Sie sind auf Seite 1von 269
A eS) <=> oo Laplace transforms applications completely explained Works with all major texts 450 fully solved problems Perfect for brushup or exam prep Use with these courses: 2 Operational Calculus [Electrical Engineering ( Mechanics | College Mathematics SCHAUM’S OUTLINE OF THEORY AND PROBLEMS or LAPLACE TRANSFORMS MURRAY R. SPIEGEL, Ph.D. Former Professor and Chairman, Mathematics Department Rensselaer Polytechnic Institue Hartford Graduate Center SCHAUM’S OUTLINE SERIES McGRAW-HILL New York San Francisco Washington, D.C. Auckland Bogoté Curacas Lisbon Londen Madrid Mexico City Milan Montreal New Dethi San Juan Singapore Sydney Yokyo Toronto Copyright © 1965 by McGraw-Hill. Ine. All Rights Reserved. Printed in the United States of America. No part of this publication may be reproduced, stored in a retrioval-system, or transmitted, In any form or by any means, lectronie, mechanteal, photocopying, recording, or otherwise, without the prior written permission of the publisher 60281 1213 14 15 SH SH 754321069 Preface ‘The theory of Laplace transforms or Laplace transformation, also referred to as operational calculus, has in recent years become an essential part of the mathematical background required of engineers, physicists, mathematicians and other scientists. This is because, in addition to being of great theoretical interest in itself, Laplace transform methods provide easy and effective means for the solution of many problems arising in various fields of science and engineering. The subject originated in attempts to justify rigorously certain “operational rules” used by Heaviside in the latter part of the 19th century for solving equations in electro- magnetic theory. These attempts finally proved successful in the early part of the 20th century through the efforts of Bromwich, Carson, van der Pol and other mathematicians who employed complex variable theory. This book is designed for use as a supplement to all current standard texts or as a textbook for a formal course in Laplace transform theory and applications. It should also be of considerable value to those taking courses in mathematics, physics, electrical engi- neering, mechanics, heat flow or any of the numerous other fields in which Laplace transform methods are employed. Each chapter begins with a clear statement of pertinent definitions, principles and theorems together with illustrative and other descriptive material. This is followed by graded sets of solved and supplementary problems. The solved problems serve to illustrate and amplify the theory, bring into sharp focus those fine points without which the student continually feels himself on unsafe ground, and provide the repetition of basic principles so vital to effective learning. Numerous proofs of theorems and derivations of formulas are included among the solved problems. The large number of supplementary problems with answers serve as a complete review of the material in each chapter. Topics covered include the properties of Laplace transforms and inverse Laplace transforms together with applications to ordinary and partial differential equations, integral equations, difference equations and boundary-value problems. The theory using complex variables is not treated until the last half of the book. ‘This is done, first, so that the student may comprehend and appreciate more fully the theory, and the power, of the complex inversion formula and, second, to meet the needs of those who wish only an introduction to the subject. Chapters on complex variable theory and Fourier series and integrals, which are important in a discussion of the complex inversion formula, have been included for the benefit of those unfamiliar with these topics. Considerably more material has been included here than can be covered in most first courses. This has been done to make the book more flexible, to provide a more useful book of reference and to stimulate further interest in the topics. I wish to take this opportunity to thank the staff of the Schaum Publishing Company for their splendid cooperation. M. R. SPIEGEL Rensselaer Polytechnic Institute January, 1965 CONTENTS THE LAPLACE TRANSFORM .. ae Definition of the Laplace transform. Notation. Laplace transforms of some elementary functions. Sectional or piecewise continuity. Functions of ex- ponential order. Sufficient conditions for existence of Laplace transforms. Some important properties of Laplace transforms. Linearity property. Firs translation or shifting property. Second translation or shifting property. Change of scale property. Laplace transform of derivatives. Laplace trans- form of integrals. Multiplication by t*. Division by & Periodic functions. Behavior of f(e) as ¢—®. Initial-value theorem. Final-value theorem. Generalization of initial-value theorem. Generalization of final-value theorem. Methods of finding Laplace transforms. Direct method. Scries method. ‘Method of differential equations. Differentiation with respect to a parameter. Miscellaneous methods. Use of Tables. Evaluation of integrals. Some special functions. The gamma function. Bessel functions. The error function, The complementary error function. Sine and cosine integrals. Exponential integral. Unit step function. Unit impulse or Dirac delta function, Null fonctions. Laplace transforms of special functions. THE INVERSE LAPLACE TRANSFORM . seeeeeeeees 42 Definition of inverse Laplace transform, Uniqueness of inverse Laplace tran: forms. Lerch’s theorem. Some inverse Laplace transforms. Some important properties of inverse Laplace transforms. Linearity property. First transla- tion or shifting property. Second translation or shifting property. Change of scale property. Inverse Laplace transform of derivatives. Inverse Laplace transform of integrals. Multiplication by #", Division by s. The convolution property, Methods of finding inverse Laplace transforms, Partial fractions method, Series methods, Method of differential equations. Differentiation with respect toa parameter. Miscellaneous methods. Use of Tables. The complex inversion formula. ‘The Heaviside expansion formula. The beta function, Evaluation of integrals. Chapter 3 APPLICATIONS TO DIFFERENTIAL EQUATIONS. 8 Ordinary differential equations with constant coefficients. Ordinary differen- tial equations with variable coefficients, Simultaneous ordinary differential equations. Applications to mechanics. Applications to electrical circuits, Applications to beams. Partial differential equations. APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS : Integral equations. Integral equations of convolution type. Abel equation. The tautochrone problem. Integro-differential equations. Difference equations. Differential-difference equations. 112 Chapter 5 COMPLEX VARIABLE THEORY .. 136 The complex number system. Polar form of complex numbers. Operations in polar form. De Moivre's theorem. Roots of complex numbers. Function: Limits and continuity. Derivatives. Cauchy-Riemann equations. Line in- tegrals. Green’s theorem in the plane. Integrals. Cauchy's theorem. Cauchs’s integral formulas. Taylor's series. Singular points. Poles. Laurent’s series. Residues. Residue theorem. Evaluation of definite integrals, CONTENTS Chapter 6 FOURIER SERIES AND INTEGRALS 173 Fourier series. Odd and even functions, Half range Fourier sine and cosine series. Complex form of Fourier series, Parseval’s identity for Fourier series. Finite Fourier transforms. The Fourier integral. Complex form of Fourier integrals. Fourier transforms. Fourier sine and cosine transforms. The convolution theorem. Parseval’s identity for Fourier integrals, Relationship of Fourier and Laplace transforms. Chapter 7 THE COMPLEX INVERSION FORMULA ... ‘The complex inversion formula. The Bromwich contour. theorem in finding inverse Laplace transforms. A sufficient conditi the integral around T to approach zero. Modification of Bromwich contour in case of branch points. Case of infinitely many singularities. 201 Chapter 8 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS....... 219 Boundary-value problems involving partial differential equations. Some im- portant partial differential equat One dimensional heat conduction equa- tion, One dimensional wave equation. Longitudinal vibrations of a beam, Transverse vibrations of a beam, Heat conduction in a cylinder. ‘Transmission lines. Two and three dimensional problems. Solution of boundary-value problems by Laplace transforms. APPENDIX A. TABLE OF GENERAL PROPERTIES OF LAPLACE TRANSFORMS . a ; APPENDIX B. TABLE OF SPECIAL LAPLACE TRANSFORMS APPENDIX C. TABLE OF SPECIAL FUNCTIONS ... 255, 257 Chapter 1 DEFINITION OF THE LAPLACE TRANSFORM Let F(t) be a function of ¢ specified for ¢>0. Then the Laplace transform of F(t), denoted by £ (F(t)}, is defined by 1A) = 1 = Se#roat w where we assume at present that’the parameter s is real. Later it will be found useful to consider s complex. ‘The Laplace transform of F(t) is said to exist if the integral (1) converges for some value of s; otherwise it does not exist. For sufficient conditions under which the Laplace transform does exist, see Page 2. NOTATION \ If a function of ¢ is indicated in terms of a capital letter, such as F(t), G(t), Y(t), ete, the Laplace transform of the function is denoted by the corresponding lower case letter, ie. f(8), 9(8), (8), ete. In other cases, a tilde (~) can be used to denote the Laplace trans- form. Thus, for example, the Laplace transform of u(t) is %(s). LAPLACE TRANSFORMS OF SOME ELEMENTARY FUNCTIONS we Fo LAF) = fs) 1 1 2>0 oom 2 t 2>0 The adjacent table shows Laplace transforms of various elementary funetions. For de- tails of evaluation using defini- tion (1), see Problems 1 and For a more extensive table see Appendix B, Pages 245 to 254. 5. sinat 6. [ cosat 1. sinh at 1 cosh at Lt 2 THE LAPLACE TRANSFORM (CHAP.1 SECTIONAL OR PIECEWISE CONTINUITY A function is called sectionally continuous or piecewise continuous in an interval t38 if the interval can be subdivided into a finite number of intervals in each of which the function is continuous and kas finite right and left hand limits. Fo | r Fig. 1-1 An example of a function which is sectionally continuous is shown graphically in Fig. 1-1 above. This function has discontinuities at ti, t: and ts. Note that the right and left hand limits at te, for example, are represented by lim F(te+<) = Fltz+0) = F(te+) and lim F(t:~< = F(tz—0) = F(ts-) respectively, where « is positive. FUNCTIONS OF EXPONENTIAL ORDER If real constants M>0 and y exist such that for all t>N le R()| 0. Example 2. F(t) =e is not of exponential order since | e~v ¢!"! = e®-7' can be made larger than any given constant by increasing Intuitively, functions of exponential order cannot “grow” in absolute value more rapidly than Mer' as t increases. In practice, however, this is no restriction since M and y can be as large as desired. Bounded functions, such as sin at or cos at, are of exponential order. SUFFICIENT CONDITIONS FOR EXISTENCE OF LAPLACE TRANSFORMS Theorem 1-1. If F(t) is sectionally continuous in every finite interval 0St=N and of exponential order y for t>N, then its Laplace transform f(s) exists for all s > y. For a proof of this see Problem 47. It must be emphasized that the stated conditions are sufficient to guarantee the existence of the Laplace transform. If the conditions are not satisfied, however, the Laplace transform may or may not exist [see Problem 32]. ‘Thus the conditions are not necessary for the existence of the Laplace transform. For other sufficient conditions, see Problem 145. CHAP. 1] THE LAPLACE TRANSFORM 8 SOME IMPORTANT PROPERTIES OF LAPLACE TRANSFORMS In the following list of theorems we assume, unless otherwise stated, that all functions satisfy the conditions of Theorem 1-1 so that their Laplace transforms exist. 1. Linearity property. Theorem 1-2. If c: and cz are any constants while F,(f) and F,(t) are functions with Laplace transforms fi(e) and fe(s) respectively, then La Filt) + eel) = (F(t) + oe (Fo() = crfuls) + cafrls) (2) ‘The result is easily extended to more than two functions. Example. LEAP — Bcos2t + Bet) = 40 (2) — Be feo82t) + BL {em ar 1 *(@) ~ 8(@%) * #4) 8 _ Be weed ‘The symbol 2, which transforms F(t) into f(s), is often called the Laplace trans- formation operator. Because of the property of expressed in this theorem, we say that £ is a linear operator or that it has the linearity property. 5 eri + 2. First translation or shifting property. Theorem 1-3. If £(F(t)} = f(s) then LferF()} = f(s—a) (3) Example, Since <{cos2t} = 35-7, we have att -e = 8tt Ltd Kleseos2h = Gaara ~ wee Ts 8. Second translation or shifting property. Ft Theorem 1-4. If £{F(t)) = f(s) and G(t) = i a me then LGD) * f(s) @) Example. Since ¢(#) = &{ = §, the Laplace transform of the function _ ft-2 t>2 OM = ty eee is 6e~24/e4, 4. Change of seale property. Theorem 1-5. If £{F(t)} = f(s), then eqray) = 34(8) © Example. Since ¢{sint} = iy, we have 1 2 efsin3t} = warti ~ ers THE LAPLACE TRANSFORM (CHAP. 1 Laplace transform of derivatives. Theorem 1-6. It £(F(t)} = f(s), then (FD) = f(s) — FO) (6) if F(t) is continuous for 0=t=N and of exponential order for t>N while F’(?) is sectionally continuous for 05 ¢= Example. If F(t) = cost, then £{F(1)} = jeg and we have ==? (aia) -1 = Bs ‘The method is useful in finding Laplace transforms without integration [see Problem 15). LAF) = £{-3sin3t} = Theorem 1-7. If in Theorem 1-6, F(t) fails to be continuous at ¢=0 but lim F(t) = F(0+) exists [but is not equal to F(0), which may or may not exist}, then L{F()}) = sf(s) — FO+) (” Theorem 1-8. If in Theorem 1-6, F(t) fails to be continuous at t = a, then LIF} = sf(s) — F(0) — e-* {F(at) ~ F(a-)} (8) where F(a+) — F(a—) is sometimes called the jump at the discontinuity t=a. For more than one discontinuity, appropriate modifications can be made. Theorem 1-9. If £(F(t)} = f(s), then LF} = # f(s) — 8 FO) — FO) (9) if F(t) and F(t) are continuous for 0=t=N and of exponential order for t>N while F(t) is sectionally continuous for 0=t = N. If. F(t) and F’(t) have discontinuities, appropriate modification of (9) can be made as in Theorems 1-7 and 1-8. Theorem 1-10. lf £{F(t)} = f(s), then LAF} = ot f(0) — oUF (O) ~ 88 FO) — = 8M) — FP) (10) if F(t), F(t), ..., F°-%() are continuous for 0N while F(t) is sectionally continuous for 0 = t = N. Laplace transform of integrals. Theorem I-11. If £{F(t)} = f(s), then “lf Fe aul = we . (1) Pee ae a ee | AS, tua} = ara as can be verified directly. CHAP. 1) THE LAPLACE TRANSFORM 5 7. Multiplication by ¢*. Theorem 1-12. If £{F(t)) = f(s), then Le PY = cr Se = Cyrene) (12) Example. Since £ {e%) = ‘we have Lith = eee, = 8 Division by t. Theorem 1-13. If £{F(t)) = f(s), then {FO} = S fu) du (18) provided lim F(p/t exists. Example. Since ¢({sint}) = 1 " e+. tao t fen} S By = eta 9. Periodic functions. Theorem 1-14. Let F(t) have period T > 0 so that F(t +T) = F(t) [see Fig. 1-2). Sie F(t) dt Then 4EFO) = ae (14) Fo _Period T 1 U yn » CAS Fig.1-2 10. Behavior of f(s) as s> ©. Theorem 1-15. It £{F(t)) = f(s), then lim f(s) = 0 (15) 11. Initial-value theorem. Theorem 1-16. If the indicated limits exist, then lim F(t) = lim sf(s) (16) 6 THE LAPLACE TRANSFORM (CHAP.1 12 Final-value theorem. Theorem 1-17. If the indicated limits exist, then lim F() = lim sfle) (17) 18. Generalization of initial-value theorem. If lim F()/G() = 1, then we say that for values of ¢ near t= 0 (small t, F(t) is close to G(t) and we write F(t) ~ G(t) as t>0. Similarly if lim f(s)/o(s) = 1, then we say that for large values of s, f(s) is close to g(s) und we write f(s) ~ 9(s) as s> =. With this notation we have the following generalization of Theorem 1-16. Theorem 1-18. If F(t)~ G(t) as t>0, then f(s) ~ g(s) as s> © where f(s) = L(F()} and g(s) = £ (G(E)}- 14. Generalization of final-value theorem. If lim F(/G(t) = 1, we write F(t) ~G(t) as t+ =, Similarly if lim f(e)/o(8) = 1, we write f(s) ~ 9(#) ass 0. Then we have the following generalization of Theorem 1-17, Theorem 1-19. If F(t)~ G(t) as t> =, then f(s) ~ g(s) as s>0 where f(s) = L(F(O}) and 9(s) = £{G(H}. METHODS OF FINDING LAPLACE TRANSFORMS Various means are available for determining Laplace transforms as indicated in the following list. L Direct method. This involves direct use of definition (1). 2. Series method. If F(t) has a power series expansion given by F(t) = a + at + at? + + S ant (18) its Laplace transform can be obtained by taking the sum of the Laplace transforms of each term in the series. Thus A 2t ent 4) = + Ge lee = PU A condition under which the result is valid is that the series (19) be convergent for s>y. See Problems 34, 36, 39 and 48. 9) 3 Method of differential equations. This involves finding a differential equation satis- fied by F(t) and then using the above theorems. See Problems 34 and 48. 4. Differentiation with respect to a parameter. See Problem 20. 5. Miscellaneous methods involving special devices such as indicated in the above theo- rems, for example Theorem 1-13. 6. Use of Tables (see Appendix). CHAP. 1] THE LAPLACE TRANSFORM EVALUATION OF INTEGRALS Tf f(s) = £{F(6}, then S- eR) dt = f(s) Taking the limit as s > 0, we have . S Fiat = f(0) assuming the integral to be convergent. (20) (e1) ‘The results (20) and (21) are often useful in evaluating various integrals. See Problems 45 and 46. SOME SPECIAL FUNCTIONS I. The Gamma function. If n>0, we define the gamma function by T(n) = S ur em" du The following are some important properties of the gamma function. L Natl) = nnn), n>0 Thus since T(1) =1, we have M2) =1, (8) = (22) =2,4(4) = 8! and in general T(n+1) = nl, if n is a positive integer. For this reason the function is some- times called the factorial function. 2. ra) = 3. Mp)TA—-~) = 4, For large n, =, O 0, whose graph appears in Fig. 1-4. CHAP. 1] THE LAPLACE TRANSFORM 9 It is geometrically evident that as «+0 the height of the rectangular shaded region increases indefinitely and the width decreases in such a way that the area is always equal to 1, ie. S Pit)dt = 1. . This idea has led some engineers and physicists to think of a limiting funetion, denoted by 8(t), approached by Fi(t) as «> 0. This limiting function they have called the unit impulse function or Dirac delta function. Some of its properties are L So dt=1 2, S 3(1) Gt) dt = G0) for any continuous function G(t). O 3. S &(t-a) G(t) dt = G(a) for any continuous function G(t). ° Although mathematically speaking such a function does not exist, manipulations or operations using it can be made rigorous. TX. Null functions. If 2(¢) is a function of ¢ such that for all > 0 S N(u)du = 0 (81) we call N(t) a null function. fi tae Example. The function F(t} = {-1 t=1 is a null function. 0 otherwise In general, any function which is zero at all but a countable set of points [i.e. a set of points which can be put into one-to-one correspondence with the natural numbers 1,2,3,...] is @ null function. LAPLACE TRANSFORMS OF SPECIAL FUNCTIONS In the following table we have listed Laplace transforms of various special functions. For a more extensive list see Appendix B, Page 245. Table of Laplace transforms of special functions Fw fle) = 0; (b+) £{t} = a 8>0; (0) £ fer} = P (@) £0} = ery dt = yim fo estat ens |P = jim =P = lim = 1 ite>o pan —# lo pam (b) 4) = = jim where we have used integration by parts. CHAP. 1] THE LAPLACE TRANSFORM n om © ete) = Bin fo eremeat = lim ioero or = a if s>a For methods not employing direct integration, see Problem 15, oy _ 8, Prove that (a) £(sinat) = 2a, (b) L{cosat) = weg if s>0. (@ fisinat) = fF esinatat = Jim fem sinatat . nat — acosat) = gig tere ® fteonat) = fem conat dt = jim = lim & (<8 cosat + asinat) |? = ee et at wig sit s>0 - fre * ‘We have used here the results Pee = ott (a din Bt — B cos Bt) S exe sin seat es o) Seccossear = Seeongt + pring ® Another method. Assuming that the result of Problem 1(c) holds for complex numbers (which ean be proved), we have otia £ {oe} siis, ) But et = cosat + isinat, Hence eat) =f em Ht(cosat + dina ete = fem teo8 w = fo emeosat dt + if ene sinat de = efeosath + 1 sinat) 0 7 From (8) and (4) we have on equating real and imaginary parts, Lteosat} = ata, €{sinat} = Fe 12 . THE LAPLACE TRANSFORM [CHAP.1 3. Prove that (a) £(sinhat} = ="—s, (b) £ {cosh at) = wo if >a}. @ Aisinh at) = {25} = fees =) ae = 1 g-weceat —1 [ e-ste-at = ES eee at — 5 P emenst ae 1 = Fete - fete = 1f = - {43-3} Another method. Using the linearity property of the Laplace transformation, we have at once 5 for #>|al £toinhaty = “} = Fete - dates = aig for eel (8) As in part (@), kXeoshat} = sae} = Lefer + pater 1 for 6 > |a| 4, Find £(F(0)} if F(Q) = { ae By definition, £tFO) = S etR() dt = ; THE LINEARITY PROPERTY 5. Prove the linearity property [Theorem 1-2, Page 3}. Let £(Fy()} = file) = f e-#P,(Qdt and £4F2(t)) = fale) S e-*F, (dt, Then if e, and cy are any constants, ~° for Filt) + 02 Fale)? ens fey Fy(t) + eg Fa(t)) de ek EF UO} + cok {Fat} emt Py (t)dt HR (Qdt + op = erhils) + erfals) ‘The result is easily generalized (see Problem 61}. CHAP. 1] THE LAPLACE TRANSFORM 13 6 Find ¢ {de + 6i — 8sin4t + 2cos2t}. By the linearity property [Problem 5] we have £ (de + 6 — Bsindt + 2eosat} = AL {eH} + 6.L{H) ~ BL fsindt) + 21 {eos2th _ 1 Baty 4 2 = «(G4) + 6() - Gah) +2 — 44 6 _ 12 28 = get we Fret Bia where > 6. TRANSLATION AND CHANGE OF SCALE PROPERTIES 7. Prove the first translation or shifting property: If £(F(t)) = f(s), then £ (e*F(t)) = i(s—a). We have erin = feerma = 16 ‘Then Let FO} = SI eo {eat F()} dt = Lemmy de = Me lo 8& Find (a) 2 {te*}, (b) £{e* sin 4t}, (c) £ {e" cosh Bt}, (d) .£ (e“*(8 cos 6t — 5 sin 6t)). @ <@) = Then £ {tei} = 2 (e—3* 4 erie () {sin ae) = Then £{e~® sin 4t} = 4 = 4 F216 © Has +20" 4 =4 = *- (@ £ cosh5e} = arte = 7 Then £ {et cosh 5t} = Another method. {et cosh Bt} = e{eor(S4*)} = hetrey (@ £48 cos6t — & sin6t} = 3. {cost} — 5.2 (sin6t} * 6 3e— 30 8(shas) - 6 (tes) = 36 Tun (eM @omst- Sunsyy = TE = eeaT 14 THE LAPLACE TRANSFORM (CHAP. 1 9. Prove the second translation or shifting property: F(t — t He CEO) = Fe) and Gy = {FO 28, £16) = Sf naw ae = SS Sionma + Sl care at then .£(G(t)} = * f(s). Haq de + Fema ae = f[omre-o a = Semen du em Fu) du = ema f(a) where we have used the substitution ¢ = uta. 10. Find (R10) it FQ = {PRED Pee stot, gtr) = ferme ae +f ements a 5 fl cco eee 0 = ent ( e- cosu du Method 2. Since £ {cost} = Ll. Prove the change of scale property: If £(F(t)} = f(s), then £{F(at)} = £4(2). i e—* F(at) dt -f = LG emery du £(F(at)) cule) F(x) dala) using the transformation ¢ = u/a. CHAP. 1) THE LAPLACE TRANSFORM 15 nt) net 12. Given that «(4h tan *(1/s), find efsneth, U By Problem 11, ©, so that Jim e*? F(P) = 0 for #>y. For cases where F(t) is not continuous at ¢=0, see Problem 68. 14, Prove Theorem 1-9, Page 4: If £{F(t)} = f(s) then £ {F”"(t)} = 8* f(s) — 8 F(0) — F"(0). By Problem 13, £1G()) = 8. £4GO} — GO) = e918) — GO) Let G() = FW). Then UP") = £{F'(O} — FO) = s[se{F(t)} — FO) — FO) 22 {F() — #0) — FO) = stf(s) — sF(0) — FO) The generalization to higher order derivatives can be proved by using mathematical inductio# [see Problem 65}. 15. Use Theorem 1-6, Page 4, to derive each of the following Laplace transforms: Mey =2, MHeH=% Oe) =. Theorem 1-6 states, under suitable conditions given on Page 4, that £tP(O} = 8 L4F(O} — FOO) a 16 THE LAPLACE TRANSFORM [CHAP.1 (@) Let F(t)=1. Then F'(!) = 0, F(0) = 1, and (1) becomes £0} = =e£Q}-1 or £) = Ie @ (b) Let F(O) ‘Then F(t) = F(0) = 0, and (1) becomes using part (a) 40) = Ve = eet} — 0 or eit} = Ist @) By using mathematical induction we can similarly show that ¢{¢*) = n!/s"*1 for any positive integer n. (@) Let F(t) = ect, Then F(t) = aett, F(0) =1, and (1) becomes Lact} = sete) —~ 1, ie ax (et) = egfet}—1 or L{er} = /(e—a) inat) =< 16. Use Theorem 1-9 to show that £ {sinat} = Fe Let F(t) = sinat, Then F(t) = a@cosat, F(t) = —a? sinat, F(0) , F(0) =a. Hence from the result iF") = LF} — FO) — FO) we have £{-atsinat} = s?¢{sinat} — #(0) - a ie. a g{sinat) = s£{sinat} — a or f(a) = pe LAPLACE TRANSFORM OF INTEGRALS 17, Prove Theorem 1-11: If £{F(t)} = f(s), then & w Flu) au = fies. ' 0 J Let Gi) = Flaydu. Then G'(@®) = F() and G(0) = 0. Taking the Laplace transform of both sides, we have £(GIO} = e£(G(0) — GO) = #£1GD = fe) Thus 4a) = 42 aff Foyau} = fo 18 Find «ff Sita. We have by the Example following Theorem 1-18 on Page 5, eft} = ed ‘Thus by Problem 17, CHAP. 1) THE LAPLACE TRANSFORM a MULTIPLICATION BY POWERS OF ¢ 19. Prove Theorem 1-12, Page 5: If £(F(t)} = f(s), then (t"F(t)) = (-1)" Sno = (-1)"f™%s)_ where n = 1,2,8, We have fe) = ‘Then by Leibnitz’s rule for differentiating under the integral sign, #2 py = Ef enrma = HPO at = Sf tore at = fre SEP (O)} dt = ~£{tRr(O) Thus cero = -£ = -10 a) which proves the theorem for 7 ‘To establish the theorem in general, we use mathematical induction. Assume the theorem true for n=k, ie. assume eM (MF ONdE = (IRF) ® ‘Then ES we rora = (nese or by Leibnitz’s rule, HS ew erode = CayRserng) Sf oner rena = (DEH PAH (8) ® 2 It follows that if (2) is true, ie. if the theorem holds for n= k, then ($) is true, i.e. the theorem holds for n= k+1. But by (1) the theorem is true for n=1. Hence it is true for n= 1+1=2 and n=2+1 = 8, ete, and thus for all positive integer values of 7. To be completely rigorous, it is necessary to prove that Leibnitz’s rule can be applied. For this, see Problem 166. 20. Find (a) £(tsinat}, (b) 4 {t* cosat}. (a) Since ¢{sinat}) = fay, we have by Problem 19 ctesnad) = -$ (ate) = eo 18 THE LAPLACE TRANSFORM [CHAP.1 Another method. Since eteosat} = "6 *cosat dt wie we have by differentiating with respect to the parameter a [using Leibnita’s rule), Hehe teosatat — f en" (-tsinat}dt = -x{tsinat) = de(qte) = —_2a trom which aa ae a rat At sin at} Cray Mette rah tito Letonat = ef wath. @) Since ¢{cosat} = z, we have by Problem 19 @ EiBeosaty = a(sta) We can also use the second method of part (a) by writing we cttonan = e-em} = —retonn which gives the same result. DIVISION BY ¢ [ F(t) = 21, Prove Theorem 1-13, Page 5: If “{F(t)} = f(s), then 40, F(u) du. Let Gir) =. Then F(t) = tG(t). Taking the Laplace transform of both sides and using Problem 19, we have {FO} = ~ Ee} oo -# Then integrating, we have ae) = -f fan =f" fen au « cP = Sta Note that in (1) we have chosen the “constant of integration” so that lim g(s) = 0 [see Theorem 1-15, Page 51. 2. (a) Prove that f° “ar S$” fa du provided that the integrals converge. (0) Show that 5 Sint at (@) From Problem 21, Hu) du CHAP. 1) THE LAPLACE TRANSFORM 19 Then taking the limit as ¢> 0+, assuming the integrals converge, the required result is obtained. () Let F(t) = sint so that f(s) = 1/(e*+1) in part (a). Then fa S Sas = tant PERIODIC FUNCTIONS 23. Prove Theorem 1-14, Page 5: If F(t) has period 7 > 0 then 7 5 em" F(t) dt Lt} = We have cro = fo eer ar = fomrma + femroa + f eorma + J ce In the second integral Jet ¢ = «+7, in the third integral let ¢ = «+27, ete, Then : 7 7 iy) = (em Fw) du HUD P(t 1) du + [em 8e427 Pt 27) du + oe 4Ro) = (em rw dn + f wt du + f atom) + 7 =f cm ray de + co fem rw du te BP f° om RQ da to , a a = teeny mF dn a Jf onre S o-™ F(u) du etre where we have used the periodicity to write F(u+T) = F(u), P(u+2T) = F(u), ..., and the fact that 1 Lert tt eten yo wl @ in (1), a F(t) continuous at ), we find that 0 = lim of) - F(0) or tim 6/6) = FO) = lim If F(t) is not continuous at t =0, the required result still holds but we must use Theorem 1-7, Page 4, 26. Prove the final-value theorem: lim F(t) = lim #(s). By Problem 13, Ley = S ee () dt = f(s) ~ FOO) ‘The limit of the left hand side as » > 0 F lim Jeet de = Soro = kim [rat = jim (F)~ FO} = jim FW - FO) ‘The limit of the right hand side as # > 0 is Tim, 9 f(s) — F(0) ‘Thus lim F®) — FO) lim ¢ f(s) — F(O) or, as required, tim FO tim fe) If F(®) is not continuous, the result still holds but we must use Theorem 1-7, Page 4. CHAP. 1] THE LAPLACE TRANSFORM 21 27. Illustrate Problems 25 and 26 for the function F(t) = 8e-*. We have F(t) = 3e-%, f(s) = £{F()} = a By the initial-value theorem (Problem 25), im 3e-* = tim 82. tin, 8 iim se3 or $=8, which illustrates the theorem, By the final-value theorem (Problem 26), fim Set tin are or 0=0, which illustrates the theorem. THE GAMMA FUNCTION 28. Prove: (a) 1(n+1) = nr(n), n>0; (b) Tin +1) = a}, n=1,2,8, (@) Paty = = jim { = J emteaen aul =Pre-P + af wrntene au} IF : BO nfl wrtee de = ary it n>0 J, za rm ray = [oer = lim [Ue"du = Jim a-e% = 2. 5 rm J, Am, in Pin+1) = n(n). Then 1(2) = 17) = 3, 1/8) = 20(2) = 2-1 = 21, Pld) = B4(8) = +21 = Bt In general, 1(n +1) = n! if m is a positive integer. 29. Prove: ut dy and let Let Ip = [oes de = Ip = I, the required value of the integral. Then (fire a(f" ew c) S at Slr ae dy ieee Re Fig.16 where Rp is the square OACE of side P [see Fig. 1-6]. BR 22 30. SL Prove: £(t) = THE LAPLACE TRANSFORM (CHAP. 1 Since the integrand is positive, we have SSeomady = hs Sfeormaeay a) ® % ‘where , and Ry are the regions in the first quadrant bounded by the circles having radii P and PVE respectively. jing polar coordinates (r, @) we have from (1), me 22 cov ety dr de = ty de dé SOS a So Ser ee ae ® o fa-e" s ios Ea-er o Then taking the limit as P+ © in (8), we find Jim Tp =P = 2/4 and I= Ve/2. Prove: 1() = Vx. Ng) = ft Letng aot, its integral bees on axing Protea #8 f T(n+1) roo if n>-1, s>0. ci =f" “ey = So (eye) = anf wera = Mat ‘ttm dt. Letting st=u, assuming #> 0, this becomes 82. Prove: £{t-""} = Vals, 8>0. 33. Let n= 1/2 in Problem 31. Then am 2 TDL VEL ctem = GR = = fF Note that although F(t) = ¢-1/2 does not satisfy the sufficient conditions of Theorem 1-1, Page 2, the Laplace transform does exist. The function does satisfy the conditions of the theorem in Prob. 145, By assuming r(n+1) = mr(n) holds for all n, find: (a) M(-H, (6) 1-9, (6) (9, (@ 10), (e) T(-1), (f) T(-2). (@) Letting n=—}, TG) =—Yr-P. Then 1(-J) = -2r() = —2vz. () Letting n=—-$, r—y) = -$r(. Then r(-§) = —gr(-}) = (2)@)VF = $F by part (a), (o) Letting n= -, r(-H) = -fr(-). Then r(-) = —gr(- = (2) (Vz = —,8,V7 by part (6). CHAP. 1] THE LAPLACE TRANSFORM 23 (@) Letting n= 0, 1(1) = 0+ 10) and it follows that 1(0) must be infinite, sinee T(1) (@ Letting n= —1, r(0) = —11(—1) and it follows that 1-1) must be infinite. () Letting n =—2, 1(-1) = -21(-2) and it follows that 1-2) must be infinite, In general if p is any positive integer or zero, 1(—p) is infinite and [see Problem 170], rere = eam BNE) (ots BESSEL FUNCTIONS 34. (a) Find .£ {Jo(t)} where Jo(t) is the Bessel function of order zero. (b) Use the, result of (a) to find {Jo(at)}. (a) Method 1, using series. Letting n=0 in equation (28), Page 7, we find e ® tO = 1-gt+ee- pee tO 16! Then Lol} = ~ Bae a t = _ ns (3) a ~} _ ma) a - ~ Ver using the binomial theorem {see Problem 172]. Method 2, using diferential equations. ‘The function Jo(t) satisfies the differential equation tH) + Tel) + thoy = 0 w [see Property 5, Page 8, with n=O]. Taking the Laplace transform of both sides of (1) and using Theorems 1-6 and 1-9, Page 4, and Theorem 1-12, Page 5, together with J9(0) = 1, J(0) = 0, ¥ = £ {Jol}, we have = 8 ggty ~ a(t) — -y iW = gelev ~ a(t) — 0) + fey—- 1) - B= 0 , wy | from which re . wy _ _ sae uo y 7 Teri and by integration: v= s Vert ‘Thus by the initial-value theorem [Page 5], Now lim ey(e) =e and lim Jy(t) +i 0 we have ¢=1 and so £{J(t)} = i//st+1. For another method, see Problem 165. (6) By Problem 11, a4 1 Jere 24 THE LAPLACE TRANSFORM [CHAP.1 35, Find .£ {J:(t)}, where J,(t) is Bessel’s function of order one. From Property 8 for Bessel functions, Page 7, we have Jo(t) = —J;(t). Hence £4} = LUO} = = [9 Yo} - 1) 1 et E WeEIzs ~ veri eri The methods of infinite series and differential equations can also be used [sce Problem 178, Page 41). THE SINE, COSINE AND EXPONENTIAL INTEGRALS 1 36. Prove: £(Si()} = “ff “du. Then F(0)=0 and F(t) = Method 1. Let FQ = f Smt or er = sint. 0 ‘Taking the Laplace transform,

a wt-a) = We have u(t—a) tt few Then ecue—ay = fee mae + Another method. Since {1} = 1/s, we have by Problem 9, £{U(t—a)} 41. Find £ (F< (t)} where F.(t) is defined by (80), Page 8. Ik 088 . Then ° t>e Wehave Fi) = { LEO} = Som ran at Sorta ae + f e#(0) de CHAP. 1] THE LAPLACE TRANSFORM 27 42, (a) Show that lim £{F.(t)} = 1 in Problem 41. eo (0) Is the result in (a) the same as {lim rao}? Explain. (a) This follows at once since = imi Gaeta) 2 et ao It also follows by use of L'Hospital’s rule. (6) Mathematically speaking, lim Fe(t) does not exist, so that 4 lim rio} is not defined. Nevertheless it proves usefal to consider (0) = lim F,(O to be sich that “¢(3(9) = 1. We call 8(t) the Dirac delta function or impulse function. 43. Show that £ (8(t—a)} ‘This follows from Problem 9 and the fact that {8(t)} = 1. ‘+, where 8(t) is the Dirac delta function. 44. Indicate which of the following are null functions. 1 t=1 1 ists 0 otherwise’ © FO = 19 otherwise’ (a) F(t) = (c) F(t) = a(t). (@) F(f isa null function, since { F(u)du = 0 forall t>0, (®) If t<1, we have S F(u) du o Ir ists2, wehave f° Pujdu = f" (de = ¢-1. o i isa whee [rad = fod = 1 Since f° F(w)du x0 for all ¢>0, F(t is not a null function, fj F 80) du =1 for all €> 0, 6(t) is not a null function, ve EVALUATION OF INTEGRALS 45. Evaluate (a) Sf te-* cost at, of (a) By Problem 19, Lit cost} ff tert cose dt _ a = 4 #-1 = Gg teas} = &(xt) ery 28 ‘THE LAPLACE TRANSFORM (CHAP. 1 ‘Then letting # awe and fT tetteortt : (®) If F(t) = ete then f(s) = £{F()) = SEI Fay: Thus by Problem 21, ‘Taking the limit as # + 0+, we find 46. Show that (a) Sf 10 dt=1, (b) Sf e-terf Vidt = V22. (0) By Problem 34, Senso ae = 0 ‘Then letting 8 > 0+ we find S Joltydt = 1. (®) By Problem 39, Sl cw etvea = f ‘Then letting #>1, we find {" e-t ert VEdt vir. MISCELLANEOUS PROBLEMS 47. Prove Theorem 1-1, Page 2. ‘We have for any positive number N, 7 eure d+ ["eeroa Since F(t) is sectionally continuous in every finite interval 0= t= N, the first integral on the right exists. Also the second integral on the right exists, since F(t) is of exponential order y for t>N. To see this we have only to observe that in such case Soonre a| s tf eH FO | dt a ns s S e-* |FW)] dt = Sf comer ae = ‘Thus the Laplace transform exists for s> y. CHAP. 1} THE LAPLACE TRANSFORM 29 48. Find 2 {sin V4). Method 1, using serie sinyt= yi Then the Laplace transform is Vays (iy ~ an OBR aA “eo te = Me apt gp at 31 + (3/2) (5/2) 1(1/2) 1(9/2) conv) = SUP SAB, BRB, = Vr 1 (1/2? 6) _ (1/22. = a {1- (ge) + a at = Mem = Mem Bee 287 Method 2, using differential equations. Let Y(t) = sin Vt. Then by differentiating twice we find 4ty’ + 2y'+ Y = 0 Taking the Laplace transform, we have if y = 1 {Y(t)} -tZ ry — 2 ¥@ — YO} + ey- YO + y = 0 or 4ey' + (6s-—1y = 0 Solving, v= gem For small values of t, we have sinyt ~ Vt and £{Vt} = Ve/26¥%, For large «, y ~ c/s¥/?, It follows by comparison that ¢ = Vw/2. Thus €{sinyvt} = pe cos 49, Find cj SV" me of Let F(t) = sin Vé. Then rosa F(0)=0. Hence by Problem 48, 0, this becomes vay = sf e-*dne + Ingat —ythn ‘Then differentiating with respect to k, Seemed <7 MED = et yine ° Letting k=0 we have, as required, 7 e“Intdt = g{int) Jo Supplementary Problems LAPLACE TRANSFORMS OF ELEMENTARY FUNCTIONS 51. Find the Laplace transforms of each of the following functions. In each case specify the values of for which the Laplace transform exists. (a) 20 Ans. (a) 2Me—4), ama (0) B0-% (b) 3/042), a>-2 © 5-8 (©) (6—3e)/02, *>0 @ 2B (@ (+4e—s/e%9+1), > 0 (©) Bcosbt (e) Bs/(o2 + 25), 2>0 (N 10 singe (Nh 60/(02 +36), 2>0 (9) 6 sin 2t — 5 cos 2t (o) (2-5eyet+4), 8 >0 ) (e+ 1 (hy (e+ 42+ 24/68, 8 > 0 ( Gint — cose)? () (2-22 +4/ale2+4), 9 >0 (A) Bcoshbt — 4 sinh St G) @s—20/(02-25), ee > 5 CHAP. 1] THE LAPLACE TRANSFORM 31 52, Evaluate (a) £{(5e%—3)*}, (6) £ {4 cos? 2t). 2 30 9 2, oe por ea ty 4 Wit apy #>0 Ans. (a) 53. Find .£{cosh?4t}. Ans. ‘ae = 64) o2? a OSts5 Find ctF(O) it @ FO = {? 1 us \ ) FO { Ans. (0) 4em#/s (6) G(L—e-%) ~ few 55. Prove that ¢{t"} = 56. Investigate the existence of the Laplace transform of each of the following functions. (a) 1¢+1), (B) et, (€) cos Ams. (a) exists, (b) does not exist, (c) exists LINEARITY, TRANSLATION AND CHANGE OF SCALE PROPERTIES BT. Find .¢ {8 — 218 + 4e-8t — 2 sin 5t + 8 cos 21). mm, 4 10 30 Ans go et saa amet eae 58, Evaluate each of the following. (a) £{e-8} Ans. (a) 6/(0 + 8)¢ (0) 2 {e-* cos 2t} (0) (8 +1)/(s? +28 + 5) (© (20% sin at} (©) 8s? 6s + 25) KE + 2h (d) (4s? — 48 + 2)/(2 — 18 (0) £ {6 (8 sin 4t — 4 cos 4t)} (©) (20—4s)/(st ~ 48+ 20) (A) £{e-# cosh 2¢} (+ ayiiet + 86-412) (0) £{6-*(8 sinh 2¢ — 5 cosh 21} (9) (1 5e)/Mo2 + 28-3) 59. Find (a) ¢{e-t sin? }, (6) {(1 + te), 2 1, 38 6 6 aerary Ost oriF ed wre * ora : (e-1)? t>1 = oO. Fi F( F C Bem ¥ ‘ind £{F(O} if Fl) {i ere Ans, 2e~#/a% 61. If F,(t), F2(0), ..., y(t) have Laplace transforms f,( (8) respectively and Civey «+56 are any constants, prove that Le) Fit) + ee Felt) + vos + eg Fal} = or file) + egfale) + +++ + enfale) 32 THE LAPLACE TRANSFORM (CHAP. 1 etl @ i <0) = ge’ find £(F(21)}. Ans. (6? 28 + 4)/4(o + 1)%(0 — 2) eos 68. If £{F(O) = + find £{e7 F(3t)}. 6. Tf f(s) = £{F(}, prove that for r>0, £ (rt Flat)} LAPLACE TRANSFORMS OF DERIVATIVES 65. (a) If c{F() = fle), prove that EPO} = 8 f(s) — 9° FO) ~ 8 FO) — FO) stating appropriate conditions on F(t). (6) Generalize the result of (a) and prove by use of mathematical induction. 66, Given F(t) = fe OS EST. (@ Find £CF(O}. (0) Find £(F"(). (¢) Does the result £4F()} = 8.£{F()} — F(0) hold for this case? Explain. Ans. () 5 -F- Sw 2- = J@ o tanh 5 34 THE LAPLACE TRANSFORM (CHAP.1 Fig. 1-7 Fig.1-8 88. Find .{F()} where F(t) is the periodic function shown graphically in Fig, 1-8 above. Looe Ans. 3 wate ®. Let Fi) = {e spay. Mhere F(t) has period 4, (a) Graph Fi. (6) Find £ (FO). Ane, (0) Sa See ee 9. If F() = #, 00. na, 1 e-H(s+1) Aue HI 92, (a) Show that the function F(t) whose graph is the triangular wave shown in Fig, 1-9 1 has the Laplace transform 5 tanh 5 (®) How can the result_in (a) be obtained from Problem 872 Explain, INITIAL AND FINAL*VALUE THEOREMS 98, Verify the initial-value theorem for the functions (a) 8 — 2 co (8) (t+ 3)%, (o) t+ sin Bt. 94, Verify the final-value theorem for the functions (a) 1+ e~t(sint + cost), (b) e-%, 95. Discuss the applicability of the final-value theorem for the function cos t. 96. If F(t) ~ ot as t+0 where p>~1, prove that f(s) ~ eP(p+i/e! as @+o, %. If F(t) ~ cl ast+= where p>—t1, prove that f(s) ~ eT(p+i)/er+! as so, CHAP. 1) THE LAPLACE TRANSFORM 35 THE GAMMA FUNCTION 3) rid) 18/2) T(4) $8. Evaluate (a) r(6), (6) “TD, (o) rela), ) “Ta Ana. (a) 24, (b) 1/60, (c) 8Yz/4, (d) 32/315 9%, Find (a) £84 t-¥}, (6) CEE}, (0) AL + VEN). Ans. (a) (28+ 1)Vir/2s°, (b) P(2/3)/s%/9, (0) (62 + 2¥ea8? + Go + B¥r st + D)/e8 nd) cbt 176s}, 100. Find @ fh, (e) (ee Ana. (a) VeT@ #3), (0) 105V7/16(8— 297 BESSEL FUNCTIONS 1 101. Show that £ {e~#t Jg(bt)}, = —__—____.. Ventas tae = 8 102, Show that {¢Jo(@0) = ioe" : . 1 _ 03. Find (a) £{e° Jo(40)}, (0) £ {t Jo(20)}. Ans.) are Oram 104, Prove that (a) J5(t) =—Ju(, (6) St Iy(O} = Oya 105. If Io(t) = Jolit), show that {Ig(at)) = a>. 106, Find £{tJo(t) eo“). Ams, (@—1)(a®@ — 20 +2992 IW Show that (0) "Jide = 1, SF sunae a 108, Find the Laplace transform of {et Jy(21)}. Ana, ——#——_ oie ae Cub Vea +8 - 9-2 1 108, Show that £{tJ(0} = Gea 110. Prove that £ {Jp(aVt)} = vi. Braiate fees ayn at, ane. ais 5 (vei — on . 112, Prove that ¢{J,(o} = YEEL= 9" and thus obtain £{J,(at)). veri THE SINE, COSINE AND EXPONENTIAL INTEGRALS, M8. Evaluate (a) £ (e% Si(t)}, (6) £(t Si(0}. -1 1-2 tan“te ad Ans. (a) tan-!(¢—2)/(8~2), (b) wen 36 THE LAPLACE TRANSFORM [CHAP.1 In(@+1) Bet 114, Show that 2 {# Gi(O} ot) - fete UB. Find (a) £{e™ Bi(}, (6) £{t Bi(o). ns, () M244) q met) 2 Ane) eg) Oe eT tis. (@) £{e-! Si(2H}, (0) £ {te Bi (34)). vos (a) (aEile-+ 12 dgin(#8) — Ans. (0) ay? ©) yah #*) @FHGTH THE ERROR FUNCTION 117. Evaluate (a) £{e% erf VE}, (6) £{terf @vt)). 1 3s+8 Ans. (a) —_, apap @aaVvent Fer Ow 8. Show that ¢(erfe Vi} = a . Vati(Vveti+1) 119, Find Aff et Vi au} Ans. Weer THE UNIT STEP FUNCTION, IMPULSE FUNCTIONS, AND THE DIRAC DELTA FUNCTION 120. (a) Show that in terms of Heaviside’s unit step function, the function F(t) = {°° 9<* <8 t>8 can be written as ¢#{1—U(t—8)}. (6) Use 2 {U(t—a)} = 6 */e to find 2 (FCO). te sen Ans, (o) Le Seth 121, Show that F() = see O< tS econ be written as Fite FO = Fxlo + (FO — Fy} uta) W22. IE F() = Ft) for 0a,_,, show that ro PQ + (a(t) — Fy} Ut—ay) + o> + CF) — Fyr-1() Mt—a,-2) 123, Express in terms of Heaviside’s unit step functions. 2 sint O0 Be Ans, (a) t+ (4t— ¢) W(t—2), (b) sint + (sin 2t — sin t) U(t—7) + (sin 3t — sin 2t) W(t—2z) 124, Show that ¢{#u(t—2)} = (1+ 20+ 20), 9 >0. CHAP. 1) THE LAPLACE TRANSFORM 387 125, Evaluate (a) S- cos 2t 8(¢~ w/8) dt, (b) S. ent Wt—2) dé. Ane. (a) -1/2, (6) e? 126, (a) If 5’(t— a) denotes the formal derivative of the delta function, show that Sf rose-oa = -P@ (0) Bvatuate ” e-stse—2) at, Ans, (b) 4e-8 127, Let G,(t) = We for 0S t 2/4 and 0 for t0, (0) im t F(@) = 0 for some constant m such that 0 N. a 1 146. Show that (a) < {Jo(t) sint} = —7—— sin {} tan“! (2/)) _ oy a il (6) £4olt) cost = We con (4 tant 2/9} 2a t>1 = —-4he- M1. Let F(t) {i acter: Prove that £(F(0)} dy 08 £1G(E+ D))- U8, If ¢{P"(t)} = tan-t (1/s), F(0)=2 and F'(0) = Qe — 1+ tan“11/s |, find £ (F(t). Ans. 19. Prove that fee! F(gt)} = 3 ) where a and fare constants and ¢(F(t)} = fle). 8 150, Show that the Laplace transform of e¢ does not exist, while the Laplace transform of e~ét does exist. CHAP. 1) THE LAPLACE TRANSFORM 39 151. (@) Show that, <{eme = fin( #44). t (6) Brauate ff Ans, (6) $1n5 152. (@) Find {Epo}. ®) Show that f° SA A=W gy = mn (YE*3), 158. Work Problems 87 and 88 by using Method 4 of Problem 36. 154. Suppose that .£ {F(t} existe for #=a where a is real, Prove that it also existe for all — F(®) a>a. 158. Find the Laplace transform of the periodic function F(t) shown graphically in Fig. 1-10. aa easy tan dy 156. Prove that (=1y"-! (4n = 2)! Fig.1-10 Lisin@} = ee 6! at efsinte) = st orcas FF HEE Tee F 86) and generalize [see Prob. 144], 7 > ate 158. Find ¢ {te-# Jy (tv )}.. Ans. TT ae + OR 159. Find £{¢U(t—1) + B8(t—1)}, Ame, 67 # (et +04 1)/e2 160. Find £{cost int 8(¢— =}. Ans. —e-"* Ing 161. Let F(t) and G(t) be sectionally continuous in every finite interval and of exponential order as t-+ ©. Prove that 2{F(t) G(t)} exists. 162, The Laguerre polynomials La(t) are defined by ede nl dee (t). (6) Find £{L,(t)}. Lt) = {te} n= 0,12... (@) Find Lo(t), Ly(0), «+5 168. (a) Let a,b,a,A and A be constants. Prove that Lata + bt-#} = Alas" + bs A) if and only if a+ =1 and A= +Vreoar, (®) A function F(t) is said to be its own Laplace transform if ¢ {F(t)} = Fle). Can the function F(t) = at-«+b¢-® be its own Laplace transform? Explain, 40 ‘THE LAPLACE TRANSFORM [CHAP. 1 164. If F(t) and G(t) have Laplace transforms, is it true that F(t)G(t) also has a Laplace transform? Justify your conclusion. cos (t sin 8) do to show that ¢{Jg(t)} 165, Use the result Jo(t) = + Vert 166. Prove that Leibnits’s rule can be applied in Problem 19, stating suitable restrictions on F(t). 167. (a) Prove that f ° en(Lagtt) A ©) Prove that f~ 1=E08t ae . 168. Let F(t) = 0 if t is irrational and 1 if ¢ is rational. (a) Prove that {F(t)} exists and is equal to zero. (6) Is the function a null function? Explain, te show tat "#2048 = ; 110, Prove that if p is any positive integer, reas = ena) ~( IML, Verity the entries (a) 55, (b) 61, (c) 64, (d) 65, (e) 81 in the Table of Appendix B, Pages 248 and 250. 172. Using the binomial theorem show that for |2| < 1, Gta = 1-F2+ 3a — and thus verify the summation of the infinite series in Problems 84 and 29, 173. Use infinite series to find the Laplace transforms of (a) sin t, (6) cost, (ce) e%*, (d) cos Vt. 174, Prove that £ {erf (t)} rfc (e/2) and thus find {erf (at)}. 115. (a) Find ¢{erf Ve} by using the method of differential equations. (0) Find {cos Vé/Vt} using infinite series. M76, Show that (0) [” Jo(@V¥u) cosu du = sint, A () 7 seve) sin w du = cost. CHAP. 1) THE LAPLACE TRANSFORM 41 177, Show that f"J(2Via) Jolw) du = Jolt). 5 178. Use (a) infinite series, (5) differential equations to find £{J,(t)}. See Problem 35. 119, If #>0 and n> 1, prove that eft 180. Prove that if »>1, im = ww S fe = i432 ant ‘The function f(n) is called the Riemann zeta function. 181. If f(s) = £ {F(E)}, show that f(7 mr = fins) 4), rar = sine 182. If L,(t), m= 0,1,2,..., are the Laguerre polynomials [see Problem 162|, prove that = edo(2vt) = “ cosau du. (a ie ere = X 183. Let sla,t) = few conan du. (0) Show that = —* 7 where J00,0) = Ve/2vi. (b) By solving the differential equation in (a) show that ° _ Jat) = e-™ cosaudu = \ S, avi ate 184, Use Problem 188 to find j se Problem 48, Page 2). Fh! : o7¥2t sinh t sint a 185. Prove that Chapter 2 DEFINITION OF INVERSE LAPLACE TRANSFORM If the Laplace transform of a function F(t) is f(s), ie. if 2 {F(t)} = f(s), then F(t) is called an inverse Laplace transform of f(s) and we write symbolically F(t) = 2-1 {f(s)} where £- is called the inverse Laplace transformation operator. 1 ‘! pag We can write cfg} ee UNIQUENESS OF INVERSE LAPLACE TRANSFORMS. LERCH’S THEOREM Example. Since {et} = Since the Laplace transform of.a null function N(t) is zero [see Chapter 1, Page 9], it is clear that if 2 (F(t)} = f(s) then also £{F(t)+N(t)} = f(s). From this it follows that we can have two different functions with the same Laplace transform. t= otherwise Example. The two different functions F,(t) = e-% and F,(t) = ee have the same Laplace transform, ie. 1/(s +3). If we allow null functions, we see that the inverse Laplace transform is not unique. It is unique, however, if we disallow null functions {which do not in general arise in cases of physical interest]. This result is indicated in Theorem 2-1. Lerch’s theorem. If we restrict ourselves to functions F(t) which are sectionally continuous in every finite interval 0St=N and of exponential order for t>N, then the inverse Laplace transform of f(a), ie, £-' {/(@)} = F(t), is unique. We shall always assume such uniqueness unless otherwise stated. SOME INVERSE LAPLACE TRANSFORMS ‘The following results follow at once from corresponding entries on Page 1. 42 CHAP. 2] THE INVERSE LAPLACE TRANSFORM 43 Table of Inverse Laplace Transforms 7) £719} = Fit) 1 cos at sinh at cosh at SOME IMPORTANT PROPERTIES OF INVERSE LAPLACE TRANSFORMS In the following list we have indicated various important properties of inverse Laplace transforms. Note the analogy of Properties 1-8 with the corresponding properties on Pages 3-5. 1. Linearity property. Theorem 2-2. If ¢, and cz are any constants while f,(s) and f2(s) are the Laplace transforms of F,(t) and F2(t) respectively, then Ler fils) + cofels)) = ero {fils)} + c2€ * {fa(s)} @) = aFi(t) + cF2(t) ‘The result is easily extended to more than two functions. Example. = tet deante + Sine Because of this property we can say that <~! is a linear operator or that it has the linearity property. 2. First translation or shifting property. Theorem 2-3. If £-*(f(s)} = F(t), then Lo (f(s—a)} = e* F(t) 2) a4 THE INVERSE LAPLACE TRANSFORM (CHAP. 2 eens 3. Second translation or shifting property. Theorem 24. If £-'{f(s)} = F(t), then er _ [F(t-a) t>a £7 (e-*f(a)} 3 toe (8) Example. Since ‘¢ 4h} = sint, we have [sin(t— 2/3) if ¢> 2/3 jo if t< 2/8 4. Change of scale property. Theorem 25. If £ '(f(s)) = F(t), then t oii = ERG) “ Brame, Since 1 {arfg} = edt, we have if 2% 1 2 Lage at a “ {wots = eos feos at as is verified directly. 5. Inverse Laplace transform of derivatives. Theorem 24. If £-*{f(8)} = F(t), then A ery = erfasroh = Care re 6 1} Example. Since <7") rpqp = sint and é(ata) = Gere We have fo ls eine or eft) = desin i. {ots} - oi {aston} atsine 6. Inverse Laplace transform of integrals. Theorem 27. If £-'{f(8)} = F(t), then eff ran) = 7A 6) fo es pad. “ Example. Since Stara} CHAP. 2] THE INVERSE LAPLACE TRANSFORM 45 7. Multiplication by s". Theorem 2-8. If £°'{f(s)} = F(t) and F(0)=0, then Lisi = FO ” Thus multiplication by s has the effect of differentiating F(t). It F(0) #0, then £'f8fe) - FO) = FO or Lo'{s f(s} = Ft) + F(0) a(t) (8) (9) where 8(t) is the Dirac delta function or unit impulse function {see Page 9}. Example. Since ¢-t wey} = sint and sin0 = 0, then “| Generalizations to £ '(#"/(s)), n = 2,8,. ae wap > qlsind = cose s+» are possible. 8& Division by «. Theorem 2-9. If £ *(/(s)} = F(t), then elt (4 = Sf F(w) du (10) Thus division by s (or multiplication by 1/s) has the effect of integrating F(t) from Otot. Example. Since -! {arta} =i feiss at = Generalizations to £~' (f(s)/s"}, n = 2,3, . sin2udu = 4(1 ~ cos 2) +» are possible [see Problem 70]. 9. The Convolution property. Theorem 2-10. If £-* {f(s)} = F(t) and & *{9(s)} = G(t), then £7 F(8) 9(8)} = JS Fu)G(t~u)du = F*G (uy) We call F*G the convolution or faltung of F and G, and the theorem is called the convolution theorem or property. From Problem 21, we see that F*G = G*F. aample. since c-H{ 5) = ot and ect {2g} = owe have = f ett du = — ot 46 THE INVERSE LAPLACE TRANSFORM [CHAP. 2 METHODS OF FINDING INVERSE LAPLACE TRANSFORMS Various means are available for determining inverse Laplace transforms, as indicated in the following list. Compare with Page 6. 1. Partial fractions method. Any rational function P(s)/Q(s) where P(s) and Q(s) are polynomials, with the degree of P(s) less than that of Q(s), can be written as the sum of . , ; ' « A As+B rational functions fealled partial fractions| having the form Te‘-5y> (asi be FoF where r = 1,2,3,.... By finding the inverse Laplace transform of each of the partial fractions, we can find ¢-' {P(s)/Q(s)}- 5 2-5 = 4 B piel. Be ayer ~ Sea * Ger ay + Example 2, 38 — 4842 Ag+ B Ca+D ard Bet ae ae — 8) The constants A,B,C, etc., can be found by clearing of fractions and equating of like powers of s on both sides of the resulting equation or by using special methods [see Problems 24-28]. A method related to this uses the Heaviside expansion formula [see below}. 2. Series methods. If /(s) has a series expansion in inverse powers of ¢ given by f(s) = (12) then under suitable conditions we can invert term by term to obtain nt? nt? Ft) = tat OO Gy (18) See Problem 40. Series expansions other than those of the form (12) can sometimes be used. See Problem 41. 3. Method of differential equations. See Problem 41. 4. Differentiation with respect to a parameter. See Problems 18 and 88. 5. Miscellaneous methods using the above theorems. 6. Use of Tables (see Appendix B). 7. The Complex Inversion formula. ‘This formula, which supplies a powerful direct method for finding inverse Laplace transforms, uses complex variable theory and is considered in Chapter 6. THE HEAVISIDE EXPANSION FORMULA Let P(s) and Q(s) be polynomials where P(s) has degree less than that of Q(s). Suppose that Q(s) has n distinct zeros a,, k= 1,2,8,...,n. Then {Pe _ o Pew “ {art ~ 2%) (14) CHAP. 2] THE INVERSE LAPLACE TRANSFORM 47 This is often called Heaviside's expansion theorem or formula. See Problems 29-81. ‘The formula can be extended to other cases [see Problems 105 and 111]. THE BETA FUNCTION If m>0,n>0, we define the beta function as 1 Bim,n) = S wnt (1 —uyr-t du (15) We can show the following properties [see Problems 82 and 33]: E(m) T(n) 1 Bonny) = mee a 1 2. S sin’™-eos"-19d0 = 5Blm,n) = ener EVALUATION OF INTEGRALS The Laplace transformation is often useful in evaluating definite integrals. See, for example, Problems 35-37. Solved Problems INVERSE LAPLACE TRANSFORMS 1. Prove that (a) ¢ {aa} mean =F, n=0,1,2,8,..., where 01=1, 1 ‘in at \ () <4 tz} a wife} = cosat, (e) £7 ff) ao {etal = cosh at. (a) x fort} rim 0) <{5} = det = 4(25) = gi Then <1 faa} = for n= 051,280... © «fant = Le (sina) = (@) £feosat) = hg. Then orf 48 THE INVERSE LAPLACE TRANSFORM (CHAP. 2 co f28zetl = Letaionat) = aig tim ofetal (A £feoshat} = Ay. Then et {ata} cosh at. 2 Prove that eo{al = rea for n>~1. fej. i4 = 2 ree 1 “{rcra} = Pera <0) = apy ee Tete MPT by Problem 31, Page 22. 141 tn and the result is equivalent to that of Problem 1(6). n>-1. Note that if x= then r(n+1) = at 8. Find each of the following inverse Laplace transforms. @efars} @ e{zh Owan (@) ct {as} = ae [Problem 1(6)} ® fy = det [Problem 1(a)} © {ah =#-£ [Problems 1(b) or 2] @ et{ett = con Vit [Problem 1(d) 2 {zea} = Geosh4t [Problem 1(/)} sinh V3¢ A o et fpha} = fa). 0 _ an oire {ast - 7m) = Wp = [Problem 1(e)} [Problem 2] LINEARITY, TRANSLATION AND CHANGE OF SCALE PROPERTIES 4, Prove the linearity property for the inverse Laplace transformation [Theorem 2-2, Page 43). By Problem 5, Page 12, we have Lk Pi) + PO} = ex EPi(} + eek {Fa} = efile) + ee fale) CHAP. 2] ‘THE INVERSE LAPLACE TRANSFORM 49 ‘Then Loe fil) + co fa()} = er Fi(t) + oF alt) = LAM + e2£-* fe()} The result is easily generalized [see Problem 52). . 1 {5e+4 2e~18 = 5. Find (a) ef tt aa + Bago} - 6 «BB +45 8-68 @) «€ ‘os 98-16 * fais} 20-18 24 ~ 80 # [5,4 % , 18 = ofgeg-ge tah (oy es {Stt4 + 24 ah ot Te Be + 4(t2/2!) — 2cosBt + 18(f sin Be) + 24(e8/31) — 30{E8/2/T(7/2)} 5t + 20 — 2cosdt + Gsindt + 40 — 168°%/y¥r since (7/2) = §+$+47q) = Bve B4ds , 8— ‘st w ofgty- diet tes {8 1/4 )-§ 2 1/1 = <1) -3a 7 3 ere aa) + 4(a'ons) = 8/2 — fsinh 4t/3 — $ cosh4t/S + Zsin3t/4 — § cos 3t/4 6. Prove the first translation or shifting property: If £-'{f(6)) = F(t), then £7 fls—a)) = et F(t) By Problem 7, Page 18, we have £{ettF()} = f(s—a). Then ety = try Another method, Since f(s) = f e-# F(t) dt, we have fe-a) = e-em Ri dt = emt (eet (D) dt = fet F(t)) ‘Then £1 Ele a)} et F(t) 7. Find each of the following: 3547 CC) {tty} \yV2s +3 1 = gf oa—4 Se efeSatat = of = ser{ = Gert cosdt + 2e%sindt = 2e%(Scosdt + sin de) (er {tarvish 2 eft (@) 50 ‘THE INVERSE LAPLACE TRANSFORM yf 412 | _ yfaett2) 2. faeta o £ {atti} ae {ey} 3 { rar i = rfot bgp 1 aa f=} ‘a ataa} = det — geet = beg 1 f%47] _ ny - @ <7 \e=a3) = <"le=aay * rat - my 1 = 2 = BONG aa wari} = Betcosh2t + Setsinh2t = et(3 cosh 2t + 5 sinh 2t) = tet et For another method, see Problem 24. - 1 1 a 1 = det i ‘ {rest vas {ertans} = Lew 2 Lipo 5° "Ta Ve Prove the second translation or shifting property: If 2-'{f(s)} = F(t), then -1{e°**f(s)) = G(t) where _ fF(t-a) t>e a = {o t5 \o t<5 = Alt s ae we—5) (b) Since ¢~ [oon 5¢—4r/5) > tel8 0 t< 49/5 foo St > de/5 lo t< de/5 = cos Bt Ut — 4x/5) (0) We'have s+1_) as fates Pray = erfetpt} fst) + ketal vse vit = emt con et 4 4 oom sin VE a a = E(w + oi) Thus a1 [let Dene Bs {seer} [Sartentes + ain Bee ecaenn 6 Vi {Vi cos $B e-» + sin Fe-n} ‘u(t—z) \ 7 1 eee! (@ Wehave {esa} = ety {a} At8/2 et = a 8 = “Ter = 52 THE INVERSE LAPLACE TRANSFORM [cHAP. 2 Thus = #et {Sara} det(e—aynrreerD —g = Br ° t3. Then or B=-1 Using these values we obtain the result in Method 1. See also Problem 7(c), Page 50. 2%, Find 71 fo wSho . GF He-De-B! We have ata GFHe=aHe a CHAP. 2} THE INVERSE LAPLACE TRANSFORM 59 Let us use the procedure of Method 2, Problem 24, ‘Multiply both sides of (1) by #+1 and let s>—1; then Bet — A = lim ae 7 6 Multiply both sides of (1) by #2 and let #2; then - wens Lk B= earn 3 Multiply both sides of (1) by #—8 and let #3; then C= Nim eee Thus The procedure of Method 1, Problem 24, can also be used. However, it will be noted that the Present method is less tedious. It ean be used whenever the denominator has distinct linear factors. 26, Find oe. (e+ ie— 2) Set— 158-11 | _A_ “erine—33 ~ a+i* 3 + Gay t i A procedure analogous to that of Problem 25 can be used to find A and B, Multiply both sides of (1) by #+1 and let 8 —1; then Se? — 15e —11 A = lim 7 ‘Multiply both sides of (1) by (s—2)* and let #2; then ee we know A and B, we have from (1), Bite 11-178 c D @rine—ae ~ aFT + anit = @ To determine C and D we can substitute two values for s, say 8=0 and #=1, from which we find respectively, DM _ 1,0, zp Fpctett+e-p ie, 8C~6D = 10 and 3C—3D = 11, from which C=4, D=1/3. Thus _,f5st—1be—-11) _y [=U8 = 4 a (eat Sy = OY asT + waap t Hae ts = bet — Tee 1 ot = -je geet + abet + em 60 THE INVERSE LAPLACE TRANSFORM (cHap.2 Another method. On multiplying both sides of (2) by s and letting s>, we find 0 = —}+D which gives D= 4. Then C can be found as above by letting s = 0. ‘This method can be used when we have some repeated linear factors. 27. Find g-1{__38+1_) (s-1j(e+1)J Set Bex esycrsi eT o ne im 3841 ; ultiply both sides by s—1 and let +1; then A = lim St #1 = 2 and +t = eeoy «ree aT ® To determine B and C, let s=0 and 2 (for example). then 1 2B+C p= 24 2Bee from which C=1 and B = Be41 < ‘f “ere | = Qe — 2eout + sint Another method. Multiplying both sides of (2) by » and letting s+ =, we find at once that B = 28. Find e{__ +2843 } (42s FBP +285) Method 1. jttmta Ast Bo, Cat w (e+ Qe + 2)(s% + 28 + 5) e+ 2s+2 e+ 2et 5 Multiplying by (0% + 26+ 2)(68 + 20+ 6), t+ 248 (A+ B)(s? +2845) + (Ca + Dye? +28 +2) = (A+Os + QA+B+2C4D)s? + (6A 4+2B4+2C42D)8 + 5B + 2D Then A+C =0, 2A+B+20+D=1, 6A42B+2042D = 2, 5B+2D = 8, Solving, A=0, 4, C=0, D=4. Thus an e+ 2e+3 25 + Det + De TH) fertsint + Bed gent (sin t + sin 26) CHAP. 2} THE INVERSE LAPLACE TRANSFORM 61 Method 2. Let #=0 ing: Multiply (1) by s and let s+»: 0 = A+C 3 _ A+B, C+D Let # w = 4f2+ 5? 1 Let ¢@=—1: g = A+B+—~ Solving, A=0, B=}, C=0, D=% as in Method 1. This illustrates the case of non-repeated quadratic factors. Method 3. Since #?+2e+2=0 for e=—1*%, we can write ++ 2 = (t1-Het149 ly e+ Pet 5 = (et1—2i(s+1+2i) Then et 2e43 tests F Be OE + Be +B ert 14 ie +i Beri +2) = 4 B = Fini + seta —_P oFita + + 1-8 Solving for A,B,C,D, we find A = 1/6i, B = -1/6i, C = 1/6i, D = -1/6i. ‘Thus the required inverse Laplace transform is ecaahe gator | gra-ame ga tent of a + a ~ @ an'(S58 *) + get (AG) = fertsint + gem! sin 2¢ = feo" (sin ¢ + sin 2t) ‘This shows that the ease of non-repeated quadratic factors can be reduced to non-repeated linear factors using complex numbers. HEAVISIDE’S EXPANSION FORMULA 29. Prove Heaviside’s expansion formula (14), Page 46. Since Q(s) is a polynomial with n distinct zeros a1,a2,-..,0,, We can write according to the method of partial fractions, Puy Qe) Multiplying both sides by #— a, and letting 6 aj, we find using L’Hospital’s rule, a im 2) Gg ‘i fone Ax aim Qe) @— on POL =} 8-4 Pla) = iin Pe) ie (oat) = Peo im ats = oe ‘Thus (1) can be written Pl) _ Pla) 4 Ply) 1 Ae) = Malem t+ Map emm bt 62 a THE INVERSE LAPLACE TRANSFORM (CHAP. 2 ‘Then taking the inverse Laplace transform, we have as required LPO Pld cae gong PO sant gp cee g Plt sagt {ea} = ga tt ae Ft ae 30. Find lace. I @+D(e—2)(e—8) We have P(s) = 28—4, Qls) = (s+ 1)(e—2(s—3) = ds? +846, Qe) = B8?—Be41, a1=~1, a) =2, aj=3. Then the required inverse is by Problem 29, PID)... PD PR) sae “2 een tea" tea" = B Compare with Problem 25, -2 4 ea 4, 1 +t M4 oe log 4 1 at 4 Thee = het — Get + pet Bs+1 \ (De +D!* 31. Find £ { We have P(s) = 88+1, Q(s) = (¢—1)(e +1) = at ste Qs) = 88-2041, a =1, ~i since st+1 = (e—i(e+7). Then by the Heaviside expansion formula the required =i, inverse PO) pg PO guy PD pce aa" * ea" * era io) 4 gett = get 2tai° = Bet + (-1—Jalcost + isint) + (-1+ fifeost — ising) = Ret — cost + gsint — cost + Joint = Bet — 2eost + sint Compare with Problem 27. Note that some labor can be saved by observing that the last two terms in (1) are complex conjugates of each other. THE BETA FUNCTION . 1 32. Prove that B(m,n) = S am\(L—ay ide = where m>0,n>0. Consider 4 Gy = Ff wmre-ayt rc) S (tan tds ‘Then by the convolution theorem, we have LAG = ktm tft = Lem) ry rm) PG ae ae 7 = get [LOMP@)) POm) LO) gant ‘Thus ae) £ “f rot f Tawra met(t— amide = POLO) men or ami (t—attde = Tontny ° 1 Letting t = 1, we obtain the required result, CHAP. 2] ‘THE INVERSE LAPLACE TRANSFORM 63 on 33. Prove that S sin?™-16 cos*—-16 do = Bm, n) = ree. From Problem 32, we have tm, * amat(d—ayetde = Elon) £00) Bim, n) J (acl d: Tém+ n) Letting = sin?¢, this becomes (m,n) = sin2m=1g cos?-? = Ein) rm) Bim, n) 2 2m-1 9 cost-1 9 do rim) tie} from which the required result follows. - dp int 0 cost “ 34, Evaluate wf sin‘ cost da, (b) S costo dd, (0) J” Fay: (a) Let 2m—1=4, 2n— 6 in Problem 33, Then m=5/2, n= 7/2, and we have - * = 16/2) 0(T/2)_(8/2)(1/2 W225 Be S. sint@ cost ¢ dg = eae 12 (0) Since cos is symmetric about ¢ = 2/2, we have Sf conte de = af cste do Then letting 2m—1=0 and 2n 4, ie, m=1/2 and =5/2 in Problem 33, we find 2f "Peosteds = 2[ same) 25(8) = 2fe-gianive ray sin-1/2 9 cost? 9 de eae 0S. Jeans Letting 2m—1 = 1/2 and 2n—1 = 1/2, or m=1/4 and n=8/4 in Problem 98, we find vt de _rayay (8/4) eve fo 6 -Vtane 211) 2 sin (=/4) 2 using the result T(p)r(1—p) = z/(sinps), 02 _ _ m2le—3) > Bo grog _ Ans. (a) to aD cy (t—2)wt—2). (6) {i ee 4 sin A(t — 8) Wt—3), © ‘pea 1 a @—1-veue—wive. 0 t<1 }, (b) ot, (aoa 9 £>2 59. Find (a) ¢-' Ans. (a) or (en2e-2) — DF UE) t<2 Je sin ta) > B eee ® a fog OF gt sin (¢— 2) ue 9) wo tt fF" emr (nde = fly and fe Gio.dt = fpr), where p and q are constants, Sind ionship between F(t) and Git). Ame. Gilt) = e-*/F(t/p\/p a rel e. If of z | = ert Vi, find of oral }, a>o. Ans. erf Vat/Va 1 [VFI - (=e mh 62. If 1) = Jalt), find Wh Ans. a* J,(at) oR eS Vea os 6. Find (a) { aot we { Ee i vVa(s—1)f" vero” . ‘T(Bt—6) t>2 _ put Ans, (a) et ertVt, (6) fi peg % 7o(st—6) ue 2) INVERSE LAPLACE TRANSFORMS OF DERIVATIVES AND INTEGRALS 64, Use Theorem 2-6, Page 44, to find (@) £-* {1s ~ a)8} given that £1 {1s —a)} = eat, () £71 {a/(s2 — a2)2} given that £-! {1/\s¢— a2} = (sinhat)/a. 65, Use the fact that £-1(1/e) = 1 to find ¢ 1 (1/o") where m =2,8,4,.... Thus find <-' (1/(¢~a)"}, 66. Find {atta} Ans, 4te-t sin 67. Find (@) fin (33)}. (6) f In (Ha). * du Ans. (a) (e-!—e~®/t, (6) f 68, Find -!{tan-1(2/s)}. Ang, 2 sin¢ sinh t/t CHAP. 2} THE INVERSE LAPLACE TRANSFORM 1 Sau — cos be yy, dae ff 802 MULTIPLICATION AND DIVISION BY POWERS OF « mm Poo wat cH {091 = f° 6" fry ax a am Can the integral be written asf ‘ S * f "Po det Explain, @. Find 1{tin( a1fia 1 {+2 9 oft 11, Evaluate (a) £ {aces}: m4 {ates} @ £ {etn}: Ans, (a) 1—t+4P—en% (b) B+} fe-™, () Let + t+ BEY 72 Find (a) ¢-* > () «(ara Ans. (a) $ext VE), (6) f° Jo(au) du and discuss the relationship between TTT a=aterT} these inverse transforms. 18. Find (@) olpantesa}: ) on Ans. (0) $(# We (& If Fit) = £-44/(6)}, show that (@) £71 {sf"(s)} = -tF(t) — F(t) (ce) £7¥atf"(s)) = OF) + de P() + 2F(t) ©) £eP"@) = OPH + UFO 7%. Show that £71 {s2/'(s) + F(O)} = —tF"(f) —2F(). THE CONVOLUTION THEOREM 76, Use the convolution theorem to find (a) ¢-t Ans. (a) (ete), (0) sy(o% — e-* — Ate 2) i aif ns. f(sint — a mrad A geataeg} Aim ine mee 78, Find 1 . Ans. ft cos 2t + } sin 2t eat} te Hots : fa Hot 1. Find (a) ¢ {at} we {ets Ans. (a) 4{(3—B) aint — St cost}, (6) gytlsin 2¢ — 2¢ cos24) 72 THE INVERSE LAPLACE TRANSFORM [cHAP. 2 8. Prove that F*(G*H} = (F*G}*H, ie. the associative low for convolutions. 81. Prove that (a) F*{G+H} = F*G+F*H, (6) (F+G)*H = F*H + G*H. 82, Show that 1*1*1*...%1 (nones) = tI/(n—1)! where n = 1,2,8, ss. stor oat fff moa = SSA? ro de a. show tat (f= f rman = nde So 5. Prove the convolution theorem directly by showing that 1H) 9%) = {fone au { ff “em a(n) as} SS Sen {f PW) Gt—w) aw at. G ent) Fae) Gv) du dev 86. Using the convolution theorem, verify that Sf, sinwcos(t~u) du = 4esine on 81. Show that if £ du = ole-De K(a—b)2). PARTIAL FRACTIONS 88 Use partial fractions to find (a) otf Ans. (a) Get — 2e-*, (b) 1— Hert + Jet ee a1{ tears ®. Find (@) £ {i st. we oan Hert Ama, (a) $o-t? — fe-2/8, (6) Bet — Jet? + Bert 27-128 1 J +168 —24 rae }. ors {etree} Ans. (a) Se-# — 8cos8t, (b) fin dt + cos2t — sin2t 9%. Find (a) £7 $1, Find 2-1 {emitters} Ans. 4o-#(4 cost — 8 sint) — ge-s ; 1 {_#= 2048 : 2 Find (a) ¢ eaten} ) (0) (GE+B)em% — Deer Ana, (0) 42t— et + be CHAP. 2] ‘THE INVERSE LAPLACE TRANSFORM 73 9. Find 27) 2-3 FOO — Ba? F Bet Bl” Ans. Se! — fom — dhe-tcos2t + fhe~“ sin 2 m4. —— ne. J sin ¢ sin oarieraTo}: Ane deine sinh¢ ind gt {28821 ee eo %. Find ¢ (Snttel Ans. }sint + |teost - t 9%. Use partial fractions to work (a) Problem 44, (b) Problem 71, (c) Problem 78, (d) Problem 76, (e) Problem 77. 97. Can Problems 79(a) and 79(b) be worked by partial fractions? Explain. HEAVISIDE’'S EXPANSION FORMULA 28-11 e+ 2-9 98. Using Heaviside's expansion formula find (a) ef () eo 198 +87 Ve Be + DOTS) Ans. (a) 8e-% — o8, (6) Be® — Be-! — Be Ans. Jet ~ ott + See Ans, 2e-' + Bsint — 2 cost 101. Use Heuviside's expansion formula to work (a) Problem 76(a), (b) Problem 77, (e) Problem 88, (@ Problem 89, (e) Problem 90. Compare with Problem 91. 7 = et 102. Find avatars} Compare with Problem 93. Compare with Problem 94. 105. Suppose that f(s) = P(s)/Q(s) where P{s) and Q(s) are polynomials as in Problem 29 but that Q(e) = 0 has a repeated root a of multiplicity m while the remaining roots, 61, bs, ....b, do not repeat. (a) Show that Pee) fe) = aR () Show that A, = lim mk (e) Show that ¢-1{f(@)} = of 14 THE INVERSE LAPLACE TRANSFORM [CHAP. 2 186 Une Probtem 16 nd Go) 2 { BETREEIE., ) + Ans. (a) (8t—2)et + de-¥, () let ~ et) Ans, (202 t+ Bye% + Gent 108, Use Problem 105 to work (a) Problem 26, (b) Problem 44, (c) Problem 71, (d) Problem 73, (@) Problem 76(6). 109. Can the method of Problem 105 be used to work Problems 79(a) and 79(b)? Explain. 298 G+ iPr ie 110. Find of using Problem 105, Compare with Problem 95, 111, Develop a Heaviside expansion formula which will work for the case of repeated quadratic factors. dot +503 + 602 + 8942 = DGEs e+ FF ‘112, Find of i using the method developed in Problem 111. Ans. eb + e-#{(3~2t) cost — 3 sin t} THE BETA FUNCTION 118, Evaluate each of the followings (a) f''28/2(1—eytde, 0) f ad—2-1Rae, ©) [AVI P av A Ans. (a) 16/815, (b) 4096/35, (¢) 2 116 ow hat" VT=Bas = ot 22 we 5 115, Evaluate each of the following: (a) f costo de, (6) f°” sintacoste de, (0) I sino costo ae. A e Ans, (a) 5x/S2, (b) x/32, (0) 37/128 116, Prove that if p is an even positive integer, if p is an odd positive integer. spi show that r(p)r(—p) = 57 - where 00, is e{fr acme = Se (e) Can the method of Problem 37 be used to evaluate the integral in (a)? Explain. ta} die? 126. Evaluate f To(u) J(e—u) due Ans. Jolt) — cost MISCELLANEOUS PROBLEMS 121. Find cfs} Ane G{ct— 62 ( co — Vi siny ‘)} 128, Prove that f'"(e~a)p(b—a)tde = (b—a)r***1 B+ 1.q-+1) where p>—1, q>—1 and b> a, [Hint. Let x—a = (b—a)y,) 4 os 1 Evaluate @) f° FSR, [Vea Dae, Ane (o) x, MEI sag) 130. Find ¢-1 {sass}. Ans. {1 — cos (t—1)} W(t — 1) — {1 — cos (¢- 2)) u(t 2) enn ws. show nat ct { 182. Prove that J Jn(us) sin(t—w du = ft Jy(t). 6 THE INVERSE LAPLACE TRANSFORM (CHAP. 2 Lae 183, (a) Show that the function f(s) is zero for infinitely many complex values of s. What are these values? (6) Find the inverse Laplace transform of f(s). 1 t>2e ins. (a) 9 = +i, +24,*8i,... (0) Fi) = or F(t) = Wt~2) Ans. (a) @) FO tt octeg, % Fit) = wean) 134, Find {inet We)} Ans, 170) 2s z ° 16y3 185, Show that f" ae—wu'yt dv = 168s, 136, Let F(t) = @ at all values of ¢ which are irrational, and F(t) =¢ at all values of t which are rational. (a) Prove that ¢ (F(d)} = 2/s%, @>0. (b) Discuss the significance of the result in (a) from the view- point of the uniqueness of inverse Laplace transforms. 137, Show how series methods can be used to evaluate (a) £-1{1/(e?+1)}, (6) £-? {In(1 + 1/8)}, (©) £74 {tan (1/s)}. Ans 138, Find £-1 {e—*—2¥5}, - e-e-9 UEE= 8) ve(t— 38 199, show that [MSRM dy = Fert, €>0. 1+ 2 an ‘ anion show that MO If FQ) =t-¥8, €>0 and Git) = { ° en * o0. Prove that LOMYGlMl— ee} = Fit) if O 7. em a1” iit M45. (a) Show that e{s} = #£- (®) Discuss the relationship of the result in (a) to that of Problem 127. CHAP. 2] THE INVERSE LAPLACE TRANSFORM 7 146, Can Heaviside's expansion formula be applied to the function /(o 1(e cosh s)? Explain. M1, Prove that Joe") dz = ave. 148, Show that 149, Show that 150. Find ¢-' { Ans, t/q — et erte (vt) ma 151. Show that @ yee— ws = 2 Zo where [t] denotes the greatest integer less than or equal to t. e #8 t - . — am * a” ap 182, Show that << {tal Chapter 3 ORDINARY DIFFERENTIAL EQUATIONS WITH CONSTANT COEFFICIENTS The Laplace transform is useful in solving linear ordinary differential equations with constant coefficients. For example, suppose we wish to solve the second order linear differential equation ON + aT 4 py = Ft) or ¥" +a¥' +BY = F(t) @ where a and f are constants, subject to the initial or boundary conditions YO)=4, YO) =B @) where A and B are given constants. On taking the Laplace transform of both sides of (1) and using (2), we obtain an algebraic equation for determination of ¢(¥(t)) = y(s)- The required solution is then obtained by finding the inverse Laplace transform of y(s). The method is easily extended to higher order differential equations. See Problems 1-8. ORDINARY DIFFERENTIAL EQUATIONS WITH VARIABLE COEFFICIENTS ‘The Laplace transform can also be used in solving some ordinary differential equations in which the coefficients are variable. A particular differential equation where the method proves useful is one in which the terms have the form myo) @) the Laplace transform of which is cr Eero) “ See Theorem 1-10, Page 4, and Theorem 1-12, Page 5. For details of solution see Problems 9-11. SIMULTANEOUS ORDINARY DIFFERENTIAL EQUATIONS The Laplace transform can be used to solve two or more simultaneous ordinary dif- ferential equations. The procedure is essentially the same as that described above. See Problems 12 and 13. 18 CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 79 APPLICATIONS TO MECHANICS Suppose a mass m, attached to a flexible spring fixed at O, is free to move on a friction- Jess plane PQ [see Fig. 3-1]. If X(t), or briefly X, denotes the instantaneous displacement of ‘m at time t from the equilibrium or rest posi- tion, there will be a restoring force acting on m equal to ~kX, where k is a constant depend- ing on the spring, and called the spring con- stant. This follows from Hooke’s law which, on the basis of experiment, states that the re- storing force acting on a spring is proportional to the stretch or extension of the spring from the equilibrium position. According to New- ton’s law which states that the net force acting on m is equal to the mass times the accelera- Fig 01 tion, the equation of motion is altro (0 peation max = kX or mX" 4kX = 0 (5) If in addition, there is a damping force proportional to the instantaneous speed of m, the equation of motion is ef dé m = wax - po or mX” + BX’ +kX = 0 6) where the proportionality constant f is called the damping constant. A further modification takes place when some prescribed time-varying external force FI6) also acts on m. In such case the equation of motion is nox = -kX - o& + FU) or mX" + BX + ka = F(t) ” By using Laplace transforms to solve equations (5), (6) or (7) subject to various ap- propriate initial conditions of physical interest, the displacement X(t) can be found. See Problems 14, 15, 27 and 28. APPLICATIONS TO ELECTRICAL CIRCUITS A simple electrical circuit [Fig. 3-2] con- sists of the following circuit elements con- nected in series with a switch or key K: 1. a generator or battery, supplying an elec- tromotive force or e.m.f. E (volts), 2. a resistor having resistance R (ohms), 3. an inductor having inductance L (henrys), 4, a capacitor having capacitance C (farads). These circuit elements are represented symboli- Fig.3-2 cally as in Fig. 8-2. 80 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 3 When the switch or key K is closed, so that the circuit is completed, a charge Q (coulombs) will flow to the capacitor plates. The time rate of flow of charge, given by 2. , is called the current and is measured in amperes when time ¢ is measured in seconds. More complex electrical circuits, as shown for example in Fig. 3-3, can occur in practice. Fig. 3-3 An important problem is to determine the charges on the capacitors and currents as functions of time. To do this we define the potential drop or voltage drop across a circuit element. (a) Voltage drop across a resistor = RI = R42 + #Q (b) Voltage drop across an inductor = L Lag (e) Voltage drop across a capacitor = & (@) Voltage drop across a generator = —Voltage rise = —E The differential equations can then be found by using the following laws due to Kirchhoff. Kirchhoff’s Laws 1, The algebraic sum of the currents flowing toward any junction point [for example A in Fig. 3-3] is equal to zero. 2. The algebraic sum of the potential drops, or voltage drops, around any closed loop [such as ABDFGHA or ABDFQPNMA in Fig. 3-8] is equal to zero. For the simple circuit of Fig. 3-2 application of these laws is particularly easy {the first law is actually not necessary in this case]. We find that the equation for determina- tion of Q is eo Q 190, Ro = (8) By applying the laws to the circuit of Fig. 3-8, two simultaneous equations are obtained [see Problem 17]. CHAP. 3) APPLICATIONS TO DIFFERENTIAL EQUATIONS 81 Note the analogy of equation (8) with equation (7). It is at once apparent that mass m corresponds to inducténce L, displacement X corresponds to charge Q, damping factor B to resistance R, spring constant k to reciprocal of capacitance 1/C, and force # to electro- motive force E. Such analogies are often useful in practice. APPLICATIONS TO BEAMS Suppose that a beam whose ends are at 0 and @=1 is coincident with the axis [Fig. 8-4]. Suppose also that a vertical load, given by W(z) per unit length, acts trans- versely on the beam. Then the axis of the beam has a transverse deflection Y(z) at the point 2 which satisfies the differential equation Y ey We Be o ges} (lt 2sin at @ Y at @ + FO * Plu) sin a(t —w) du Note that in this cuse the actual Laplace transform of F(t) does not enter into the final solution. 8. Find the general solution of Y”—a*Y = F(t). Let ¥(0)=c,, ¥(0)=cp. Then taking the Laplace transform, we find sty — se; — cy — aty = f(s) = mite or y= aoa + ‘Thus ¥ = cycohat + Ssinhat + 2 f Plu) sinhalt—w) du Accoshat + Bainhat + 2 f” FQ sinhatt—w aw which is the required general solution. ORDINARY DIFFERENTIAL EQUATIONS WITH VARIABLE COEFFICIENTS % Solve t¥”+¥'+4t¥ = 0, ¥(0) = 3, ¥(0) = 0. We have £0) + 20) + £4} = 0 or —Ltery — 2 ¥0) - 10) + (ey — YO) — 4% = 0 fey wos y = 0 dy ede Then wy ge so and integrating nyt EMC HO) = or VEG TInverting, we find ¥ = eden To determine ¢ note that, ¥(0) = eJ9(0) = c=, Thos ¥ = 342 86 10. 1. Solve t#¥”+2¥" Let Y'(0+) = 0. APPLICATIONS TO DIFFERENTIAL EQUATIONS +t¥ = 0, ¥(04) =1, Yim) =0. ‘Then taking the Laplace transform of each term — Atay — 2¥04) - YON) + Bev — YON) - fy Integrating, Since y+0 as #— ~aty) — 2ey +1 + ty —2—y' = 0 (+I ~ y = ~tente ta ©, we must have A=#/2, Thus = Z—ten-ts = tant} vy = Fo tants = tan ‘Then by the Example following Theorem 1-18 on Page 5, BS veer vos coe a ‘This satisfies Y(x) =0 and is the required solution. Solve Y”-t¥'+¥ =1, ¥(0) =1, ¥(0) =2 We have ie, or ‘An integrating factor is oS ("*#)«* or integrating, 4a — cory + cry = cay = Ft sy — 2¥0) - YO) + Z y—¥(0)) ty = 3 ey-s-2tetyty =} eet stealing = tee Then Been = (1 y= Rene f(242s Doe a = pew ferezer nes de 211) ne 2a date vast [aetet + 2044 4 6 1 e 1,2, 6 ue G+ a+ Sewe (CHAP. 3 CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 87 To determine o, note that by series expansion, + 2+ San yer yet 2 = i+ = = olf ett ony Then since ¢-1 {st} = 0, k = 0,1,2,..., we obtain on inverting, Y = 1+ (+2) But ¥"(0) =2, so that ¢=0 and we have the required solution a Y= 142t SIMULTANEOUS ORDINARY DIFFERENTIAL EQUATIONS aX _ ox—ay 12. Solve | subject to X(0) =8, oy = Y-2x Taking the Laplace transform, we have, if ¢{X} = ex-—8 = x — By or (1) sy 8 = y- de or) Solving (1) and (2) simultaneously, e-2 8 2 3 _ vos a2 8 - 2 an ree) = 4a) = Bent + Bett ¥ = gt) = bet — gett X” 4 48X = bet 5 13. Solve ibject to X(0)= 85, X'(0) = = (enn = is singe SUbiect to X(0)= 85, X"(0) = ~48, ¥(0) = 27, ¥"(0) = —55. Taking the Laplace transform, we have ste — 9(35) — (48) + ay ~ 27 + 82 = TE sy — 9(21) — (65) — A{er—35} + By = 4 88 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 3 or (#4524 y = se— 214+ J a) a 30 —tse + (2+ 3)y ate — 195 + ® Solving (2) and (2) simultaneously, 35e— 2 + . = 3503 — 480? + 3000 — 63 15(e? +3) _ 3 FGFS) FYGRF ITD) RFD AF +9) 45 3 28 = — ro tai t eat y= = 608 + 30(8? + 3) RFT) © FHF + 467 +9) = 2 6 3, a B+ ~ Re ati t BH Then X = £4 fe} = 80 cost — WainBt + Se-t + 2cos2t Y = £-{y} = B0c0s8¢ — 60 sint — Set + singe APPLICATIONS TO MECHANICS 14. A particle P of mass 2 grams moves on the X axis and is attracted toward origin O with a force numerically equal to 8X. If it is initially at rest at X = 10, find its position at any subsequent time assuming (a) no other forces act, (b) a damping force numerically equal to 8 times the instantaneous velocity acts. (a) Choose the positive direction to the right [see Fig. 3-5]. When X>0, the net force is to the b—— ¥—— left negative) and must be given by 8X... ———}>—-—_-x When X<0 the net force is to the right (i.e. is ° P positive) and must be given by —8X. Hence in either case the net force is —8X. Then by New- Fig. 3-5 ton's law, (Mass) + (Acceleration) = Net force ex a —8x 2X 4 ax = 0 « or oe CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 89 The itial conditions are: (2) X(0)=10, (8) X"(0)=0. Taking the Laplace transform of (1) and using conditions (2) and (8), we have, if 2 = {X}, 108 atx — 108 + doe oe ‘Then x £MGa) = 10 cos2t The graph of the motion is shown in Fig. 9-6 below. ‘The amplitude [maximum displacement from 0} is 10. The period |time for a complete cycle] is x. The frequency [number of cycles per second] is 1/5. a : Period x Fig.3-6 Fig. 3-7 (8) When X>0 and dX/dt>0, P is on the right of O and moving to the right. Then the damping force is to the left (ie, is negative) and must be given by —8dX/d¢. Similarly when X <0 and dX/dt<0, P is on the left and moving to the left so the damping force is to the right (ie. is positive) and must also be given by —8dX/dt. The damping force is also —8dX/dt for the cases X>0, dX/dt<0 and X<0, dX/dt>0. Then (Mass\(Acceleration) = Net force @x _ or aoe = -8x-8 nx, jax te, FE Ege 2 w with initial conditions (5) X(0)= 10, (6) X’(0)= ‘Taking the Laplace transform of (4) and using conditions (5) and (6), we have ste — 108 + or 2 Then X = ga} 10(s + 2) + 20 2 = a1 fin Jd woet{eha} + mefetag} = le + Bore = 106-2 (1+ 24) " a a> The graph of X vs, t is shown in Fig. 8-7 above. Note that the motion is non-oscillatory. The particle approaches O but never reaches it. 90 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 8 15, A particle of mass m moves along the X axis and is attracted toward origin O with a force numerically equal to kx, k>0. A damping force given by BdX/dt, p>0, also acts. Discuss the motion, treating all cases, assuming that X(0)=Xe, X’(0) = Vo. The equation of motion is ax _ x mae = HX — Bae PX AX oy or Ge t wat ex = 0 () where a = A/2m, o? = kfm. ‘The Laplace transform of (1), using the initial conditions, yields ste — Xqe — Vo + Qaler—X) + ote = 0 @Xy + (Vo +2aXo) (40%, Vet ake etre toa * Greta e or a= Case 1, uta? > 0. In this case, X= Leh = XyemstoosVarmakt + en at gin Vat at The motion is called damped oscillatory [see Fig. 3-8 below}. The particle oscillates about O, the magnitude of each oscillation becoming smaller with each swing, The period of the oscillations is given by 2e/Va?—e#, and the frequency is Vat—a2/2e. The quantity w/2r (corresponding to ie. no damping) is called the natural frequency. Case 2, ut — a? In this case, X= ghey Xo | Vetaks a A < { Fat ere = Kye + (VotaXeenat Here the particle does not oscillate indefinitely about O. Instead, it approaches O gradually but never reaches it. The motion is ealled critically damped motion since any decrease in the damping constant f) would produce oscillations [see Fig. 3.9 below]. Case 3, ut a2 <0. In this case, X = got = The motion is called overdamped motion and is non-oscillatory. ‘The graph i critically damped motion [see Fig. 3-10 below!. CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS a1 Damped oscillatory motion Critically damped motion Overdamped motion Fig. 3-8 Fig. 3-9 Fig. 3-10 APPLICATIONS TO ELECTRICAL CIRCUITS 16. An inductor of 2 henrys, a resistor of (=) 16 ohms and a capacitor of .02 farads are connected in series with anemf.of 2h& ean E volts. At t=0 the charge on the 9 capacitor and current in the circuit are zero. Find the charge and current at any time t>0 if (a) E = 800 (volts), a (6) E = 100 sin 3¢ (volts). Fig. S41 i Q and I be the instantaneous charge and current respectively at time t. By Kirchhoff’s laws, wwe have a Q ae + ier + or = F @) or since I = dQ/dt, &Q aQ _ aoe + 6+ 0e = B @) with the initial conditions Q(0) = 0, 1(0) = Q'(0) (@) If E = 800, then (2) becomes @Q , dQ - Se + BSE + 25q 150 ‘Then taking the Laplace transform, we find (#4 ~ #Q(0) — QO} + Blea~ QO) + 25q = 82 150 S(e® + Be + 25) 69+ 48 or 5 — 68448 ’ PF Bet 2B = So+4) +24 arr 9 —_6+4) at (+4 +9 (+42 +9 Then Q = 6 — Ge-tcosst ~ 8e~* sin Bt = x 506-4 sis r= Go = b0e~* singe 92 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 3 (6) If E = 100 sin3¢, then (2) becomes eg e+ 882 + 259 = sosinae Taking the Laplace transform, we find (480425 = ZS = 150 and i CeCe Tes . = Bi Bos ,% 1 wat ~ 26 a2 +9 52 +9 26 +4249 52 (@+4e+9 25 fa 25 gs Bg ‘Thus @ = Basinse — Banat + Bewsinse + Bet conse = Besinat — 3 cosa) + e+e cosat + 2 sind) = % con8t) + 2 e-s(@ const + 2 sin and 1 = @ = Becosse + ssinay) — Be-s 7 sinst + 6 e083!) For large ¢, those terms of Q or I which involve eM are negligible and these are called the transiont terms or transient part of the solution. The other terms are called the steady-state terms or steady-state part of the solution. 17. Given the electric network of Fig. 8-12, 110 volts, determine the currents in the various branches if the initial currents are zero. Kirchhoff's second law [see Page 80] states that the algebraic sum of the voltage or poten- tial drops around a closed loop is zero. Let us traverse loops KLMNK and JKNPJ in a clock- fashion as shown. In traversing these mt 000 L loops we shall consider voltage drops as positive when we travel against the current. A voltage Fig. 8-12 rise is considered as the negative of a voltage drop. Let I'be the current in NPJK. This current divides at the junction point K into I, and I, so that I+ Iq. This is equivalent to Kirchhoft’s first law [see Page 80]. Applying Kirchhoff’s second law to loops KLMNK and JKNPJ, we then have respectively dl, aly 10, ~ 25 + 4G2 + 20h = 0 aor — 0 + 292 A+ 10 = 0 | aly | th _ -G+2gt wh = 0 1 S | Ht Mh + Bh = | subject to the conditions 1,(0) = I5(0) CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 93 Taking the Laplace transform of the system and using the initial conditions, we find Bir — (ei, — 1,0)} + 2Asi— 120) + 10 = 0 tei, 1,0) + 20i, + 15% = 55/6 or (#45), — (204+ 10)ig = 0 (+20) + 15i = B5/s From the first equation, é, = 2%, so that the second equation yields aro, = Boo = i = Then I, = 1-082 = 2h = 2-20-80 1 = +h = 8— ses APPLICATIONS TO BEAMS 18. A beam which is hinged at its ends 2=0 and 2=1 [see Fig. 3-13] carries a uni- form load Wo per unit length. Find the deflection at any point. The differential equation and boundary con- ditions are ay _ Wo SA= Br OO<#e ‘This can be written-in terms of Heaviside’s unit function as Wee) = Wo {Ula) — We — 12)) Taking Laplace transforms of (1), we have, if y = y(e) = £(¥(e)}, sty — 8 ¥(0) — (0) — #¥"0) -— 0) = From the first two of conditions (2) and the unknown conditions Y’(0) =e, Y’’"(0) a1, Wo v= tat pete) Inverting, we find Ye = ‘This is equivalent to Y@) = ‘We now use the conditions ¥"(I)=0, ¥’"()=0 to find } Wo (a — 1/2) — gp Gi we - wy o0. Find @ {et = Sen Wat, () {2} = Se en Dae assuming suitable restrictions on U = U(z,t). (@) Integrating by parts, we have wi, _ ¢” 0 - P wt OU {i} = J owe e = fim J, at + 0 e# Ue, dt J 9 } = sf nve0 dt — U(e,0) * Ue, t lo = sufz,s) — U(e,0) = su — Ule,0) where u = u(z,s) = £{U(z,t)}. We have assumed that U(e,t) satis as a function of t. the restrictions of Theorem 1-1, Page 2, when regarded (®) We have, using Leibnitz’s rule for differentiating under the integral sign, af} = [ena - gfewe - 23. Referring to Problem 22, show that ie = (@) fe SE} = sulz,s) — 8U(2,0) — Uel2,0) FU au m «fel = where Us(z,0) = 2) and uw = w(z,8) = £(U(2,0)- Let V = aU/at. Then as in part (a) of Problem 22, we have 220) av «fa = {it} = 2£{V} — V(x,0) 8 [s.£{U} — U(e,0)] — U4(2,0) = su — 6 U(e,0) — U4(x,0) Note the similarity of the results of this problem and part (a) of Problem 22 with Theorems 1-6 and 1-9, Page 4. Extensions are easily made. CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 97 24, Find the solution of av _ au ea ce 7 2a + Ur ~~ Ule,0) = be which is bounded for > 0, t > 0. Taking the Laplace transform of the given partial differential equation with respect to ¢ and using Problem 22, we find SE = _26u-Ue@,O) + u o 4 arte = ates from the given boundary condition. Note that the Laplace transformation has transformed the partial differential equation into an ordinary differential equation (1). To solve (1) multiply both sides by the integrating factor oS “+P = g-caens, Then (1) ‘ean be written Integration yields Re eer ee et ae meee a a2 = aa Now since U(z, t) must be bounded as z+, we must have u(z,8) also bounded as z= and it follows that we must choose ¢=0. Then 8 ante ri and 80, on taking the inverse, we find Ue.) = bem is easily checked as the required solution, a _ eu 25. Solve Sp = Gap t>0. U(e,0) = 8sin2rx, U(0,t) = 0, U(l,t) = 0 where 0<2<1, Taking the Laplace transform of the partial differential equation using Problems 22 and 28, we find veo) = FE or FR w= ~ Bain dee w where u = u(z,s) = £{U(2,#)}. The general solution of (1) is w= eee + cere + sin 2ra ® ae Taking the Laplace transform of tho: boundary conditions which involve t, we have £{U0,0} = us) = 0 and =k {U(1, 1} = uta) = (3) 98 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 8 Using the first condition [u(0,2) = 0] of (8) in (2), we have ate = 0 (a) Using the second condition [u(1, «) = 0] of (8) in (2), we have ee + ee~ ° (6) From (4) and (6) we find ¢,=0, e2=0 and so (2) becomes (6) from which we obtain on inversion Ul@,t) = Bes sin 2ex ca) This problem has an interesting physical interpretation. If we consider a solid bounded by the infinite plane faces #=0 and = 1, the equation wy ca ao oat is the equation for heat conduction in this solid where U = U(z,t) is the temperature at any plane face x at any time t and k is a constant called the difusivity, which depends on the material of the solid. The boundary conditions U(0,t) = 0 and U (1,1) = 0 indicate that the temperatures at = and x=1 are kept at temperature zero, while U(z,0) = 3 sin2ex represents the initial temperature everywhere in 0<2<1. The result (7) then is the temperature everywhere in the solid at time ¢>0. Further applications are considered in Chapter 8. . 26. Find the bounded solution of z x z>0,t>0 suchthat U(0,t) = 1, U(z,0) ‘Taking the Laplace transform of the partial differential equation and the condition U(0,t) = 1, we find respectively m- Ue, = £4 oo FS-m = 0 on and u(0, 8) (@) From (1), « = u(t,e) = oyeVi= + ege-V52, Since U(e,t) must be bounded as z-> =, u(z,s) £{U(@,t)} must also be bounded as 2+. ‘Then we must have ¢=0, assuming Vé> 0, so that ulz,e) = ege-Ver (8) From (2) and (8) we find ¢, = 1/s, so that eis u(a,e) = Thus using Problem 43, Page 67, we find ® 2¢7 Ue,t) = erfe @ = ot dv ave, Ve Maeviy Physically, this represents the temperature at. any point of a semi-infinite solid *>0 whose face is kept at unit temperature and whose initial temperature is zero [see Problem 25). CHAP. 3} APPLICATIONS TO DIFFERENTIAL EQUATIONS 99 MISCELLANEOUS PROBLEMS 27. Suppose that in Problem 14, Page 88, an external force F(t) acts on the particle but there is no damping force. (a) Find the position of the particle at any time if F(t) = Focoset. (b) Discuss the physical significance of your results. (a) If the external force #(t) is taken into account, the equation of motion becomes ex _ 20% = ex + FH % or ax” + 8X = Ft) (2) ‘As before, the initial conditions are X(0) = 10, X0) = 0 ) It Flt) = Fycoset, (2) becomes ex” + 8X Fy cos ut wo ‘Taking Laplace transforms and using conditions (s), we find, if 2 = £(X}, Betz — (10) — 0) + Bx Then if ot #4, 6 or = ey Cy adam X = gla} = 1 cos2e + ag ott — cos wt) ” If of =4, then (6) becomes = + ® and s0 using Problem 18, Page 58, XK = ctf) = 1 cos2e +e sinze (9) 8 (®) If o®=4 or w=2, ie. if the frequency of the applied external force is equal to the natural frequency of the system, it is seen from (9) that the oscillations about the equilibrium position increase indefinitely. This phenomenon is called resonance and the frequency corresponding to © =2 is called the resonant frequency. If in such case the particle is attached to a spring, the spring will break. 28. Work Problem 27 if (a) F(t) = FoU(t—a), (b) F(t) = Fo8(t). (a) In this case the equation of motion is [equation (2) of Problem 27) 2X" + 8X = Fyu(t—a) where X(0) = 10, X"(0) = ‘Then taking Laplace transforms, we find Foe~ 2[ata— 108) + 8x 100 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 3 10s | Fyenot era * merD = 10s | Poems fa = Bed + eee - ata} 10 cos2t + £Fo{1—cos2(t—a)) if t>a | 10 cos 2¢ if t1 and ¥() = fsin2e + 4(1— cost) for t<1 where F(t) = { 43, Solve Problem 42 if: (a) F(t) = U(t~2), [Heaviside's unit step function]; (6) F(t) = a(t), [Dirac delta function]; (c) F(t) = 6(t~2). Ans. (a) Y(t) = $sin2t if t<2, Jsin2e + [1 —cos(@t—4)} if t>2 () Y= sin, t>0 (©) ¥() = foin2e if 1<2, flsin 2+ sin(2e—4)}? if e>2 ORDINARY DIFFERENTIAL EQUATIONS WITH VARIABLE COEFFICIENTS Solve each of the following by, using Laplace transforms and check solutions. Moy" +ey'-¥ = 0, ¥()=0, ¥()=1. Ans Y=t 45. t¥" 4 (L—29¥’—2¥ = 0, ¥0)=1, YO) =2 Ans, Y = ot 46. 1Y" + (t-)¥'—¥ = 0, ¥(0)=5, Y(m)=0. Ans, Y= be-t 47. Find the bounded solution of the equation ey” + ty + (@-HY = 0 which is such that Y(1)=2 Ans. 2J,(#)/J; (1) 104 APPLICATIONS TO DIFFERENTIAL EQUATIONS [cHAP. 3 SIMULTANEOUS ORDINARY DIFFERENTIAL EQUATIONS subject to the conditions Y(0) = 48. Solve { » ¥"(0) = 2, Z(0) = 0. Ans. Y = 24404 4o-t—Psint+ foost, Z = 1- fet + Paint — deost Y'—2'-2¥ +20 yr +20'+¥ =0 Ans, Y = fet + sett — feost—g sine + pte sint ig ¥(0) = ¥'(0) = 20) 49. Solve { Zz © fet + tent 4 2¥" = et 50, Sol Ne lx +ex—¥ if X(0) = ¥() = ¥"(0) =0. Ang, X= 1+e-t—ena— te, Y= 1tet—oenst—ae-™ where a = H2—V3), b= H2+V2) 51. Solve Problem 49 with the conditions Y(0) = 0, ¥"(e)=1, Z(0) =0. WY +24 = (t—Aert Y-Z=se Ans. Y=4(t), Z= given that Y(0)=1, Z(0) = 52. Solve { sa Solve {-8¥" + 82" s jw 2 = given that Y(0) =—1, (0) = Z(0) = 4, 20) Ans. Y = B+§t— pr fert, Z = Fe+$+ fort + pent + cost 4, Find the general solution of the system of equations in Problem 49. Ans. ¥ = e+ opsint + ccost + Jt + ge-t Z = 1-esint — cycost — Je-t APPLICATIONS TO MECHANICS 55. Referring to Fig. 3-1, Page 79, suppose that ma damping forces are present. m has a force F(t), >0 acting on it but that no (a) Show that if the mass starts from rest at a distance X=a from the equilibrium position (X = 0), ‘then the displacement X at any time t>0 can be determined from the equation of motion mX" +X = Fl), XO)=a, XO) =0 where primes denote derivatives with respect to t. (®) Find X at any time if F(0) = Fy (a constant) for t > 0. (c) Find X at any time if F(t) = Fye—%t where a > 0. v(m (B) Poa Falk aE (0M — cos VE) + OPT sin Vem t 4 Ans, (bt) X ( X= at 86. Work Problem 56 if #(t) = Fo sinwt, treating the two cases: (a) o + VE/m, (b) « = Vk/m. Discuss the physical significance of each case. 51. A particle moves along a line so that ita displacement X from a fixed point O at any time ¢ is given by XM) + AX) + BX) = 80 inde CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 105 (a) If at ¢=0 the particle is at rest at X=0, find its displacement at any time t>0. (6) Find the amplitude, period and frequency of the motion after a long time. (0) Which term in the result of (a) is the transient term and which the steady-state term? (@ Is the motion overdamped, critically damped or damped oscillatory? Ans. (a) X(t) = 26-# (cost +7 sin#) — 2sin 5t + cost) (8) Amplitude = 2V2, period = 22/5, frequency = 6/27 (c) Transient term, 2e-% (cost +7 sin); steady-state term, —2(sin St + cos 5t) (@ Damped oscillatory 58, Suppose that at £=0, the mass m of Fig. 31, Page 79, is at rest at the equilibrium position X Suppose further that a force is suddenly applied to it so as to give it an instantaneous velocity Vo in a direction toward the right and that the force is then removed. Show that the displacement of the mass from the equilibrium position at any time t>0 is ' veltan E if there is no damping force, and ® Yee-mam where y= [2-2 if there is a damping force of magnitude X'(t) where p < 2Viem. 59. Work Problem 55 if: (a) F(t) = Fy Wt— 7), [Heaviside’s unit step function}; (6) F(t) = Fy8(t— 7) (Dirac delta function). Discuss the physical significance in each case. Ans. (a) X = oFycosvkimt if t7 (o) X = aFycosvi/mt if tT Suppose that at ¢=0 the mass m of Fig. 8-1, Page 79, is at rest at the equilibrium position and that a force F, 8(t) is applied. Find the displacement at any time ¢>0 if (a) the aystem is undamped, (0) the system is critically damped. Discuss the physical significance of each case. Py Pe Ana. (a) 72 sin Vint, (0) site-am 61, A ball of mass m is thrown upward from the earth’s surface with velocity Vp. Show that it will rise to a maximum height equal to V3/29, where g is the acceleration due to gravity. ‘A mass m moves along the « axis under the influence of a force which is proportional to its instan- taneous speed and in a direction opposite to the direction of motion. Assuming that at t= the particle is located at X=a and moving to the right with speed Vs, find the position where the mass comes to rest. 68. A particle moves in the zy plane so that its position (X,Y) at any time is given by xX" +4kY = 0, YY" + 16x 0 If at time ¢=0 the particle is released from rest at (a,5), find its position at any time ¢>0. dan x = (St) ve + (BE) wt vt y= (Zo) conve — (2M) coo VERGE Oe 106 APPLICATIONS TO DIFFERENTIAL EQUATIONS (CHAP. 3 APPLICATIONS TO ELECTRICAL CIRCUITS 64, A resistor of R ohms and a capacitor of C farads are con- neeted in series with a generator supplying B volts [see Fig. 3-17]. At t=0 the charge on the cap: Find the charge and current at any time ¢> 0 a constant; (5) B= Eye-, a >0. R c Ans. (a) Q = CEg(l~ eR), I = (Eg/R)e-t/Rc CB (eat — en tine 0) @ = Eee » 1 = ob ee Fig. 3-17 at at) if «# URC 6, Work Problem 64 if F = By sinat snd the initial charge on the eapacitor is Qy An. Q = £0. + gp tetame etn — Be {const nat} 1 = dQiat 0 RFF TREC | ® , 66, An inductor of L henrys and a capacitor of C farads are in series with a generator of E volts. At t the charge on the capacitor and current in the uit are zero. Find the charge on the capacitor at any time t>0 if: (a) B= Eq, a constant; (b) B= Eye-*, @ >0. Ans. (a) Q = CE {1 ~cos(t/VEG)} Ey () @ = Terr TEG aBy e+e {e-# — cos (t/VEC)} + in (t/VLC) 61, Work Problem 66 if # = Ey sinet, discussing the cases (a) «VEC and (b) w= W/VEC and explaining the physical si 68, Work Problem 66 if E(t) is (a) Eq Uta) where U(t— a) is Heaviside’s unit step fonction, (6) By 8(t) where 6(t) is the Dirac delta function. Ans. (a) Q=0 if ta (&) Q = EyVCTE sin (VLC) 69, An inductor of 3 henrys is in series with a resistor of 80 ohms and an e.m4. of 150 volts. Assuming that at t=0 the current is zero, find the current at any time t>0. Ans. I = B(1—e=10) 7. Work Problem 69 if the em: jn 20t — 2 cos 20¢ + 2e~10" is given by 150 sin20t. = Ana. I = 1. Find the charge on the capacitor and the current in the circuit (Fig. 3-18] at any time t after the key K is closed at t=0. Assume that L,i,C and B are con- stants and that the charge and current are zero at t=0, Treat all cas 72. (a) Work Problem 71 if Z = Ep sin ot. (b) Show that 1_ resonance occurs if we choose o (6) Discuss the case Fig. 3-18 7B. An electric circuit consists of an inductor of L henrys in series with a capacitor of C farads, Att an emf. given by ae spe oT is applied. Assuming that the current and charge on the capacitor are zero at t= at any time t>0. , find the charge CHAP. 3] APPLICATIONS TO DIFFERENTIAL EQUATIONS 107 Ans. Q = Fiu- VLC sin(t/VLC)} if 0 0. Ans. Q = sin 10t — 2 cos 10 + e7 1" (sin 10t + 2 cos 10t) Fig. 3-19 APPLICATIONS TO BEAMS 75. A beam which is clamped at its ends x=0 and 2=I carries a uniform load W, per unit length. Show Wo atl aye that the deflection at any point is ¥(#) = —“yy7— 1%. Work Problem 76 if the end «=0 is clamped while the end «=I is hinged. TT. A cantilever beam, clamped at 2=0 and free at #=I, carries a uniform load We per unit length. Wo 2? Show that the deflection is Y(x) = pyje7 (2? — dle + 62), 78. A beam whose ends are hinged at #=0 and x=l has a load given by 0 o0 of the boundary-value problem ov ay ees ¥(@,0) = @ Ans, Y(x,t) = #+1—er# 85, Solve the equation . #y _ 2 oe = FE a> otro subject to the conditions ¥(0,t) = 10sin2t, ¥(#,0) = 0, ¥, (2,0) 0, lim ¥(e,t) CHAP. 3) APPLICATIONS TO DIFFERENTIAL EQUATIONS 109 MISCELLANEOUS PROBLEMS 96. Show that the solution of the differential equation yo — BY@ Fe) subject to Y(0) =a, ¥(0)=8 is YO = aeoshkt + (b/k) sinh ke + Ef” Flu) sinh Mew) du 0 M. Solve YH) 4¥") = Bsine, ¥(0)=¥"0)=0, ¥"O)=1, ¥™%0) = Ans, Y = @—2+e-t+sint + cost 98, Find the general solution of the differential equation of Problem 45, ans v= ete f Goats ot 99, Find that solution of the equation 1" — (4 9¥48¥ = t-1 which has a Laplace transform and is such that Y(0) = 0. 100, What is the general solution of the differential equation in Problem 99? 101, (a) Use Laplace transforms to show that the solution of Y Get Y = Acoset, ¥(0)=«, YO)=8 is Y() = KY) + a coskt + (A/k) sin kt. ee (®) Give a physical interpretation of the results of part (a). Ey = ae 102, Solve for X: 4 ¥/4Z' = X+Z if X(0)=2 YO)= 3, 2(0) Xa X+Y¥ Ans, X = ge-t2{3 cos (Vi t/2) — 2V3 sin (V3t/2)} 108. Work Problem 29 by letting Y = VY,, where V is a new dependent variable. 104, Can the method of Laplace transforms be used to find the general solution of Y"+¥ = sect Explain, 105. (a) Find a bounded solution of (t-¥" + (Any — ay = 0 such that ¥(0) (b) What is the general solution of the equation in (a)? ‘Ana, (a) ¥ = Bett, (6) ¥ = eet f ae + eget 110 APPLICATIONS TO DIFFERENTIAL EQUATIONS [CHAP. 3 106. (a) Show that satisfies the differential equation al G- 1(0) = =/2. (®) By solving the differential equation in (a), show that Mt) = Feterteve 107. A particle moving on a straight line (the 7 axis) is acted upon by a force of repulsion which is pro- portional to its instantancous distance from a fixed point O on the line. If the particle is placed at a distance a from O and is given a velocity toward O of magnitude Vp, find the distance of closest approach to 0. 108. If the ball of Problem 61 encounters air resistance proportional to its instantaneous velocity, show that the maximum height reached is Th (Ve + mg — keg) — ™E FateVy + mg — kg) — EP where & is a constant of proportionality. 109, In the cireuit of Fig. 8-18, Page 106, suppose that the emf. H is a function of ¢ while L,R and C are constants. At the instant ¢ = 0 that the key K is closed, assume that the charge Q on the capacitor and current 7 are zero, Show that if R? < 4L/C, then the current at any time ¢>0 is given by 1¢ Rush, R MW) = 5 S E(t—w) em + (cone -£= where a@ = VI/LC — R*/4L2, 10, Work Problem 109 if (a) R® = 4L/C, (b) R® > 4Lic. in oo) du M1, Present a mechanical analog to (a) Problem 64, (6) Problem 66, (c) Problem 71. 112. Give an electrical analog to (a) Problem 55, (®) Problem 57. FO 118. Give 2 mechanical analog to Problem 74 involy- ing masses connected by springs. 114. A particle of mass m moves along the x axis under the influence of a force ¥(t) as indicated Sa t in Fig. 8-20. If the particle starts from rest at ¢=0 determine its position and speed at any time £>0, Fig. 3-20 115. A beam which is clamped at 2=0 and @ a where O0 , they are given by A Bly _/ est — on! Ez * rete @— a2 R p EM ( =e ) > Tyla MP ag—ay where ; and az are the roots of the equation (Lib, — Mat + (LyRy + LgRi)a + RyRy = 0 120, Discuss Problem 119 if LyL, = M%. Chapter 4 INTEGRAL EQUATIONS An integral equation is an equation having the form Y(t) = F(t) + S K(u,t) ¥(u) du (1) where F(t) and K(u,t) are known, a and 0 are either given constants or functions of t, and the function ¥(t) which appears under the integral sign is to be determined. The function K(u, t) is often called the kernel of the integral equation. If a and } are constants, the equation is often called a Fredholm integral equation. If a is a constant while b=t, it is called a Volterra integral equation. It is possible to convert a linear differential equation into an integral equation. See Problems 1-3 and 25. INTEGRAL EQUATIONS OF CONVOLUTION TYPE A special integral equation of importance in applications is Y(t) = Plt) + S K(t~w) Y(u) du @ This equation is of convolution type and can be written as YQ) = Ft) + K)*¥(t) ‘Taking the Laplace transform of both sides, assuming £ (F(t)} = f(s) and ¢(K(t)) = k(s) both exist, we find W@) = 1) + He)ve) or 4(e) = LO. ‘The required solution may then be found by inversion. See Problems 6 and 6. 112 CHAP. 4! APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS 113 ABEL'S INTEGRAL EQUATION. THE TAUTOCHRONE PROBLEM. An important integral equation of convolution type is Abel’s integral equation Si where G(t) is given and @ is a constant such that 0 7H Pes fa-areol + firma} + § w-9 ie, cad bee ¥® — ¥0) + a-H¥O — Yo) + fF vwdut [evade = 2% () — YO) + A-H¥® — YO) S (w) S we) 77% ‘ * cal bee or Y(t) + 1- Y(t) + Y(u)du + eeY(u)du = F- > +1 Sime + J m3 Another integration from 0 to t yields q 6 58 YQ - YO) + f G-wvwdu t+ Ff e-w Vaan t+ [ t-wervwan = £- Bre J J J af which can be written rH * ., _ 8 Be (t) + (t+t— ut (t—ue} ¥(u)du = woe tt7s O 3. Express as an integral equation the differential equation Y(t) — AY"(t) + 6Y"(t) — AY) + Y(t) = 8 cost subject to the conditions Y(0)=~-1, Y(0)=4, ¥"(0)=0, ¥"(0)=2. Method 1. Let Y(t) = V(t). Then as in Problems 1 and 2, we find yn = Sf vandet2, yy = fC Ve) du + 2 ° yy = S « we Voojdn + Se ara vod sery ry = fl 5 116 APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS (CHAP. 4 Substituting these into the given differential equation, it becomes Vit) = 25 — 16t + 4 — fo + Seosat + f {4 = 6{t—u) + 2(t— mw)? — ftw) Ven) de ively from 0 to t as in the second methods of Problems 1 and 2, we find the integral equation y(t) -S (4 6(¢—u) +2¢—we- Ye-} Vegdu = - 12 + oe — BE 4 50 + Zcosae ‘These integral equations, as well as those obtained in Problems 1 and 2, are Volterra integral ‘equations; the limits of integration are from 0 to t. In general this type of integral equation arises from linear differential equations where conditions are specified at one point. For an example of a Fredholm integral equation which arises from linear differential equations in wl i specified ut two points, see Problem 25. 4, Convert the integral equation Y(t) = 8t-4-2sint + S (tw)? — Btu) +2) Yu) du ” into a differential equation. We make use of Leibnitz’s rule, ac vee) ae a ee aS, Knee Spdu + KN, OS — Kole), OFF a ‘Thus we have on differentiating both sides of the given integral equation, YQ = 8 = Qeost + f 2t-w rendu - 3° Ya) du + 27H «) , Another differentiation yields, yr" 2sint + af” Y(uydu — s¥(t) + 2¥') O) and a final differentiation yields the required differential equation yt) = 2eost + 2¥() — 8¥() + 2¥%G) wo or y" — ay" + 8Y' — 2¥ = 2eost ‘The initial conditions obtained by letting ¢= (2) and (8), are in the given integral equation and also in equations Y0) = -4, ¥@) = -% YO) = —2 ‘Note that the initial conditions are contained in the integral equation. It is possible to convert every linear differential equation into an integral equation, However, not every integral equation can be converted into a differential equation, as, for example, Yo = cost + f Infuto Yu du 0 CHAP. 4] APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS 117 INTEGRAL EQUATIONS OF CONVOLUTION TYPE 5. Solve the integral equation Y(t) = # + S ¥(u) sin (t—u) du. . The integral equation can be written YO = @+ Y@*sine ‘Then taking the Laplace transform and using the convolution theorem, we find, if y= £{¥), 2 v vie gta solving, £ + & and 50 = e+e ‘This can be checked by direct substitution in the integral equation. 6. Solve the integral equation S ¥(u) ¥(t—u) du = 16sin4t. 7 The equation can be written as Y@*Y) = 16sinat Taking the Laplace transform, we find Wor = am uo = TS Then YQ = £-1{y(9} = *8J(40) Thus Y(t) = 8Jo(4t) and Y(t) = —8Jo(4t) are both solutions. ABEL'S INTEGRAL EQUATION. THE TAUTOCHRONE PROBLEM. 7. Solve Ske = l+tte 0 Vt=u The equation can be written yore = 1+e+e Then taking the Laplace transform, we find AY) LE = HET BY or and v 118 APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS (CHAP. 4 - _ a few, om | pam Inverting, Y= yay {am + rer * feat . = Lew + 20 + form) = HP erer+se The integral equation is a special case of Abel's integral equation. 8. A bead is constrained to move on a fric- tionless wire which lies in a vertical plane. If the particle starts from rest at any point of the wire and falls under the influence of gravity, find the time of descent to the low- est point of the wire. Assume that the bead has mass m and starts from rest at point P with coordinates (u,v) as shown in Fig. 4-1. Let point Q, having coordinates (,y), be some intermediate point in the motion and suppose that the lowest point of the wire is taken to be the origin O. Let e be the are length Fig 41 0Q. From the conservation of energy, we have Potential energy at P + Kinetic energy at P = Potential energy at Q + Kinetic energy at @ do? yoy + 4m (#) where do/dt is the instantaneous speed of the particle at Q. Then (ay = %v-» or using the fact that o decreases as time ¢ increases, = —vie~— 9) C2) nov +0 The total time 7 taken for the bead to go from P to O is given by 7 , e Soe = Soe > Sees When the shape of the curve is given, the are length can be expressed in terms of y and we find @ do = Fly)dy (% ef fins i Seuss ® In general T is a function of v, ic. of the starting position. Thus (2) becomes Tr 9. Find the shape which the wire of Problem 8 must have if the time taken to reach the lowest point is a constant, i.e. is independent of the starting position. In this case we have to find F(y) such that = FW) = Lf fo, ie S wa” @ CHAP. 4] APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS 119 where T is a constant. This integral equation of convolution type is a special case of Abel's integral equation [see Page 113] and can be written vigT = Fw) ty? (2) Taking Laplace transforms and noting that £ (F(y)} = f(s), ¢{v-"/*) = r(p)/s'? = Va/s?, we have y2gT ve _ yi 7 = Og or) = The inverse Laplace transform is given by . = TV% [1] ~ Tye _ Tvi, Py = Es {a} = Te ram = Since we have ® If we let ) (8) can be written since the slope must be positive. From this we find on integrating, a= feta te 6) Letting y = bsin®¢, this can be written - [f= B sin? 6 = 2 f costeds +o = bf C+eos20d0 +o = F(2e+sinds) + 0 2b sine cose de + © ‘Thus the parametric equations of the required curve are $ (20+ sin2e) +0, y = Bsints = $(1—cos2e) Since the curve must pass through the point «=0, y=0, we have Then letting the parametric equations are 2 =a(g+sing), y = a(1—cos¢) ‘These are the parametric equations of a cycloid (see Fig. 42 below]. For a given constant T, the wire has the shape of the curve shown heavy in the figure. The cycloid is the path taken by a fixed point on a circle as it rolls along a given line [see Problem 44]. 120 APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS (CHAP. 4 Fig. 4-2 INTEGRO-DIFFERENTIAL EQUATIONS 10. Solve ¥“(t) +5 f" cos 2(t—u) ¥(u) du = 10 if ‘The equation can be written Y(t) + Beos2t*¥(e) = 10 Then taking the Laplace transform, we find 5 10 ay — ¥0) + fe eta 258+ 100 + 8s + 40 or y= a(8? + 9) ‘Hence by Problem 44, Page 67, Y HA + 120¢ + 80.con se + 50 sinse) Note that by integration from 0 to ¢ using ¥(0) = converted into the integral equation , the given integro-differential equation can be Yi + 5S (tu) cos2(e—w) Yew) du = 10 +2 ° DIFFERENCE AND DIFFERENTIAL-DIFFERENCE EQUATIONS IL Solve 8Y(t) - 4¥(t-1) + Y(t-2) = ¢ if ¥()=0 for ta Hef) = {s et ‘Thus by Problem 15, for nStt The integral equations obtained here are examples of a Fredholm integral equation with a sym- metric kernel. 180 APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS (CHAP. 4 Supplementary Problems INTEGRAL EQUATIONS Convert each of the following differential equations into integral equations. m8. Y(t) + BV") — 8¥(t) = 5-34, (0) =-2, YO) =3. Ans. V(t) + f 2-8t+8u) Viu) du = 5+ 2t—22, ViQ=Y"H 0 or YC) +f (2-84 8u) ¥(u) du = —2— e+ sea 8 ° 2M, BY") — BY") — AV) = det + 2eost, YC) =4, YO) =-1. Ans, 2V(0) + f° (2u—2t—3) Vow) du = det + 2eost+ 5-2 VID= YH A oe m1 + femme du = 6106+ wet Beat 2. Y(t) + 8Y() = 3sint + 2cost, ¥(0)=0, ¥(0)=-1, ¥"(0)=2. Ane. V(t) +4 { (tut V(u) du = Bsint+2eost~ae+ 4, Y= Y(t) or Y(t) + S (tw? ¥(u) du = 50/2 + t—8 + Beost — Qeint 2. YM + cost YM) = et 0) =—2 Y4O)=0. Ans. VQ + f (tw) cont Va) du = ot + 2e08%, vw =¥"H or YO + fF e-w cou Yew du = t- 3 bert mY") —tYO@+ EY = 14% YO=4, YO) =2 Ans. v(t f° (@= tut Van du = 14 91-4828, VY = YH or YO -f (t—2u+ tut—ud) Yu) du = 8/2 + H/6 + Ot + 4 o a. ¥) — 2H + -A)¥() = 14+ 4t~ 20+ 4, ¥(0)=1, YO) =0, YO) =-2, YO) Ane. Vit) + S {A(t—w) (1) - 2e(t—w} Ve) du = 0, Vi) = ¥™(t) oY — f° ule—a) + 2(¢-we + Hew — a) Kew de eo ee : 8 e1-e@4%48 £,4 1-8 + y+ a4 — i80 * F680 Convert each of the following integral equations into differential equations and asso- ciated conditions. a ¥() = seost + ff (t—u) ¥(u) de Ans. ¥"(t)—¥() = —Bsint, ¥(0)=5, ¥(0) =0 CHAP. 4] APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS 131 BY) = ea att ea few YE) dw 0 Ans. Y(t) +6¥(t) = 0, ¥(0)=4, YO) =—8, ¥"(0) =2 aX) + f {(t— wt + 4(t- w) — 8} ¥(u) du 0 Ans, Y(t) ~ 3") + 4H + 2¥() = —e-!, YO)=1, ¥"0)=2, ¥"(0) =8 35. Yo - [ @-w oe ¥en au =t ¥'0) Ans. YM"(t)— 2 tant ¥"(® — (1+ sect) ¥() = —t—2tant, ¥(0) 36. vos f (24 dt ut—u—2) ¥(u) du = 0 5 Ans, Y(t) + (8t~2) Y(t) + (+10) Y(H + YC) , ¥(0) = 0, ¥(0) =0, YO) =0 INTEGRAL EQUATIONS OF CONVOLUTION TYPE 3. Solve Y(t) = c+af cos (¢~ x) ¥(a) du. 0 Ans. YQ) = t+2+ 2(t—let 38, (a) Show that the integral equation Yt) = t+ (t—u)§ ¥(u) di 7 + S rH has solution Y(0) = (sine + sinh, (®) Is the solution in (a) unique? Explain, 89. Find the continuous solution of the integral equation Ans. Y(t) =1 Yu) ¥(t-u) du = 2¥() + t-2 40. Show that the only solution of the integral equation S Y(u) sin (tu) du = Y(t) is the trivial solution Y(t) = 0. 41, Discuss the solutions of the integral equation S Yu) Ge-u) du = ¥(. O ABEL'S INTEGRAL EQUATION AND THE TAUTOCHRONE PROBLEM 42, Solve the intogral equation = Ve Ans. YQ=4 48, Show that the solution of the integral equation f” Peta = a4) ie 8 wes, 182 APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS [CHAP. 4 44, A circular wheel of radius a [sce Fig. 4-8) rolls on a straight line, taken to be the « that a fixed point 0” on its rim, or contact with the line at O describes the cycloid @ = a(y—sing), y = a(1—cosg) shown dashed in Fig. 4-3, 45, Prove that the curve in the tautochrone prob- Jem, Page 118, is a eycloid and discuss the rela- tionship to the curve of Problem 44, Fig. 43 46, Show that the time required for the bead of Problems 8 and 9 to slide from the top P of the wire to ‘the bottom O [lowest point on the eycloid] is »Va/g. YW) du = F(, assuming F(0) =0, is 41, If 0 Gu) yy) = snot ("ry (tet du 48. Discuss the solutions of the integral equation in Problem 47 if F(0) %0. Illustrate your remarks by considering YW) a, = (em = te INTEGRO-DIFFERENTIAL EQUATIONS 49, Solve Sf ¥(u) cos(t—a) du = YQ) if ¥(0)=1. Ang, ¥() = 1+48 5a. Save f' YW) Vt—w du = m4 it YO) Ans. Y(t) = +160%/yz 51. (a) Show that the integral equation of Problem 49 can be expressed as the integral equation 1+ J (t—w Yu) cos(t—a) du = YH (b) Solve the integral equation in (a). ¥'0) 52 Solve SY) VUE-w du = YQ - YOY it YO) Ana. Y(t) =0 DIFFERENCE AND DIFFERENTIAL-DIFFERENCE EQUATIONS 53. Solve Y(t) — 8Y(¢—1) + 2¥(t-2) = 1 if Yi) =0, #<0. Ans, Y(t) = 2042 — []—8 CHAP. 4) APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS 183 54. Show that the solution of Y(t) = 2¥(¢-1) +e if Y(=0, t<0 is yo = Game —njat? =o (wr 2)t 55, Solve ¥"(t)—¥(t—1) = F(t) where Y(t)=0, ¥(f)=0 for #50, and Fr) = { ts0 at t>o g _ (wanes Ans YO) = 2 2 Gea 56, Solve 8Y(t) — 5Y(t—1) + 2¥(t—2) = F() if ¥()=0, t<0, and ry = {0 #<0 @ t>0 w Ans. YQ) = BL — pyrene mt 57. Solve the difference equations (0) 80,42 Sayer + 20y = 0 if ay =1, «=O. 0) tuse+ Zager Bay = 0 if ap =0, y= 1. Ans. (a) 3(2/3)"— 2, (6) 3{1 — (—3)"} 58 The Fibonacci numbers do. dy,d,... are defined by the relation Gus. = Gayi+ a, where ay=0, @,=1. (a) Find the first ten Fibonacci numbers. (6) Find a formula for a,. An (0) 3,1,2,85,8,28,218 0) a = L{(HYS)- (EB) 59. Solve the equation a,,2—4a,4;+ 4a, = 0 where ay=1, a=4. Ans. ay = 2Mn+1) 60. Solve the equation ayyq— 20,4; +20, = 0 where a Ans. a, = ((149"— (192i G1. (@) Solve Gyyy—20y43—Oyyy +20, = 0 if t=0, 0,1, q=1. (b) Find ay. Ans. (a) ay = Y{2"—(-1)"}, (8) ay = 841 62, (a) Show how a solution to aay — 6,41 + 8d, = 0 can be obtained by assuming a, =r" where 7 is an unknown constant, (6) Use this method to solve Problems 57-61. MISCELLANEOUS PROBLEMS 63, Show that the non-linear differential equation ¥"@) + {VO} = teint, — ¥()=1, YO) =-1 ‘ean be written as the integral equation vo + Sew era ae 3 — t — 2eost — taint ‘ 64. Salve f YQ ¥it—w du = a¥(O + ye — 26 Ans, Y(t) =¢ or Y(t) = 20()-t 184 APPLICATIONS TO INTEGRAL AND DIFFERENCE EQUATIONS [cHaP. 4 65, Express as an integral equation: Y(t) — Y(t) = Seost -sint, Y¥(r)=1, ¥(2)=~2 Ans, V() = 2 +124 Bont —sint + f° (t—w) Viw) du, where Vie) = Ye 65. Solve YQ = 6+ f Ye A(e-w) du An. ¥() = O41) f" Jou du + Heol — JeIKO) J, 61. Find G(2) such that Glu) Gle—u) du = 8{ein x — x cos). Ans. G(x) = +4 sine 68. Solve f ¥(w) ¥(t—u) du = t+ 2¥(t). Ans. ¥) = AO-f'Hwdu or YQ = 23-HO+ Sf Iwan 0 ° 69. Solve the following difference equations using Laplace transform methods. (@) yso— 5aysy+ 6a, = 2n+1, a9 =0, y= 1. (8) Oyy24 Mity44— 5a, = 2-8, y= 3, Ans, (a) a, = $3"— 5+ n+h (6) ay = Bn? —dn+ 2+ (—H)* ]. Solve (a) Ouén + 2tysy tay = R42 Gy =0, Gy =0. () nse — 6ayy1 + Say = 2% a= Ans. (a) ay = 4(8n— 10-1)" + Het 1) (6) ay = E+ gy — ed a = 0, TL. Solve Gyss~ 2ys2— Ouest 2d, = n2+2%, Go =0, a =1, a =1, Ans. Oy = f+ fm — gd + fneae — Beer — Ye 72, (a) Show how a particular solution to Problem 69(a) can be found by assuming a, = A+n where A and B are unknown constants, (b) Using the result of part (a) and the method of Problem 62, show how to obtain the solution of Problem 69(a). (c) How can the method indicated in parts (a) and (6) be revised to enable solution of Problems 69(6), 70(a), 70(6) and 71. 18, Find all continuous fanetions F(®) for which f° wF(u) con (¢—w) du = tovt ~ aint. Ans. F(t) = ~—2e-* . 7A. Show that the non-linear differential equation yr +2 = YD, YO) =0, Ya) =0 can be written as the integral equations yy = S (2-2) ¥(u) du + S20 au + f Kew vw du 7 oro = Si eosnee-o ri ae — [nemo ranae + "eRe YH de ult-1) u0, there exists a 8>0 such that |f(z)—I| << whenever 0 < |z—‘| < 8 Similarly, f(z) is said to be continuous at 2 if, given any «>0, there exists a §>0 such that. |f(2)—f(eo)| < « whenever |z—zsl < 8. Alternatively, f(z) is continuous at zo if lim f(z) = f(z). me DERIVATIVES If f(2) is single-valued in some region of the 2 plane the derivative of f(z), denoted by #'(@), is defined as fle + 2) ~ f@) a ae an provided the limit exists independent of the manner in which Az>0. If the limit (12) exists for z= zo, then f(2) is called differentiable at zo. If the limit exists for all z such that Jez < 8 for some §>0, then f(2) is called analytic at 2. If the limit exists for all z in a region &, then f(z) is called analytic in ®. In order to be analytic, f(z) must be single- valued and continuous. The converse, however, is not necessarily true. We define elementary functions of a complex variable by a natural extension of the corresponding functions of a real variable. Where series expansions for real functions f(x) exist, we can use as definition the series with x replaced by 2. eB sw Example 1. -H+8-fee, From these we can show that e* = ett = e#(cosy + isiny), as well as numerous other relations. Example 2. We define a? as ¢>!4 even when a and 6 are complex numbers. Since e%*'=1, it fol- lows that o! = et(@+2=) and we define Inz = In (re) = Inr + i(9+2ke), Thus Inz is a many-valued function. ‘The various single-valued functions of which this many- valued function is composed are called its branches. CHAP. 5) COMPLEX VARIABLE THEORY 139 Rules for differentiating functions of a complex variable are much the same as for those of real variables. Thus te = ner, Atein 2) = cose, ete. CAUCHY-RIEMANN EQUATIONS A necessary condition that w = f(z) = u(z,y) +év(z,y) be analytic in a region R is that u and 0 satisfy the Cauchy-Riemann equations ou _ av ua ae = yy = Tae (12) (see Problem 12). If the partial derivatives in (12) are continuous in &, the equations are sufficient conditions that f(z) be analytic in R. If the second derivatives of u and v with respect to z and y exist and are continuous, we find by differentiating (12) that ou, Fu ev | ey get = % set oge = 0 (18) Thus the real and imaginary parts satisfy Laplace’s equation in two dimensions. Func- tions satisfying Laplace’s equation are called harmonic functions. LINE INTEGRALS Let C be a curve in the zy plane joining points (21,41) and (22, ys). The integral ton) SPdz+Qdy or Pdz + Qdy c um) where P and Q are functions of z and y, is called a line integral along curve C. This is a generalization of the integral of elementary calculus to curves. As in elementary calculus it can be defined as the limit of a sum. Two important properties of line integrals are: coun Guu 1 Pde+Qdy = — Pdz + Qdy y) Gy 2. If (s,ys) is any other point on C, then or tsa) enn SO Par + ody SO" Paw + Qdy + Pde + Qdy itn) en) GoW If C is a simple closed curve (one which does not cross itself anywhere) as in Fig. 5-8, the line integral around C, traversed in the positive or counterclockwise direction, is de- noted by § Pas + Qdy c For evaluation of line integrals, see Problem 15. 140 COMPLEX VARIABLE THEORY (CHAP. 5 GREEN’S THEOREM IN THE PLANE Let C be a simple closed curve bounding a re- gion R [see Fig. 5-3]. Suppose that P,@ and their first partial derivatives with respect to x and y are continuous in R and on C. Then we have § Par + Ody = J (2 &) avay which is often called Green’s theorem in the plane. INTEGRALS If /(2) is defined, single-valued and continuous in a region R, we define the integral of fle) along some path C in ® from point 2: to point 2, where 2: = 21+ iyi, 2 = 22+ iy, as a uno tonto See = L°" wrionds rian = f°" ude — ody + if" ode + nay With this definition the integral of a function of a complex variable can be made to depend on line integrals. An alternative definition based on the limit of a sum, as for functions of a real variable, can also be formulated and turns out to be equivalent to the one above. The rules for complex integration are similar to those for real integrals. An im- portant result is IS. fe) az| f if@)ilde] = MS as = ML (14) where M is an upper bound of |/(2)| on C, ie. |/(2)| SM, and L is the length of the path C. CAUCHY’S THEOREM Let C be a simple closed curve. If f(z) is analytic within the region bounded by C as well as on C, then we have Cauchy's theorem that § fladz = 0 (15) See Problem 19. . Expressed in another way, (15) is equivalent to the statement that S . f(z)dz has a value independent of the path joining 2: and 2. Such integrals can be evaluated as Fe) — F(z) where F’(2) = f(@). Example. Since f(s) =2= is analytic everywhere, we have for any simple closed curve C G de = 0 145 1+ Also, foo ue = 2 ls = (+o -@n = +4 CHAP. 5] COMPLEX VARIABLE THEORY 141 CAUCHY’S INTEGRAL FORMULAS If f(z) is analytic within and on a simple closed curve C and a is any point interior to C, then fe 7 ) fe) = 35 mi § £2. de (16) where C is traversed in the positive ee sense. Also, the nth derivative of f(z) at z=a is given by on = fw F(2) f(a) = 25 pln ae a ‘These are called Cauchy's integral formulas. They are quite remarkable because they show that if the function f(z) is known on the closed curve C then it is also known within C, and the various derivatives at points within C can be calculated. Thus if a function of a complex variable has a first derivative, it has all higher derivatives as well. This of course is not necessarily true for functions of real variables. TAYLOR’S SERIES Let f(z) be analytic inside and on a circle having its center at =a. Then for all points # in the circle we have the Taylor series representation of f(2) given by fe) = fa) + rea) + LO e— ae + Me aye + (18) See Problem 29. SINGULAR POINTS A singular point of a function f(z) is a value of z at which f(z) fails to be analytic. If f(2) is analytic everywhere in some region except at an interior point z=a, we call z=a an isolated singularity of f(2). 1 Example. If f(z) = ey then z= 8 is an isolated singularity of f(z). POLES fe) =7 2=a, and if n is a positive integer, then f(z) has an isolated singularity at z=a which is called a pole of order n. If n=1, the pole is often called a simple pole; if n=2 it is called a double pole, ete. (2) ye #(a) #0, where ¢(2) is analytic everywhere in a region including Example 1. /(2) = G—aig gay has two singularities: a pole of order 2 or double pole at = and a pole of order 1 or simple pole at = = 3=1 _ 32-1 __ has two sim jes at 2 = 224 Example 2. f(e) = $54 = Gy Fag gy has to simple poles at 25. ‘A function can have other types of singularities besides poles. For example, f(z) = has a branch point at z=0 (see Problem 45). The function f(z) = sing has a singularity at 2=0. However, due to the fact that lim “22 ig finite, we call facet singularity a removable singularity. me 142 COMPLEX VARIABLE THEORY [oHAP. 5 LAURENT’S SERIES If f(z) has a pole of order n at z=a but is analytic at every other point inside and on a circle C with center at a, then (z—a)*f(z) is analytic at all points inside and on C and has a Taylor series about z= a(z—a)t as a Laurent series where the terms with k <0 constitute the principal part. A function which is analytic in a region bounded by two concentric circles having center at z=a can always be expanded into such a Laurent series (see Problem 119). It is possible to define various types of singularities of a function f(z) from its Laurent series. For example, when the principal part of a Laurent series has a finite number of terms and a 10 while a-y-1,@-»-2,... are all zero, then z=a is a pole of order n, Tf the principal part has infinitely many non-zero terms, z=a is called an essential singularity or sometimes a pole of infinite order. Example, The function e@ = 14449154 +++ has an essential singularity at 2 = 0. RESIDUES The coefficients in (19) can be obtained in the customary manner by writing the coeffi- cients for the Taylor series corresponding to (z—a)*/(z). In further developments, the coefficient a-1, called the residue of f(z) at the pole z=a, is of considerable importance, It can be found from the formula a w= praes (- 91) (20) ay = lim where m is the order of the pole. For simple poles the calculation of the residue is of particular simplicity since it reduces to . a1 = lim(e~a)fie (22) RESIDUE THEOREM If f(e) is analytic in a region R except for a pole of order n at z=a and if C is any simple closed curve in R containing z=a, then f(z) has the form (19). Integrating (19), using the fact that a _ (fo ifnat §. @-ay {on ifn=1 (2) (see Problem 21), it follows that § feyae = Qria-1 (23) i.e, the integral of f(z) around a closed path enclosing a single pole of f(2) is 2ri times the residue at the pole. CHAP. 5} COMPLEX VARIABLE THEORY 143 More generally, we have the following important Theorem. If f(z) is analytic within and on the boundary C of a region R except at a finite number of poles a,b,c, ... within R, having residues a-1,b-s,¢-1,... respectively, then § fod = @riar+ brteat--) (24) ie. the integral of f(z) is 2xi times the sum of the residues of f(z) at the poles enclosed by C. Cauchy’s theorem and integral formulas are special cases of this result which we call the residue theorem. EVALUATION OF DEFINITE INTEGRALS The evaluation of various definite integrals can often be achieved by using the residue theorem together with a suitable function f(z) and a suitable path or contour C, the choice of which may require great ingenuity. The following types are most common in practice. S F(a) dz, F(a) is an even function. 5 Consider § F(z) dz along a contour C consisting of the line along the « axis from —R to +R and the semi-circle above the 2 axis having this line as diameter. Then let R+«, See Problems 37, 38. 2. S G(sin 6, cos 6) d0, @ is a rational function of sin@ and cosé. 0 Bae ogg = ete ’ a} let z=e%. Then sing and dz=ieda or da= 2 dafiz. The given integral is equivalent to § F(z)dz where C is the unit circle with center at the origin. See Problems 39, 40. 3 i Fe {ogame ae, F(a) is a rational function. Here we consider f° F(z)e™dz where C is the same contour as that in Type 1. See Problem 42. bi 4. Miscellaneous integrals involving particular contours. See Problems 43, 46. 144 COMPLEX VARIABLE THEORY (onar. 5 Solved Problems COMPLEX NUMBERS 1. Perform the indicated operations. (a) (4-2) + (645) = 4-21-6451 = 4-64 (245i (6) (“7+ 38) — (2-4) = -74+ 81-2440 9+ : (e) (83-2143) = 31 +3i) — 201 +3) = 34+ 9i- 21-6? = 34+ 91-2+6 = 9471 243i +38i _ (-5+5i)4+8i) _ —20~15i+ 201+ 157% a¥Bi 1~9% = +O = Se14) 2 td, = 8 3 iPr Petri 1+ + @P+ PPE it 14i ) 1 =6i -3 © lH sé =5 2. If 2 and z are two complex numbers, prove that |z:z=| = |e:| zs). Let 2) = 2+ ivy, % = tyt iv, Then Vestal = [Coy ivsdeee tive) | = Leite — vive + ewe t+ end | = V@m— ne tent = Vat it pis ae = VeRT HEF = VERVE = letiviie tive! = teal lel 8. Solve 2°-22~4 = 0. The possible rational roots are 1,2, +4. By trial we find 2=2 is a root. Then the given equa- tion can be written (z—2)(2*+2z+2) = 0. The solutions to the quadratic equation azt+bz+e = 0 b+ Tae wae Vi=8 2+ Va are = EVOE EE ay gat gen oo2 whe ghee 2 = EYE a arr “i 3 Lei ‘The set of solutions is 2, -1+4, -1-i. POLAR FORM OF COMPLEX NUMBERS 4. Express in polar form (a) 8 + 3i, (b) -1+ Bi, (c)—1, (d) -2 — 2V3i. [See Fig. 5-4.) VFR = 3¥2 Then 3+ Bi = r(cose t+ ising) = 3V2(cose/4 +isinz/4) = SV2cise/4 = syZer/ (2) Amplitude @ = 45° = 1/4 radians. Modulus r (6) Amplitude ¢ = 120° = 27/3 radians. Modulus r = (1)? + (V5) = Va = 2. Then -1 + vie (cos 2e/3 + isin2x/3) = 2eis2x/3 = gens CHAP. 5] COMPLEX VARIABLE THEORY 145 Fig.5-4 () Amplitude @ = 180° = radians. Modulusr = Vii)? + OF = 1. Then -1 = loose + isinz) = cise = ot (@ Amplitude ¢ = 240° = 4¢/8 radians. Modulus r =, 4{(-2)? + (-2VB)? 4. Then 2 — 2VB = Alcosdr/3 + isindy/3) = Acis4x/3 = detms 5. Evaluate (a) (-1+ VBi)", (b) (1+. (@) By Problem 4(b) and De Moivre’s theorem, (1+ VB! = [2(co2x/3 + isin 2x/3)}! = 230(cos 20x/3 + i sin 2/2) = 102A [cos (2r/8 + 6s) + isin (2x/B + 6x)] = 1024(cos2e/3 + isin 25/8) = 102-g+ yVFQ = 12 + stays “() -1+% = V2(cos 135° + isin 135°) = V2 [cos (135° + k+360°) + isin (135° + k+360°)] Then Catone = f8ye[ an (82+ Eat) + i (188 + este) 3 The results for k=0,1,2 are VB (cos 45° + isin 46°), Vii (cos 165° + isin 165°), ‘V2 (cos 285° + i sin 285°) The results for k = 8,4,5,6,7,-.. give repetitions of these. ‘These complex roots are represented geometrically in the com Fig. 5-5 plex plane by points P,,P2,P; on the circle of Fig. 5-5. 6. Determine the locus represented by (@) |z~2| = 3, (b) 2-2) = |e +4], (€) 2-3) + [2 +8) = 10. (0) Method 1. |z—2| = |z + iy—2| with center at (2,0) and radius 3. le-2+iy| = Ve—BP TH = 8 or (2-2 +42 =9, acircle Method 2. \z—2| is the distance between the complex numbers z= z+iy and 2+0%. If this distance is always 3, the locus is a circle of radius 3 with center at 2+ 01 or (2,0). 146 1 COMPLEX VARIABLE THEORY (CHAP. 5 (®) Method 1. e+ iy—2| = |etivtd| or V@—BPF = VETO + ee Squaring, we find «= —1, a straight line. Method 2, The locus is such that the distances from any point on it to (2,0) and (—4,0) are equal. ‘Thus the locus is the perpendicular bisector of the line joining (2,0) and (~4,0), or #=—1. (@) Method 1._ The locus is given by V@=3P Fv + V@TSP +H = 10 or Vest e 10 — V@F8"+¥% Squaring and simplifying, 25+ 3x = 5y(e+3)> +. Squaring and ¥ simplifying again yields 32+ 4% = 1, an ellipse with semi-major and semi lengths 5 and 4 respectively. Method 2. The locus is such that the sum of the distances from any point on it to (3,0) and (-3,0) is 10, Thus the locus is an ellipse whose foci are at (3,0) and (8,0) and whose major axis has length 10, jinor axes of Determine the region in the z plane represented by each of the following. @ b<1. Interior of a circle of radius 1, See Fig. 5-6(a) below. () 1<|e+24 52. |z+ 24] is the distance from 2 to —2i, so that [z+ 2i at —! (0,—2); and |z+2i] =2 is a circle of radius 2 wi represents the region exterior to |z+2i] = 1 but interior to or on [z+ below. 1 is a circle of radius 1 with center center at 24. Then 1 <|z+2i] 52 2. See Fig. 5-6(6) (0) 8 < args = */2. Note that args = ¢, where z = ref. The required region is the infinite region bounded by the lines @ = 7/3 and 9 =/2, including these lines. See Fig. 5-6(¢) below. @) Fig. 5-6 Express each function in the form u(z,y) + ¢v(z,y), where w and v are real: @ 2, (0) W—2, (0 & W@ nz (a) w = B= (etiy)® = 2 + S2%(iy) + Belin)? + (iy)> = a? + Binty — Bay? — ay? a8 — Bay? + i(Ba%y— ¥) Then = u(a,y) = a — Bxy?, v(x,y) = 3aty — y. ee 1 tne tty tne tiy erm imety ~ Gnatee Then ule.) = GAG Mee) = Goa CHAP. 5] COMPLEX VARIABLE THEORY 147 (@) et = ciety = trey = e8x(cos3y + isin Sy) and uw 8 cosy, v = of sin dy @ Ing = Infret) = Inet io = mV FR + itan-ty/e and w= Fln@t+y), v= tant y/e Note that Inz is a multiple-valued function (in this case it is infinitely many-valued) since ¢ can be increased by any multiple of 2. The principal value of the logarithm is defined as that value for which 05 ¢ <2» and is called the principal branch of Inz. 9 Prove (a) sin(z+iy) = sine coshy + icosasinhy (b) cos(z+iy) = cosecoshy — ising sinhy. We use the relations = cose +ising, e-# = cosz — 2, from which Then a destin — gues ty = sing coshy + ‘cosa sinhy Similarly, , eetw 4 eat 2 = cos(etiy) = cos. cos (2 + iy) 5 = Hee-yteretyy = ferv(cos + isina) + o (cosz — isinz)} = (eon (24E*) — iin = coseeoshy — isine sinhy = DERIVATIVES. CAUCHY-RIEMANN EQUATIONS 10. Prove that #22, where 2 is the conjugate of z, does not exist anywhere. vim Heda) = fle) aro ry in which Az = x + iAy approaches zero. Then by detain, a) = i this tnt exits independent of the manner 4 2 tim BEBE = tim Eee Tiny - 7TH ae* ~ aro Az te aa + fay = lim ST wt Ae idy—(@— i) _ yyy AB = tt ar) ‘Ba + iy arao Ae + aay ase Fras] - ‘ned limit ie im 2% = If Ay =0, the required limit is Jim > = ‘red limit ie im —282 If Ae = 0, the required limit is im, =A! = —1. ‘These two possible approaches show that the limit depends on the manner in which 42-0, so 148 COMPLEX VARIABLE THEORY [CHAP. 5 IL (a) If w = f() = find ae {b) Determine where w is non-analytic. (a) Method 1. 1+(tas) _1te dy jig Lea) me dz Ae ‘az = tga Provided <1, independent of the manner in which Ae ~ 0 Method 2. The usual rules of differentiation apply provided 2 #1. Thus by the quotient rule for differentiation, ‘7 e 1 a cd (1 ad(ize) _ Godgdt) ~UtaGC-) _ a-nw-atacn . _2 a@e\imz) = dar = (=a = 1 , where the derivative does not exist; ie. the (0) ‘The function is analytic everywhere except at = function is non-analytie at 2—1. 12, Prove that a necessary condition for w = f(z) = u(ay)+iv(z,y) to be analytic in a region is that the Cauchy-Riemann equations 2% = 2%, 34 _% be satisfied in the region. ey Since f(z) = fletiv) = ulz,y) + iv(e,y), we have fetas) = fie taetiytay] = wetany tay + ivle tax, y tay) Then tim 22442) Ff) jig MEA yay) — ule) + Hole + ae, y+ Ay) — view) aro ‘Ae remed ax + iy If Sy =0, the required limit lim Meta) amo ‘az ae az + fae fev) i{retang= se} = Hye If Ax =0, the required limit is im Wey + Ay) — ule, y) uevtay) — wel _ Lau , av am tay * av tay ty If the derivative is to exist, these two special limits must be equal, ive., ou, av _ Law, _ _ jam, av ae + foe = fay tay = tay * ay au. _ aw to that we must have 2 = ae Conversely, we can prove that if the first partial derivatives of u and v with respect to z and y are continuous in a region, then the Cauchy-Riemann equations provide sufficient conditions for f(e) to be anal} 18. (a) If f(2) = u(z,y) + iv(z,y) is analytic in a region R, prove that the one parameter families of curves (x,y) = Ci and v(2,y) = Cz are orthogonal families. (b) Illustrate by using f(z) = 2. (@) Consider any two particular members of these families u(z,y) = ug, v(2,y) = vp which intersect at the point (co, ¥o)- . CHAP. 5] COMPLEX VARIABLE THEORY c 149 Since du = u,dz+uydy = 0, we have Also since dv = v,de + v,dy = 0, ‘When evaluated at (2p, yo), these repre- sent respectively the slopes of the two curves at this point of intersection. By the Cauchy-Riemann equations, Ui = ty, ty =z, we have the product of the slopes at the point (2, ys) equal to %) _ so that any two members of the respective families are orthogonal, and thus the two families are orthogonal. () If f(a) =24, then u = at y2, v=22y, The graphs of several members of 2*— y= Ci, 2ey = Cy are shown in Fig. 5-7. Fig.5-7 14, In aerodynamics and fluid mechanics, the functions ¢ and y in f(z) = $+iJ, where f(e) is analytic, are called the velocity potential and stream function respectively. If @=2+4e-y+2y, (a) find y and (b) find f(2). 7 i 6 ey _ _ oe (@) By the Cauchy-Riemann equations, 5¢ = %, % = —%¢. then eer) Hs gy o Ba wee wy Ba we Method 1. Integrating (1), y = 2ay + 4y + F(a). Integrating (2), y = 2zy —2e + Gly). These are identical if F(e) = —2e +e, Gly) = 4y-+e where ¢ is any real constant. Thus y = 2ay + dy —2ete. Method 2. Integrating (1), y = 2xy+4y+F(e). Then substituting in (2), 2y+ Fe) = 2y-2 or Fi) = 2 and Flz) = —2e+e. Hence y = 2ey+dy—2e+e. () From (a), We) = gt ip = att de — yt + By + ieytdy—2e +0) (oy? + Diey) + dle tiy) — 2ile + iy) + ie = 24 de Wet ey where o, is a pure imaginary constant, san also be accomplished by noting that 2 = 2+iy, 2 = 2-iy so that 2 = . The result is then obtained by substitution; the terms involving 2 drop out. 150 COMPLEX VARIABLE THEORY [CHAP. 5 LINE INTEGRALS a 7 15. Evaluate S. (a? y)dz + (y+2)dy along (a) a straight line from (0, 1) to (1,2), (®) straight lines from (0,1) to (1,1) and then from (1,1) to (1,2), (¢) the parabola zat y=P+l (@) An equation for the line joining (0,1) and (1,2) in the ey plane is y = 2+1. Then dy=dz and the line integral equals 1 Sf. etme + etm ade = zo (®) Along the straight line from (0,1) to (1,1), y=1, dy=0 and the line integral equals Siena +0490 = Sf pas = -2/8 =o Along the straight line from (1,1) to (1,2), #=1, dz =0 and the line integral equals 2 Sf. c-n0 + ttn = f'wrng = 00 ‘Then the required value = 2/3 + 10/3 = 8/3. (©) Since £=0 at (0,1) and ¢=1 at (1,2), the line integral equals 1 1 f @-@+pa + (erp rored = f esrassoerae-nae = 2 m0 ‘ GREEN’S THEOREM IN THE PLANE 16. Prove Green's theorem in the plane if C is a simple closed curve which has the property that any straight line parallel to the coordinate axes cuts C in at most two points. Let the equations of the curves AEB and AFB (see adjoining Fig. 5-8) be y = Y,(z) and y = Y2(z) respec- tively. If is the region bounded by C, we have Si 2 te dy i es ay] Fig. 5-8 (2) 1 > 2 = ' oar ? 2 i z & ie e ee o fre = Sf Paw Similarly let the equations of curves HAF and EBF be «= X,(y) and 2 = Xu) respectively. Then CHAP. 5} COMPLEX VARIABLE THEORY . 161 Sf Bava = Lee Bala = SL" 100% ~ 0 ay ean = faxunds + faand = § oa ‘Then ) feu = s andy 7 = 2Q _ ap Adding (1) and (2), Pde + Qdy = IS 2) aay, Extensions to other simple closed curves are easily made. 11. Verity Green's theorem in the plane for § Qny—a) dz + (a+y)dy where C is the closed curve of the region bounded by y=2? and y*=a. ‘The plane curves y=2* and y?=2 intersect at (0,0) and (1,1), The positive direction in traversing Cis as shown in Fig. 5-9. Along y= 2%, the litte integral equals S_ (eave — addr + e+ Mae = =o Along y?=< the line integral equals ° ° S., 2090) — dw) + Wty dy = f (44-298 + 29%) dy = -17/16 Then the required line integral = 7/6 — 17/15 = 1/30. SF (Q-B)aa = [f{gern-fer-o}en § (= 28) de dy = LS (1-22) dy de S. @~22y) Hence Green’s theorem is verified. de = (eit —9a8/2—a2 4208) de = 1/80 ve INTEGRALS, CAUCHY’S THEOREM, CAUCHY’S INTEGRAL FORMULAS 2+ 18. Evaluate S eda he (a) along the parabola 2 =t, y=t? where 15t (b) along the straight line joining 1+ and 2+ 4i, 2, (c) along straight lines from 1+% to 2+4 and then to 2+4i. 152 COMPLEX VARIABLE THEORY (CHAP. 5 We have a Sree Go wt iomae stay = fo eae 4 tay ca 1D on 2.0 Sew de 2ovdy + 6 f Boyde + vey nD ony Method 1. (a) The points (1,1) and (2, 4) correspond to =1 and t=2 respectively. Then the above line integrals become near +(e@—wanay = -%8 — 6 f {(e2— eh) de — A(y(ep2Edey + : 4-1 1 = $54@-0) or y = 32-2 Then (®) The line joining (1,1) and (2,4) has the equation y we find Set (Ge -299 ae ~ 22(80- 298d} 2 +if (0) Prom 14% to 2+4 [or (1,1) to (21) v=, dy =0 and we have S {2x(82— 2) de + [2% — (8x —2)2]8 day 2 @t-1de + if" gods = $+ 31 From 2+i to 244i (or (2,1) to (2,4)], =2, de =0 and we have mtydy + if G-wdy = -20 - 96 _ f 86 ‘Adding, K + 3) + (-30—91) = 37 Gi. Method 2. The line integrals are independent of the path [see Problem 19], thus accounting for the same values obtained in (a), (6) and (c) above. In such case the integral can be evaluated directly, as for real variables, as follows: fi'ee = ap etapa tae net Shree s 19. (a) Prove Cauchy's theorem: If f(z) is analytic inside and on a simple closed eurve C, then (z)dz = 0. § 1 (0) Under these conditions prove that f” f(2)dz is independent of the path joining P, and P». a @ § tous = £ (w+ ioyide tidy) = By Green's theorem, where & is the region bounded by C- CHAP. 5] COMPLEX VARIABLE THEORY 158 Since se) analy, 2 = 22, 2 = 2 problem 10, and a0 the above integrals are zero, ‘Then f /(s)dz = 0. We are assuming in this derivation that /"(e) (and thus the partial I derivatives] are continuous. This restriction can be removed. (b) Consider any two, paths joining points P, and P, (see Fig. 5-10). By Cauchy’s theorem, fade = 0 PAP EP, Then Sf $e) de + f fade = 0 PIAPy “BP or feds = - ff fia)de = ff fede Bree fe" L, “ie. the integral along P,APs (path 1) P,BPy (path 2), and so the integral is path joining P, and Py. This explains the results of Problem 18, sinee f(2) =? is analytic. integral along \dependent of the Fig. 5-10 20. If (2) is analytic within and on the boundary of a region bounded by two closed curves C; and C2 (see Fig. 6-11), prove that § terae = § eras As in Fig. 6-11, construct line AB (called a cross-cut) connecting any point on Cz and a point on C;. By Cauchy's theorem (Problem 19), s fede = 0 Fig. 5-1 Aardansroa since /(c) is analytic within the region shaded and also on the boundary. Then He) dz + ff fede + fade + {feds = 0 Oy blows fous Suess But ff fede = - f fds, Hence (1) gives feds = - ff fade = He) de be. = § oe ‘Note that f(z) need not be analytic within curve Cz. where C is a simple closed 21, (a) Prove that f oor {ri if n=1 0 if n=2,3,4,... curve bounding a region having z= as interior point. (b) What is the value of the integral if m = 0,—1,—2,-8,...? 154 COMPLEX VARIABLE THEORY (CHAP. 5 (a) Let C, be a circle of radius « having center at «= (see Fig. 5-12). Since (¢—a)-* is analytic within and on the boundary of the region bounded by C and C;, we have by Problem 20, Siete - Sater To evaluate this last integral, note that on Ci, le—a| = « or z-a = ee! and dz = iceds. The integral equals 1 ; Sf ‘et -006 do SSS If n=1, the integral equals if de = 2ri. 2, () For n=0,-1, the integrand is 1, (z~a), (2—a)*, ... and is analytic everywhere inside C1, including Hence by Cauchy's theorem the integral is zero. 22. Evaluate § de where Cis (a) the circle |z| = , (b) the cirele jz +i] = 4. (a) Since z= 8 is not interior to , the integral equals zero (Problem 19). (®) Since 2=8 is interior to |z+4] = 4, the integral equals 2ri (Problem 21). 23. If f(z) is analytic inside and on a simple closed curve C, and a is any point within C, prove that a - 1 ¢ 1 fe) = sf LO a Referring to Problem 20 and the figure of Problem 21, we have GMa = § Ma fo, B= 2 Letting 2—a = eel, the last integral becomes f(a + «e*) da, But since f(z) is analytic, it is continuous. Hence lim oo and the required result follows. a. 2 er f flat)de = i f lim fla+ 2. If |s|=2 the ratio test fails. However, the series of absolute values 3 converges if |2| = 2, since 1 a converges. Thus the series converges (absolutely) for 2 ° = din lawn] = t Then the series, which represents sin z, converges for all values of 2. = Gi im | nth i fe=antt 3. @ 3,252. wenave tim tim [Seo d>3. ‘The series converges if |2—i] <8, and diverges if |z i = Se and the series becomes Sc, This series diverges since If [2—i] = 3, then 2— the nth term does not approach zero as n=. 8 but not on the boundary. ‘Thus the series converges within the circle |z— 156 COMPLEX VARIABLE THEORY [CHAP. 5 27. 1f 3S az" is absolutely convergent for |e| SR, show that it is uniformly convergent for these values of z. ‘The definitions, theorems and proofs for series of complex numbers and functions are analogous to those for real serie: In particular, a series & ug(z) Ft is said to be absolutely convergent in a region R if 'uq(2)i converges in ®. We can also show that if !u,(z)| converges in R, then so also does Sue, an absolutely convergent series is convergent. Also, a series 3 us(e) convergent to a sum function S(s) in a region is sald to be wniformly convergent in ® if for any «>0, we can find N such that Sylz) — S@)| < « for all n>N where N depends only on « and not on the particular 2 in &, and where Sy2) = ule) + mle) + + + le) An important test for uniform convergence is the following. If for all in R we can find constants M, such that lup(@)| 5S My o1, and 3M, converges then 3 u,(z) converges uniformly in R. This is called the Weierstrass M test. For this particular problem, we have laqa*| S fa,(Re = M, n= 0,1,2, Since by hypothesis & M, converges, it follows by the Weierstrass M test that 3 aye" converges uniformly for |a| = R. 28. Locate in the finite z plane all the singularities, if any, of each function and name them. @ ar #=-1 is a pole of order 3. 41 & De-1+8H poles of order 1 (simple poles). is a pole of order 2 (double pole); z =i and 2 = 1-21 are sin mz (0) PERG m0. Since #42242 = 0 when z fo = -14i, we can write 2+ 22+2 = {2—(-1+i)2—(-1—a) = @+1-Het1+9. The function has the two simple poles: = —1+4 and = @ + #=0 appears to be a singularity. However, since lim 9, 2 mye removable singularity. CHAP. 5) COMPLEX VARIABLE THEORY 157 Another method. Binge LTS 2=0 is a removable singularity. vert a 1 mt This is a Laurent series where the principal part has an infinite number of non-zero terms. ‘Then z=1 is an essential singularity. Oe This function has no finite singularity, However, letting 7=1/x, we obtain e!/* which has ‘an essential singularity at u=0. We conclude that z= is an essential singularity of In general, to determine the nature of a possible singularity of f(z) at z=, we let #=1/u and then examine the behavior of the new function at u=0. 29. If f(z) is analytic at all points inside and on a circle of radius R with center at a, and if a+h is any point inside C, prove Taylor’s theorem that fat+h) = fla) + hf'a) + - (a) ove Hea) + By Cauchy’s integral formula (Problem 23), we have - i) de Kat) = gf Sede @ By division, a = — ~ank Vit — Wea] if h he cee he att watt eat ap tt Gat wow ® Substituting (2) in (1) and using Cauchy's integral formulas, we have ary = LG Me yh IG fede sory = gG Bot 25 § la +m = fa) + kf + F F(a) + Ry where r= MG fade et J, att ah) Ste) Now when 2 is on C, | ae and R, so that by (14), Page 140, we have, since 2eR is the length of C, | s BEM. oeR As n+, |R,|>0. Then R,-0 and the required result follows. If f(@) is analytic in an annular region r, 5 |z~a| 5 rp we can generalize the Taylor series to a Laurent series (see Problem 119). In some cases, as shown in Problem 30, the Laurent series ean be obtained by use of known Taylor series. 158 COMPLEX VARIABLE THEORY (CHAP. 5 30. Find Laurent series about the indicated singularity for each of the following functions. Name the singularity in each case and give the region of convergence of each series. ( oF Let 2- Then # = 1+u and . _ ot Le www = = eG = Stuur Fe Feit = ¢ 4 =), deme, 3 3 cy is a pole of order 2, or double pole. The series converges for all values of 2 1. wet aa oat ) = =0 is an essential singularity. ‘The series converges for all values of «¥ 0. Let 2- sinuts) _ _sinw = Mute) simu we ai ~1+ z= isa removable singularity. ‘The series converges for all values of 2. © Geyeryi tanh let ett oe = @FDEFD uu HN guts tuto) Lae ou tat — m+ . = ppp t 2 - ety + ery - 1 is a pole of order 1, or simple pole. The series converges for values of ¢ such that 0 < jz++1| <1. (© gpg 1 = 0-2. ©) aerapi <= 0-2 Case 1, 2=0. Using the binomial theorem, — ae +2" (GY + ENG PERE + } 2 =0 is a pole of order 1, or simple pole. The series converges for 0 < |z| <2. CHAP. 5] COMPLEX VARIABLE THEORY 169 Case 2, 2=~2, Let s+2=u. Then a Bera = -1.-—4,-1,-1l-sde aoe = “ae+aR ~ TesaR aera ~ 16 ~ e+? #=-2 is a pole of order 8. ‘The series converges for 0 < |z +2 < 2. RESIDUES AND THE RESIDUE THEOREM 31. If f(z) is analytic everywhere inside and on a simple closed curve C except at z=@ which is a pole of order n so that fe) = fa $oves + dy + ae) + ae(z—a)? + @-ay where a-»~0, prove that (a) § f(@dz = 2ria.s lim yrds te a"F2)- 0) anv = Tim ay (a) By integration, we have on using Problem 21 fay + ay(z— a) + age— a)? + +++} de I Since only the term involving a , remains, we call a_, the residue of f(2) at the pole (b) Multiplication by (¢— a)" gives the Taylor series af) = ig + Ogg ee) Hee Fayed toe ‘Taking the (n—1)st derivative of both sides and letting 2->a, we find + (mle, = tim ear fey aaa dem from which the required result follows. 82, Determine the residues of each function at the indicated poles. (0) Gapttapay! = 2% These are simple poles. Then, Residue at 2=2 is tim (2-2) {eastern} =f Residue at 2=4 is lim (2-9 eats} S| ‘i is 2 Residue at lim (e+) facut ra! 160 COMPLEX VARIABLE THEORY (CHAP. 6 1 ©) ara 0,-2 2=0 isa simple pole, 2 =—2 is a pole of order 3. Then: Residue at 2=0 is = om LB) = tim 1/2) 2 2d = im, 2 aa(z) as stim, 3(3) = 3 Note that these residues can also be obtained from the coefficients of 1/z and 1/(¢+2) in the respective Laurent series [see Problem 30(e)]- (Gage += 8 a pole of order 2 oF double pole, Then: Residue is tim {e -3+ cn} = lim Steer) = him (eet +2ter) = et + Stet (@) cots, 2 =5r, a pole of order 1. Then: Residue is lim (@~6r)+ = lim cos) = ( lim 2) oy = ey = 1 where we have used L’Hospital’s rule, which can be shown applicable for functions of a complex variable. 38, If fle) is analytie within and on a simple closed curve C except at a number of poles a,b,c,... interior to C, prove that § fea = Pri {sum of residues of /(2) at poles a,b, ¢, ete. f Refer to Fig. 5-18. By reasoning similar to that of Problem 20 (ie. by constructing cross cuts from C to Cy, C2, Cy, ete), we have (z)dz = (2) dz (z)dz + ++ $1 f 10 + fr ia. 513 For pole a, co He) ay tt + meme) te ba Similarly for pole b, f(z) = oe + By t bo + bile b) + so that GF tou = ais, Continuing in this manner, we see that § fevas = Qeifa_y + bay + -++)) =. Bei (sum of residues) CHAP. 5} COMPLEX VARIABLE THEORY 161 edz Gera 34. Evaluate § where C is given by (a) |z| =3/2, (b) |z| = 10. c Residue at simple pole #=1 is lim {e-» fe Residue at double pole = —8 is “im 2 ima {errg va} (@) Since |z| = 8/2 encloses only the pole = jired i = mi(#) = the required integral = 2ri ( a) ~ (8) Since |2|=10 encloses both poles #=1 and z=—8, the required integral = EVALUATION OF DEFINITE INTEGRALS 4 35. If (fa) =H for 2= Rev, where k>1 and M are constants, prove that lim ff fede =0 nae Jp where I is the semi-circular arc of radius R shown in Fig. 5-14, By the result (14), Page 140, we have M aM | = fella = Mer = 2M Sr Svolial = a Fig 5-14 since the length of arc L = sR. Then im [fires] = 0 ants yim f fords = 0 M1 it fo = mam: 36. Show that for z= Re, |f(z)| = Re If 2= Rel, |fle)| = (say R> 2, for example) so that Mf Note that we have made use of the inequality ay +] = |21| — |zo| with 21 =Rtet™ and zy =1. if R is large enough dz 87. Evaluate fo. Consider § 45, where C is the closed contour of Problem 35 consisting of the line trom “R to R and the semi-circle 1, traversed ini the positive (counterclockwise) sense. Since +1 = 0 when z= orl, cami, eSeU4,"ctet/4, these are simple poles of 1/(x4+1). Only the poles e*/* and ot lie within C. ‘Then using L'Hospital’s rule, 162 COMPLEX VARIABLE THEORY [CHAP. 5 : Residue at on = tim fie orb eer ett je Le Leet = dims = ae asta = — any 1 Residue at ev = mua ony a = tm 3 = Lone anvedisg 42 Thos § sty = mitgermn s jemi = rm . ati ie. Oy a = of Ba’ ade _ - @ IPG +2242 50° the required integral has the value = 38, Show that” The poles of TREE and z= —1+i of order 1. enclosed by the contour C of Problem 35 are z= i of order 2 5 a 2 — o-12 Residue at == is tin {oe GFE TET} 100 i , 2 - Resdueat ¢=-1+i stim HOG Ea = ade = & me § aavetrars - & . sf" aide de =i ° |, Siete + J, STIS RID 50 Taking the limit as R~® and noting that the second integral approaches zero by Problem 35, ‘we obtain the required result. 38. Evaluate f" dt. = le = ie do = iedo 50 tha Let 2=e% Then si z Ga) 4 do = ido 50 that ee = deliz = Qdz atm > fey > fants B+ a( 3 ) where C is the cirele of unit radius with center at the origin, as shown in Fig. 5-15 below, CHAP. 5] COMPLEX VARIABLE THEORY 168 The poles of TET iere wt the simple poles 101 + V=100F 86 Only —i/3 lies inside.C. jeResidue at —i/8 = lim, im ae Te ; by L'Hospital’s Qde tan f artitims = 40. Show that , Ig z=el, cose = dz = izde. A+ 7 cos ae) f CO Y, Be — We where C is the contour of Problem 39. The integrand has a pole of order 3 at z=0 and a simple pole 2= } within C. ‘ sow wmi ls. #t1 | _ a Residue at 2 =0 is ty Bo Fcenicers =F . are —peqe tt} = - Residue at 2 = } is a, {e b aaane=5} ee _ BH = A pey {2d = ap require: Then f attitaye = -heeo{ ews 41. Tt (2S zx for z= Re®, where k>0 and M are constants, prove that lim {om fe)dz = 0 mJ where ris the semi-circular are of the contour in Problem 85 and m is a positive constant. If 2 = Re®, S eine f(g)de = femme ike de. =f" ‘Then [oem "(Rel Re” de jeimnel 1(Re'9) iRe'4| Io do W S- Jrsoer- ese je se , = fe mind |/(Re®)| R de Mo ( 9-maan 5 Sect J oomtinede 164 COMPLEX VARIABLE THEORY (CHAP. 5 Now sing = 20/r for 05 than or equal 7/2 (see Problem 3, Chapter 7). Then the last integral is less = Semen = Bika an As R= this approaches zero, since m and k are positive, and the required result is proved. * cosma —m 42. Show that S emt ds = Se", m>0. Consider FF de where C ia the contour of Problem 35. The integrand has simple poles at z= i, but only == lies within C. ; An © ome Residue at 2 =i is Tim {e-og=Fsa} Then nem or ren . de = em i wri and 80 rem Taking the limit as R-> and using Problem 41 to show that the integral around 1 approaches zero, we obtain the required result, 43. Show that S Stas = 5, no ® 2 The method of Problem 42 leads us to con- sider the integral of ¢'*/z around the contour of Problem 35. However, since z= 0 lies on this path of integration and since we eannot inte- grate through a singularity, we modify that contour by indenting the path at z= 0, as shown in Fig. 5-16, which we call contour C’ or ABDEFGHJA. Since 2 is outside C’, we have or Replacing = by —= in the first integral and combining with the third integral, we find, Stat or CHAP. 5) COMPLEX VARIABLE THEORY 165 Let r>0 and R> =. By Problem 41, the second integral on the right approaches zero. The first integral on the right approaches 0 ie 0 ~ tim (Sp rete de = — tim fie do = wi ind, re mJ, since the limit can be taken under the integral sign. Then we have 5 ° tim aif SR@dz = wi or Raw Dy MISCELLANEOUS PROBLEMS 44, Let w= 2? define a transformation from the z plane (zy plane) to the w plane (uv plane). Consider a triangle in the 2 plane with vertices at A(2,1), B(4,1), C(4,8).. (a) Show that the image or mapping of this triangle is a curvilinear.triangle in the uv plane. (b) Find the angles of this curvilinear triangle and compare with those of the original triangle. (a) Since w= <2, we have u = z*—y%, v = 2ay as the transformation equations. Then point A(2,1) in the ay plane maps into point A’(8, 4) of the uv plane (see figures below). Similarly, points B and C map into points B’ and C’ respectively. “The line segments AC, BC,AB of triangle ABC map respectively into parabolic segments A’C’, B’C’, A'B’ of curvilinear triangle A’B‘C’ with equations as shown in Figures 5-17(a) and (8). ° a Fig. 517 (8) The slope of the tangent to the curve vf = 4(+u) at (4) is m= 42} = 2] = 3. Htles.0 Phew The slope of the tangent to the curve w= 20+1 at @d)is my = | =u = 3, lao Then the angle @ between the two curves at A’ is given by tang = Mom 2 SA, ado itm ~ TF Op Similarly we can show that the angle between A’C’ and B’C’ is «/4, while the angle between AB’ and B’C’ is 7/2. ‘Therefore the angles of the curvilinear triangle are equal to the correspond- ing ones of the given triangle. In general, if w= f(2) is a transformation where f(z) is analytic, the angle between two curves in the z plane intersecting at 2= zp has the same magnitude and sense (orientation) as the angle between the images of the two curves, so long as f'(z) #0. This prop- erty is called the conformal property of analytic functions and for this reason the transformation w = f(z) is often called a conformal transformation or conformal mapping function. 166 COMPLEX VARIABLE THEORY (CHAP. 5 45. Let w= V2 define a transformation from the z plane to ‘the w plane. A point moves counterclockwise along the circle |z|=1. Show that when it has returned to its start- ing position for the first time its image point has not yet returned, but that when it has returned for the second time its image point returns for the first time. Let z=, ‘Then w = Vz = e#/2, Let #=0 correspond to the starting position. Then z= w=1 [corresponding to A and P in Figures 5-18(a) and (6)].. and @ Fig. 5-18 ‘When one complete revolution in the 2 plane has been made, so the image point has not yet returned to its starting position. However, after two complete revolutions in the z plane have been made, ¢ w = ci0/2 = edvi=1 0 the image point has returned for the frst time. 2x, 2=1 but w = #2 = oi i, 2=1 and It follows from the above that w is not a single-valued function of z but is a double-valued function of =; ie. given =, there are two values of w. If we wish to consider it a single-valued function, we must restrict ¢. We can, for example, choose 0 @<2r, although other possibilities exist. This represents one branch of the double-valued function w =z. In continuing beyond this interval we are on'the second branch, eg. 2r 26 <4r. The point z=0 about which the rotation is taking place is called a branch point. Equivalently, we can insure that f(z) = V7 will be single-valued by agreeing not to cross the line Oz, called a branch line. 2 get . : 46. Show that f" Fdz = sZ, O CHAP. 5] COMPLEX VARIABLE THEORY 169 64. Prove that fle 2|e| is not analytic anywhere. 1 65. Prove that (2 pig 3s analytic in any region not including « 66. If the imaginary part of an analytic function is 2x(1 ~ y), determine (a) the real part, (6) the function, Ans, (a) y2— 22 2y +0, (0) 2is~a+0, where ¢ is real 67. Construct an analytic function f(z) whose real part is e~*(x cosy + y siny) and for which f(0) = 1. Ang, 207 +1 68, Prove that there is no analytic function whose imaginary part is a? —2y. 68. Find f(z) such that f"(2) = 42-8 and f(l+i=—3i, Ane. fle) = 22 8243-4 LINE INTEGRALS 4.2) 70. Evaluate S (e+y)dz + (y—2)dy along (a) the parabola y*= 2, (6) an lines from (1,1) to (1,2) and then to (4,2), (d) the curve # = 2+t+1, y= @+1, Ane, (a) 84/3, (B) 11, (e) 14, (@) 32/8 traight line, (¢) straight 11, Evaluate f (@x~y+4)de + (Gy +82—6)dy around a triangle in the 2y plane with vertices at (0,0), (3,0), (8,2) traversed in a counterclockwise direction. Ans. 12 72, Evaluate the line integral in the preceding problem around a circle of radius 4 with center at (0, 0). Ans, 642 GREEN'S THEOREM IN THE PLANE. INDEPENDENCE OF THE PATH 18. Verity Green’s theorem in the plane for ff (2!—ay)de + (yP—2ay)dy where C is a square with vertices at (0,0), (2,0), (2,2), (0,2). Ams. common value = 8 1. (a) Let C be any simple closed curve bounding a region having area A. Prove that if a,, dy, ,;, by, by are constants, § (oye-+ ayy + 0x) de + (bye+ bay + by) oy (b- aA (®) Under what conditions will the line integral around any path C be zero? Ana. (8) aq =, 75. Find the area bounded by the hypocycloid 2#/9+ y2/ = 2, (Hint. Parametric equations are x = acos't, y= asintt, 05¢S2r) Ans. ra/B TK If z= recone, y= raina, prove that 4 f rdy—ydr = 4 f r%do and interpret. 72. (a) Verity Green's theorem in the plane for (29—aty) dz ++ aytdy, where C is the boundary of a the region enclosed by the circles =*-+y? = 4 and a%+y* = 16. (b) Evaluate the line integrals of Problems 71 and 72 by Green’s theorem. Ans, (a) common value = 1207 18, (@) Prove that f°" (pey—yk +8) de + (et 4av®) dy Is independent of the path joining (1,0) and 0 (2,1). (b) Evaluate the integral in (a). Ana. (b) 5 170 COMPLEX VARIABLE THEORY (cmap. 5 INTEGRALS, CAUCHY’S THEOREM, CAUCHY’S INTEGRAL FORMULAS 1a, Bvaluate f°" 22+) de: ts (a) along the path # = 2¢+1, y = 4t2?-t—2 where O=t=1. (©), along the straight line joining 1-21 and 8 +i, (c) along straight lines from 1~ 2i to 1+ and then to 3+ i. Ans. 17 + 19% in all cases 80, Evaluate S (2—2+2)dz, where C is the upper half of the circle '2| =1 traversed in the positive c sense. Ans. ~14/3 a1, Evaluate § Geeeg) where C is the circle (@) [el = 2, (8) [2-3] = Ans. (a) 0, (b) ri/2 # : seh vert 82, Evaluate § TERT te where Cis: (a) a square with vertices at — (eo) the circle [i= V2, Ans. (a) —8ri/8 (6) —2ni (¢) 2ri/3 1+i,-B4i,-3-% (0) the circle z+ ii dz, (b) § i <+5 ae where C is any simple closed curve enclosing # = 1. Evaluate (a) f $8 fine to 2=1 Ans, (a) —2ri (b) rie/8 84. Prove Cauchy's integral formulas. (Hint. Use the definition of derivative and then apply mathematical induction. SERIES AND SINGULARITIES 85. For what values of 2 does each series converge? “ar Oo) 3, (Hayat (28 + 22 + 2 Ans. (a) allz (8) jz () 2=-1+ 85, Prove that the series & is (a) absolutely convergent, (b) uniformly convergent for |z| 1. S,anFn 81. Prove that the series etd (1 ens ton f° £8822 ae = Wy, ea cos Qnty = 108, f SO828% ge = Sats * sin wx = 109. s snr de = =e us. f wosha * ~ Zeosh@ Py" [ at (RB, 0), (R, 0), (Rx), (—Ryx). ‘Then let R=] int. Comite § retante with verte Jo cosh 2 172 COMPLEX VARIABLE THEORY (CHAP. 5 MISCELLANEOUS PROBLEMS M4, If z= rel and f(z) = u(r, Riemann equations are + iv(r,0), where 7 and ¢ are polar coordinates, show that the Cauchy- au or ae 115, If w = f(z), where f(2) is analytic, defines a transformation from the z plane to the w plane where z= 2+iy and w= u+iv, prove that the Jacobian of the transformation is given by = rer 116, Let F(z, y) be transformed to Glu, ) by the transformation w= f(z). Show that i¢ P+ SE = 0, then at all points where /"(2) #0, S34 #¢ = 9, aid 117. Show that by the bilinear transformation w = SET4, where ad—be % 0, circles in the + plane are transformed into circles of the w plane. 118. If f(z) is analytic inside and on the circle |z—a| = R, prove Cauchy's inequality, namely, vera = 5M where |f(2)| SM on the circle, [Hint Use Cauchy’s integral formulas.| 119, Let Cy and Cz be concentric circles having center a and radii r, and rp respectively, where 7, The finite Fourier cosine transform of F(z), 0 o Ft) = {5 20 feo (26) ‘Then from (16), Page 176, with a replaced by y, we see that the Fourier transform of F(t) is * {F(t} = * ptetint a(t) dt = * em w(t dt (27) FRO = Sf (t) Sere 27) where we have written s = x+iy. The right side of (27) is the Laplace transform of 4(t) and the result indicates a relationship of Fourier and Laplace transforms. It also indicates a need for considering s as a complex variable z + iy. 178° FOURIER SERIES AND INTEGRALS (CHAP. 6 To indicate the relationship even further, note that if F(t) and G(t) are zero for t<0, the convolution of F and G given by (22) can be written F*G = F(u) G(t—u) dr 8) J, Fe @(t—w) au (28) and (24) corresponds to L(F*G) = AF} L(G} (29) in agreement with (11) on Page 45. In view of the fact that there is an inversion formula (17) corresponding to (16) for Fourier transforms, one would feel that there ought to be a corresponding inversion formula for Laplace transforms. Such an inversion formula is obtained in Chapter 7. Solved Problems FOURIER SERIES 1. Prove S sin E* de =f, cos 2 de = 0 if k=1,2,8,.... = —beo et qe coske + zh cos(-kx) = 0 = goin! = sinter — Linke) = 0 ‘ Mat tee OF) gg, Mat nt Oo ma¥n 2 Prove (a) So TT cosy de = JS, sin 8 sin ae = {' nen (0) S sin "7 cos "7 de =0 where m and n can assume any of the values 1,2, 3, (a) From trigonometry: cos A cos B = 4{cos(A—B) + cos(A+8)}, sinA sinB = §{cos(A—B) — cos (A +B)}. Then, if m *n, by Problem 1, cos ™ ogg 3 FE de = = 0 S275 Similarly if m* n, Sf) sin sn de = am CHAP. 6] FOURIER SERIES AND INTEGRALS 179 If m=n, we have 1g MH ong Mat ay, J co BE coe IE ae 3S (1+ oo Lf (tent) ae = Note that if m=n=0 these integrals are equal to 2! and 0 respectively. (0) We have sinA cosB = }{sin(A—B) + sin(A+B)}, Then by Problem 1, if mn, " fsx cos de nSMet 4 sin meth de = 0 T T a j Mr net 1! | anez _ . soon ttae = Ef" ain Las = 0 The results of parts (a) and (6) remain valid even when the limits of integration —1,1 are replaced by ¢, ¢+21 respectively. 3 Iftheseries 4+ > (a cos" + by sin ) converges uniformly to f(2) in (-1,), 4 show that for n = 1,2,8,..., (@) an = iS, F(z) cos de, (b) b, = f F(a) sin? dz, (@) A= 3. (a) Multiplying Fa) = A+ 3, (% cos 2 + by sin ™ a by con" and integrating fram =I tI, using Problem 2, we have SF) con de = ® -1 sin BE as} if oe on = FS ree Bee it m= 49,... 1 6 +S on Saini cont ae + 8, fan TE ant as ‘Thos ba = i¢ Mere + _ SF) sin™ 2 de if m= 1,2,8,.. 180 FOURIER SERIES AND INTEGRALS (CHAP. 6 (©) Integration of (1) from —I to J, using Problem 1, gives 1 1 S rea = at or 4 = BS rear Putting m=0 in the result of part (a), we find ay = af F(a)dz and so A =<. The above results also hold when the integration limits —1,1 are replaced by c, 0+ 2. ‘Note that in all parts above, interchange of summation and integration is valid because the series is assumed to converge uniformly to F(#) in (1,1). Even when this assumption is not warranted, the coefficients am and bm as obtained above are called Fourier coeficients corresponding to F(z), and the corresponding series with these values of aq and by, is called the Fourier series corresponding to F(2). An important problem in this case is to investigate conditions under which this series actually converges to F(x). Sufficient conditions for this convergence are the Dirichlet conditions established below [see Problems 12-17). 4, (a) Find the Fourier coefficients corresponding to the function 0 -B<2<0 Dy = iod = (2) { denes Period (b) Write the corresponding Fourier series. (&) How should F(2) be defined at 2=—5, «= 0 and 2=6 in order that the Fourier series will converge to F(z) for —5 S 2 <5? The graph of F(z) is shown in Fig. 6-4 below. F(a, — Paid Fig. 64 (a) Period = 21=10 and 1=5. Choose the interval c to ¢+2I as —5 to 5, so thate=—5. Then = 1 6 ey cos BE de 5S, Pe cos 8 de = BY S00 om tet ce + f° 9 os de } = bfoomtes = i(eam = 0 ifnno cos FE de = 2 (ue =B 5 5 J, MO (sie a Met _u Mra b= PL Ream tte = 5S resin ae a * (0) sin 22 ae + S$ @sin™Z ae) = 3 f sin" ae Sforza} = 2 fat wtz)f = 3d = conne) 5 mr CHAP. 6) FOURIER SERIES AND INTEGRALS 181 () The corresponding Fourier series is m4 § nr ne $+ 3 (ccm + ban) ries wore 8 (sin 2% 4 V gin 82% 4 2 gin Sz + (sink 4 Lene + Lan BE.) (©) Since F(#) satisfies the Dirichlet conditions, we can say that the series converges to F(z) at all points of continuity and to L@+0) + FZ) 44 points of discontinuity. At x= -5, 0 and 5, which are points of discontinuity, the series converges to (8 +0)/2 = 8/2 as seen from the graph. If we redefine F(x) as follows, 32> o=-5 0 -B<2<0 Fa) = {3/2 2=0 Period = 10 8 00, fel) -f FO) cos de = Sf 2 cost ae fen) = 60) = ff tede = 16 UW. Find F(z) if: (@) F,{F(x)) = 16(-1)"/, n = 1,2,3, (0) F, (F(a)} = sin (mx/2)/2n, n = 1,2,3,... and s/4 if n= (@) From equation (8) of Problem 9(a) with I= 8, we have +» where 0<2<8; , where 0<2<2n, (®) From equation (6) of Problem 9(5) with = 2r, we have Fa) = rot{a Lee, sim 2 +o CONVERGENCE OF FOURIER SERIES 12. Prove that (a) zt cost + cos2t+--- +cosMt = Sin(M+d)t 2sin yt Le*sint+tg, — 1 1 sn(M+aty _ 1 @ if 2 sin Ht dt => 1s ~2singt at = 3. (a) We have cosntsin}t = f(sin(n+J)t — sin(n—4)0). Then summing from n= 1 to M, sin §t{cost + cos2t+ +++ +cosMt} = (singt—sin}t) + (sin je — sin Je) +o + fein + He - sin Py = Hsin t+ pe — sing On dividing by sin jt and adding 4, the required result follows. (0) Integrate the result in (a) from —rto0 and Otow respectively. This gives the required results, since the integrals of all the cosine terms are zero, 186 FOURIER SERIES AND INTEGRALS (CHAP. 6 18, Prove that lim S F(2)sinnzde = lim § F(x)cosnzdz = 0 if F(z) is sec- tionally continuous. This follows at once from Problem 8 with J=-, since if the series a + 3 (az + b3) is con- vergent, lim o, = lim b, The result is sometimes called Riemann’s theorem. 14 Prove that Jim f F(e)sin(M+4)zde = 0 if F(2) is sectionally continuous. We have S7 Fe sinar+pede = * (r¢e) sin $2) cos Ma de + f {Fl@) cos 42} sin Me da Then the required result follows at once by using the result. of Problem 13, with F(2) replaced by F(a) sin fx and F(z) cos x respectively which are sectionally continuous if F(z) is. The result can‘also be proved when the integration limits are a and 6 instead of —r and x. 15. Assuming that | = z, i.e. that the Fourier series corresponding to F(x) has period 21=2z, show that wh + Slacosme + besinns) = 1g F(t+2) 2) SOD at in ft Sy(e) = iMs Using the formulas for the Fourier coefficients with [= 7, we have a, cosne + b, sine = CS. Foy) conn du) conne + es. FQ) sinned) = 267 FU (com me cosne + sin ma sin na) du =} S F(u) cos n(u— 2) du Also, $e Then Sue) = B+ 3 (a, cos ne + 8, sin nz) = eS Fe ut 23 Sf Fw) cos n(u—2) du = 26" Fw oof + Sonn} au 1a” ain (M+ Pu 2) Banju-ay re nee gpe su) = 2G rer OD F Fe Since the integrand has period 2r, we can replace the interval —r — of length 2z, in particular —7,7. Thus we obtain the required result. x — by any other interval CHAP. 6] FOURIER SERIES AND INTEGRALS 187 16. Prove that ° Syl) - (Feo Fe oy = f EY eegige+m) ana ) sin(M+4)t dt From Problem 12, sue) = 2" Fern MAD a + bf" re+ 9 EP ae « Multiplying the integrals of ‘Problem 12(b) by F(x —0) and F(x +0) respectively, F@+0)+F@-0) _ 16° py 82 Ol+ He 1 eto me pe ; 2 Leo treo eo Subtracting (2) from (1) yields the required result. 17. If F(z) and F(z) are sectionally continuous in (—z, x), prove that 7 = Fle+0)+F(«-0) dim Si) = FEROS TEND The function nero ete is sectionally continuous in 0< Sx because F(z) is section- ally continuous. Also, in Pitta) ~ Fle +0) im Fitta)- Floto), ¢ 9 y Fit-+2) - Fle +0) ime, 9 ain fe = in ¢ Tange = ER 7 exists, since by hypothesis F"(z) is sectionally continuous so that the right hand derivative of F(z) at each « exist, Thus Fitt#)=Flet0) 2 ainft is sectionally continuous in 05 t 5 x. similarly, Fet2)—Fe+o) Simitarly, FOO Tet ectionally continuous in —» St $0. ‘Then from Problems 14 and 16, we have lim Sy(e) — [Beste o} wore z or jim Sy(e) = E@+0)+ F@—o) aC an, Sue) ; THE FOURIER INTEGRAL AND FOURIER TRANSFORMS 1 |tia° (b) Graph F(z) and its Fourier transform for a= 1. 18, (a) Find the Fourier transform of F(z) = { (@) The Fourier transform of F(2) is Si Fine de = f° ema ~te sine = 2 xo a) = (o = ( nie, 10) 188 FOURIER SERIES AND INTEGRALS [CHAP.6 (6) The graphs of F(z) and f(X) for a=1 are shown in Figures 6-8 and 6-9 respectively. Fe fo) Fig.6-9 18. (a) Use the result of Problem 18 to evaluate f SimA¢.c0s.Ax gy, (®) Deduce the value of f 2" du. (@) From Fourier's integral theorem, if 10) =f Fede then Fe) = Ef moa 1 ile ‘Then from Problem 18, 2d f ° ae J. ‘The left side of (1) is equal to 7S ‘The integrand in the second integral of (2) is odd and so the integral is zero. ‘Then from (1) and (@), we have ° sin Na cos Aw * ale amd 1) = Si rae = - fae Lo Sta = 3 By letting Na =u and using Problem 111, Page 171, it is seen that this is correct. The method can also be used to tnd FS" du directly where F(z) = { This is equivalent to PROOF OF THE FOURIER INTEGRAL THEOREM 22, Present a heuristic demonstration of Fourier’s integral theorem by use of a limiting form of Fourier series. tat Fey = B+ 3 (ocoME + nant) : : FU) cm ™tde and b, = 26" PA sin du ition (see Problem 15), Fay = 5 roydu + 7, Pe ont (ua) du ® If we assume that Sf LF(w)| du converges, the first term on the right of (2) approaches zero as [>©, while the remaining part appears to approach in |S, fF con ua) da ® 190 FOURIER SERIES AND INTEGRALS (CHAP. 6 This last step is not rigorous and makes the demonstration heuristic. Calling 44 = =/l, (8) can be written F@) = he a BA F(n Bd) “) where we have written 10) = 2 fF" Fe) cosnu= 2) du ) But the limit (4) is equal to Fa = Sf, 1090 = 1 (Max fF cosa) du which is Fourier’s integral formula. This demonstration serves only to provide a possible result. To be rigorous, we start with the integral Lfas” F(w) cos Nu 2) de and examine the convergence. This method is considered in Problems 23-26, ie Prove that: (a) lim Sf Sa ay = dee (@) Let w=» Then Jim f , = Eby Prob Jem 43, Page 164, *"" () Let x= —y. Then lim Awe 24, Riemann’s theorem states that if G(x) is sectionally continuous in (a,6), then . lim f{ G(e)sinazdz = 0 with a similar result for the cosine (see Problem 81). Use this to prove that @) tim f “Ret 0) 8 ay = £F(e+0) ©) Tim r Fe+o yO a dv = ZF(e-0) where F(z) and F"(z) are assumed sectionally continuous in (0,1) and (~1,0) respectively. () Us Problem 23(a), seen that a proof of the given result amounts to proving that tim [e+ - Ferd = 0 This follows at once from Riemann’s theorem, because G(x) = FE+Y—FEF9 4, sectionally continuous in (0,1) since lim, F(o) exists and f(z) is sectionally continuous. (0) A proof of this is analogous to that in part (a) if we make use of Problem 23(6). CHAP. 6) FOURIER SERIES AND INTEGRALS 191 25. If F(a) satisfies the additional condition that SI. \F(@)|dz converges, prove that sin dv (@) lim [re +0) dy = ZF@+0), (b) lim f" Fle+ ea dv = EF(@~0). We have 1 ory ay = sinnw “FP inne f, Fern av =f Peto a+ f (e+ 0) 22 dy ® . : : ; a : fret Ma =f reso May + [rer SM ae ® Subtracting, St (Fle+») — Fle +0)} sind 4 @ = Poets) — (e402 aw + f Ret yay — fret ao Sf Fete ~fe+09 So rena — free oY Denoting the integrals in (8) by 1,1,,J, and Is respectively, we have J = I, +1, +1; s0 that is IM + Und + th) ® sind 1¢* Now In) = Sf [roto inde | ay iS lF(e +0] do Also Il = Fe +0) iS Since S \F(@)| de and S SMX gy both converge, we can choose 1 so large that [fal 5 Ul. y Qn Fig. 6-10 By Problem 6, 5,=0 and a = FF Pecos" de = Ef sins coune ax = 1i¢" ‘i = Lf_cosmtiye , cos(n =i (e+ ne) + sin(e—na)}de = 2 {- se 4 sont 7S, aS = ifisamectie y seemsentl 1 fttagnee _ tem Fl Fi T Fleet wet | = eee) fnwt, Ao Oo Then =F) = 28, Show that [ Za, = Zes, 2x0. Let F(x) = e~* in the Fourier integral theorem Fe) = 2 Jf cos hw da J Flu} cos hu du ‘Then af cons an f e-* cosnu du = e# coe coat de = gla, we have 2 (core = ge "cosh re CHAP. 6] FOURIER SERIES AND INTEGRALS 198 1-a 0SAS1 0 A>." 29, Solve the integral equation Sf ” F(a) cosan da = { 7 Let Fie) conne de = f{0) and choose (0) = integral theorem, ' Fa) = 2¢ foyeosreds = 2f" a-neonrea = 2Aa0e2 J, ed 21 — conn) {> OSAS1 i Noa: Then by Fourier’s 80. Find (a) the finite Fourier sine transform and (b) the finite Fourier cosine transform of aU/ox where U is a function of and t for 0<20. (@) By definition the finite Fourier sine transform of aU/az is, on integrating by parts, "aU nee nef! ome nee Bain de = Ue,0 inl Tf Ue co de oe or . #.{iz} () The finite Fourier cosine transform is ‘e{U} Ch aU nee _ nee , OT dx = U(a,t) cos] on Fe {z} = -2ERW) — (U0, — UGH cose) 81. Work Problem 80(a) and (b) for the function a°U/a2*, Replacing U by #U/az in the results of Problem 30, we find eul _ _azy jeu ws fee) = rf = SE pW) + 7FW0.9 — 0,0 cosm) ® 7{2a} S -99.{ = (U0, — Ut) conne) = —ME FW) — 040, — Us,0 comme} where U, denotes the partial deriv: ive with respect to =. 82. Use finite Fourier transforms to solve au eu Se = Ger Ut) = 0, U(A,t) = 0, U(e,0) = 22 where 0<2<4, t>0. Take the finite Fourier sine transform (with = 4) of both sides of the partial differential equa- tion to obtain 194 FOURIER SERIES AND INTEGRALS (CHAP. 6 WU Mp = (1 OU nae arsine de = Sf 2S sin ae Writing u = ¥,{U} and using Problem 31(a) with the conditions U(0,1) = 0, U(4,t) = 0, we find au Lm a ~ — ie" ® where u = u(n, t). ‘Taking the finite Fourier sine transform of the condition U(z,0) = 22, we have as in Prob. 10(a) ‘uln0) = Fy {2x} = 32d —cosne) ® Solving the differential equation (1), we find if ¢ is an arbitrary constant u = wn.) = cemns oy Since ¢ = u(n,0), we have from (2) and (3) = 821 —cosne) Thus from Problem 9(a), the inverse Fourier sine transform is 2 § 92 — cosne) Ul, t) ik a u en atetenie 6 (: = cos 2) eonteti6 Physically, U(e, 1) represents the temperature at any point « at any time ¢ in a solid bounded by the planes «=0 and 2=4. The conditions U(0,t) = 0 and U(4,t) = 0 express the fact that the ends are kept at temperature zero, while U(z,0) = 2x expresses the initial temperature as a function of 2. Equivalently, the solid can be replaced by a bar on the = axis with endpoints at «=0 and # ‘whowe surface in Tnsulated. 33. solve Y= Y, 250, t>0, subject to the conditions fl O0 0 [see Problem 32]. Supplementary Problems FOURIER SERIES, ODD AND EVEN FUNCTIONS, FOURIER SINE AND COSINE SERI 34. Graph each of the following functions and find their corresponding Fourier series using properties of even and odd functions wherever applicable. 8 0e (b) Determine the limit of this transform as «0+ and discuss the result. Ans (a) SB, (6) 1 52. (a) Find the Fourier transform of F(z) { . ; -2 |ala- (®) Evaluate J, (z2e52™2) con Z dz. an ce) a(S28.=HMA), gy te F 0 221 In each case obtain the graph of F(z) and its transform. Ama. (a) 4. It Fle) = { °S*<1 nd the (a) Fourier sine transform, (b) Fourier cosine transform of F(2). 1 cosa sind Te &) 55. (a) Find the Fourier sine transform of e-*, 2 = 0. eerT de = Ze-m, m>0_ by using the result in (a). (0) Explain from the viewpoint of Fourier’s integral theorem why the result in (4) does not hold for m=0. Ans, (a) ML+X4) (©) Show that f” 56. Solve for ¥(2) the integral equation . 1 ose0.) Ans. (a) 2/4, (6) 2/4 58. Use Problem 64 to show that (a) S- (att) ae =3 oF 58. Show that, somes sin 2) dy ° 198 FOURIER SERIES AND INTEGRALS MISCELLANEOUS PROBLEMS 60, If —r<20 U(0,t) = 0, Ul4,t) = 0, Ule,0) = (b) Give a possible physical interpretation of the problem and solution. Ans, (a) Ule,t) = 8e-2" sin gx — 2e-84 sin Box at. seve = BY, 906, 10, saber to the conditions UO) = 0, UGH =0, Ue, = {i and interpret physically. A . vey = §2ft Sn il 2 ae 3<2<6 CHAP. 6| FOURIER SERIES AND INTEGRALS 199 68. nL. 72. 2B. m. B. 16. 1. (a) In solving the problem aw ay at ax? U,A0, t) = 0, U,(6,t) = 0, U(a,0) = 2x O0 which transform [sine or cosine] would you expect to be more useful? Explain. (8) Find the solution to the problem in (a). (<5 22) eninge ana (6) 64 S (080 A fiexible string of length r is tightly stretched between points «= 0 and @ =x on the = axis, its ends fixed at these points. When set into small transverse vibration the displacement ¥(z, t) from the » axis of any point « at time ¢ is given by SF = q2%¥ where o?=T%p, T=tension, » = mass per unit length. ad (a) Using finite Fourier transforins, find a solution of this equatior (sometimes called the wave equa- tion) with a? =4 which satisfies the conditions Y(0,t) = 0, Y(t) = 0, Y(z,0) = 01 sine + 0.01 sin dz, Y,(2,0)=0 for 0<20. (®) Interpret physically the boundary conditions in (a) and the solution. Ans. (a) Y(z,t) = 0.1 sina cos 2t + 0.01 sin 4z cos 8¢ 2 (6) Solve the boundary-value problem 2 = 92% ¥(en0) = 0.052(2—2), ¥; (2,0) =0, where 0<@<2,t>0, (6) Interpret physically. subject to the conditions ¥(0,t) = 0, ¥(2,t) 16 zt Ane. (a) Y(z,t) = Solve the boundary-value problem 20 = 2Y, yoo,y = 1, Ue, = 8, Ule,0) = 2 where O0, = Ana Ula.) = 1+ Me Ae08R 6a gin ne , Give a physical interpretation to Problem 71. Solve Problem 70 with the boundary conditions for ¥(#,0) and Y,(z,0) interchanged, i.e. Y(#,0) Y;(z,0) = 0.05x(2— 2), and give a physical interpretation. 32g 1g, Det Dee gp, 812M. 804 2, Gn Ans. Y(@,t) = Prove the results (4) and (5) on Page 174, 1 [el < 0 jei>1° Verify the convolution theorem for the functions F(z) = G(z) = { Write Parseval’s identity in complex form using the results (4) and (5) on Page 174, Prove the result (15) on Page 176. Prove the results (19) and (21) on Pages 176 and 177 respectively. 200 FOURIER SERIES AND INTEGRALS (CHAP. 6 19. Prove the results (28) or (24) on Page 177. [ne iff) = S- 0" Fw) du and g() = s e-Bv G(x) dv, then FO) 00) = f S eH AG+0) F(x) G0) du dv Now make the transformation uv = =| 80. If f(\) and g(\) are the Fourier transforms of F(x) and G(x) respectively, prove that Sf rae = LS miu where the bar signifies the complex conjugate. 81, Prove Riemann’s theorem (see Problem 24). 82. (a) Show how to use Fourier transforms to solve au 2u oe = the eo if U(0,t) = 0, Ule,0) = 7%, and U(e,t) is bounded. () Give a physical interpretation. aan Uien = Ef MER NA ay a. (a) Solve 2 = 22, 10,0 =0, Ule,0) =e #>0, Ule,t) is bounded where 2>0, t>0. (®) Give a physical interpretation. 2 yet gin Ne Ans. U(e,t) = 2 f Mee tsine J, FT 2 0SeS1 + U(e,t) is bounded where »>0, t>0. o a>. Ans. 85. (a) Show that the solution to Problem 38 can be written snvi aroave Ue) = #S, oe de — Tf ‘dv ee a-wnavt 2 aU FU ana the conditions of Problem 83. () Prove directly that the function in (a) satisfies $Y = 27 Chapter 7 THE COMPLEX INVERSION FORMULA If f(s) = £(F(t)}, then 2 (f(s)} is given by Le FQ = 94 S.. et f(s) ds, t>0 @ and F(t)=0 for <0. This result is called the complex inversion integral or formula, It is also known as Bromwich’s integral formula. The result provides a direct means for obtaining the inverse Laplace transform of a given function f(s). The integration in (1) is to be performed along a line s=y in the complex plane where s=a+iy, The real number y is chosen so that 8 = lies to the right of all the singu- larities (poles, branch points or essential singularities) but is otherwise arbitrary, THE BROMWICH CONTOUR In practice, the integral in (1) is evaluated by considering the contour integral 1 ai § e* 10) as @ where C is the contour of Fig. 7-1. This con- tour, sometimes called the Bromwich contour, is composed of line 4B and the are BJKLA of a circle of radius R with center at the origin O. If we represent arc BJKLA by f, it follows from (1) that since T = /R?—9, Fi) = tin ty Se f(6) ds (8) naw oat = lim fen § e f(9) ds ~ ad, e* f(s) as} USE OF RESIDUE THEOREM IN FINDING INVERSE LAPLACE TRANSFORMS Suppose that the only singularities of f(s) are poles all of which lie to the left of the line s = for some real constant y. Suppose further that the integral around I in (8) approaches zero as R>. Then by the residue theorem we can write (8) as 201 202 THE COMPLEX INVERSION FORMULA (CHAP.7 F(t) = sum of residues of e* f(s) at poles of f(s) () = residues of e* f(s) at poles of f(s) A SUFFICIENT CONDITION FOR THE INTEGRAL AROUND 1 TO APPROACH ZERO ‘The validity of the result (4) hinges on the assumption that the integral around T in (8) approaches zero as R>. A sufficient condition under which this assumption is correct is supplied in the following Theorem 7-1. If we can find constants M>0, k>0O such that on T (where #= Re"), M Wel < Re (6) then the integral around 1 of e%f(s) approaches zero as R>=, ie, Jim S. etf(s)de = 0 (6) ‘The condition (5) always holds if f(s) = P(s)/Q(s) where P(s) and Q(s) are polynomials and the degree of P(s) is less than the degree of Q(s). See Problem 15. ‘The result is valid even if f(s) has other singularities besides poles. MODIFICATION OF BROMWICH CONTOUR y IN CASE OF BRANCH POINTS If (8) has branch points, extensions of the above results can be made provided that the Bromwich contour is suitably modified. For example, if f(s) has only one branch point at s=0, then we can use the contour of Fig. 7-2. In this figure, BDE and LNA represent arcs of a circle of radius R with center at origin O, while HJK is the arc of a circle of radius ¢ with center at O. For details of evaluating inverse Laplace transforms in such cases see Prob. 9. CASE OF INFINITELY MANY SINGULARITIES If we wish to find the inverse Laplace transform of functions which have infinitely many isolated singularities, the above methods can be applied. In such case the curved portion of the Bromwich contour is chosen to be of such radius R» so as to enclose only a finite number of the singularities and so as not to pass through any singularity. The required inverse Laplace transform is then found by taking an appropriate limit asm. See Problems 13 and 14. CHAP. 7) THE COMPLEX INVERSION FORMULA. 208 Solved Problems THE COMPLEX INVERSION FORMULA 1. Establish the validity of the complex inversion formula. We have, by definition, fo) = fem Pw) du Then ver Xi fee Fw aude tate Bei Letting s = y+ iy, de = idy, this becomes tim gh en S ewe af =O [o-™ F(w)] du te en [2re-t() > 0 ° to jo" t20 by Fourier's integral theorem (see Chapter 6]. ‘Thus we find t>o as required. In the above proof, we assume that ¢-”F(u) is absolutely integrable in (0,*), ie, SJ” e*iF(| du converges, so that Fourier’s integral theorem can be applied. To insure this ~ condition it is sufficient that F(@) be of exponential order 7 where the real number 7 is chosen so that the line ¢=y in the complex plane lies to the right of all the singularities of f(s). Except for this condition, y is otherwise arbitrary. 2, Let r denote the curved portion BJPKQLA of the Bromwich contour [Fig. 7-3] with equation s = Re, 0,0 = 27-0, ie. Tis the are of a circle of radius R with center at O. Suppose that on Tr we have M Weil < Fe where k>0 and M are constants. Show that im fo 4 do = 0 Fig. 7-3 It Typ 1) and 14 represent ares BJ, JPK, KQL and LA respectively, we have S oneyae L S ett) an + J, ert y(s)ds + f r 7 ert fle) de + S ett f(s) ds ‘Then if we can show that each of the integrals on the right approach zero as R~= we will have proved the required result, To do this we consider these four integrals. Cas Integral over 1, or BJ. Along T we have, since #= Rel’, 659 5 x/2, he S et f(s)de = 204 THE COMPLEX INVERSION FORMULA (CHAP. 7 2 Then Th SM oemeonnn jon 6 i [f(iRe'®)| [iRe'*| do =f 77 acon \f(Re¥)| do . Mu Bef emeroide = gis f* etrmoray where we have used the given condition |f(s)| = M/R on 1 and the transformation @ = 1/2—¢ «" B where o = 2/2 = sin (VR). Since sing < sin gy & cos = y/R, this last integral is less than or equal to ow ‘ey ‘ Mer to Met oy Bo from = MI 2 Mey But as R-, this last quantity approaches zero [as can be seen by noting, for example, that sin-"(y/R) = Y/R for large R). Thus lim 1, Case 2. Integral over P, or JPK. Along (, we have, since # = Rel, r/2S 0S, te = fet pteyde =f" eRe jume ints ao J, 2 Then, as in Case 1, we have i upon letting @ = r/2+ 4. Now sing = 2y/z for 0 = y ©, ‘Thus lim Ip = je 3. Integral over ['y or KQL. ‘This case can be treated in a manner similar to Case 2 [see Problem 58(a)]. Case 4. Integral over 1, or LA. This case can be treated in a manner similar to Case 1 [see Problem 58(6)}. 3. Show that sing = 2g/r for 0552/2, Method 1. Geometrical proof. From Fig. 7-4, in which curve OPQ represents an are of the sine curve y @ and y = 26/7 represents line OP, it is geometrically evident that sing = Mle for 05 9 5 e/2 Method 2. Analytical proof. Fig. 7-4 Consider Fig) = “22, We have ae dp CD CHAP. 7) THE COMPLEX INVERSION FORMULA 205 If Gig) = gcosg — sing, then Ge) = -esing @ Thus for 0 = # < 2/2, G(s) $0 and Gig) is a decreasing function. Since G(0) = 0, it follows that Gig) 20. Then from (1) we see that F'(s) = 0, or F(g) is a decreasing function. Defining F(0) = Tim F() = 1, wo seo that F(g) decreases from 1 to 2/» as 4 goes from 0 to x/2. Thus sing 2 2 $ 7 1 w from which the required result follows. USE OF RESIDUE THEOREM IN FINDING INVERSE LAPLACE TRANSFORMS 4. Suppose that the only singularities of f(s) are poles which all lie to the left of the line 8=7 for some real constant y. Suppose further that f(s) satisfies the condition given in Problem 2. Prove that the inverse Laplace transform of f(s) is given by F(t) = sum of residues of e* f(s) at all the poles of f(s) yr af ejede = ib S, et foyde + ah J. et fle) ds where C is the Bromwich contour of Problem 2 and I is the eircular are BJPKQLA of Fig. 7-3. By the residue theorem, & § et f(a) ds = sum of residues of e* f(a) at all poles of f(s) inside C wt So = & residues inside C Thus mr a S 7, Ot fle) de = resides inside @ — & Sf ot #(0) de ‘Taking the limit as R-> ©, we find by Problem 2, Fi) = sum of residues of e* f(e) at all the poles of /(e) 5. (a) Show that f(s) % satisfies the condition in Problem 2. ‘ (0) Find the residue of 5 at the pole #=2. (©) Evaluate «fh by using the complex inversion formula. (a) For #= Rel, we have i {Re 2 = laa 206 THE COMPLEX INVERSION FORMULA (CHAP. 7 for large enough R (e.g. R>4). Thus the condition in Problem 2 is satisfied when k=1, M=2. Note that in establishing the above we have used the result 2; ~ za! & jz! — [ea] [see Problem 49(¢), Page 167]. (0) The residue at the simple pole # (0) By Problem 4 and the results of parts (a) and (6), we see that cad { 5 Note that the Bromwich contour in this case is chosen so that y is any real number greater than 2 and the contour encloses the pole # = 2. a} = sum of residues of o fle) = o 6. Evaluate ef 1 } by using the method of residues. (6+ 1)(e-2) Since the function whose Laplace inverse is sought satisfies condition (5) of the theorem on Page 202 [this can be established directly as in Problem 5 or by using Problem 15, Page 212), we have ej LL ede Gripe mf = aid, @r nea 1 etds ari J, eT Dea? wry = Z residues of ORT ED FE a, 104 (ag/agR3)e-20 + - FQ plo ge H + (aylagh™)e— m0 + FBR Let A denote the maximum of |ay/agy las/aq, . ..» lay/agl. Let B denote the maximum of |b;/bo|, |by/bo!, - --, |by/bol- Then a, os 1+ shew 4 2. | + aR | ene] os We ee for R>A+1. Also, fs ' for R > 2B+1. ‘Thus for R larger than either A+1 or 2B +1, we have wot = [e|-ats- ah where M is any constant greater than 2lap/by! and k = n—m = 1, This proves the required result, , _, {cosh avi} 16. Find £ {aan aval where 0<2 1. (®) Give an interpretation of your answer as far as Fig. 7.9 Laplace transform theory is concerned. 216 THE COMPLEX INVERSION FORMULA [cHaP.7 25, Use the inversion formula to evaluate ¢ f } where a and 6 are any positive constants. = 28. Use the inversion formula to work: (a) Problem 18, Page 63; (b) Problem 25, Page 68; (c) Problem 28, Page 60; (d) Problem 110, Page 74. INVERSE LAPLACE TRANSFORMS OF FUNCTIONS WITH BRANCH POINTS 2, Find £-1{e-¥%} using the complex inversion formula. 28. Find ¢-1 {3} by the inversion formula. erf Vt by using the inversion formula. 29. Show that ¢-1 {—L_| leat 30, Find got {5} by using the complex inversion formula. BL, (a) Use the complex inversion formula to evaluate £1 {¢ method, 9} and (b) check your result by another 32, Evaluate £-1{In (1+ 1/s)} by using the inversion formula, Ana. (1 — e~9/t 38, Evaluate -!{In (1 + 1/e%)} by the inversion formula, Ama. 2(1 — cos #)/t INVERSE LAPLACE TRANSFORMS OF FUNCTIONS WITH INFINITELY MANY SINGULARITIES 34 Find g-t } using the complex inversion formula. 1), e 1 ft i Prom iat ctf 38, Find tf 87. By using the complex inversion formula, prove that fd = Hema) _ 20? J 1 net ns ‘sinh as ~ 6a a co a _ 1 sinu(t-+a) , 1} _cos(2n—1)rt/2a eG et Tea) Be Gatto — (On 14a” MISCELLANEOUS PROBLEMS 89. Evaluate (a) £-'{1/(@—1)4}, (6) £-1{e-*/(¢—1)8}, by using the complex inversion formula. 40, Find efge} by contour integration. Ams, t cost 1 {2 2 (80? cos 84) sin 41, Evaluate {rts} - Ana, Z{3t® cost + (t®— 84) sin t) CHAP. 7] THE COMPLEX INVERSION FORMULA 217 2, a. BL. 52. Find 71 aves} by the complex inversion formula and check your result by another method. satisfies the conditions of Theorem 7-1, Page 202. 8S ue (nat 1 n= DEMME t) xt (a) Prove that the fonction fle) = yA © Proweme erfadag} = 6+ Discuss the relationship between the results of Problem 43(b) and Problem 35. Evaluate ¢-! {ria} by the inversion formula, justifying all steps. Ans, (sin t cosh t — cos t sinh t) (a) Prove that if © >0, _, [aen2¥F alae (0) Prove that for large values of ¢ the integral in p@rt (a) can be neglected. sinh @coshe Prove that for 0< 2 <1, Find 71 {sate}. 4 sin (2n — tre /2 Prove that for 0<2 <1, Vi cosh Vo e+ Qn— 1P/e Show that -1fin(1 + ve) _ 1 ~costt+e) » 1S yy (y eofageueth = Lneetee 4 F3,m(1 Show that for 0<2 0; Oi t<0 87. Evaluate ds where y>0. (6) How can you check your answer? Complete the proofs of (a) Case 3 and (6) Case 4 of Problem 2. 59. A periodic voltage E(t) in the form of a half-wave rectified sine curve as indicated in Fig. 7-10 is applied to the electric cireuit of Fig. 7-11. Assuming that the charge on the capacitor and current are zero at t=0, show that the charge on the capacitor at any later time t is given by * Bo | sinut — sin w(t +7) inat — sin a(t +7) | ey = Bo { sinut — sinw(t+7) , sinat ~ sino(t+7) 2rBy cos 2ent/T + or wet (a? = 4r?n?/T?)(a? — de?n®/T?) where o® = 1/LC, o? = 22/T? and’ wXa. BO) 7/2 Tr 37/2 ar Fig. 7-10 Fig. 7-11 ‘Work Problem 59 in case 2 = and discuss the physical significance of your results, 61, Verity Theorem 7-1, Page 202, for the function «~¢¥i/e, a>0 [sce Problem 9}. 62. ;\ where a > 0, by use of the inversion formula and check by another method. 68, Prove that cotferwy =f" vtert-wrtsin VBP ae. 5 G4, Generalize the result of Problem 63. 65, A spring of stiffness k and of negligible mass is suspended vertically from a fixed point and carries a mass m at its lowest point. The mass m is set into vibration by pulling it down a distance =, and releasing it. At each time that the mass is at its lowest position, starting at ¢=0, a unit impulse is applied, Find the position of the mass at any time ¢>0 and discuss physically. Chapter 8 BOUNDARY-VALUE PROBLEMS INVOLVING PARTIAL DIFFERENTIAL EQUATIONS Various problems in science and engineering, when formulated mathematically, lead to partial differential equations involving one or more unknown functions together with certain prescribed conditions on the functions which arise from the physical situation. The conditions are called boundary conditions. The problem of finding. solutions to the equations which satisfy the boundary conditions is called a boundary-value problem. SOME IMPORTANT PARTIAL DIFFERENTIAL EQUATIONS 1, One dimensional heat conduction equation au eu ot ae Here U(z, t) is the temperature in a solid at position x at time t. The constant k, called the diffusivity, is equal to K/ep where the thermal conductivity K, the specific heat ¢ and the density (mass per unit volume) p are assumed constant. The amount of heat per unit area per unit time conducted across a plane is given by — K U.(a,t). 2 One dimensional wave equation This is applicable to the small transverse vibrations of a taut, flexible string initially located on the « axis and set into motion [see Fig. 8-1]. The variable Y(z, t) is the dis- placement of any point « of the string at time t. The constant a*=T/p, where T is the (constant) tension in the string and p is the (constant) mass per unit length of the string. Fig. 8-1 3 Longitudinal vibrations of a beam ay ot ey oat This equation describes the motion of a beam (Fig. 8-2) which can vibrate longitudi- nally (i.e, in the @ direction). The variable ¥(z,t) is the longitudinal displacement from the equilibrium position of the cross section at z. The constant c= gB/p, where g is the 219 Fig. 8-2 220 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 acceleration due to gravity, E is the modulus of elasticity (stress divided by strain) and depends on the properties of the beam, p is the density (mass per unit volume) of the beam. Note that this equation is the same as that for a vibrating string. ey ay 4, Transverse vibrations of a b oad verse vibrations of a beam oe tS = 0 This equation describes the motion of a beam (initially located on the z axis, see Fig. 8-3) which is vibrating transversely (i.e. perpendicular to the z direction). In this case Y(c, 1) is the transverse displacement or deflection at any time ¢ of any point 2, The constant b* = Elg/p where E is the modu- lus of elasticity, I is the moment of inertia of any cross section about the x axis, g is the acceleration due to gravity and p is the mass per unit length. In case an external trans- verse foree F(z, t) is applied, the right hand side of the equation is replaced by b* F(c, t)/EI. au 5. Heat conduction in a cylinder at Here U(r,t) is the temperature at any time t at a distance r from the axis of a cylindrical solid. It is assumed that heat flow can take place only in the radial direction. 6. Transmission lines aE _ py pt an ot a oF je = GE - OF These are simultaneous equations for the current J and voltage E in a transmission 0} —________—« ae 1 line [Fig. 8-4] at any position « and at any "7 time t. The constants R,L,G and C are re- Generator, spectively the resistance, inductance, conduct- Mi ance and capacitance per unit length. The end 2=0 is called the sending end. Any Fig. 8-4 other value of x can be considered as the receiving end. TWO AND THREE DIMENSIONAL PROBLEMS Many of the above partial differential equations can be generalized to apply to problems in two and three dimensions. For example, if Z(x,y,t) is the transverse displacement of any point (x,y) of a membrane in the zy plane at any time ¢, then the vibrations of this membrane, assumed small, are governed by the equation #2 _ 9p (#2 4 2) oF oat * oy ® CHAP. 8] APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 221 ae (2 eo | PP =@ Similarly, = ata =) = ays 2) where V’¢ is called the Laplacian of (x,y,2,t), is the equation for the transverse vibra- tions of a pulsating membrane in three dimensions. The general equation for heat conduction in a three dimensional solid is, assuming constant thermal conductivity, specific heat and density, au eu | FU | #U A oe (B+ &) RU (3) The equation for steady-state temperature {where U is independent of time so that aU/at = 0] is uv oo al Fl wet apt me vu ‘= 0 @) which is called Laplace’s equation. This is also the equation for the electric (or gravita- tional) potential due to a charge (or mass) distribution at points where there is no charge (or mass). SOLUTION OF BOUNDARY-VALUE PROBLEMS BY LAPLACE TRANSFORMS By use of the Laplace transformation (with respect to t or 2) in a one-dimensional boundary-value problem, the partial differential equation (or equations) can be transformed into an ordinary differential equation. The required solution can then be obtained by solving this equation and inverting by use of the inversion formula or any other methods already considered. For two-dimensional problems, it is sometimes useful to apply the Laplace transform twice [for example, with respect to t and then with respect to 2] and arrive at ordinary differential equation. In such case the required solution is obtained by a double inversion, The process is sometimes referred to as iterated Laplace transformation. A similar tech- nique can be applied to three (or higher) dimensional problems. Boundary-value problems can sometimes also be solved by using both Fourier and Laplace transforms [see Prob. 14]. Solved Problems HEAT CONDUCTION 1. A semi-infinite solid 2 >0 [see Fig. 8-5] is initially at temperature zero. At time t =0, a constant temperature Us >0 is applied and maintained at the face e=0. Find the temperature. at any point of the solid at any later time t> 0. The boundary-value problem for the deter- mination of the temperature U(x,t) at any point « and any time ¢ is APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 222 au. aU = eZ a >0, t>0 1U@,6| 0. -value problem in this case is the same as in the preceding problem except that the Us is replaced by U(0,t) = G(t). Then if the Laplace transform of o(s) and 50 The boundar boundary condition U(0, ¢) G(t) is g(s), we find from (8) of Problem 1 that ey ule,a) = g(e)e~Vare= Now by Problem 11, Page 209, 1 {9-VTRay = 1-8/2 g-st/aKt otk } Wa Hence by the convolution theorem, # un 8/ e-28/4k Gt — u) du Sowa U@,t) : z fa a oval Ve sav on letting v = 2%/dku. CHAP. 8} APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 223 3. A bar of length | [see Fig. 8-6] is at constant temperature Us. At t=0 the end x=1 is suddenly given the constant temperature U; and the end 2=0 is insulated. Assuming that the surface of the bar is insulated, find the temperature at any point x of the bar at any time t>0. #2u at Wage Oo Ue) = Ue U0) = 0, Ube) = 0; Taking Laplace transforms, we find ‘ Pu du _ sw U dT ae ® eu — U(e,0) = u(0,8) = 0, wll ® ‘The general solution of (2) is 7 U u = cycosh Velk2 + cesinh Vek + —* From the first condition of (¢) we find o:=0 and so U u = ecosh Valk + —° From the second condition of (2) we find Ue uy U,— Uo eyeosh Vekt + 2 = + or og = —* + . * * ‘cosh Va/kL cosh Valk x # cosh Vek The inverse of the first term is Uy. By the complex inversion formula, the inverse of the second term is, apart from the constant factor U,—Us, given by Thus un) = 2+ w-00 A 77, cosh Valls Bei J, te @ cosh Valk As in Problem 18, Page 210, this is easily shown to be equal to the sum of all the residues of the integrand at the poles which are all simple poles and occur at #=0, sikt = (n~ Brim = 0,41, 22... or Gray = 12,8, 224 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 Now: Residue at # = 0 =u Residue at ¢ = — Gacy = 4 is lim (s—2) Bn, 90 oavat) i et cosh Volk x = lim lim ——— {a8 eva (ae “= } = { im ——_1 H{ et (sinh Valk D(U2V ke), = MEI can aitatee4t eos (28 = Dee = ASD ec urat cos using L’Hospital’s rule. Thus we obtain . 4(U,~U) = Ue = Uy + THE VIBRATING STRING 4. An infinitely long string having one end at x=0 is initially at rest on the x axis. The end z= 0 undergoes a periodic transverse displacement given by Aysinwt,t>0. Find the displacement of any point on the string at any time. Y BN . an ot Fig. 8.7 If Y(z,1) is the transverse displacement of the string at any point at any time ¢, then the boundary-value problem is 2 eM asaeso aw ¥(#,0) = 0, ¥e(«,0) ¥0,0 = Aysinot, |¥(@,0| 2/0 _ me Yen {° t0. The boundary-value problem is zy ay Ls po O<20 YO.) = YU) = 0, ¥(z,0) = uxll—2), YY, (2,0) = 0 ‘Taking Laplace transforms, we find, if y(z,s) = £{¥(x,t)), sty ~ 8¥(2,0) — ¥, (2,0) my =s or Gb-Sy = -aaize w where (0,3) = 0, y(s) = 0 (2) The general solution-of (1) is v= eycosh®= + cp sinh + #22) _ 2aty ) ‘Then from conditions (2) we find ame cg = 2/1 = cosh sl/a _2a%y a= Hg = Hh ) = eames Py so tha (memes y= ACoA — Nite all) _ Bae By using residues [see Problem 17, Page 213] we find Yeo = onfe + (327 _ (4)} #) SAW gg BR = Dele =D 6g, On = peat sr Bn — 1 8 a 226 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (omar. 8 VIBRATIONS OF BEAMS 6. A beam of length 1 which has its end 2=0 fixed, as shown in Fig. 8-8, is ini- tially at rest. A constant force Fo per I unit area is applied longitudinally at the free end. Find the longitudinal displace- ment of any point « of the beam at any Fig.8-8 time t > 0. If Y(q, ¢) is the longitudinal displacement of any point x of the beam at time ¢, the boundary-value problem is ey _ ey att O0 ¥(@,0) = 0, ¥,(2,0) = 0, ¥@,8) = ¥.(,t) = PvE where E is Young’s modulus. Taking Laplace transforms, we have, if u(z,s) = £{¥(2,0)}, @y cd 42,8) — 8¥(20) — Ye(e0) = Ee or EE Gy = 0 . W0,8) = 0, alla) = Fo/Bs ® Solving the differential equation, we find u(e,8) = cycosh(sz/c) + ey sinh (sx/e) From the first condition in (1), ¢; and so y(z,8) = eg sinh (sa/e) us(®,8) = (#/e) cosh (a/c) From the second condition in (1), we have e,(s/e) cosh (el/e) = Fy/Ee or c = A _ Fo, sinh (s/c) Then vee) = SP Feosh (o/c) (2) Hence by Problem 14, Page 211, _ Fo BLS _(" , n—Are | (2n— 1 )eot von = Ble + SS geting em Bape] © 7. In the beam of the preceding problem, determine the motion of the free end z=1 as a function of time t. For « =1 we obtain from (2) of Problem 6, eFo_ sinh (el/c) Wes) = ER cosh (all) But from Problem 92, Page 34 or entry 134, Page 258, this is the Laplace transform of the triangular wave of Fig, 8-9 below which describes the motion of end =I as a function of t. CHAP. 8| . APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 227 Yt) Fig. 89 8& A semi-infinite beam which is initially at rest on the x axis is at time t=0 given a transverse displacement h at its end «=0, Determine the transverse displacement ¥(q,t) at any position «>0 and at any time t>0. The boundary-value problem is yuh =o o Y¥(z,0) = 0, ¥,(x,0) = 0, ¥(0,t) = |¥(z,t)| < M (2) Taking Laplace transforms, we find even — 0%) — 10.0408 = 0 we SeEy = 0 (0,8) = h/s, ‘zx(0,8) = ‘y(a,8) is bounded (3) The general solution of the differential equation is vl2,s) = eVFRBx(o, cos Va7Eb x + eq sin VaR 2) + 0-72 (ey cos Vax + cy sin Va/2 2) om the boundedness condition we require ¢,= ey =0 so that u(x.) = eV (05 cos Vs/2b x + cy sin Vs/2b x) From the first and second boundary condition in (8), we find ¢,=0 and ¢,=h/s so that (2,0) = ber Virb: cos Veit x The inverse Laplace transform is, by the complex inversion formula, Yea) = gly FRAN s con VarBb a dy ‘To evaluate this we use the contour of Fig. 8-10 since 5 =0 is a branch point. Proceeding as in Problem 9, Page 207, we find, omitting the integrand for the sake of brevity, that Yeo = “ayaa f Se $} ° 228 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 Along BH, # =e, V#= iV@ and we find So = SOMA co Vi ay Along KL, #=ue-™, Vé=—ivw and we find (Thee NTHRe enh VEE 2 gy f > [ene Along HJK, 6 = S OT pettitt= VeRO Go ecb x do HK i ‘Then (4) becomes Yet) = af _ 1¢ enue AVM econ VAAPS aut Letting 1/2 = v®, this can be written Ye) = fa - 26" 2M sin cosh ve «| The result can also be written in terms of Fresnel integrals as |see Problem 66 and entries 10 and 11, Page 255) ING Ye = af = fF? eon we + sinh del reo TRANSMISSION LINES 9. A semi-infinite transmission line of negligible inductance and conductance per unit length has a voltage applied to it at the sending end, x=0, given by 0,1) = {e oT Find the voltage E(z, t) and current I(x, t) at any point «>0 at any time t>0. If we take L=0 and G= ), the transmission line equations are given by oe 3n a ‘The boundary conditions are By o, Fan =f LR ane nosed 2, ¢eVRCANET = — ot dv (letting RCz*/4u = v2) Ve VevRCravE _ a{2 [mn Oe 2 2 fe a} Veo verve = Pafert (BEE) - 0t(22)} avi=T ave while if 00,t>0 (2) Give a heat flow interpretation of the problem. 230 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 ‘The problem arises in considering a semi-infinite conducting solid whose initial temperature is Uy in which radiation into a medium 2 <0 at temperature zero can take place. This radiation is assumed to be such that the flux at the face «=0 is proportional to the difference in temperatures of the face #=0 and the medium # <0, ic, Uz0, = —a[U0,)-0] = —aU(0,t) To obtain the solution, we take Laplace’ trans- forms and find eu su — Uy = oo mp zoe «) 4,(0,8) = —au(0,e), —ultya) is bounded ey ‘The general solution of the differential equation is U ule,8) = cyeV7Re + ce Velkx 4 —* From the boundedness condition, o=0. ‘Then Ur uez,e) = ge Viex 4 Ue 2 sti + ay From the first condition of (2), we find * ey ——— _ so that avE— =) view) = 22 prone 4 Ue aVa~@) is ‘Then using the complex inversion formula, fx Ule,t) = Uy + ay otf a = aUy (7+! gst Virz Bettie elVs—a) As in Problem 8 we have, omitting the integrand, a pa ae . ~ peak J + fe st @ coo Ud RL. = 0+ Along EH, #=ue™, ¥e=iVu and we find gut ~iVa7ee f= (ote in ulivu — a) Along KL, #=ue-™, ¥s=—ivu and we find (= fe «wiv — a) Along HJK, # =e and we find Using these results in (8), we see that (1 gst Valz Bri, on a(VE— a) CHAP. 8] APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 231 Hence vey = 2 Se [eee =p sinav ee 2] du 2ae = if u=ot, 11. A taut, flexible string has its endpoints on the x axis at e=0 and a=1. At time t=0, the string is given a shape defined by F(z), 0<2<1, and released. Find the dis- placement of any point « of the string at any time t>0. ‘The boundary-value problem is ae = eth o0 Oy YO, = va, ¥(@,0) = Fle), ¥;(e,0) = 0 «) It is convenient to consider, instead of equation (1), the equation av Lay ae > Oat and after the final solution is obtained to replace ¢ by at [see Problem 49]. ‘Taking Laplace transforms, we find Py — 8¥(2,0) ~ ¥ete,0) = Sh or GU sty = —0FQ@) ® 0,2) = 0, y(1a) = 0 @) The general solution of (8) is [see Problem 8, Page 85] ues) = ecoshsx + ep sinhax — 5 Pu) sinh a(— 1) du From the first condition in (4) we finde; =0, ao that vie,s) = opsinhex — S Flu) sinh a(e~ u) de 6 From the second condition in (4) we find 0 = egsinhe = ("FQ sinha. —19 ds o a = fr HRY ay ‘Thus (5) becomes wea) = F(u) sinh o(2—u) du The first integral can be written as the sum of two integrals, one from 0 to =, the other from « to 1. ‘Then 2,9) = J Foy {sha inh sinnete-u9} au + f Poste = fre 82 ay 1 du +f Few sitbet. 7 sinh 282. APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 ‘We must now find the inverse Laplace transform. By the complex inversion formula, the inverse of the first term is cae sinh s(1 2) sinh ew maf eff Peo aa pst aa a Since this is equal to the sum of the residues at the simple poles # = nzi, we find the required inverse Son Sf” Poy Same 2) sinew gy = 3 {fr sin neu a sin nee cos nat f atte = cos ni oh Similarly, the inverse of the second term is 3 { SF sinew a} sin nz cos nzt Yen) = 3 {fire If now we replace t by at, we have Adding these we find nae ax} sin nea cos nxt =f Yeo = af jr) sinew sinnrz cosnrat & An infinitely long circular cylinder of unit radius has a constant initial temperature T. At t=0 a temperature of 0°C is applied to the surface and is maintained. Find the tem- perature at any point of the cylinder at any later time t. If (r, 9,2) are cylindrical coordinates of any point of the cylinder and the cylinder has its axis coinci dent with the « axis [see Fig. 8-12], it is clear that the temperature is independent of g and z and can thus be denoted by U(r,t). The boundary-value problem is i. w ) eerer a UG, =0, Ur, = 7, [Urb] mttaer and then to replace ¢ by kt. u(1,2) = 0, u(r,a) is bounded The general solution of this equation is given in terms of Bessel functions as urs) = eJoiVar) + 2 Yoli Since ¥o(iVS*) is unbounded as +> 0, we must choose cz = 0. CHAP. 8} APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 233 Then ura) = edotiver) + F From u(1,e) = 0, we find eJotive) + © 0 or 4 = T _ TIoliver) Thos ur, 0) 2 Gay By the inversion formula, sple zeros where V3 = AuAgy-.-Aye-+++ ‘Thus the integrand has simple poles and also at 6=0. Furthermore it can be shown that the integrand tatisfies the conditions of Problem 2, Page 203, so that the method of residues can be used. We have: Residue of integrand at #=0 is e lim ¢ r0” a dyl Residue of integrand at ¢ = —2y is ety(iVer) { 3m poy = Apa Yeas olive) f eas = 1 \{staoarh {e ag Jo (ive) /2V' | NM _ Be-¥e Zor) olin, © var FiO) where we have used L’Hospital’s rule in evaluating the limit and also the fact that J,(u) = —J,(u). Then = ra rdy— & 20 mon = refs 3 eben) Sot To (ar) = 83K Replacing ¢ by kt, we obtain the required solution vey = aS ‘sae 18. A semi-infinite insulated bar which coincides with the « axis, x>0, is initially at temperature zero. At t=0, a quantity of heat is instantaneously generated at the point «=a where a>0. Find the temperature at any point of the bar at any time t > 0. ‘The equation for heat conduction in the bar is au _ eu We rE ara tro ® 284 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS (CHAP. 8 ‘The fact that a quantity of heat is instantaneously generated at the point 2 ‘by the boundary condition @ can be represented Ula,t) = Qa(t) ® where Q is a constant and 4(t) is the Dirac delta function, Also, since the initial temperature is zero and since the temperature must be bounded, we have U(@,0) [U@,9| 0, t>0, Fig. 8-13, CHAP. 8] APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 235 ‘Taking the Laplace transform of equation (1) and using condition (4), we find, if « = u(z,y,8) = £10 (2,y. OD}, ma = o am + oy ™ ‘Multiplying (7) by sinnx and integrating from 0 to » [ie. taking the sine transform; see Page 175], we find " Pu at _ Pesinnede + { ¥sinnzde = [ susinne de Sf Bhaneete + [Bhan iia or, if 7 usin na de, eH + nulz,ys) cosms + nul0,y,8) + ie aa ® Since from the Laplace transforms of conditions (2) and (3) we have 40,458) = 0, u(ryv8) = 0 (6) becomes SE - wernt = 0 ‘This has the solution GB = Ae 4 Bens ‘From the boundedness of % as y >, we require A =0 so that @ From condition (6) Timo = f Hence letting y= 0 in (9), we find B 100 (L= conn) or a= By the Fourier sine inversion formula [see Page 175), we have 2 $10 (Le sme) «ine «1 wat ” ‘We must now obtain the inverse Laplace transform of this. We know that y sere Lor fewey 2x8 Warm prise gant so that ONT = evwiatenate ave Hence = WH 9m dy AOD ay on letting y%/4v =p * 236 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS [cHAP. 8 ‘Then inverting term by term in (10) and using this result, we find (eset) anne fete ty ™ wane Unt) Supplementary Problems HEAT CONDUCTION 15. A semi-infinite solid « > 0 has its initial temperature equal to zero. A constant heat fiux A is applied at the face r=0 so that —K U,(0,t) = A. Show that the temperature at the face after time ¢ is A [ke K 16, Find the temperature at any point > 0 of the solid in Problem 15. Ans, 4 (kth #1 — 4x exte (2/2VHt)} 17, A solid 0= 21 is insulated at both ends 2=0 and #=l. If the initial temperature is equal to az(l—z) where a is a constant, find the temperature at any point 2 and at any time &. e grikstetyt ‘ cos 2882 Ans, FE me co T 18. (a) Use Laplace transforms to solve the boundary-value problem O<2<10, t>0 U(10,) = 2, (0,4) = 0, U(x,0) = 50 (0) Give a heat flow interpretation to this problem. age 4 $400 B= canta r1800 ggg 2H = Dee Ans. (a) U(x,t) = 220 222 #7 e en a Can—18e41/1600 oy (2m ~ Luxe cos Ba 19, (a) Solve =2>0,>0 U,(0,0) = e-%, U(a,t) is bounded (®) Give a heat flow interpretation to this problem. ve Bet fate Ans. Ulet) = et — fo et atdy 7S 20. (a) A semi-infinite solid z>0 has the face z=0 kept at temperature Uy cosut, t>0. If the initial temperature is everywhere zero, show that the temperature at any point x >0 at any time t>0 Ule,t) = Uye- YORK eos (wt — VaR 2) - 2s" wer erat sin eval de (®) Show that for large ¢, the integral in the result of part (a) is negligible. CHAP. 8] APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 237 21, A semi-infinite solid 22 0 is initially at temperature zero. At t=0 the face #=0 is suddenly raised to a constant temperature 7) and kept at this temperature for a time t,, after which the temperature is immediately reduced to zero, Show that after an additional time ty has elapsed, the temperature is a maximum at a distance given by 2 = 2VktyIn2 where k is the diffusivity, assumed constant. 22, At t=0, a semi-infinite solid z>0 which is at temperature zero has a sinusoidal heat flux applied to the face r=0 so that —KU,(0,t) = A+Bsinot, t>0. Show that the temperature of the face at any later time is given by BE wn 4 PVE (cman de) inet = ("snes dr) xcs} 28, Find the temperature of the solid in Problem 22 at any point = > 0. THE VIBRATING STRING 24, (a) Solve the boundary-value problem ey _ ey Sea Bh ocece, too ¥, (0,9 =0, Yet =A, Ye.) =0, ¥y(n0) =0 () Give a physical interpretation of the problem in (a). Ans, Yanq = SS CB sing pe sin anne 25, Solve the boundary-value problem Ye = Vato 00 ¥0,0=0, Ye) and interpret physically. » ¥(@,0) = nale—2), ¥,(2,0) =0 42ux~9) 3 __1 =z mi @n— 1 Ans. Ye = 4ge(e—2) + in (2n—1)x cos (2x —I)t 26, A tightly stretched flexible string has its ends fixed at z=0 and e=1. At t=0 its midpoint is dis- Placed a distance h and released. Find the resulting displacement at any time t>0. An. Yet) = $3 Gone © cog An — Neat 2Y _ BY ae Gee ¥,(0,t) = Asinet, Yiz,0) = 21. (a) Solve (6) Give a physical interpretation to this problem. Ane (0) Yee.) = AX (ems(t—2/0) = 1) if t> aia and it t 3 fe VIBRATIONS OF BEAMS 28, A beam of length I has the end =0 fixed and 2=/ free. Its end is suddenly displaced longi- tadinally a distance @ and then released. Show that the resulting displacement at any point x at time ¢ is given by Vey = 4 MS CD yg Be ogee 238 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS [cHAP. 8 29, A beam has its ends hinged at 2-0 and x=2 At t=O the beam is struck so as to give it a trans. verse velocity Vy sin ra/l. Find the transverse displacement of any point of the beam at any later time. 80, Work Problem 20 if the transverse velocity is Vy (I~ 2). 31, A beam of length I has its ends hinged. Show that its natural frequencies of transverse oscillations are xiven by a nh = Br fFe s = oP, 1,2,3, 32, A semi-infinite elastic beam is moving endwise with a velocity —v, when one end is suddenly brought to rest, the other end remaining free. (a) Explain with reference to this problem the significance of each of the following and (6) solve the resulting boundary-value problem. Fue (zt) = a Yee(x,t) z>0,t>0 ¥(z,0) = ¥,(2,0) = vy ¥(0,f) = 0, lim ¥.(z,8) = 0 Ans. (6) Y(t) = —vgr/a if t>2/a and ~ugt if tS 2/a TRANSMISSION LINES 33. A semi-infinite transmission line of negligible inductance and conductance per unit length has its voltage and current equal to zero. At t=0, a constant voltage Ey is applied at the sending end x= 0. (a) Show that the voltage at any point « >0 at any time t > 0 is given by Bl,t) = Eq ere (eVRC/2V1) and (b) that the corresponding current is = BO IC yaa ncete ted = Ser 34, In Problem $3 show that the current at any specific time is a maximum at a pos the receiving end, ion V20/RC from 35. A semi-infinite transmission line has negligible resistance and conductance per unit length and its initial voltage and current are zero, At t=0 a voltage E,(t) is applied at the sending end 2=0. (a) Show that the voltage at any position x >0 is {i —aVEC) t>xVLC ° Be) = t<2VE6 and (6) that the corresponding current is VCILE(t-2VEC) t > VEC \ teams Ket) = such that R/L = G/C. Show that the voltage is aVEG) t> avi t<2ViC and compare results with that of Problem 85. What is the current in this case? 386, Suppose that the transmission line of Problem 35 i given by {; VRE Bt El,) = 87, (a) A transmission line of negligible resistance and conductance has its sending end at #=0 and its receiving end at «= 1. A constant voltage Ey is applied at the sending end while an open circuit is maintained at the receiving end so that the current there is zero. Assuming the initial voltage and current are zero, show that the voltage and current at any pi given by CHAP. 8] APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 239 Ep cosh VLIC (t- 2) Be, ad ‘cosh VEE tay = ENTE sinh VEIC— 29 ° cosh VLIC (b) Discuss the significance of the fact that the voltage and current in (a) are independent of time. 38. (a) Work Problem 37 if the line has negligible resistance and capacitance but not negligible inductance and conductance, showing that in this case n= I)ez (n= Let) He,t) = sf eon 28 TME ,, n= Uist coe 2 ave () What is the current in this case? Discuss the convergence of the series obtained and explain the significance, MISCELLANEOUS PROBLEMS BR.) Sale the bonndargeato probien WU ae ocecyere UO) = 0, UA,) = 0, Ue, (®) Give a physical interpretation to the problem in part (a). Ans. Ule,t) = 21-2) -4 sin nex 40." Work Problem 39 if the con Ans, Ul.) = § — at + § 41, A solid, 0<2<1, is initially at temperature zero. The face 2=0 is given a temperature UO, t) = G(e), > 0, while the end «=Z is kept at 0°C, Show that the temperature at any point « at any time t is 2s FS Ul, = FF n (2, t) 3 { i 42, Work Problem 41 if the end «= 1 is insulated. maetul Ge — u) as in 43. Show that in solving a boundary-value problem involving the equation 90 = 420 it is equivalent to solve the problem by replacing the equation by 20 = 22 and then replacing ¢ by kt. 44. A solid, 0 0,t>0 where Y(0,t) = 0, Y(#,0)=0. Ans. ¥(t,t) = x(t — 6%) ea —— jon 2% = ge ®® it te cou 49. Show that in solving a boundary-value problem involving the equation SF = a £2, it is equivalent to solve the problem by replacing the equation by 2 = 2X and then replace ¢ by at. 50. Show that a transmission line problem in which re- stance and conductance are negligible, is equivalent to a problem in the vibration of a string. 51. A string is stretched between 2 =0 and x=1. The end @ = 0 is given a transverse displacement accord- ing to Y(0, 0) = F(t) where F(t) is a prescribed func- tion of time, while the end 2 = remains fixed. Find the transverse displacement. 52, A semi-infinite plate having width r [see Fig. 8-14] has its faces insulated. The semi-infinite edges are maintained at 0°C, while the finite edge is insulated, If the initial temperature is 100°C, find the tempera- Fig. 814 ture at any point at any time. 53. A solid, 0 0 is Ue@,t) _ wr 2wE wx/l | sinSrx/l | sin5we/t Ye) = opp h—=) — at wot ge tet +} if 0 << U2, while the corresponding result for 1/2 < = <1 is obtained by symmetry. 55. Show that the boundary-value problem ww at VEO, O0 Us, U(@,0) has the solution U; sinh a(t—2) + Us sinh az anbal % nenMetl (Uy cos ne — Us). nee fF eile ant Ule,t) = CHAP. 8} APPLICATIONS TO BOUNDARY-VALUE PROBLEMS 241 Show that Problem 55 cau be interpreted as.a heat fiow problem in which a bar of length J can radiate heat into its surroundings. ST. A transverse force given by F(2) = #(r—2) acts at each point = of a beam which is hinged at its ends £=0 and «=z. If the initial transverse displacement and velocity are zero, find the transverse displacement at a later time ¢. 58. A semi-infinite transmission line of negligible inductance and conductance per unit length has a voltage applied to its sending end «=0 given by B(0,f) = Epcosut, > 0. Assuming the initial voltage ‘and current to be zero, (a) show that after a long time the voltage at any point > 0 is given by Elz, t) = By e~ Vous cos (wt — VoRC/B 2) and (b) show that the corresponding current is given by T(z,t) = By VoC7R e-VORCR cos (wt — YoRC2x — z/4) 59, A semi-infinite string is initially at rest on the « axis, and its end #=0 is fixed. At ¢=0 each point 2 of the string is given an initial velocity defined by F(z), «>0. Find the resulting displacement of each point # at time ¢> 0. 60, A concentrated transverse foree F = Fy sinwt, t>0, is applied at the midpoint of a beam hinged at its ends «=0 and #=l, Show that the resulting transverse displacement is vey = Mesinet fh { ineVolb sinh " 4EI ® |eostVa/2Vb cosh WWVu/2V, 2bFol 3 sinny/2 BET ay na? — venta) if 0 <2 < 1/2, with a result obtained by symmetry for 2 0 is given by Fon = v{ue- ste = u(e-BE#) + w(e Be) — where U is Heaviside’s unit step function. Discuss this solution graphically. 242 APPLICATIONS TO BOUNDARY-VALUE PROBLEMS [cHaP.8 65, ‘Two semi-infinite conducting solids, x <0 and @>0 [see Fig. 8-16], have constant thermal conductivities and diffu- sivities given by Ky, ky atid Ko, ke respectively. The initial temperatures of these solids are constant and equal to U, and U; respectively. Show that the temperature at any point of the solid x >0 at any time ¢ is vey = Uy + where a = K,Vig/KoVki j au _ aU [rine The next conduction equations are SP = yz, 2<0 and De ZU e>0 and we must have lim Ul) = lim Ule,t) and Tim K,Uele,t) = lim, KUe(2, 6 66. Verify the result at the end of Problem 8, Page 228, 67. An infinite circular eyinder of unit radios has its initial temperature zero, A constant Sox A fs applied to the convex surface. Show that the temperature at points distant r from the axis at any time ¢ is given by 2A S prt Jar) _A rf Ue) = Ge ske— any + A BivO) where A, are the positive roots of Jo(A) 68. A cylinder of unit radius and height has its circular ends maintained at temperature zero while its convex surface is maintained at constant temperature Uy, Assuming that the cylinder has its axis coincident with the z axis, show that the steady-state temperature at any distance r from the axis and z from one end 400 & sin @n—1)wx Iol(@n— ter} 7 at Qn~1 Ty{@n— 1x} U2) = 69. (a) Solve the boundary-value problem YO.) =0, YUH =0, ¥l2,0) = 0, ¥,(,0) 0, Yee(ht)=0, EL¥,,(0,0) = Py sinut (®) Interpret the problem in (a) physically. bPo sin et (sinh (1-2) Vorb - save inh 1 Vu/b sin o/b jin mara/t sin brat /2 nat — Binte/E) An. (a) Y¥@t) = 70. A semi-infinite transmission line of negligible inductance has its initial voltage and current equal to zero. At t=0 a constant voltage Ep is applied at the sending end ©=0. Show that the voltage at any point z>0 at any time {> 0 is, Ee) = a8 ot (VE- #2) = eV ot (VE+ (2)} = By cosh 2VGR ‘What is the corresponding current? Appendix A TABLE OF GENERAL PROPERTIES OF LAPLACE TRANSFORMS fe) Fo 1 afy(s) + bfa(s) aFy(t) + bF a(t) 2 vo) aF(at) 3. He-0) re a mye) We-a) = iaie fee 5. #f(@) ~ FO) ro 6. #f(e) — #F(0) - FO) ry 7. | fe) — #1 FOO) — 2 PO) — + — FO) FO 8. re) —tF(t) 9. ro) aro 10. F() Carer "W. i * Poy du 12. fa Sf Ss Fw dus = Ss or Few) au 13. 10) 018) S Flu Gt) du 243, 244 TABLE OF GENERAL PROPERTIES OF LAPLACE TRANSFORMS [Appendix A (a) F(t) 4, * yeu) du Fo 4, S100 t 15. S e- Fu) du Fw = Fe+7) 16. 7 44am S oevid) Fw due a govt He) we SI u-n2 5, (2Vat) Pu) dee 19. S Tel2Vule =) ) Flu) du 0 20. Fie’) 21. fQns) sing Pay Plow) pox 22. aw Ted polynomial of degree less than n, Qe) = (e-anlo— ay) ++ (0-5) where ay,a2,...,, are all distinct. Appendix B TABLE OF SPECIAL LAPLACE TRANSFORMS fo) Fy ; : : s ; tS n= 12,3)... 4 4 n>o ce eo n= 1,2,3, ri 7. n>o 8 e + a * ete cos at 10. ' re o* conat 12. wo Fa cosh at “4 hat 245 246 TABLE OF SPECIAL LAPLACE TRANSFORMS [Appendix B He) Fw 15. | ot cosh at 16. weaesy 78 a ae be age 18. = cae sinat 5a cos at 19, ma bie 20. iz tm in at Seaton 2. cos at — Jat sin at 22. t cos at 23. at cosh at — sinhat ae 25. sinh at + at cosh at te 26. iz 2a coshat + Jat sinhat 27. teosh at 28. Ta Ga otP) ninst ~ tet const 29. wa tsin at =e cos at ator G4e0) sin at — at ent 31. ie Pon St sin at + at? cos at Appendix B] TABLE OF SPECIAL LAPLACE TRANSFORMS 247 Ais) Fit) st Cs (e+ a2) ms a we 3s*— a? f sinat 34. ear ote 2a 8 — 30% , 35. eae 48 cos at 26. sae tet pe conet oS — ate Bsinat Le (e+ att 24a 1 (3+) sinhat ~ Sat cosh at Lo (a) 8a! 5 at? cosh ot — ¢ sinhat " Ba! 40. 8t sinh at + at? coshat 41. a ee 42. (8+ .%#2) sinhat + Bat cosh at * 8a | yt (8+ 2) coshat + Tat sinh at 3 # sinh at | bs to 45. 4 cosh at 46. Af® cosh at @ sinh at az 24a 248 TABLE OF SPECIAL LAPLACE TRANSFORMS [Appendix B fa) Fi) 51. ee? | esate — Foret = cos 52. ena VBat _ 3a V3 2 at 53. a 3 («= 4 Ben aue con 5 *) 54. zm Galvin at cosh ot — cos at sinh at) at inh at 55. atid aa 2 ae wie gyloin at cosh at + cos at sinh at) 37. cos at cosh at ioe ; 58. gialsinh at — sin at) * 1 59. wea pipleosh at — cos at) Legs is gy sinh at + sin at) 61. woe ‘Heosh at + cos at) Oo, 1 ent e-at Vetat Verb 206 — a)Vet® 1 ert Vat 8 wate Va 64. —— et ert Vat Vea—a va be*terfe ova} Appendix B] TABLE OF SPECIAL LAPLACE TRANSFORMS 249 fo) | Fo 66. e Jo(at) : Vara ee 67. Iy(at) Wita-on |S ; Were a a> onda 6. Veta (at) Vst—a?)e 9. a>-t I, (at ‘ Vine arta(at) on FTE) —_ . “ JolaV He 28) 70. aca lav HEF 26) ) . oT we IloVF—B) e>d . V8? +a? oO t 6) Jn(2Vat) erf (a/2Vt) erfe (a/2Vt) NOU orfe (we + a) avi, 89. Fea Soe deg OV) du 90. oe a ne + a*)/a?} Ci (at) 2. tate + ae Ei(at) 93, -&thne Int y = Enler’s constant = .5772156... 4. 7 (S48) 2(con ot — con) oy + otme Intt 7 = Buler’s constant = .5772156... 96. Eo —(nt+y = Enler’s constant = 5772156. 97. aoe (int + 7)? — gat -y = Euler's constant = .6772156. Appendix B) TABLE OF SPECIAL LAPLACE TRANSFORMS 251 1 i) Fo —_|_—. an Pint) —eiat Dine gs 4 int ren sinat 99, tan“! (a/s) t 100, fanmt tale) Silat) 101. SS erte (Vale) ve 102. oF" /"0* erfe (s/2a) nos Verte ult erf(at) 104. es erfe Vas 105. e%* Ei (as) | 106. 3 [omer {5 - si cea} — sinae Ci coo] s . « 2 me 107. vines {p- si coo} + cons cian za 108. cosas { - Si cea} — sin as Ci (as) + _ tan—! (t/a) 109. sinas Si(ae)> + cosas Ci (as) | Ln (£42) e 2 a? no. [5 - sie Tt + Git(as) 1, (£32) i. 2 t a m1. ° xO 112. 1 a(t) 113. ee 8(t—a) 114, ‘u(t—a) See also entry 189, Page 254. 252 TABLE OF SPECIAL LAPLACE TRANSFORMS [Appendix B fi) Fw) 115. ame 8 cosh sa na. a cos A DEE cog (ee et - ns. ithes = i 120, aa 121. 122. wee, 1+%3 123, wee e+e) — 125, ae a 3 (a9 1 m= 1) eC He gog OHNE 126. me, 2S arte nena in n= ter 128, @ Mata? gin MEE 129. i+4 SE einen Qn 180. Bee a a ed i 131. sie germane = BES Og erst son 2" at Appendix B] TABLE OF SPECIAL LAPLACE TRANSFORMS 253 Ke) FO Joli HUA Jy (Ng 132, (ev) ) FI To(dyx/a) ‘sJoliavs) rE OR) where Ay, Ag. are the positive roots of Jo(A) = 0 JnlieVa) = eNU Tela) ae t— at) + t+ ge SO 133, PJo(iaV) a) 2.00) where AyAe «+. are the positive roots of Jo(X) = 0 ‘Triangular wave function Fw 134, A io ta 4a a ‘ ‘Square wave function Fey 135. y i — — sei ae te fae hia ae a t | oo Lu Rectified sine wave function Fe 136, atta (¥) 1 0 0 7 o @ 26 ‘Sa ‘ Half rectified sine wave function Fe 137. @2taGS ea 14 7 7 @ 2a ‘Me 4 ‘ Saw tooth wave function Fo 1 ad 1 : 1 H 1 : ' ' « 2a a ae ‘ 254 TABLE OF SPECIAL LAPLACE TRANSFORMS [Appendix B 7) FO 's unit fanetion u(t a) co 139. ial 1 also entry 114, Page 251. a Pulse function Fo emt — ea a — 140, segres) i H t ' bo ° « a ‘ Step fanetion Fo ‘ 2 — 141. 2 —_ if See also entry 78, Page 249. oe _——______ * te ae te Fan nSta naa za(l + 67) bie . ae? + oF 1 Na Appendix C TABLE OF SPECIAL FUNCTIONS 1. Gamma funetion . Tin) = f wrteudu, 0 >0 0 2. Beta function 1 Bm, n) Serta de = Fare, m, n>0 3. Bessel function - _ a ale) = aoe! Hansa * BaontDHOntH } 4. Modified Bessel function I(t) = "J qCiz) 5. Error function 6. Complementary error function erfe (t) 7. Exponential integral 8, Sine integral 9. Cosine integral 10. Fresnel sine integral 11, Fresnel cosine integral cy = Sf cosut du “ Sf 12, Laguerre polynomials ry = Sees, n= 01,2, 255 INDEX Abel's integral equation, 118, 117-120 Absolute convergence, 155, 156 definition of, 156 Absolute value, 136 Acceleration, 79 Aerodynamics, 149 Amplitude, 89 of a complex number, 187 Analytic function, 138 necessary and sufficient conditions for, 148 Analytic part, of a Laurent series, 142 Argund diagram, 187 Argument, 137 Battery, 79 Bains, applications to, 81, 93-96 cantilever, 94 deflection curve or elastic curve of, 81 on elastic foundation, 111 vibrations of, 219, 220, 226-228 Bending moment, 81 Bessel functions, 7, 8, 23, 24, 28, 282, 238, 255 generating function for, 7 integral representation of, 67, 68 Laplace transforms of, 9, 23, 24 modified, 8, 255, Bessel’s differential equation, 8 Beta function, 47, 62, 63, 255 relation of, to convolution theorem, 62 Bilinear transformation, 172 Boundary conditions, 81, 219 Boundary-value problems, 81, 219 ‘one dimensional, 219, 220 solved by Fourier transforms, 198-195, 221, 234-236 solved by Laplace transforms, 81, 96-98, 102, 221 two and three dimensional, 220, 221 Brachistochrone problem, 136 Branches, of a many-valued function, 138, 166 Branch line, 166 Branch points, 141, 166 complex inversion formula and, 202, 207, 208 Bromwich contour, 201, 210 modification of, 202, 227 Built-in end, beam with, 81 Cantilever beam, 94 Capacitance, 79 ‘of transmission line, 220 Capacitor, 79 Cauehy-Riemann equations, 189, 147-149 proof of, 148 Cauchy's inequality, 172 Cauchy’s integral formulas, 141, 151-155 proof of, 154 Cauchy's theorem, 140, 151-155 proof of, 152, 153 257 Change of scale property, 3, 18-16, 44, 48, 52 for inverse Laplace transforms, 44, 48, 52 for Laplace transforms, 44, 48, 52 Charge, 80, 221 Circuit, electric [see Electrical circuits clements, 79 primary, 111 secondary, 111 Clamped end, beam with, 81, 94, 95 Complementary error function, 8 Complex conjugate, 136 Complex inversion formula, 46, 201, 203-205 branch points and, 202, 207, 208 conditions for validity of, 202, 203-205, 212 for functions with infinitely many singularities, 202, 209-211, 212, 213 proof of, 203 residues and, 205-207 Complex numbers, 136, 144 amplitude of, 187 argument of, 137 axiomatic foundations for, 136 equality of, 136 imaginary part of, 136 polar form of, 187, 144, 145 real part of, 136 roots of, 187, 145 Complex number system, 186 Complex plane, 187 Complex variable, functions of a, 188 Concentrated load, 95 representation of by Dirac delta function, 95 Conductance, of transmission line, 220 Conduction of heat [see Heat conduction} Conductivity, thermal, 219, 221 Conjugate, complex, 136 Continuity, of fanctions of a complex variable, 138 sectional or piecewise continuity 2, 4, 28, 42, 178, 186, 187, 190 Contour, 148 Bromwich [eee Bromwich contour] Convergence, absolute, 155, 156 of Fourier series, 185-187 uniform [sce Uniform convergence] Convolutions, 45 [see also Convolution theorem] associative, commutative and distributive laws for, 4, 56 integral equations and, 112, 117 Convolution theorem, 45, 55-58 (see also Convolutions] beta function and, 62 for Fourier transforms, 177 proof of, 55, 56 Coordinates, cylindrical, 282 polar, 187 rectangular, 136 Cosine inteyral, 8, 24, 25, 256 Laplace transform of, 10, 25 258 INDEX Cosine seri Coulombs, 80 Critically damped motion, 90, 91 Cross-cut, 153 Current, 80 Cycles per second, 89 Cycloid, 113, 119, 120, 192 tautochrone problem and, 118, 117-120 Cylinder, heat conduction in, 220, 282, 288 Cylindrical coordinates, 232 [sce Half range Fourier series] Damped oscillatory motion, 90, 91 Damping constant, 79 Damping force, 79, 88-90 Definite integrals, evaluation of, 148, 161-165 Deflection curve, 81 Deflection of beams [eee Beams, applications to] Delta function (see Dirac delta function} De Moivre's theorem, 137 Density, 220, 221 Derivatives, inverse Laplace transform of, 44, 52, 53 Fourier transform of, 193 Laplace transform of, 4, 15, 16, 96 of functions of a complex variable, 188, 139, 147-149 . Difference equations, 118, 120-125, 127, 128 differential-, 118, 114, 120-125 Differentiable function, 138 Differential-difference equations, 118, 114, 120-125 Differential equations, applications to, 78-102, 219-236 for finding inverse Laplace transforms, 46, 65, 66 fi for finding Laplace transforms, 6, 28, 29 general solutions of, 88-85, 100, 101 integro-, 118, 120 ordinary (sce Ordinary differential equations] partial [see Partial differential equations} relation of, to integral equations, 114-116, 128, 129 solution of, by Fourier transforms, 193-195, 221, 284-236 solution of, by Laplace transforms, 78, 81-87, 96-98, 102 Differentiation, rules for, 139 with respect to a parameter, 6, 18, 46, 53, 65 Diffusivity, 98, 219 Dirac delta function, 8, 9, 26, 27, 45 Laplace transform of, 10, 27 use of, in applications to beams, 95 Dirichlet conditions, 173 Displacement, longitudinal, 219, 220, 226, 227: of a spring, 79 of a string, 199, 219, 220, 224, 225, 231, 232 transverse, 81, 220 Division, by t, 5, 18, 19 by powers of 6, 45, 53-55 Elastic constant, 111 Elastic curve, 81 Elastic foundation, beam on, 111 Electrical circuits, applications to, 79, 80, 91-93, 214, 215 complex, 80, 92, 93 7 simple, 79, 91, 92 Electric ‘potential, 221 Electromotive force, 79 Elementary functions, Laplace transforms of, 1, 10-12 of a complex variable, 138 emf, 79 Equilibrium position, 79, 219 Error function, 8, 26, 28, 208, 209, 255 complementary, 8, 208, 209, 255 Laplace transform of, 10, 26 Essential singularity, 142, 157 Euler's constant, 29, 250 Eoler’s formula, 137 Even extension, 183 Even functions, 178, 174, 182-184 Existence of Laplace transforms, sufficient ‘conditions for, 2 Expansion formula of Heaviside [eee Heaviside’s expansion formula| Exponential integral, 8, 24, 25, 255 Laplace transform of, 10, 25 Exponential order, functions of, 2, 4, 28, 42 External force, motion of a spring under, 79, 99, 100 Factorial function [tee Gamma function] Faltung [see Convolutions| Farads, 79 Fibonacci numbers, 133 Final-value theorem, 6, 20, 21 generalization of, 6 proof of, 20 Fixed end, beam with, 81 Flexural rigidity, 81 Flaid mechanics, 149 Force, damping, 79 electromotive, 79 external, 79, 99, 100 restoring, 79 Fourier integrals, 175, 176, 187-198 complex form of, 176 Parseval’s identity for, 177, 189 Fourier integral theorem, 175, 176, 189 [see also Fourier integrals] proofs of, 189-191 Fourier series, 173-176, 178-184, 185-187, 192 coefficients in, 178, 179, 180 complex form of, 174 convergence of, 185-187 Dirichlet conditions for, 173 half range, 174, 182, 188 Parseval’s identity for, 174, 183, 184 Fourier transforms, 176, 187-195 convolution theorem for, 177 cosine, 176, 177 finite, 175, 184, 185 inverse, 175-177 of derivatives, 193 INDEX 259 Fourier transforms (cont.) partial differential equations solved by, 193-195, 221, 284-236 relation of, to Laplace transforms, 177, 178, 203 sine, 176, 177 symmetric form of, 176 Fredholm’s integral equation, 112, 116, 129 differential equation expressed as, 128, 129 Free end, beam with, 81, 94, 226 Frequency, 89 natural, 90, 99 of damped oscillatory motion, 90 resonant, 99 Fresnel integrals, 228, 255 Functions, analytic [see Analytic function) many-valued, 138, 166 of a complex variable, 138 of exponential order, 2, 4, 28, 42 single-valued, 188 table of special, 9, 255 Gamma function, 7, 21-28, 255 Stirling’s formula for, 7 General solution, of @ differential equation, 83-85, 100, 101 Generating function for Bessel functions, 7 Generator, 79 Gravitational potential, 221 Greatest integer less than f, 121, 122 Green’s theorem in the plane, 140, 150, 151 proof of, 150, 151 Half range Fourier series, 174, 182, 183 Half wave rectified sine curve, 20, 218, 253 Harmonic functions, 139 Heat conduction, 98, 194, 220-224, 280, 282-286 general equation for, 221 in a cylinder, 220, 282, 283 in an insulated bar, 223, 224, 283, 284 in a semi-infinite plate, 234-296 in a semi-infinite solid, 221, 222 Heat flow problems, 98 [see also Heat conduction] involving radiation, 230 Heat source, 234 Heaviside’s expansion formula, 46, 47, 61, 62 extensions of, 78, 74 proof of, 61, 62 Heaviside’s unit function, 8, 26, 60, 254 Laplace transform of, 10, 26 Henrys, 79 Hinged end, beam with, 81, 98 Hooke’s law, 79 Hospital’s rule [see L’Hospital’s rule] Hypocycloid, 169 Image, 165 Imaginary part, 136 Imaginary unit, 136 Impulse functions, 8, 9, 26, 27, 95 [see alto Dirac delta function} Independence of the path, 140, 152, 158 Inductance, 79 mutual, 111 of transmission line, 220 Induetor, 79 Initial-value theorem, 5, 20, 21 generalization of, 6 proof of, 20 Insulated bar, heat conduction in, 223, 224, 233, 234 Integral equations, 112, 118, 114-120, 126 Abel, 113, 117-120 Fredholm, 112, 116, 129 kernels of, 112, 129 of convolution type, 112, 117 relation of, to differential equations, 114-116, 128, 129 solved by Fourier transforms, 193 Volterra, 112 Integral formulas, Cauchy, 141, 161-155 Integrals, evaluation of, 7, 27, 28, 47, 63, 64 Fourier [see Fourier integrals] Freel, 228 inverse Laplace transform of, 4, 16 Laplace transform of, 44, 52, 53 Tine, 189, 140, 150 of functions of a complex variable, 140, 151-155 Integro-differential difference equations, 114 Integro-differential equations, 113, 120 Inverse Fourier transforms, 175-177 Inverse Laplace transforms, 42-77 complex inversion formula for [see Complex inversion formula] definition of, 42 methods of finding, 46 of derivatives, 44, 52, 58 of functions with infinitely many singularities, 209-211, 212, 218 of integrals, 4, 16 operator, 42 properties of, 43-45 uniqueness of, 42 Inversion formula, complex [see Complex inversion formula] for Fourier transforms, 175-177 for Laplace transforms, 46, 178 [see also Complex inversion formula] Isolated singularity, 141 Iterated Laplace transformation, 221 Jacobian, 56, 172 Jump, at a discontinuity, 4 Kernel, of an integral equation, 112 symmetrie, 129 Key, in an electrical circuit, 79 Kirchhoff’s laws, 80, 91, 92 Laguerre polynomials, 39, 255 Laplace's equation, 139, 221 Laplace transform operator, 3 Laplace transforms, 1-41 behavior of, as 0, 5 260 INDEX Laplace transforms (cont,) definition of, 1 existenco of, 1, 28 inverse [sce Inverse Laplace transforms] iterated, 221 methods of finding, 6 notation for, 1 of derivatives, 4, 15, 16, 96 of elementary functions, 1, 10-12 of integrals, 44, 52, 58 of special functions, 9, 10 properties of, 3-6 elation of, to Fourier transforms, 177, 178, 203, solution of differential equations by, 78, 81-87, 96-08, 102 . Laurent’s series, 142, 158, 159, 172 classification of singularities by, 158, 159 Laurent’s theorem, 172 [see also Laurent’s series} Leibnite’s rule, 17 Lerch’s theorem, 42 L'Hospital’s role, 161, 162 Limits, of functions of a complex variable, 138 right and left hand, 2 Linearity property, 3, 12, 18, 43, 48, 49 for inverse Laplace transforms, 43, 48, 49 for Laplace transforms, 43, 48, 49 Linear operator, inverse Laplace transformation as, 43 Laplace transformation as, 3 Line integrals, 139, 140, 150 Longitudinal vibrations, of a beam, 219, 220, 226, 227 Many-valued functions, 138 Mapping, 165 ‘Mathematical induction, 15-17 Mechanics, applications to, 79, 88-91 Membrane, vibrations of, 220, 221 Modified Bessel functions, 8 Modulus of elasticity, 220 Moment, bending, 81 Motion, ‘non-oscillatory, 89 Multiple-valued functions, 138 Multiplication, by s*, 45, 53-55 by t, 5, 17, 18 ‘Mutual induction, 111 Natural frequency, 90, 99 Newton's law, 79, 88 Null functions, 9, 27, 42 Laplace transforms of, 10 relation of, to inverse Laplace transforms, 42 Odd extension, 182 Odd funetions, 173, 174, 182-184 Ohms, 79 Operator, inverse Laplace transform, 42 Laplace transform, 3 Ordered pair, 136, 187 Order of a pole, 141 Ordinary differential equations, applications to, 78-96, 99-102 general solution of, 83, 84 Ordinary differential equations (cont,) simultaneous, 78, 87, 88 solution of, using convolutions, 85 with constant coefficients, 78, 82-85 with variable coeficients, 78, 85-87 Orthogonal families, 148, 149 Oscillatory motion, 90, 91, 99 damped, 90, 91 Overdamped motion, 90, 91 Parallelogram law, 167 Parseval’s identity, for Fourier integrals, 177, 189 for Fourier series, 174, 183, 184 Partial derivatives, Fourier transform of, 198 Laplace transform of, 96 Partial differential equations, 81, 96-98, 219-236 important list of, 219-221 solved by Fourier transforms, 198-195, 221, 234-286 solved by Laplace transforms, 81, 96-98, 102, 221 Partial fractions, 46, 58-61 Heaviside’s method for [see Heaviside’s expansion formula] with distinct linear factors, 59 with non-repeated quadratic factors, 61 with repeated linear factors, 60 Period, 89 of damped oscillatory motion, 90 Periodic functions, Laplace transform of, 5, 19, 20 Piecewise continuity, 2, 4, 28, 42, 173, 186, 187, 190 Plates, heat conduction in, 284-236 Polar coordinates, 137 Polar form, of complex numbers, 137, 144, 145 operations in, 187 Poles, 141 of infinite order, 142 Potential drop, 80 Potential, electric or gravitational, 221 velocity, 149 Primary circuit, 111 Principal branch, 147 Principal part, of a Laurent series, 142 Principal value, 147 Pulse function, 254 Quadratic equation, 144 Radiation, 280 Ratio test, 155 Real part, 136 Receiving’ end, of transmission line, 220 Rectangular coordinates, 136 Rectified sine wave, 253 Recursion formula, 124 Removable singularity, 141, 156-158 Residues, 142, 159-161 and the complex inversion formula, 205-207 Residue theorem, 142, 143, 159-161 Proof of, 160, 161 INDEX 261 Residue theorem (cont,) use of, in finding inverse Laplace transforms, 201, 202, 205-207 Resistance, 79 of transmission line, 220 Resistor, 19 Resonance, 99 Resonant frequency, 99 Restoring force, 79 Rest position (see Equilibrium position] Riemann’s theorem, 174, 186, 190 Riemann zeta function, 41 Roots of complex numbers, 137, 145 geometric representation of, 145 Saw tooth wave function, 253 Secondary circuit, 111 Sectional continuity, 2, 4, 28, 42, 173, 186, 187, 190 Semi-infinite, beam, 227, 228 Plate, 234-236 string, 224, 225 transmission lines, 220, 228, 229 Sending end, of transmission line, 220 Series, convergence of, 155 Fourier [see Fourier series} Laurent’s [see Laurent’s series] of funetions of a complex variable, 155-159 Taylor's, 141, 157 Series electrical circuit, 79, 91, 92 Series expansions, 138 [see also Series] Series methods, for finding inverse Laplace transforms, 46, 65, 66 for finding Laplace transforms, 6, 28, 24, 29 Shear, vertical, 81 Simple closed curve, 139 Simple pole, 141 Simply-supported end, beam with, 81 Simultaneous differential equations, 78, 87, 88, 220, 298, 220 Sine integral, 8, 24, 25, 265 Laplace transform of, 10, 24, 25 Sine series (sce Half range Fourier series} Single-valued function, 138 Singularities, 155-159 [see also Singular points] and the complex inversion formula, 202, 205-218 essential, 142, 187 isolated, 141 Singular points, 141 [see also Singularities) Source of heat, 234 Specific heat, 219, 221 Spring constant, 79 Spring, vibrations of, 79 Square wave, 214, 253 Steady-state temperature, 221 Steady-state terms, 92 Stirling’s formula, 7 Strain, 220 Stream funetion, 149 Stress, 220 String, vibrations of, 199, 219, 220, 224, 225, 231, 232 Sufficient conditions for existence of Laplace transforms, 1 proof of, 28 Switch, in an electrical circuit, 79 Symmetric form of Fourier transforms, 176 Symmetric kernel, 129 Tables, of inverse Laplace transforms, 43, 245-254 of Laplace transforms, 1, 9, 10, 248-254 of special functions, 9, 265 Tautochrone problem, 118, 117-120 Taylor's series, 141 Taylor's theorem, 157 ‘Temperature, 98, 219 [see also Heat, conduction} steady-state, 221 Tension, in a string, 219 ‘Thermal conductivity, 219, 221 Transient terms, 92 ‘Transmission lines, 220, 228, 229 Transverse defiection of a beam, 81 ‘Tranfverse vibrations, of a beam, 220 of a string, 219, 224, 225 Triangular wave, 226, 227, 253 Uniform convergence, 156 Fourier series and, 179, 183 Weierstrass M test for, 156 Uniform load, 98 Uniqueness of inverse Laplace transforms, 42 Unit impulse function, 8, 9, 26, 27, 95 [see also Dirac delta function] Unit step function, 8 [see also Heaviside’s ‘unit function] Vectors, 167 Velocity potential, 149 ‘Vertical shear, 81 ‘Vibrations, of a beam, 219, 220, 226-228 of a membrane, 220, 221 of a spring, 79 of a string, 199, 219, 220, 224, 225, 281, 232 Voltage drop, 80 Volterra’s integral equation, 112 Volts, 79 ‘Wave equation, 219 ‘Weierstrass M test, 156 2 axis, 136 xy axis, 136 ‘Young's modulus of elasticity, 81, 220 Zeta function, Riemann, 41 Master Laplace transforms with, Schaum’s—the high-performance study guide. It will help you cut study time, hone problem-solving skills, and achieve your personal best on exams! Students love Schaum's Outlines because they produce results. Each year, hundreds of thousands of students improve their test scores and final grades with nese indispensable study guides. Got the edge on your classmates. Use Schaum's! Hf you don't have a fot of time but want to excel In class, this book helps you: Brush up before tests Find answers fast Study quickly and more effectively Get the big picture without spending hours poring over lengthy textbooks ‘Schaum's Outiines give you the information your teachers expect you to know in a handy and succinct format—without overwhelming you wath unnecessary details. You get a complete overview of the subject. Plus, you get plenty of practice exercises to test your skill. Compatible with any classroom text, Schaum's let you study at your own pace and remind you of all the important facts you need to remember—tast! And Schaum's are so complete, they're perfect for preparing for graduate or protessional exams. Inside, you wil find 450 problems. including step-by-step solutions. Hundreds of addtional praca problems, with answers supplied Clear explanations of applicatons of Laplace transforms Understandable coverage of Founer series and the complex inver- sion theory Hf you want top grades and a thorough understanding of Laplace trans- forms, this powerful study toot is the best tutor you can have! Chapters inciude: The Laplace Transform + The Inverse Laplace Transform + Applications to Differential Equations + Applications to Integral and Difterence Equations * Complex Variable Theory * Founer Series and Integrals + The Complex Inversion Formula + Applications to Boundary-Value Problems + Appendixes: A: Table of General Properties of Laplace Transforms B: Table of Special Laplace Transforms C: Table ‘of Special Functions McGraw-Hill kof The Me $15.95USA ISBN 0-07-Ob0232-x A I 928007 OVER 30 MILLION SOLD PSR) Bre Eee uy Mathematics & Statistics cs Cd Crete Cent ed eens