have (0) = 1. Consequently, from f(a — z) = f(0) we have 1 f= FH: For any natural number » we have S(nz) = f(r - YzIf@), from which it follows that len) = flor and combining this with the results above we see that this relation holds for all integers n, no matter what sign. In particular, taking x = 1, we have S(n) =", where a = f(1). For a rational number p/q we have He) Hj-e Finally, since f is continuous and a* is the continuous extension of a* from the domain of rational values r to all real values x, S@) =a. 2. Take the derivative with respect to y aty = 1. 3. Since f(z) = f(z), we have f(x) = Ce*; but f(0) = 0. (Compare the solution of Problem 2.3, No. 2.) < and, usin} ‘) sf (;Ch. 3: Techniques of the Calculus 131 EXERCISES SECTION 3.5, page 228 1. Given the value of one of the following, sinh x, cosh , tanh x, determine the possible values of the others. (a) coshx = §. (®) sinhz = e (c) tanh 2 = 12. 2. Derive the following hyperbolic identities. (a) sinh 2% = 2 sinh x cosh x, (6) cosh 2x = 2cosh?z — 1 =1 +2:sinh*z. © tanh 3 = SRE =H (d) tanh (a + 6) = eee © coche = (/) sinh x - tae. 3. Prove (cosh x + sinh x)" = cosh nx + sinh nz. PROBLEMS SECTION 3.5, page 228 1. Prove the formula sinh a + sinh 6 = 2sinh (—F*)oosn(* > *) 2 Obtain similar formulas for sinh a— sinh b, cosh a + cosh 6, cosh a— cosh b. 2. Express tanh (a + 5) in terms of tanh a and tanh b. Express coth (a + 6) in terms of coth a and coth 6. Express sinh 3a and cosh 3a in terms of cosh a.132 Sec. 3.5: Solutions and Hints to Problems 3. Differentiate (a) cosh x + sinh x, (6) etanhz+cotha, (c) log sinh (x + cosh®x), (d) ar cosh x + ar sinh z, (e) ar sinh (2 cosh 2), (f) ar tanh (22/(1 + 2%), 4. Calculate the area bounded by the catenary y = cosh z, the ordinates « =aand x = b, and the z-axis, Answers to Exercises SECTION 3.5, page 228 1. (a) sinha = +4, etc. (6) cosh x = 4, etc. (c) coshz = 48, sinh x = 32, etc. 2. (a) 2 sinh x cosh x = Het — e-*)\(e* + e-*) = 3 (e* — e-*) = sinh 2x. The hyperbolic identities, like the corresponding trigonometric identities, may also be derived from the addition theorems. 3. Proceed by induction. True for n = 1. If true for m =k then (cosh x + sinh z)**! = (cosh x + sinh x\cosh kx + sinh kx) = (cosh x cosh kx + sinh 2 sinh kz) + (cosh x sinh kx + sinh x cosh kx) = cosh (k + 1)x + sinh (k + Iz. Solutions and Hints to Problems SECTION 3.5, page 228 1. Apply the addition theorems to the right side. 6 -6 sinh a — sinh b = 2cosh (+ ) sian (7 ). a+b a—b cosh a + cosh b = 2 cosh z cosh . b -b cosh a — cosh b = 2sinh (+ ) inn (* ).Ch. 3: Techniques of the Calculus 133 tanha@ + tanhb 1 + tanhatanhb 1 + cothacothb cotha + cothb * ha —1 hat] sine = /(@22—); cosh fa - (=). etanh x4 coth x cosh 4x — 1” 2. tanh (a + 6) = coth (a + b) = 3. (a) sinh x + cosh 2, (6) —8 (c) (1 + sinh 2x) coth (w + cosh? x), (d) sinh x ©) a? cosh? x +1” 4. sinh b — sinha. EXERCISES SECTION 3.6, page 236 1, Find the maxima, minima, and points of inflection of the following functions. Graph them and determine the regions of increase and decrease, and of convexity and concavity: (a) 3 — 6x +2, (6) 21 — 2), © Tye On (©) sin?x. 2. Which point of the hyperbola y? — $x? = 1 is nearest to the point 0,y =3? 3. Let P be a fixed point with coordinates 2, yp in the first quadrant of a rectangular coordinate system. Find the equation of the line through P such that the length intercepted between the axes is a minimum. 4. A statue 12 ft high stands on a pillar 15 ft high. At what distance must a man whose eyes are at an elevation of 6 feet be in order that the statue may subtend the greatest possible angle at his eye? 5. Two sources of light, of intensities a and 5, are at a distance d apart. At which point of the line joining them is the illumination least? (Assume that the illumination is proportional to the intensity and inversely proportional to the square of the distance.)134 Sec. 3.6: Problems 6. Of all rectangles with a given area, find (a) The one with the smallest perimeter. (6) The one with the shortest diagonal. 7. In the ellipse x*/a? + 27/6? = 1 enscribe the rectangle of greatest area. 8. Two sides of a triangle are a and 6. Determine the third side so that the area is a maximum. 9. A circle of radius r is divided into two segments by a line g at a distance A from the center. In the smaller of these segments inscribe the rectangle of greatest possible area. 10. Of all circular cylinders with a given volume, find the one with the least area. 11. A dairy farmer wishes to enclose a rectangular pasture using 3000 ft of available fencing for three sides and an already existent 1000 ft stone wall for the fourth side. How can this be done so as to maximize the area? 12. A swimmer wishes to reach a particular point along the shore in the least possible time. Assuming the shore to be straight and that he swims at constant speed, then runs along shore at a faster constant speed, find the path of least time (discuss all possible cases). 13. For what bases a can the inequality x < log, z be satisfied? PROBLEMS SECTION 3.6, page 236 1. Determine the maxima, minima, and points of inflection of x3 + 3px + q. Discuss the nature of the roots of z* + 3px + q = 0. 2. Given the parabola y® = 2pz, p > 0, and a point P(x = &, y =») within it (7? < 2pé), find the shortest path (consisting of two line segments) leading from P to a point Q on the parabola and then to the focus F(x = 4p, y = 0) of the parabola, Show that the angle FOP is bisected by the normal to the parabola, and that QP is parallel to the axis of the parabola (principle of the parabolic mirror). 3. Among all triangles with given base and given vertex angle, the isosceles triangle has the maximum area. 4. Among all triangles with given base and given area, the isosceles triangle has the maximum vertex angle. *5. Among all triangles with given area, the equilateral triangle has the least perimeter. *6. Among all triangles with given perimeter the equilateral triangle has the maximum area.Ch. 3: Techniques of the Calculus 135 *7. Among all triangles inscribed in a circle the equilateral triangle has the maximum area. 8. Prove that if p > 1 and x > 0, x”? — 1 > p(x — 1). 9. Prove the inequality 1 > (sin 2)/z > 2/n,0 2,0 <2 < 2/2. (6) cos > 1 = 2/2. *11. Given a, > 0, a, > 0,..., a, > 0, determine the minimum of ate +a, +2 n Wayds >> Ay yt for « > 0. Use the result to prove by mathematical induction that (cf. Problem 13, page 109) LEG ag eet aa im pr’ n 12. (a) Given n fixed numbers a,,..., 4, determine x so that > (a; — x)? is a minimum. i= *(6) Minimize ¥ Ja; — 2|. a *(c) Minimize > A, |a; — |, where 2; > 0. ia 13. Sketch the graph of the function y =)", yO) = 1. Show that the function is continuous at x = 0. Has the function maxima, minima, or points of inflection? *14, Find the least value « such that 1\r+2 (143) >e 2, for all positive x. (Hint: It is known that [1 + (1/x)}*1 decreases mono- tonically and [1 + (1/x)¥ increases monotonically to the limit e at infinity.) *15. (a) Find the point such that the sum of the distances to the three sides of a triangle is a minimum. (6) Find the point for which the sum of the distances to the vertices is a minimum. 16. Prove the following inequalities: (a) > 11 +2), >0. (b) e* > 1 + log(1 +2), > 0. (c) & >1+(1 +2) log (1 +2),2 >0.136 Sec. 3.6: Problems 17. Suppose f(x) <0 on (a, 6). Prove: (a) Every arc of the graph within the interval lies above the chord joining its end points. () The graph lies below the tangent at any point within (a, 6). *18, Let f be a function possessing a second derivative on (a, 6). (a) Show that either condition a or 6 of Problem 17 is sufficient for f'@) <0. (6) Show that the condition 1) 2 fu) for all x and y in (a, 6) is sufficient for f“(z) <0. *19. Let a, bbe two positive numbers, p and g any nonzero numbers p 0,0 <@ <1, without reference to Jensen’s inequality, and show that equality holds only if @ = 6. (This inequality states that the 6, 1 — 6 geo- metric mean is less than the corresponding arithmetic mean.) *24, Let f be continuous and positive on [a, 6] and let M denote its maxi- mum value. Prove M = lim /f Yer de. noCh. 3: Techniques of the Calculus 137 Answers to Exercises SECTION 3.6, page 236 1. (a) Maximum for « = — V2, minimum for x = V2, inflection for =0. (6) Maximum for « = 0, inflection for x = —3. (c) Maximum for x = 1, minimum for 2 = —1, inflection for z = 0, £V3. (d) Maximum for z = 13, minimum for « = — 3, inflection for 2 =0,4V6 4 V33. (e) Maximum for = (n +})x, minimum for x = am, inflection for 2n+1 4 2. The point (0, 1). 3, Equation of line is (y — ya)/(z — %9) = — VWlea)- 4. -V189 ft. 5. The point dividing the line ab in the ratio Va: 75. 6. The square. 7. The rectangle with corners « = +a/V2,y = +6/V2. 8. The right-angled triangle, that is c? = a? + 5°. 9. The side of rectangle opposite to g must be at the distance }{ V(8r? + 4?) + A} from the center. 10. The cylinder whose height is equal to the diameter of its base. 11. Take each side equal to 1000 ft. 12. Suppose a is the swimming speed and 6 the running speed where a < b. Let P be the ultimate destination of the swimmer and let 0 be the nearest point on shore. If OP subtends an angle ¢ at his eye and sin ¢ < a/b, least time is consumed by the straight path to P. If sin g > a/b, the path of least time is a broken path whose first segment extends to the point Q along the shore for which OQ subtends the angle arc sin a/b, at the swimmer’s eye, thence along the shore to P. 13. If a <1, the inequality can clearly be satisfied by taking x =a. If a > 0, we must satisfy the condition that the minimum of x — log, z <0. The minimum occurs at x = I/loga. Take a —4p%. 2. We have for the total distance 2 2 “4 2 2 % Le) = lg = ) +u- | + [Eg -4) +e y “Typ -[(E-J+o-] +(£ +5]. 1. An inflection at x (See figure.) OF F=(,0) N=(p+x,0) For L to be a minimum, Slo =F] Fy 7 ; a noe y 1 O = pp Hu = tp =” The equation is clearly satisfied by y = 7. Let 6 and ¢ be the angles made by the segments PQ and FQ, respectively, with the normal QN to the parabola. Note that Pap Ly) = SFE (sin + sn 9 Thus, at a minimum, 6 = ¢. Furthermore, it is easily verified that triangle QFN is isosceles, |OF| = |FN| =x + p/2 so that y = 6 and PQ is again seen to be parallel to the z-axis, 3. Let a denote the base, 2, y the legs of the triangle and 6, &, 7 the corre- sponding opposite angles, From the law of sines we have __asing “sind ’ asin y= sin 6 *Ch: 3: Techniques of the Calculus 139 Consequently, the area A is given by sin sin sin 0 1 asin £ sin (0 + 8) 2 sin 0 2 where we have employed £ + 7 + 4 = 7. The derivative with respect to é is A’ = a® sin (2 + 6)/2 sin 8, from which it follows that the minimum occurs at § = }(7 — 6) and hence that the triangle is isosceles. 4. If the base and area are fixed, then the altitude also is fixed. Let (a, 0) and (—a, 0) be the end points of the base and (a, A) the vertex. If 8 is the vertex angle then w+ — cos 8 = @2)8 + 2h(a? + x) + Ht v@ and cos @ can then be shown to have a minimum (hence 9, a maximum) at a =0. 5. The proof employs the result of Example | in the text. Let T be the triangle of given area and least perimeter, and let b be any side of it. Then, keeping 6 fixed, T must be the triangle of given base 6 and given area having the least perimeter. Hence 7 must be isosceles, and the two sides of T other than b are equal to one another. But b is any side, and T is therefore equi- lateral. 6. First show as claimed in the text that of all triangles with a given perimeter and a given base the isosceles triangle has the greatest area. For this purpose, in the notation of Problem 4, let be the length of the side joining the vertices (—a, 0) and (x, h), and v the length of the side joining (a, 0) and (z, h). The length « + v = 2b is fixed and A =ah =aV2 —(a@—a2 =aVe —G@4+ay. Eliminate u and v to obtain A =avV@— a1 — 2/6), from which it is easily verified that A has a maximum at x = 0. Finally, if 7 is a triangle of maximum area for a fixed parimeter, then T must be equilateral, for if it has base b, then by the preceding argument the other two sides must be equal. Since 6 may be any side, the triangle is equilateral. 7. Given a fixed base, the triangles inscribed in a circle have a constant vertical angle. Apply the result of Problem 3. 8. The function f(z) = (2? — 1) — p( — 1) has a unique minimum at v=,140 Sec. 3.6: Solutions and Hints to Problems 9. Find the minima of x —sinzand sinx —(2/z)x in the interval 0 f(0), which yields the result. (b) Set f(z) = cos x + 2/2 — 1. From Problem 9, f(z) = 2 —sinz 20 for 0 << 7/2, and the inequality is obviously valid for x > n/2 > 1. Consequently, f(z) > f(0) for all x. 11. The minimum is (x +a, ++ +a, : ym n-1 Waa, —v Set A, =- 5 Bae G, = Vaya, if Ay 2a" For » = 1, G, < A. Furthermore, * then by the result just obtained a> (ee) "> (ee) -» 1 12. (a) 2a da, & vb (b) Take a; < a2 < +++ a, — ay and that the value a, — a, is taken when a, 0, then =a, yields the minimum.Ch. 3: Techniques of the Calculus 141 13. Show that lim x log x? = 0; hence that lim (z2)* = lim e#!8=* 2-0 2-0 20 Maximum at x = —1/e, minimum at x = 1/e, inflections for x = 0, and for the value of x satisfying (2 + log 2%)® + 2/x = 0. 14, We have lim (1 + 1/z*#* =e. For f(z) = (1 + 1/2)*** we have a0 f@) =fOg@, where g(x) = log (I + 1/z) — (@ + a)/(x? +2). At infinity, g(2) = 0. Since “ Qa -lr +a &@=-—Gya? we see that g’(x) > 0 and g(x) increases to 0 at infinity if « > }, and decreases to Oif « <4. It follows that f(x) decreases to e for « > 4, and if « < } that f(z) has values less than e for large x. 15. (a) Let the vertices of the triangle be located at P = (a, 6), Q = (—c, 0), and R = (c, 0), where 6 > O and c > 0. We also assume a > 0(if a < 0 we replace the triangle by its reflection in the y-axis). For any point (x, y) set p= br — (a — oy — be _ be —(a t+ Oy + be i VFe@cor Vea ee where p,q, and r denote the directed distances from the sides (extended if necessary) QR, RP, and PQ, with signs positive for (z, y) above, to the left and to the right of the respective sides. Restrict (x, y) to a line parallel to the base, y = constant. On this line minimize f(x) = |p| + Igl + |r|. Observe that on the line y = b the minimum occurs at P and if y > 6, the value of the minimum must be greater. It is therefore necessary to consider only y < b. We have f@ b = 5g 7 + ————— sgn gq. Ve +(a +c iid VB? + (a —c)* eng142 Sec. 3.6: Solutions and Hints to Problems It follows that the minimum on the line y = constant occurs on the line RP, that is, |g| = 0. The value of the minimum is bry =y + 2c-——————.. so) ey Ve + (a +c? If the side RP is greater than the base QR, the over-all minimum must then occur at the vertex R. If the side RP is less than the base, the minimum occurs at the vertex P. Thus the minimum falls on the vertex opposite the longest side of the triangle. If the two longer sides are equal in length, then the minimum is reached at any point of the shortest side. If the triangle is equi- lateral, the minimum is reached at any point within or on the triangle. B (6) Let the vertices be located at the points O, A = (a, 0), B = (bcos 0, b sin 8). We minimize the sum of the distances from a point P on a circle of radius r about O. The sum of the distances of P from the vertices is given as a function of the angle ¢ of inclination of OP: S(¢) =r + V(rcos $ = a)? + (rsin 6)? + V(rcos $ — bcos 6) + (rsin $ — bsin 8; hence aq) _arsing — brsin(@ — 4) LO =a pay If P is inside the triangle, then by the law of sines (see figure) Sf) = r(sin « — sin B) and the condition for a minimum on the circle is sin x = sin 6. Thus if P yields the over-all minimum, the angle subtended by two of the vertices at P is bisected by the line joining P to the third vertex. It follows that the angle subtended by the three sides at P are all equal and equal to 2/3.Ch: 3: Techniques of the Calculus 143 If any of the vertex angles A, B, or C is equal to or greater than 2z/3, then there is no interior minimum. Thus suppose C > 27/3. The minimum point P cannot be interior to the triangle. Furthermore, it cannot lie on the side of AB opposite to C since a lesser value would be given by its image in AC. If p =|PC| >, the line PC must then lie outside the angle subtended by AB at P. The only possibility left is that p = 0 or the minimum point is C itself. 16.(a) The derivative of f(x) = (1 + xe? is positive for x >0; conse- quently, fe) > f() = 1. (6) Differentiate and use (a). (c) Differentiate and use (6). 17. Essentially the same as Miscellaneous Problems, Chapter 2, Nos. 4 and 5. (a) The statement that the arc lies above the chord is expressed by fl9a + (1 — 66] > Of(%) + (1 — 0) f(8), where 0 <6 <1. Set w= 6x +(1 —6)8. Take « O for x > Consequently, g(z) has a strict minimum at x = , or g(x) > g() =0. 18. (a) Suppose « LO=fo ~fM , £8) -f@ pou” Ban * Taking the limits as € — « and - B we obtain for « < 8, f(@) > f(A). It follows that f"(x) < 0. Suppose that the arc lies below all its tangents; that is, for all f@ < f®™ +f Oe - ). By the mean value theorem S@) -{® =f (Me - §, for some 7 between z and &. Then f'(n) > f’(é) for x < € and f(n) < f'(é) for z > é. Taking lim fo-fo » we obtain f“(g) < 0. at (® Induction on » for n = 2° shows that (a, +a, + + ay) feted te), t a) + fla) +--+ + flan]. For natural numbers p,q with p +q = 2°, then Ae + #) > Pf@) + 4fW) pt+q P+q * Now let @ be any real number satisfying 0 < @ <1. Take p as the integer part of 62°, p = [62°], and g = 2" —p, and employ the continuity of f to derive the condition a of Problem 17, 19. To show that $(@) = [6a* + (1 — 6)b*}/* = w=Ch. 3: Techniques of the Calculus 145 is monotonically increasing, note on applying the result of 17a to the convex function u log u that yi-we 4°) = [6a log a* + (1 — 6)6* log 6 — wlogw] is nonnegative. 20. For g(u) =u log u we have strict inequality, ¢*(u) = u-! > 0, for positive x. It follows in the solution of Problem 19 that $x) = 0 only if a = 5. 21. Observe for the inequality proved in Problem 19 that the proof imposes no restriction on the sign of x other than x # 0. Consider log (6a? + (1 — 6)67] x a lim log ¢(z) = lim at 04 d = Flog [oat + (1 — 061, oy = Ologa +(1 — 4)logd. 22. Clear from the remarks above. 23. Since log x is a concave function, the arc lies above its chord. Con- sequently, log a%5!-* = 6 loga + (1 — 6) logs Slog [0a + (1 — 6)5). 24, In some subinterval [, f] of [4, 6] we have f(z) > M — «. It follows that fy (M — 9B — a) 0) (6) (logz)*, — (B > 0) (©) sinx, (a) sinh x, 2 eos? (e) 24 sin x - arc tan x, (f) 24 sine + = = . (A) -1, (i) log (x log x).146 Sec. 3.7: Answers to Exercises 2. Compare the functions of Exercise | with e”*, e*, and (log z)* for a > 0. 3. Compare the functions of Exercise 1 with powers of x as « + 0. 4.(a) What are the limits, as x + «0, of e(-*) and ee)? (6) Does the limit lim e#"e(-**) exist? zo 5. If ¢(x) + © as x + ~, show that log ¢(x) is of a lower order and e#(=) of a higher order of magnitude than (x). PROBLEMS SECTION 3.7, page 248 1. Let f(x) be a continuous function vanishing, together with its first derivative, for « =0. Show that f(x) vanishes to a higher order than zasx—+0 age” + az") +--+ +a, boe™ + BH + By, the same order of magnitude as 2"-™, when x -» co. 2, Show that f(z) = , when ag, by #0, is of *3. Prove that e* is not a rational function. *4, Prove that e* cannot satisfy an algebraic equation with polynomials in x as coefficients. 5. If the order of magnitude of the positive function f(x) as 7 + © is 2 higher, the same, or lower than that of 2™, prove that (8) d§ has the corresponding order of magnitude relative to x™*?, a 6. Compare the order of magnitude as x -> 0 orf SCS) dé relative to f(@) for the following functions /(x): e VE @ =. (b) e, z (c) xe**, (d) log x Answers to Exercises SECTION 3.7, page 248 1. (a) Higher than x”, (6) lower than 2¢, (c) 0(2*), (d) higher than 2%, (e), (f) not comparable, since the functions do not tend to infinity or to definite limits. (g) same as 2, (i) higher than x, (j) lower than x*.Ch. 3: Techniques of the Calculus 147 2, Higher than e, (log 2), lower than e%, higher, same, or lower than e" corresponding to positive, zero, or negative sign of 6 — a, (b) lower than e**, e2*, (c) bounded, (d) same as e*, lower than e*", higher than (log 2)", (e), (/), (g) lower than e**, e*", higher than (log x), (A) higher than e*"§, lower than e”*S, higher than e%, (log 2), (j) same as log x, lower than e%, e2", 1\ 3. (a) Same as x, (b) lower than eB) , (¢)same as z, (d) same as x, (e) same as 2°, (f) same as «*, (g) higher than 2, (h) higher than 2!~€, lower than z, 1\c (j) lower than () ; 2, 4.(a) 0; 1. (6) Yes; 0. Solutions and Hints to Problems SECTION 3.7, page 248 1 tim 2 = f'@) =0. ano % FO) _ % +alz + om by + ble + hence 2 3. e* has higher order of magnitude at o than any power, hence the result of Problem 2 could not be satisfied. a 4. If > pie = 0, then io nd Pala) = — D pr(zde, 0 Consequently, lim p,(z) = 0. It follows that the polynomial p,(z) is identi- cally zero, and sequentially that p,_(2), py_o(2), ... , po(2) are all identically zero. 5. If f(z) has the same order of magnitude as does 2”, then Be™ < f(a) < Cx™,148 Sec. 3.8: Problems It follows that = c F(x) -[ fede <5 Sele 1 x (emt — gmt) 2C mit "aa for « > |al. Similarly B lal\m+¥ B mi [ — (lat BG FO)? a7 [! (2) ? m+)” for x > 2al. Comparisons of the same type establish the other two cases, For example, if f(z) has lower order than 2", then given any positive «, there exists a («) such that f(z) < «x™ whenever 7 > b(«). We have FO) =F) + fre de K < Fb) + fe dz; . hence m+) 2h — gett 40 Fe) © of a, iets t *2. Find the limit ofCh. 3: Techniques of the Calculus 149 *3, If « is any real number greater than —1, evaluate 1S 204 3th 2 spat lim aa 3 Solutions and Hints to Problems SECTION 3.8, page 264 1. The number a, is the lower sum for 1/z in the interval (1, 2] for a sub- division into n equal parts (cf. Problems S.2, No. 1). an — log 2. 1 2. Compare with = 1 3th EXERCISES SECTION 3.9, page 264 Evaluate the following integrals and verify the results by differentiation. 1. fretae. 2. foe dz, 3. fe Vig@dr. 4. 108? ae dz . |. fev: . 5. lam 6 |e ‘z+ Ide 1. fevew + bdr. 8. faces +a) de. 62 at %. |r 10. fi — de. _—— larctanx =a 1. feo a) de. wf Tear 4 i xde 13. J cos" x sin x dx. lo og 15. aaa Jo V1 +32?) ° a8 16. f dz (1 1 Ba 19. 41 + e)4 — 401 + 7% 20, 1 1 . = arc COS = « on er) @ SECTION 3.10, page 271 Ch, 3: Techniques of the Calculus 151 a-l 6-1" EXERCISES Evaluate the following integrals and verify the results by differentiation: x f 3de * Son? — 6c + 2° a+] “Jvi-2 3. a “Sv3-2- 1 f adx “J vit—4e +1" 9. da * Jat+a +1" 11. a * Ja? + 2ax + b° 13. si cos de, zt+l 15. ———— dr. Vi¢zt 2 f dz “Svea +5) f dz 4 |———.. V5 +2x+22 adx =. —aztl @ +1)de “JvV242 —32" 40 de * Jatoztl” 12, foo x cos! x dx.152 Sec. 3.10: Answers to Exercises Sg ® wn. [ . us. f ae 0 Va? + a? 02? + a? 1 1 19. evi — ade, 20. 21 — 22% de. 0 Answers to Exercises SECTION 3.10, page 271 1. arc tan 3x — 1), of or ine ar si 5 3. arc sinw —V1 — 2+ ‘a + 1\? 4. tog [> +Jr+( =) | or etl ar sinh atl 8. are sin. 2-1 v3 1 1 6. ~ log (x? — x + 1) + —0; — if 1 ata - arte ag ee aeCh, 3: Techniques of the Calculus 153 cos’ x ui cos? x 5 po 12. — sinks sin’ | sin? 3." 2G 14, farcsine — }evi 15, VI +2 + arsinhz. 16. jeV1 +2? — Jarsinh2. 17. ar sinh 1 = log (1 + V2). 1 2 18, -arctan=. a a 19.5. 20. = a. EXERCISES SECTION 3.11, page 274 Evaluate the integrals in Exercises 1-14. x COS x a gan fa 3. f x* cos x dz. 4. foew dz, 5. / 2? cos nz dz (n a positive integer). 6. / 22 sin nx dz (n a positive integer). 7 fe cos 2? de, 8. fantede 9. footed 10. few aah de, 1 i. free. 2. feo ¥1). 13. fon log x dx (m # 1). 14. f (log 2)? de.154 Sec. 3.11: Problems 15, Prove the formula feves dx = e[p(z) — p(x) + p@) —+-- 1, where p(2) is any polynomial. Evaluate the integrals in Exercises 16-19. 2/2 0/6 16. [ cos" x dr. 17. f cos? 30 sin* 66 49, . 0 cating 1 game w. [S. w. fo ual) jo VI = 2? jo V1 — 2? Obtain recurrence formulas for the integrals of Exercises 20-22. 20. forces x)” de, 21. foes sin bx dx, 22. | 2" cos bx dr, Evaluate the integrals of Exercises 23 and 24. 23. Je sinh bx dx, 24. fe cosh bx dx. 1 25. Evaluate} 2%e~** cos 2x dz. 1 PROBLEMS SECTION 3.11, page 274 1. Show that for all odd positive values of n the integral f e~**2" dz can be evaluated in terms of elementary functions. 2. Show that if n is even, the integral f e~* '<" dx can be evaluated in terms of elementary functions and the integral f e-** de (for which tables have been constructed). 3. Prove that {[fro aia . f is Supe *4, Problem 3 gives a formula for the second iterated integral. Prove that the nth iterated integral of f(x) is given by * pudle = wr} du. (mn —1)Ch. 3: Techniques of the Calculus 155 5. Prove for the binomial coefficient (;) that a 1 a ( ) = [« + vf ah — ade] : k. 0 6. Obtain a recursive formula for force +b) de and use this relation to integrate foc + 1tdz. a 2enl do" *7. (a) Let P,(z) = (@ — 1)". Show that a f P,(2)P p(x) dx =0, if mn. 1 x 2 (6) Prove that fiptcoae Sore 1 (c) Prove that J a™P,(x) dx =0,if m

You might also like

• Hourglass Workout Program by Luisagiuliet 2

  

  76% (21)

  Hourglass Workout Program by Luisagiuliet 2

  51 pages
• 12 Week Program: Summer Body Starts Now

  

  87% (46)

  12 Week Program: Summer Body Starts Now

  70 pages
• Read People Like A Book by Patrick King-Edited

  

  57% (82)

  Read People Like A Book by Patrick King-Edited

  12 pages
• Livingood, Blake - Livingood Daily Your 21-Day Guide To Experience Real Health

  

  77% (13)

  Livingood, Blake - Livingood Daily Your 21-Day Guide To Experience Real Health

  260 pages
• Cheat Code To The Universe

  

  94% (79)

  Cheat Code To The Universe

  34 pages
• Facial Gains Guide (001 081)

  

  91% (45)

  Facial Gains Guide (001 081)

  81 pages
• Curse of Strahd

  

  95% (467)

  Curse of Strahd

  258 pages
• The Psychiatric Interview - Daniel Carlat

  

  91% (34)

  The Psychiatric Interview - Daniel Carlat

  473 pages
• The Borax Conspiracy

  

  91% (57)

  The Borax Conspiracy

  14 pages
• The Secret Language of Attraction

  

  86% (108)

  The Secret Language of Attraction

  278 pages
• How To Develop and Write A Grant Proposal

  

  83% (542)

  How To Develop and Write A Grant Proposal

  17 pages
• Penis Enlargement Secret

  

  60% (124)

  Penis Enlargement Secret

  12 pages
• Workbook For The Body Keeps The Score

  

  89% (53)

  Workbook For The Body Keeps The Score

  111 pages
• Donald Trump & Jeffrey Epstein Rape Lawsuit and Affidavits

  

  83% (1016)

  Donald Trump & Jeffrey Epstein Rape Lawsuit and Affidavits

  13 pages
• KamaSutra Positions

  

  78% (69)

  KamaSutra Positions

  55 pages
• 7 Hermetic Principles

  

  93% (30)

  7 Hermetic Principles

  3 pages
• 27 Feedback Mechanisms Pogil Key

  

  77% (13)

  27 Feedback Mechanisms Pogil Key

  6 pages
• Frank Hammond - List of Demons

  

  92% (92)

  Frank Hammond - List of Demons

  3 pages
• Phone Codes

  

  79% (28)

  Phone Codes

  5 pages
• 36 Questions That Lead To Love

  

  91% (35)

  36 Questions That Lead To Love

  3 pages
• How 2 Setup Trust

  

  97% (307)

  How 2 Setup Trust

  3 pages
• 100 Questions To Ask Your Partner

  

  78% (36)

  100 Questions To Ask Your Partner

  2 pages
• The 36 Questions That Lead To Love - The New York Times

  

  91% (35)

  The 36 Questions That Lead To Love - The New York Times

  3 pages
• Satanic Calendar

  

  25% (56)

  Satanic Calendar

  4 pages
• The 36 Questions That Lead To Love - The New York Times

  

  95% (21)

  The 36 Questions That Lead To Love - The New York Times

  3 pages
• 14 Easiest & Hardest Muscles To Build (Ranked With Solutions)

  

  100% (8)

  14 Easiest & Hardest Muscles To Build (Ranked With Solutions)

  27 pages
• Jeffrey Epstein39s Little Black Book Unredacted PDF

  

  75% (12)

  Jeffrey Epstein39s Little Black Book Unredacted PDF

  95 pages
• 1001 Songs

  

  70% (73)

  1001 Songs

  1,798 pages
• The 4 Hour Workweek, Expanded and Updated by Timothy Ferriss - Excerpt

  

  23% (954)

  The 4 Hour Workweek, Expanded and Updated by Timothy Ferriss - Excerpt

  38 pages
• Zodiac Sign & Their Most Common Addictions

  

  63% (30)

  Zodiac Sign & Their Most Common Addictions

  9 pages
• Chapter4s Unlocked

  

  No ratings yet

  Chapter4s Unlocked

  32 pages
• Marek Capinski, Tomasz Jerzy Zastawniak - Probability Through Problems (2000, Springer) PDF

  

  No ratings yet

  Marek Capinski, Tomasz Jerzy Zastawniak - Probability Through Problems (2000, Springer) PDF

  133 pages
• m337 1 Workbook

  

  No ratings yet

  m337 1 Workbook

  39 pages
• (Lecture Notes) Andrei Jorza-Math 5c - Introduction To Abstract Algebra, Spring 2012-2013 - Solutions To Some Problems in Dummit & Foote (2013)

  

  No ratings yet

  (Lecture Notes) Andrei Jorza-Math 5c - Introduction To Abstract Algebra, Spring 2012-2013 - Solutions To Some Problems in Dummit & Foote (2013)

  30 pages
• Set Theory HT20

  

  No ratings yet

  Set Theory HT20

  52 pages
• How To Learn Math and Physics: John Baez January 2, 2016

  

  No ratings yet

  How To Learn Math and Physics: John Baez January 2, 2016

  12 pages
• Proof, Logic, and Conjecture - The Mathematician's Toolbox

  

  100% (2)

  Proof, Logic, and Conjecture - The Mathematician's Toolbox

  481 pages
• Mathematics Reading List Cambridge

  

  No ratings yet

  Mathematics Reading List Cambridge

  8 pages
• SHURMAN Multivariable Calculus

  

  No ratings yet

  SHURMAN Multivariable Calculus

  523 pages
• Great Thinkers, Great Theorems Starter

  

  100% (1)

  Great Thinkers, Great Theorems Starter

  3 pages
• Algebra I Honors

  

  No ratings yet

  Algebra I Honors

  47 pages
• Differential and Integral Calculus (Edmund Landau)

  

  No ratings yet

  Differential and Integral Calculus (Edmund Landau)

  377 pages
• Conic Section PDF

  

  No ratings yet

  Conic Section PDF

  78 pages
• Mathematical Proofs by Examples

  

  100% (1)

  Mathematical Proofs by Examples

  6 pages
• Numbers, Sets and Axioms - Hamilton PDF

  

  83% (6)

  Numbers, Sets and Axioms - Hamilton PDF

  265 pages
• MAT157 Term Test U of T 2014

  

  No ratings yet

  MAT157 Term Test U of T 2014

  2 pages
• 5 - Vector Calculus PDF

  

  No ratings yet

  5 - Vector Calculus PDF

  87 pages
• Integration in Finite Terms - Maxwell Rosenlicht

  

  100% (1)

  Integration in Finite Terms - Maxwell Rosenlicht

  11 pages
• (Golden Maths Series) N.P. Bali-Real Analysis-FirlNknlknewall Media (2005)

  

  100% (1)

  (Golden Maths Series) N.P. Bali-Real Analysis-FirlNknlknewall Media (2005)

  420 pages
• Madhava MC Paper 12

  

  No ratings yet

  Madhava MC Paper 12

  2 pages
• Algebra Part II by G Chrystal

  

  100% (2)

  Algebra Part II by G Chrystal

  640 pages
• Functions of Several Variables2

  

  No ratings yet

  Functions of Several Variables2

  5 pages
• Cauchy-Schwarz Inequality

  

  No ratings yet

  Cauchy-Schwarz Inequality

  9 pages
• Burkill A Second Course in Mathematical Analysis PDF

  

  No ratings yet

  Burkill A Second Course in Mathematical Analysis PDF

  268 pages
• Before Calculus: September 30, 2011 18:19 Ftoc Sheet Number 1 Page Number Xiii

  

  No ratings yet

  Before Calculus: September 30, 2011 18:19 Ftoc Sheet Number 1 Page Number Xiii

  5 pages
• Text Book of Conic Section, Sharma PDF

  

  No ratings yet

  Text Book of Conic Section, Sharma PDF

  152 pages
• Section 1.1

  

  100% (1)

  Section 1.1

  519 pages
• A Second Course in Elementary Differential Equations PDF

  

  No ratings yet

  A Second Course in Elementary Differential Equations PDF

  201 pages
• Integer Sequences and Circle Chains Inside A Hyperbola

  

  No ratings yet

  Integer Sequences and Circle Chains Inside A Hyperbola

  6 pages
• APEXPreCalculus PDF

  

  No ratings yet

  APEXPreCalculus PDF

  164 pages
• Continued Fractions

  

  No ratings yet

  Continued Fractions

  8 pages
• Calculus - A Complete Course

  

  100% (3)

  Calculus - A Complete Course

  1,116 pages
• Proofsss

  

  No ratings yet

  Proofsss

  30 pages
• Linear Algebra: Ho Ffman & Kunze

  

  0% (1)

  Linear Algebra: Ho Ffman & Kunze

  83 pages
• Multivariable Mathematics - Theodore Shifrin - Linear Algenra, Multivariable Calculus & Manifolds

  

  No ratings yet

  Multivariable Mathematics - Theodore Shifrin - Linear Algenra, Multivariable Calculus & Manifolds

  505 pages
• Teach Yourself Logic 2016

  

  No ratings yet

  Teach Yourself Logic 2016

  94 pages
• Morris Tenenbaum, Harry Pollard Ordinary Differential Equations Dover Books On Mathematics 1985

  

  100% (10)

  Morris Tenenbaum, Harry Pollard Ordinary Differential Equations Dover Books On Mathematics 1985

  819 pages
• (BROOKS) CalcLabs With Mathematica For Stewart's Multivariable Calculus-Brooks Cole (2007)

  

  No ratings yet

  (BROOKS) CalcLabs With Mathematica For Stewart's Multivariable Calculus-Brooks Cole (2007)

  292 pages
• Examples of Proofs

  

  No ratings yet

  Examples of Proofs

  4 pages
• FFTContent PDF

  

  No ratings yet

  FFTContent PDF

  8 pages
• ComplexAnalysis Notes

  

  100% (2)

  ComplexAnalysis Notes

  119 pages
• Elementary Statistics Triola 11th Edition Test Bank all chapter instant download

  

  100% (24)

  Elementary Statistics Triola 11th Edition Test Bank all chapter instant download

  31 pages
• Munkres J Analysis On Manifolds

  

  No ratings yet

  Munkres J Analysis On Manifolds

  190 pages
• (Silberberg, E.) The Le Chatelier Principle As A Corollary To A Generalized Envelope Theorem

  

  No ratings yet

  (Silberberg, E.) The Le Chatelier Principle As A Corollary To A Generalized Envelope Theorem

  10 pages
• Spiegel M.R. Real Variables, Lebesque Measure With Applications To Fourier Series 1990

  

  100% (3)

  Spiegel M.R. Real Variables, Lebesque Measure With Applications To Fourier Series 1990

  201 pages
• Hille E. - Analytic Function Theory. Volume 2 (2002) PDF

  

  No ratings yet

  Hille E. - Analytic Function Theory. Volume 2 (2002) PDF

  511 pages
• Line Integrals

  

  No ratings yet

  Line Integrals

  33 pages
• Theory of Equations Vol. II Contents Only by Burnside & Panton

  

  No ratings yet

  Theory of Equations Vol. II Contents Only by Burnside & Panton

  12 pages
• ST102 Michaelmas Term Revision LSE London School of Economics

  

  No ratings yet

  ST102 Michaelmas Term Revision LSE London School of Economics

  40 pages
• IARI PH.D - Water Science &

  

  No ratings yet

  IARI PH.D - Water Science &

  13 pages
• Ets 2022 - RC 7

  

  No ratings yet

  Ets 2022 - RC 7

  29 pages
• Estatica I

  

  No ratings yet

  Estatica I

  6 pages
• Basic Ceremonial Music FSW (No Score)

  

  No ratings yet

  Basic Ceremonial Music FSW (No Score)

  121 pages
• Vue en Eclaté Alpine A310

  

  No ratings yet

  Vue en Eclaté Alpine A310

  251 pages
• Old QP

  

  No ratings yet

  Old QP

  22 pages
• Telugu Bhasha Charitra by Bhadriraju Krishnamurti

  

  60% (5)

  Telugu Bhasha Charitra by Bhadriraju Krishnamurti

  267 pages
• Tarot Marsella Hadar PDF

  

  100% (1)

  Tarot Marsella Hadar PDF

  22 pages
• 2024 0239 P Signed Investigation - Redacted

  

  No ratings yet

  2024 0239 P Signed Investigation - Redacted

  31 pages
• Dcs Dapil 1 Terbaru

  

  No ratings yet

  Dcs Dapil 1 Terbaru

  46 pages
• Geronimo Stilton Journey Through Time 6 - Geronimo Stilton

  

  100% (7)

  Geronimo Stilton Journey Through Time 6 - Geronimo Stilton

  636 pages