712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf laa aceite Equations . Orthogonal Functions and Fourier Series 13. 14. 15. 16. Boundary-Value Problems in Rectangular Coordinates Boundary-Value Problems in Other Coordinate Systems Integral Transform Method Numerical Solutions of Partial Differential Equations _chrome-extension:simhnlakainojrnhinhkckipnepbhabphipagesipdtiwablviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443.,. 90/104771224, 10:15AM 9781284105971.pdt - Denns-G.Zil-Advanced.Eng neering Mathomatles-2016-Jones-Barlott pf Vat ag Orthogonal Functions and Fourier Series ove cartn nds otros CHAPTER CONTENTS partial differential equations in faa cts Age . donot Sole any PDES in thts 12.1 Orthogonal Functions chapter, the material covered sets 12.2 Fourier Series esta ioe re pipes is 12.3 Fourier Cosine and Sine Series discussed later. 12.4 Complex Fourier Series In calculus you saw that a sufficiently differentiable See eee function fcould often be 12.6 Bessel and Legendre Series ‘expanded in a Taylor series, which T2164 Fourier Beseel Series essentially isan infinite series consisting of powers of x. The 12.6.2 Fourier-Legendre Series principal concept examined in Chapter 12 in Review this chapter also involves expanding a function in an infinite series. In the early 1800s, tthe French mathematician Joseph Fourier (1768-1830) advanced tthe idea of expanding a function f ina series of trigonometric functions. Tt turns out that Fourier series are just special cases of a more general type of series representation of a function using an infinite set of ‘orthogonal functions. The notion of orthogonal functions leads us back to eigenvalues and the corresponding set of eigenfunctions. Since eigenvalues ‘and eigenfunctions are the linchpins of the procedures in the two chapters that follow, you are ‘encouraged to review Example 2 in Section 3.9. _chrome-extensionjminlakailnojrnhinhkckpnepbhabphipagesipdtiwablviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443..692/1047712124, 1015 AM 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf El Orthogonal Functions INTRODUCTION In certain areas of advanced mathematics, a function is considered to be a ‘generalization of a vector. In this section we shall sce how the two vector concepls of inet, or dot, product and orthogonality of vectors can be extended to functions. The remainder of the chapter isa practical application of this discussion, I Inner Product Recall, if u = wt + wo) + usk and v = vib + va) + vk are two veotors in Ror 3-space, then the inner product or dot product of u and visa eal number, called asealar, defined as the sum of the products of their corresponding components: In Chapter 7. the nse product was denoted hy W- (ay) = aim + aa + ay = Dn ‘The inner product (u, v) possesses the following properties: @ (ay =.u) (a) (ku, v) = k(u, v), ka scalar Gi) (uu) =Oifu=0 and (u,u)>Oituso (iy) (+ v9) = (Uw) + (yw). We expect any generalization ofthe inner product to possess these same properties. ‘Suppose that f, and fe are piecewise-continuous functions defined on an interval [a,b].* Since a definite integral on the interval of the product /(x)/:(2) possesses properties ()—() of the inner proxluct of vectors, whenever the integral exists we are prompted to make the following definition [Ree es ‘The inner product of two functions f; and f, on an interval (a, 6 isthe number css [rene I Orthogonal Functions Motivated by the fact that two vectors u and v are orthogonal whenever their inner product is zero, we define orthogonal functions ina similar mannct. ee ee “Two functions; and fae suid 6 be orthogonal onan interval a, if inf = | Heap de=0. a EEX _ orthogonal Functions ‘The functions f(x) = x” and 4x) = «are orthogonal on the interval [—1, 1]-This fact follows from (1); Unlike vector analysis, where the word orthogonal is a synonym for perpendicular, inthis present context the term orthogonal and condition (1) have no geomettie significance. I Orthogonal Sets We are primarily interested in infinite sets oF onhogonal functions. “The interval could also be (~2e, 22), (0, ), and s0 on. 672 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? =htpsieSAG2FOL2Felasticoeanstals-east-2-3443,.. 69311047712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf ee A set of real-valued functions (0X) ¢ (2), (0)... issaid to be orthogonal onan interval la blif ba(0)8,(0) de = 0, mr by) = men. @ Bj Orthonormal Sets The norm, or length lu), of a vector u can be expressed in terms of the inner product. The expression (u, u) = [1 is called the square norm, and so the norm is ul = Vo, a). Simitary, the square norm ofa function i | 200)? = (by dq) and so the norm, orits generalized length, is | J4(0)| = V(Gq.d,)- In other words, the square norm and norm ofa function din an orthogonal set {J,(x)} afc, respectively, lesen = [sioner and co IF {2,(0) isan orthogonal set of functions on the interval (a, b] with the property that | (2)|| = 1 form=0, 1, 2,... then {¢,(2)) is suid (0 be an orthonormal set on the interval Orthogonal Sot of Functions 08, 608 2k.) is enthogonal on the interval [— ‘Show that the set (1 SOLUTION Ir we make the identification (x) = 1 and d,(x) = cos nx, we must then show that JF_ by(a)e,(0) de = 0, n= 0, and JF, dby(2)6,(8) de = 0, m* n. We have, in the first cease, forn #0, cos mx cos nx de | Leos(m + mx + cos(m — nx] dx < trigonometric - ident uw a [ane + aye | sinim ay 9 2l mtn mon |, BEC Norms Find the norms of each funetion in the orthogonal set given in Example 2. SOLUTION. For g(x) = 1 we have from (3) eae dx=20 so that | 400) = V2ir. For J,(4) = cos nx, n> 0, it follows thet [2.09 cos nx de al “U1 + cos 2a dr Thus form > 0, |b,(2)) = Vr. 121 Orthogonal Functions | 673 _chrome-extension:siminlakainojrnhinhkckipnepbhabphipagesipdtiwebviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443.,. 94/1047712124, 1015 AM 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf Amortogonlsetca ve pp An orthogonal set of nonizeso functions (4, (1)}.m = 0,1, 2, .. eam be normalized thats ‘naieinioun oahonomal set, P made into.an orthonormal set—by dividing each Function by its norm. It follows from Examples 2 ‘and 3 thatthe set { 1 cosx cos 2x } Van Vn Vn is orthonormal on the interval [7 I Vector Analogy _Weshail make one more analogy between vectors and functions. Suppose ¥j.¥5 and V, are three mutually orthogonal nonzero Vectors in 3-space. Such an orthogonal set can be used asa basis fr 3space; that is, any tree-dimensional vector can be writen asa linear combination Ua Cy + eas FEW (a) Where the cj, i= 1, 2,3, are scalars called the components of the vector. Each component cj can be expressed in terms of u and the corresponding vector ¥,, To see this we take the inner product of (4) with vy (u, v1) = (0%, 91) + Ca¥. HY) + Cs(¥s. FH) = el Vil? + 2-0 + 65-0. (uy) Hence a Iv In like manner we find that the components ¢, and c, ae given by ayy vl? (u,v) (U. ¥n) Gy, = DON y (5) I Orthogonal Series Expansion Suppose (¢,(x)] is an infinite onhoponal set of func- ons on an interval [a,b], We ask: I" y = f(x) isa function defined on the interval [a,b], is it possible to determine a st of coefficients ¢,. = 0,1, 2 for which fo GabihX) + e4by(0) + + gba) +7 (6) As in the foregoing discussion on finding components of a vector, we can find the coefficients ¢, by utilizing the inner product. Multiplying (6) by d(x) and integrating over the interval [a,b] gives [fon Go| 60k) det er[ dea) de + +e] alt) (2) de + kb, Bn) + (61. bg) + °° + Gus Gr) + By orthogonality, each term on the right-hand side ofthe last equation is zero except when m In this case we have Slab) dt = cy It follows that the required coefficients c, are given by — Eefiedb,tx) de Se2@) dx” 674 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? https eSANG2FOL2Felasticoeanstalus-east-2-3443... 695/047712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf In other words, fa) = Deb, a S276, de where == BF ) With inner produet notation, (7) becomes ‘91 Thus (9) is seen to be the fu En A set of real-valued functions {¢(0), (1), 0300), ..] is said tobe orthogonal with respect ‘toa weight function w(x) on an interval [a,b] if ction analogue of the vector result given in (5). [wos.t96.09de=0, man ‘The usual assumption is that w(2) {1, cos, cos 2x, ...} in Example 2 is orthogonal with respect to the weight function w(x) con the interval [~7. 7] If [6,()) is orthogonal with respect to a weight function w(x) on the interval (a, 6], then ‘multiplying (6) by w(x)é,(x) and integrating yields 0 on the interval of orthogonality [a, b]. The set 1 fayw(yd,(0) de =e 10 outa) |? ie w(x dz(n) dr. (1) where Ile, lF The series (7) with coefficients c, given by either (8) or (10) is said to be an orthogonal series expansion of for a generalized Fourier series Ii Complete Sets The procedure outlined for determining the coefficients c, wis forma that is hase questions on whether an orthogonal series expansion sich a (7) satay posible ere ignored, Also, o expand in series of orthogonal Tunetions, itis certainly necessary that {Foot be orthogonal to each &, ofthe orthogonal set (2,(1)]. (If were orthogonal to every 4, then =O, = 0, 1,2...) To avoid the later problem we shall assume, forthe remainder oF the discussion, that an orthogonal set is complete. This means thatthe ony continuous function orthogonal to each member ofthe sti the 720 function. REMARKS Suppose that (Jo. fi)... ivan infinite set of real-valued functions that are continuous, ‘onan interval {a b. If this seis fineariy independent on {al (see page 357 forthe definition ‘of an infinite linearly independent set, then it can always be made into an orthogonal set ad, as described earlier in this seetion, can be made into an orthonormal set. Sce Problem 27 in Exercises 12.1 121 Orthogonal Functions | 675 _chrome-extensionsiminlakainojrnhinhkckpnepbhabphipagesipdtiwabviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443.,.696/1047712124, 1015 AM 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf BEG BSI tswers to selected odd-numbered problems begin on page ANS-29. In Problems 1-6, show that the given functions are orthogonal ‘on the indicated interval 1. fi) =x, Aly = 2 2HWM=" A= +1; [-1,1] 3 fi) =e flay = xe" — 5 (0,2) 4 fi) = cos, fi2) oz] B. fil) = fa) = cos 2x; [—7/2, 72) 6 fiw ef) = sin: In Problems 7-12, show that the given set of functions is orthog- ‘onal on the indicated interval. Find the norm of each function in the set 7. (sing, sin 3x, sinSr,...}; (02/2) B {cos x, 608 3x, €08 Sx, ..-} 8 (sinnx}, wo. | sin” i,2)3; 505, Oy; { P 10. pl] Mf 1 cosa}, 2 =1,2.3,..5 10, { Pp } [0.pl we facta ata) na 2. r P m= 123.05 (Ppl In Problems 13 and 14, verify by direct integration thatthe functions are orthogonal with respect to the indicated weight function on the given interval 18. Hx) = 1, Hy(@) = 2x, Hy) = 4x? — 2; wn) (20,00) Lg) = 1,0) = 4+ 1, LG) 10,00) 45. Let {d4(4)} be an orthogonal set of functions on fa, 01 such that g(x) = 1. Show that {2 4,(x) dr = 0 for n = 12, 16. Let (4,(0)} be an orthogonal set of functions on [aD] such Show that (ax-+ B)d,(1) de=0 and & 00) and any constants a and B. 7. Let {6,(2)} be an orthogonal set of functions on [a,b]. Show that b(t) + d,(0) [nC]? + 6,00)?, meen. 18. From Problem 1 we know that f(x) = x and js(x) = x? are 2, 2]. Find constants ¢, and ¢; such that = x+ ex? + er" is orthogonal to both f, and fz on the orthogonal on So ‘sae interval 19. The set of functions {sin nx}, 1 fon the interval [—m, 7]. Show t te set is not complete. 20. Suppose ff and f, are functions continuous on the interval [a] Show that (f+ fy fi) = fini + Sis 676 | CHAPTER 12 Orthogonal Functions and Fourier Series Qt; way=es, 2, 3, 118 orthogonal {A real-valued function is sad tobe periodte with period T= Oitf(x + 1) = fx) forall vin the domain of Tis the smallest positive value for which f(x + 7) = fx) holds then Ts called the fundamental period of f. In Problems 21-26, determine the fundamental period T ofthe given function. a1. fia) = cos sm pmabntoo B. fix) = sinx + sin2x 2. fix) = sin 2x + cos 4x 2. fx) = Sin3v-+ cos2x 28. fla) = si Discussion Problems 21. The Gram-Sehmlat process for constructing an orthogonal sel that was discussed in Seetion 7.7 carries over to a linearly independent set {2 fi(¥)-/00. .-} of real-valued functions continuous on an interval [a, b1. With the inner product (2406.00) dx, define the functions in the set (Gof, B4(X), 8360), -.} tobe 0) = fe) 4@) =f — data) = fay - ESP, and soon, (a) Write out d(x) inthe set. (b) By construction, the set B = {6}, dy(¥).6200), ..} is orthogonal on (a, b]. Demonstrate that dof), (0, and 64(4) are mutually orthogonal. Consider the set of functions {1.72 ...} defined on the interval [~1, 1]. Apply the Gram-Sehmidt process ‘ven in Problem 27 to this set and find d(x), d(2)-8300), and 6,() ofthe orthogonal set B" (b) Discuss: Do you recognize the orthogonal set? 2. Verify that the inner product (fi, f=) in Definition 12.1.1 satisfies properties () (iv) given on page 672. 30. In’, give an example of a set of orthogonal vectors that is not complete. Give a set of orthogonal vectors that is complete 31. The function 2. (a) 19 Avs (aces sit). ‘where the coefficients A, and B, depend only on 1, is period. Find the period Toff. _chrome-extensionsimhnlakainojrnhinhkckpnepbhabphipagesipdtiwab/viewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443... 97/1047712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf EE Fourier Series INTRODUCTION We have just soon in the preceding section that if (Jy). (8). 68) {ssc of real-valued functions tha is orthogonal onan interval a,b] and if fis. function defined con the sume interval, then we ean formally expand fin an orthogonal sris éUe(2) + e\i(8) + é:64(2) + +. In this scetion we shal expand functions in terms ofa special orthogonal st of trigonometric functions Trigonometric Series In Problem 12 in Exercises 12.1, you were asked to show that ig x } ; (om isorthogonal on the interval [=p, p]- This set will be of special importance later on in the solution Of certain kinds of boundary-value problems involving linear partial differential equations. ta ‘those applications we will need to expand a function f defined on [=p p] in an orthogonal series consisting of the trigonometric functions in (1); that is, wy = 24 3 (ajcos x + bysin = 109 = 84 3 (acs + hans) " The coefficients dy, day on Dy By «ns €an be determined in exactly the same manner as in the general discussion of orthogonal series expansions on pages 674 and 675. Before proceeding, note that we have chosen to write the coefficient of | in the set (1) a8 $y rather than dps this i for ‘Tis iawty Lay isunsd > convenience only because the formula ofa, will then reduce to. for instead ‘Now integrating both sides of (2) from ~p to p ives Joy de Since cos(vmip) and sin(nmx/p), n = 1, are orthogonal {0 1 on the interval, the right side of 3) reduces to a single term: fay de 1 a wis z Solving for dy yields il “flay. « [Now we multiply (2) by cos(m7./p) and integrate: By orthogonality we have cos de m>0, P 122 Fourier Series | 677 _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? hips eSAGG2FO;2Felasticoeanstals-east-2-3443,.. 698/047712124, 1015 AM FIGURE 1221. Function fin Example 1 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf = a sent sen {% *A8 coop a= fh ee ‘Thus (5) reduces to f(x) cos xd = Gy [ forenipeee= oo ata 8 i Fecal {r men vena 1 aa m Ply P The trigonometric series (2) with coefficients dd, and b, defined by (4. (6) and (7) respectively is sid 0 be the Fourler series ofthe function f The coefficients obtained from (4. (6), and (7) are refered to 2s Foutrler coefficlents of In finding the coefficients ay, a, and b,, we assumed that f was integrable on the interval and that (2), as well asthe series obtained by multiplying (2) by cos(am-/p) converged in such manner a8 to permit term-by-lerm integration. Until (2) is shown to be convergent for a given function f, the equality sign isnot to be taken in a strict or literal sense. Some texts use the symbol ~ in place of =. In view of the fact that most functions in applications are ofa type that guarantees convergence ofthe series, we shall use the equality symbol. We summarize the esuls Definition 1221 Fourier Ser Gris eran See oe Ras Yorn oat mes eee fe) B43 (sem tr tains), @) where w= +f" pra ® “ Bf fo coe xa (0) $f 109 sn a) HEEXXIEESEIMA Expansion in a Fourier Series Expand Ja) = {° . (2) in a Fourier series. SOLUTION. The graph of fis given in FIGURE 122.1. With p = = we have from (9) and (10) that “sonar 2 [fous [ve ~ oa] -2 [or 678 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extensionsminnlakailnojrnhinhkkipnepbhabphipagesipdtiwab/viewer himIPfle=hiips%3A%2F%:2Felasticbeanstalk-us-east-2-3443, 90/1047712124, 1015 AM Note that iene te > _chrome-extensionsminnlakailnojrnhinhkkipnepbhabphipagesipdtiwab/viewer himIPfle=hiips%3A%2F%:2Felasticbeanstalk-us-east-2-3443, 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf a2) poroomcar=2[f oar [1 — necenr sinnx woof «1 soma Leos. a] cosine +1 —cosme = (-1 ‘Therefore fix) = (13) = Note that 4, defined by (10) reduces to da given by (9) when we set m = 0. But as Example 1 shows, this may not be the ease after the integral fora, is evaluated, Hi Convergence of a Fourier Series ‘The following theorem gives sufficient conditions for convergence of a Fourier series at « point. ce ae Cc ‘onvergence Let fand ’ be piecewise continuous on the interval [~p, pI; thats, let fand ’ be continuous. ‘except ata finite number of points inthe interval and have only finite discontinuities at these points. Then forallrin the interval (—p, p) the Fourier series of fconverges to f(x) ata point ‘of continuity. At a point of discontinuity, the Fourier series converges tothe average fist) + fia-) ~¥ where /(x-+) and f(x) denote the limit offat.x from the right and from the left, respectively. For a proof of this theorem you are referred to the classic text by Churchill and Brown." SESE Convergence of a Point of Discontinui ‘The function (12) in Example 1 satisfies the conditions of Theorem 12.2.1. Thus for every x in the interval (—7, =), except at x = 0, the series (13) will converge to f(x). Atx = O the ‘function is discontinuous, and so the series (13) will converge to fO+) + fO-) 2 am “In other words, for point in the interval and > 0, 0°) = fy For W.-Y = fim fos "Ruel V. Churchill and James Ward Brown, Fourier Series and Boundary Value Problems (New York ‘McGraw-Hill, 2000), 122 Fourier Series | 679 7011047712124, 1015 AM We may assume tat the given funtion is pened 680 | _chrome-extensionsminnlakailnojrnhinhkkipnepbhabphipagesipdtiwab/viewer himIPfle=hiips%3A%2F%:2Felasticbeanstalk-us-east-2-3443, > 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf Il Periodic Extension Observe that each ofthe functions in the basie st (1) has a different fundamental period,* namely, 2p/n, a = 1, but since a positive integer multiple of a period is also 2 period we see that all of the functions have in common the period 2p (verify). Hence the right- hand side of (2) is 2p-periodic; indeed, 2p is the fundamental period ofthe sum. We conclude that «Fourier series not only represents the function on the interval (—p,p) but also gives the periodic extension of f outside this interval. We can now apply Theorem 12.2.1 to the periodic extension of f,or we may assume from the outset that the given function is periodic with period 7 = 2p: that is f(x + T) = fla). When fis piecewise continuous and the right- and left-hand derivatives ex alx= —pand.x= p, respectively, then the series (8) converges tothe average [/(p—) + (—p+)V2 at these endpoints and to this value extended periodically to +3p, +Sp, *7p, and so on. The Fourier series in (13) converges to the periodic extension of (12) on the entire x-axis. AtO, =2n, “and at 7, 3, Sz, .. the series converges tothe values te fO+) + f0-) 2 y+ fer) = at 2 respectively. The solid dots in FIGURE 1222 represent the value 7/2. “in oe lr pa oF FIGURE 1222 Peviodic extension of the function shown in Figure 12.2.1 I Sequence of Partial Sums is intresting tosce how the sequence of partial sums {Sy(2)) of a Fourier serie approximates a function, For example, the fist Urce pasta sums of (13) are coos tnx 849'= 2+ Bows snr + Line 4 In FIGURE 1223 we have used a CAS to graph the partial sums Ss(x), $,(2), and S\s(x) of (13) on the interval (—z, 7). Figure 12.2.3(d) shows the periodic extension using S,s(x) on (—477, 47), 4 4 z 4 | | be a ie a Tas a4 apy (@) Sys) om -m, 2) (0) 5) on Cm.) 0S (4) 5s(0) on (48,4) =f roy (6) Sige) on Cx, a) FIGURE 1223 Pactial sums of « Fourier series * Soe Problems 21-26 in Exercises 12. CHAPTER 12 Orthogonal Functions and Fourier Series 7011047712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf FEE BETIS answors'o sctected odd-numbered problems begin on page ANS-29. In Problems 1-16, find the Fourier series of the function fon the given interval, Give the number to which the Fourier series converges ata point of discontinuity off. In Problems 17 and 18, sketch the per indicated function extension of the —rex<0 1 10-{8 -2p. p). the graph of an even function possesses symmetry with respect othe y-axis, whereas the graph of an odd function possesses symmetry with respect to the origin, Il Even and Odd Functions tris tikely he origin of the words even and odd derives from the fat hat the praphs of polynomial functions that consist of al even powers of x are symmetsic 123 Fourier Cosine and Sine Series | 681 7211047712124, 1015 AM 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf With respect tothe y-axis, whereas graphs of polynomials that consist of all odd powers of rare ‘symmetric with respect to the origin. For example, oenineee fla) = x? iseven since f(—2) odd integer integer JO) = is odd singe f(—) = (—x)° “so. ‘See FIGURES 12:31 and 1232. The trigonometric cosine and sine functions are even and odd func tions, respectively, since cos(—x) = cos x and sin(—x) = —sin x. The exponential functions ‘fes) = e* and f(x) = e* are neither even nor od. I Properties ‘The following theorem lists some propertics of even and odd functions. [eke Proper Cor (a) The product of two even functions is even. () The product of two odd functions is even. (©) The product of an even function and an odd funetion is odd. (@) The sum (ference) of two even functions is even. (@) The sum (ference) of two odd functions is odd. (©) Ifiseven, then J", f9 de = 2/5 fa) de. (@) Iyis odd, then J", f(x) dx = 0. FIGURE 1232 Oud function PROOF OF (b): Let us suppose tha fand gare odd functions. Then we have (x) = ~FC0) and g(—) = AU, If we define the product offand g as Fx) ~ (x)g() then Fo=x) = fl) g(—) = (— f(a) = ORO) = FE). ‘This shows that the product Fof two odd functions is an even function. The proofs ofthe remai ing properties are left as exercises. See Problem 56 in Exercises 12.3 I Cosine and Sine Series _1/is an cven function on the interval (—>p.p). then in view of the foregoing properties, the coefficients (9), (10), and (11) of Section 12.2 become 2 fam poe Fe0 0s Tx dr = nin [109 sn %@- ax = 0 ue oda ‘Similarly, when fis odd on the interval (~p. p), n=0,1,2, ‘We summarize the results in the following definition, 682 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extensionsiminnlakainojnhinhkkjpnepbhabphipagesipdtiwebviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443... 703/1047712124, 1015 AM FIGURE 1233 Odd function Fin Example 1 FIGURE 1235 Oxi function fin Example2 _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf Cie (The Fourier series of an even function on the interval (—p, p) isthe cosine series “, 3 = f= F + Secon, a) ar where 2f pay as a) 4% Al 2) ai fo) on ds, @) i) The Fourier series ofan odd funetion on the interval (=p, p) isthe sine series so) = So, sin ae 4) af, a where 2f pep sin x 1 he pL fersinge xa 6) CUES Expansion ina Sine Series Expand /(x) = x, 2 0, note that if We extend the graph of the function in 22 perio mannet to the negative taxis, we obtain an odd function. In practical terms this means that we need only find the half ange sine expansion of f() = ml O <1 <1. With = 1itfollows from (5) and integration by pats that = a[ mesin mea ae FGURE 2312 Pesovietacag sion OM) the diferent equation of metion seen tobe ‘Fin Example Lae 4 Sap! wetted sin we (13) To find a particular solution x,(f) of (13), we substitute the series (12) into the differential equation and equate coefficients of sin rt. This yields (-teert +4) 686 | CHAPTER 12 Orthogonal Functions and Fourier Series 3-1"! _chrome-extensionsminlakainojnhinhkckpnepbhabphipagesipdtiablviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443... 707/047712124, 1015 AM Thus 0-5 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf @ 30-pe! west, 4) 5 sin et. (14) Fi not — Observe inthe solution (14) that there is noinleger n= 1 for which the denominator 64 — nx? of B, is zero. In general, if tere luc of, say.N, for which Nip = w, where o = VAJm. then the system described by (11) is ina state of pure resonance. In other words, we have pure resonance if the Fourier series expansion of the driving force f() contains a term sin(Wz/L). (or e0s(N7/L)) that has the same frequeney as the free vibrations. Of course, ifthe 2p-periadie extension of the driving force fonto the negative -axis yields an even funetion, then we expand fin a cosine series. BEER BESS tswers to selected odd-numbered problems begin on page ANS-29. In Problems 1-10, determine whether the given function is even, ‘oxi, oF neither. 4. fo = sin 3x 2 fix) = xeosx 3 fix) +x A f= — 4x 5 fay=e" 6 figre—e rae {en ES 3 ee -2 ce", a where Lf poemran, (8) 124 Complex Fourier Series | 689 hips eSAN2FOL2Felasticoeanstalus-east-2-3443,.. 71011047712124, 1015 AM 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf If fsatisfies the hypotheses of Theorem 12.2.1, a complex Fourier series converges to f(a) at 4 point of contisuity and to the average fot) +f 2 aa point of discontinuity. EQUETEE Complex Fourier Series Expand (a) 7 0 = ¢, cos ax +e sin ax y= e,coshax + (Cauchy-Euler equation General solutions, x > 0 Py tay’ —e'y=0, @20 2 ee ter, etd y=etelnx, a=0 Parametric Bessel equation (» = 0) eral solution, x > 0 a ty +e =0 Y= eyholax) + oY deer) Legendre’s equation Particular solutions are polynomials = dy! + n(n + by =0 Regarding the two forms of the general solution of y" — ay = 0, we will, in the future, employ the following informal rule: Use the exponential form y = cye"®* + exe when the domain of x is an infinite or semi-infinite interval; use the hyperbolic formy = ¢,coshax + ¢,siah axwhen the domain of x isa finite interval. I Eigenvalues and Eigenfunctions Orthogonal functions arise in the solution of ei ferential equations. More to the point, an orthogonal se of functions canbe generated by solving 2 two-point boundary-value problem involving a linear second-order differential equation con- taining a parameter A. In Example 2of Section 3.9 we saw thatthe boundary-value problem y'Fay=0, O=0, yL)=0, i) possessed nontrivial solutions only when the parameter A took on the values A, = 122/12, |. 2, 3... called elgenvalues. The corresponding nontrivial solutions y = ¢, sin(nsra/L) or n(nzrx/L) are called the elgenfunetions ofthe problem. For example, for (I) we have not an eigen : BYP. y'+5y 0) Solution is trivial: y = 0. is an eigenvalue (n = 2) ae ey yoo. 9 Solution is nonarivial: y= sin(2naiL) 692 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? hips eSAtG2FI;2Felasticoeanstalus-east-2-3443,.. 71311047712124, 1015 AM _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf For our purposes in this chapter it is important to recognize the set of functions generated by this 1,2,3,..., isthe orthogonal set of functions on the interval 0, L] BYP: thatis, (sin(nzxiL)) used as the basis For the Fourier sine series. jenvalues and Eigenfun« ns Itis left as an exercise to show, by considering the three possible cases for the parameter A ie; that is, A = 0,4 = a? <0, e¢ > O,and A = a? > 0, «> 0), that (zero, negative, orp the eigenvalues and eigenfunctions for the boundary-value problem ytay=0, y@=0 yu=0 are, respectively 0, to(1), Ap = Oisan eigenvalue for this BVP and latter comes from solving y" = 0 subject to the same boundary conditions y’(0) = 0.y"(L) Note also thaty = 1 can be incorporated into the family y = cos(nr-x/L) by permitting n The set (c0s( in Exercises 12.5. andy Il Regular Sturm-Liouville Problem The problems in (1) and (2) are special case can ‘important general two-point Boundary-value problem, Let p,q, r, and be real-valued functions continuous on an interval [a,b], and let r(x) > O and p(x) > 0 for every xin the interval. Then sone: Liaay-+09-+ apy 0 Subject to: A,y(a) + Byy'(a) = 0 Asx) + Bay'(D) is said to be a regular Sturm-Liouville problem. The coefficients in the boundary conditions (4) and (5) are assumed to be real and independent of A. In addition, A, and B, are nt both zero, and A> and Bare not both zero. The boundary-value problems in (1) and (2) are regular Sturm-Liouville in the differential equation (3) inboundary condition 4) we identify a= 0,4, = 1,8, = 0, and in (5), b= L.A; = 1, By = 0. From problems. From (1) we can identity r(x) = 1, g(x) =O, and pox) (2) the identifications would be a ‘The differential equation (3) is ©. 0,8, = Lin @), and b= L.A, each condition involves only a single boundary mixed ifeach condition involves both boundary points, boundary conditions y(a) = y(b),y'(a) = y'(b) are mixed boundary conditions. Because a regular Sturm-Liouville problem isahomogeneous BVP. always possesses the trivial 0. However, thissolution sof no inerest tous. Asin Example I, nsolving such problem solution ‘we seek numbers A (eigenvalues) and nontrivial solutions y that depend on A (eigenfunctions) Hl Properties. ‘Theorem 12.5.1 isalistofsome ofthe more important of the many properties of the regular Sturm-Liouvitle problem. We shall prove only the last property. Theorem 125.1 Properties of the Regular Sturm-Liouville Problem (a) There exist an infinite number of real eigenvalues that can be urranged in increasing order A, < Ay < Ay <~""< Ay =" such that Ay —> 00 a8 1+ 6. (b) For each eigenvalue there is only one eigenfunction (except for nonzero constant multiples) (©) Eigenfunctions corresponding to different eigenvalues are linearly independent. (d) The set of eigenfunctions corresponding to the set of eigenvalues is respect 0 the weight function p(x) on the interval [a, D]. 125 Sturm-Liowille Problem | 693 cos (renal) €, # 0. Incontrast 1s the corresponding eigenfunction. The /L)},n = 0, 1, 2,3, ..,{8 orthogonal on the interval [0, L]. See Problem 3 linear and homogeneous. The boundary conditions in (4) and (5), both a linear combination of y and y' equal to zero al « point, atc also called homogeneous. A boundary condition such as A,y(b) + Bsy'(b) = Cs, where C; is a nonzero constant, is non homogeneous. Naturally, a boundary-value problem that consists of a homogencous linear dif ‘erential equation and homogeneous boundary conditions is said to be homogeneous; otherwise itis nonhomogeneous. The boundary conditions (4) and (5) are said to be separated because Boundary conditions are referred to as a and .x = b, For example, the periodic ‘orthogonal with hips eSAN2FIL2Felasticoeanstalus-east-2-3443,.. 71411047712124, 1015 AM FIGURE 1251 Positive oots of tan.x = —in Example 2 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf PROOF OF (d): tt y, and y, be eigenfunctions corresponding to eigenvalues Ay and Ay, respectively. Then Eire +00) + Aare = 0 6) ” 4 sca ge VOOM + GCI + APODY, Multiplying (6) by y, and (7) by y, and subtracting the two equations gives On ~ AdPCYad A tres) - w A breve Integrating this last result by parts from.x = tox = b then yields ¥5 (0) — YD) (OD Aq —Ad| PO nn Ae = POM (D) ~ —ralys@ysta) — ys@yala)l. (8) No pani ihe eigenfunctions y,, and y, must both satisfy the boundary conditions (4) and (5). In lr, from (4) we have Aya(@) + Biyg(a) = 0 Apy(a) + Byys(a) = 0. For this system to be s ust be zero: fied by A, and B,, not both zero, the determinant of the coefficients Yal@YG(@) ~ yela)u(a) = 0. ‘A similar argument applied to (5) also gives Yalb)ye(b) ~ yx(b)y5 (0) = 0. Using these last two resus in (8) shows that both members ofthe right-hand side are zero. Hence ‘we have established the orthogonality relation {praia de =O, Aye 9) It can also be proved that the orthogonal set of eigenfunctions {y(X), Ys(X) Ys), «--) of ‘regular Sturm-Liouville problem is complete on [a, 6]. See page 675, EDEN 4 Regular Sturm-tiouville Problem Solve the boundary-value problem 0 (i) <0, where « > 0, the BYP tay SOLUTION You should verify {in (10) possesses only the trivial solution y = 0. For A = a> 0, « > 0, the general solution of the differential equation y” + ay = 0 is y = cye0s ax + c.sin ax. Now the condition 340) = 0 implies ¢; = 0 in this solution and so we are left with y = cy sin ax. The second boundary condition (1) + y"(1) = Os satisfied if x+y eysin a + exacosa = Choosing ¢# 0, we see that the last equation is equivalent 10 a. a) If we let x = a in (11), then FIGURE 125.1 shows the plausibility that there exists an infinite ‘number of roots of the equation tan x = —x, namely, the 1-coordinates of the points where 694 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extensionsminnlakailnojrnhinhkkipnepbhabphipagesipdtiwab/viewer himIPfle=hiips%3A%2F%:2Felasticbeanstalk-us-east-2-3443, 71511047712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf the graph of y = —x intersects the branches of the graph of y = tan x. The eigenvalues of problem (10) are then A, = a2, where a, 1 = 1, 2,3, .... are the consecutive positive roots 4, a, 5, ... Of (11). With the aid of a CAS it is easily shown that, to four rounded decimal places, «7, = 2.0288, a = 4.9132, a; = 7.9787, and a, = 11.0855, and the corresponding solutions are y, = sin 2.0288x, y> = sin 4.9132x, y, = sin 7.97874, and y, = sin 11.0855x In general, the eigenfunctions of the problem are |sin ,.}.1 = 1, 2,3, ‘With identifications r(x) = 1, g(x) = 0, p(a) = 1,4, = 1,8, = 0, Ay = 1, and By = 1 we see that (10) isa regular Sturm-Liouville problem. Thus {sin a,x}, = 1, 2,3, .. isan or thogonal set with respect to the weight function p(x) = 1 on the interval [0, 1] In some circumstances we can prove the orthogonality of the solutions of (3) without the necessity of specifying a boundary condition at x = a and at.x = bh. Ii Singular Sturm-Liouville Problem There are several other important conditions under Which wo seek nontrivial solutions of the differential equation (3): r(q@) = 0 and a boundary condition of the type given in (5) is specified atx=b; (1k r(b) = 0 and a boundary condition of the type given in (4) is specified atx=a; (13) (a) = r(0) = Oand no boundary condition is specified at either x= a oratx= bh; {18} (a) = r(b) and boundary conditior (15) ‘The differential equation (3) along with one of conditions (12) oF (13) is said to be a singular boundary-value problem. Equation (3) with the conditions specified in (15) is said to be a Periodic boundary-value problem because the boundary conditions are pride, Observe tha 0, then.x = a may be a singular point of the differential equation, and consequently a solution of (3) may become unbounded as.x—> a. However, we see from (8) that if r(a) = 0, then no boundary condition is required at x = a (o prove orthogonality of the eigenfunctions provided these solutions are bounded a that point. This latter requirement guarantees the existence ‘of the integrals involved. By assuming the solutions of (3) are bounded on the closed interval [a,b] ‘we can see from inspection of (8) that ‘+ Ifr(a) = 0, then the orthogonality relation (9) holds with no boundary condition atx = a (16) + IF r(®) = 0, then the orthogonality relation (9) holds with no boundary condition atx = b*; (7) + IF F(a) = r(b) = 0, then the orthogonality relation (9) holds with no boundary conditions specified at either x = a orx (18) + If r(a) = r(b), then the orthogonality relation (9) holds with the periodic boundary conditions ya) = y(b), ya) = y'(). (19) I Self-Adjoint Form if we carry out the different ion r(x)y’], the differential equation in (3) isthe same as a reny’ + Feoy" + Ga) + Apadyy = 0. (20) For example, Legendre’ differential equation (1 ~ x*jy” ~ 2ay’ + nln + Dy = Oisexactly ofthe form given in (20) with r(x) = 1 — x* and re) = —2x. Inother words, another way of ‘writing Legendre's DE is Sta - yy] + nln + Dy = 0. (20) But if-you compare other second-order DEs (say, Bessel’s equation, Cauchy-Euler equations, and DEs with constant coefficients) you might believe, given the coefficient of "isthe derivative of the coefficient of y", that few other second-order DEs have the form given in (3). On the con- trary, ifthe coefficients are continuous and a(x) # 0 for all xin some interval, then any second- order differential equation gay" + aay’ + (eG) + Adeyy (22) *Conelions (16) and (17) are equivalent to choosing Ay = 0, By respectively, 125 Sturm-Liowvlle Problem | 695 _chrome-extension:simhnlakiinojnhinhkkjpnepbhabphiipagesipdiwebiviewer mI? hips eSAN2FIL2Felasticoeanstals-east-2-3443,.. 716/047712124, 1015 AM 978128410597 pdf - Dennis-G.Zil-Advancod Eng neering Mathomatles-2016-Jones-Bartott pdf can be recast into the so-called self-adjoint form (3). To see tis, we proceed as ia Section 2.3, ‘where we rewrote a linear fi order equation ay(x)y' + aiix)y = O inthe form [py] = O by dividing the equation by a,(x) and then multiplying by the integrating Factor gr = e/"* where, assuming mo common factors, PUx) = ay(x)/a,(x). So fist, we divide (22) by a(x). The first two ‘Second, we mul- tiply this equation by the integrating factor", where a(x) and 6(x) are assumed to have ‘6 common factors é y+ “a0 Y+ [ y+ a ‘]+ a derivative of a product In summary, by dividing (22) by a(x) and then multiplying by e™*!™* we get p 4 Wemony 5 (eae 4 mn) corey HD coaye 4 (grain 4 A grain Y + ay" ac aa) P (23) ‘Equation (23) isthe desired form given in (20) and isthe same as (3): 8 evocy| 4. (Sty 5 MO re. A [ewoty] + (amon + Bhawan), <0 aw) an) Lage —— ra, ge) me For example, to express 39" + 6y' + Ay = Oin self-adjoint form, we write y" + 2y" + aby =0 and then multiply by e?# = e%, The resulting equation is, of) pe Ll 7 y+ ety’ + ates or A fey aad tery tazety=0 or Eley] +z ‘ote, P_Thiscertanly nt necessary o put a second-order dierential equation 22) nto the self-adjoint form (3 in order to solve the DE. For our purposes We se the form given in (3) to determine the veg function p(x) needed in the orthogonality relation (9) The next two examples illustrate cohogonality relations for Bessel functions and for Legendre polynomials. Ea Te Parametric Bessel Equation Tn Section 5:3 we saw thatthe general solution of the parametric Bessel differential equation 2y"-+ ay! + (2x? — Wy = 0.0 = 0, 1,2... 8 ¥ = Cys(ax) + C2¥,(ax), After dividing the pprameitie Bessel equation by the lead coetTicient.x” and multiplying the resulting equation by the integrating factor e!"" = elt = x, x > 0, we obtain the self-adjoint form y sy +(@-2)) oo Luvs (er-2)y where we idemify rx) =x, g(x) = ~r"lx, pla) = x, and A = a°, Now 1(0) = 0, and ofthe {wo solutions Ja) and (ax) only J,(ax) i bounded atx = 0. Thusin view of (16) above, the set {J,(@4)],1= 1,2,5,... isorthogonal with respect to the weight function p(x) = xon an inierval [0,2]. The orthogonality relation is [ wisteonntayo ds = 0, Ay#ay ny provided the a, and hence the eigenvalues A,= a 4 boundary condition at x = b of the type given in (5) |.2,3, .., ae defined by means of Ag (AD) + Byelfab) = 0. (25) 696 | CHAPTER 12 Orthogonal Functions and Fourier Series _chrome-extension:simnnlakailnojrnhinhkckpnepbhabphipagesipdtiwebviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443...717/1047712124, 1015 AM 978128410597 pdf -Dennis-G.Zil-Advancod ng neering Mathomatles-2016-Jones-Barttt pdf “The extra factor of e in (25) comes from the Chain Rule: = Ifax) = Jy (ax) 4 a= cd, Ge slen) = Ji(an) Fax = a (0x) For any choice of A, and 8, not both zero, its known that (25) hasan infinite number of roots 1;b. The eigenvalues are then A; = a? = (x;/6). More will be said about eigenvalues in the ext chapter. Logendre’s Equation From the result given in (21) we ean identify q(x) = 0, pCx) = 1, and A = n(n + 1). Recall from Section 5.3 when n = 0, 1,2, .. Legendre’s DE possesses polynomial solutions P,() Now we can put the observation that r(—1) = r(1) = 0 together with the fact that the Legendre polynomials P,(t) are the only solutions of (21) that are bounded on the closed interval [=1, 1], o conclude from (18) that the set [P,(2)}.n = 0, 1,2, ..i6 orthogonal with respect to the weight function p(2) = Ton [—I, I] The orthogonality relation is J Patra a 0, men, = REMARKS (@ A Sturm-Liouville problem is also considered to be singular when the interval under ‘consideration is infinite. See Problems 9 and 10 in Exercises 12.5. Gi, Even when the conditions on the coefficients p,q, rand r’ are as assumed in the regular ‘Sturm-Liouville problem, if the boundary conditions are periodic, then property (b) of ‘Theorem 12.5.1 does not hold. You are asked to show in Problem 4 of Exercises 12.5 that corresponding to cach eigenvalue of the BVP Ytay=0 y-D=KD, yiob=yo there exist wo linearly independent eigenfunctions. HELE ETE Arswors to selected odd-numbered problems begin on page ANS-20. In Problems 1 and 2, find the eigenfunctions and the equation 5. Find the square norm of each eigenfunction in Problem 1 that defines the eigenvalues for the given boundary-value 6. Show that for the eigenfunctions in Example 2, problem, Use a CAS to approximate the first four eigenvalues ‘Ais Aay As, anid Ay. Give the eigenfunctions corresponding (0 sin ag = 21 + c0s%o,]. these approximations 1. y'4ay=0, YO) =O, ya) ty) = 17. (a) Find the eigenvalues and eigenfunctions ofthe houndary- 2 y+Aay=0, 0) +90) =0, 1) =0 value problem 3. Consider y’ + Ay = O subject to y'(0 0. Show ‘that the eigenfunctions are yay T= 9, WO= 0 1G) [iewtien ts.) ose Seer 8. (a) Find the eigenvalues and eigenfunctions ofthe boundary- ‘This set, which is orthogonal on [0, LI, isthe basis for the value problem Fourier cosine series. 4 Considery’ + Ay = Osubject to the periodic boundary condi- I +y' +ay=0, 0 tions(—L) = y(L),y"(-D) = (L). Show thatthe eigenfune- (b) Put the differential equation in sel tions are (©) Give an orthogonality elation 9. Laguerre’s differential equation ay + (Ixy ty, 0 ‘This set, which is orthogonal on [—L, LI, is the basis for the thas polynomial solutions (2). Put the equation insel: Fourier sores, form and give an orthogonality relation, 125 Sturm-Liowille Problem | 697 _chrome-extensionsiminlakainojrnhinhkckpnepbhabphipagesipdtiwablviewer himIPfle=hiips%SA%2F%:2Folasticbeanstalk-us-east-2-3443... 718/1047