Discussion Problems In Problems 31 and 32, discuss how the methods of undetermined coefficients and variation of parameters can be combined to solve the given differential equation, Carey oot your ideas 4B, 3)" — 6)" + 30y = ISsinx +e tan dx BR y= 29’ bya a3 ale 33. What ate the intervals of definition ofthe general solutions in Problems 1, 7, 15, and 18? Discuss why the interval of definition of the general solution in Problem 28 is not (0, 0). 34, Find the general solution of x'y" + 'y’ —4x"y = 1 given that yj = atisa solution of the associated homogeneous equation. Computer Lab Assignments In Problems 35 and 36, the indefinite integrals of the equations in (5) are nonelementary. Use a CAS to find the first four nonzero terms of a Maclaurin series of each integrand and then integrate the result, Find a particular solution of the given differential equation Bay El Cauchy—Euler Equations INTRODUCTION The relative ease with which we were able to find explicit solutions of linear higher-order differential equations with constant coefficients in the preceding sections does not, in general, carry over to linear equations with variable coefficients. We shall sce in Chapter 5 that when a linear differential equation has variable coefficients, the best that we ean usually expect is to find a solution in the form of an infinite series. However, the type of dif- ferential equation considered in this scction is an exception to this rule; itis an equation with variable coefficients whose general solution can always be expressed in terms of powers of x, sines, cosines, logarithmic, and exponential functions, Moreover, its method of solution is quite similar to that for constant equations. Il Cauchy-Euler Equation Any linea differential equation ofthe form ae where the coetlicients 4,, a equation, an Kuler-Cauchy equation, an Euler equation, or an equi Fao + ax + ay = gt, ra aye tay = 8) are constants, is known diversely as a Cauchy-Euler imensional equation, ‘The differential equation is named in honor of two of the most prolific mathematicians of all, lime, Augustin-Louis Cauchy (French, 1789-1857) and Leonhard Euler (Swiss, 1707-1783). ‘The observable characteristic of this type of equation is thatthe degree k = n,n — 1, 1, 00f the monomial coefficients x* matches the order k of differentiation d*y/dx* td ay As in Section 3.3, we start the discussion with a detailed examination of the forms of the general solutions of the homogeneous second-order equation “The solution of hig! atta tT toy =o. a) ‘order equations follows analogously, Also, we can solve the nonhomo- geneous equation ax’y" + bry" + _8(2) by variation of parameters, once we have determined the complementary function .(a) of (I) Lead cect en 10 x = O could cause a problem, > “The coeificient of d°y/ds? is zero at x = 0, Hence, in order to guarantee that the fundamental results of Theorem 3.1.1 are applicable to the Cauchy juler equation, we confine our attention to finding the general solution on the interval (0, oe). Solutions on the interval (20, 0) can be obtained by substituting 1 = —x into the differential equation. § Exercises 3.6 Problems 49 and 50 in 36 Cauchy-Euler Equations | 141142 I| Method of Solution We ty a solution ofthe form y = x", where m isto be determined ‘Analogous to what happened when we substituted e int a linear equation with constant coeffi cients, after substituting» each term ofa Cauchy-Euler equation becomes «polynomial in tines since 12 ston 1m —2)-~ n+ Do ant Sy = atmo — 1m —2) 4m Dx = agn(mm = Ion = 2) (m = + Da For example, by substituting y = x" the second-order equation (1) becomes Dr ke Dx" + be" + ex = 1) +bm+ex"=0. a Ta + be Ft oo = aim — Dx" + bmx + ex" = (amin m+ oe = Thus y = 2"is a solution ofthe differential equation whenever m isa solution ofthe auxiliary equation amm—1)+bm+e=0 of am*+(b—am+e=0. (2) There are three different cases to be considered, depending on whether the roots ofthis quadratic equation are real and distinct, real and equal, or complex. In the last case the roots appear as a conjugate pai. Case 1: Distinct Real Roots Let m, and m, denote the real roots of (2) such that ‘m, # m;,Then y, = x"tandy, = x" form a fundamental set of solutions. Hence the general solution of (1) is ys ent + ee, () BEC Distinct Roots dy _ 49 ye ae a OO Solve x SOLUTION Rather than just memorizing equation (2), itis preferable to assume y = x" as the solution a few times in order to understand the origin and the difference between this new form of the auxiliary equation and that obtained in Section 3.3. Differentiate twice, and substitute back into the differential equation: By aD yet mim De” aT eT mtn = De 2 ax matt = ay = x"(m(m = 1) = 2m = 4) = x"(ne = 3m — 4) = 0 if m? — 3m ~ 4 = 0, Now (m + 1)(on ~ 4) = O implies m, = 1, m, = 4 and so (3) yields the general solution y = ex"! + ex" Case I: Repeated Real Roots Ifthe roots of 2) are repeated (that is, m, =m), then ‘we obtain only one solution, namely, y = x", When the roots of the quadratic equation am? + (b ~ a)m + ¢ = O are equal, the discriminant of the coefli- cients is necessarily zero. It follows from the quadratic formula that the root must be m, = ~(b — a2a. Now we can construct a second solution y, using (5) of Section 3.2. We frst write the Cauchy-Euler equation inthe standard form CHAPTER 3 Higher-Order Differential Equationsand make the identifications P(x) = b/ax and J(blax) dx = (b/a) In x. Thus OM dy centage ine Be ming ‘The general solution of (1) is then vse +e Inx “ BEDUEEM _epeated Roots Solve o. SOLUTION ‘The substitution y = x" yields + y= a"Amim = 1) + 8m + 1) = xn? + 4m + 1) = 0 when 4m? + 4m + 1 = O.or 2m + 1)*= 0. Since m, = ~4 is a repeated root, (4) gives the general solution y = ¢,x7!” + ex"? In.x = ‘or higher-order equations, if m, is a root of multiplicity &, then it can be shown that Ine, xn) xm dnxy are k linearly independent solutions. Correspondingly, the general solution of the differential equation must then contain a linear combination of these k solutions. Case IH: Conjugate Complex Roots If the roots of (2) are the conjugate pair im; = e+ iB, m, = a iB, where « and B > 0 ate real, then a solution isy = Cyt" + Cy, But when the roots of the auxiliary equation are complex, as in the case of equations with constant coefficients, we wish to write the solution in terms of real functions only. We note the identity y= (gayi = gems, which, by Euler's formula, isthe same as = cos(B nx) + i sin(B In) Similarly, 27 = cos(B nx) ~ i sin In). ‘Adding and subtracting the last two results yields YP x= 2eos(Bina) and x =a“ = 2i sin(B In, respectively. From the fact that y = Cx" + Cy‘ is a solution for any values of the constants, we see, in turn, for C= C; = Land C,= 1 C= ~Lthat yee x x78) andy, = 2%" — 2) or y= 2x" cos(Inx) and yy = Ziv" sin( In») are also solutions. Since W(x" cos(8 In 3), x* sin( In.) = Bx"! #0, B > 0, on the interval (0, os), we conclude that yy =xeos(Binx) andy, = x*sin( Ina) constitute a fundamental set of real solutions of the differential equation, 36 Cauchy-Euler Equations | 143Hence the general solution of (1) is “ey cos(B In x) +c sin(B In x)] (3) ECE 40 initial-Value Problem Solve the initial-value problem 4y" + 17y = 0, y(l) = —1,y"() = =f SOLUTION. The y’ term is missing in the given Cauchy-Euler equation; nevertheless, the substitution y = 2" yields 4ey" + Ly = x"m(m — 1) + 17) = Gn? — 4m +17) = 0 when dn! — 4mm + 17 = 0, From the quadsatic formula we find thatthe roots are my = $+ 2i fy and m = } — 2i. With the identifications a = } and f= 2, we see from (5) thatthe general solution of the differential equation on the interval (0, 00) is y=!” [ej eos(2In x) +e sin(2 In 39} So 75 18" By applying the intial conditions y(1) = —1, y’(1) = 0 0 the foregoing solution and using In 1 = 0 we then find, in turn, that cy ~ ~1 and c, ~ 0. Hence the solution of the initial- value problem is y ~ ~x"cos (2 In.). The graph of this function, obtained with the aid of FIGURE 361 Graph of solution of VP computer software, is given in FIGURE 3.6.1 The particular solution is seen to be oscillatory and in Example 3 “unbounded as xo. = ‘The next example illustrates the solution of a third-order Cauchy-Buler equation, ETE third-order Equation 5 ret dx? dx? ay jolve x + 1x2 + gy Soh wats SOLUTION The first three derivatives of ay DL nye » a . smfon — 1x2, 5 = mon = 1Yom — 2)? so that the given differential equation becomes ay a dy 4 Tx + By 8mm Lom — 28" + Semen — D2 Dae Be = x"(m(m = Dow = 2) + Stn = 1) + Tm + 8) =x (m+ Ini + Amb 8) = Cm + Dom + 4) = 0. In this case we see that y = x" will be a solution of the differential equation for m, = 2, ‘m, = 2i, and m, = —2i, Hence the general solution on the interval (0, c0) is y= ex? + e,c05(2 In x) = ¢, sin(2 In), Il Nonhomogeneous Equations The method of undetermined coefficients as described in Section 3.4 does not carry over, in general, to linear differential equations with variable coef ficients. Consequently, in the following example the method of variation of parameters is employed EEE Variation of Parameters Solve xy’ ~ 3x" + 3y= xe SOLUTION Since the equation is nonhomogeneous, we first solve the associated homogeneous equation. From the auxiliary equation (m ~ 1)(m ~ 3) = Owe find y, = e,x + ex’. Now before using variation of parameters to find a particular solution y, = m3 + ways, recall that 144 | CHAPTER 3 Higher-Order Differential Equationsthe formulas uj = Wy/W and us = Wy/W, where W,, W, and Ware the determinants defined ‘on page 137, and were derived under the assumption thatthe differential equation has been pputinto the standard form y" + P(x)y" + O@oy = f(a). Therefore we divide the given equation by», and from 33 . yt oy! toy = arte! wwe make the identification fx) = 2x%e*. Now with y; = x,y2 = 2° and bx 4 _| 0 ae — ps _|x 0 wale sof a fate Sf mare mf ace wetind wf EE tet andy = BES 2x? 2x? The integral ofthe later function is immediate, but in the ease of wwe integrate by parts twioe. The results are wy = —ie" + te" ~ 2e" and u; = e#, Hence Jp = Mays Hasyn = (Cate* + 2xe* — De + etx! = Be" — Deel Finally the general solution of the given equation is wx to? + Oto = Deel = Yr Vet Dy T) AGeneratization The second-order diferentil equation ad’y ay ate — a) Fr + We + = 0 (6) is a generalization of equation (1). Note that (6) reduces to (1) when x0 = 0, We can solve Cauehy-Buler equations of the form given in (6) exactly as we did in (1), namely, by seeking solutions y = (x — 49)" and using B= mx = xy"! and x xo" See Problems 39-42 in Exercises 3.6 REMARKS ‘The similarity between the forms of solutions of Cauchy Euler equations and solutions of linear equations with constant coefficients isnot ust a coincidence. For example, when the roots ofthe auxiliary equations foray’ + by’ +- ey = Oandar’y’ + bay’ + ey = Oare distinct and real, the respective general solutions are y= cet + ee and y= ea + eg, x>0. a In view of the identity e™* = x, x > 0, the second solution given in (7) can be expressed in the sae form as the frst solution Pane y= ce™™* + ce = ge + cae, where f= In x, This last result illustrates another fact of mathematical life: Any Cauchy— Euler equation can always be rewritten as a lincar differential equation with constant cocf- ficients by means of the substitution x = e'. The idea is to solve the new differential ‘equation in terms of the variable f, using the methods of the previous sections, and once the general solution is obtained, resubstitute f = In x. Since this procedure provides a good review of the Chain Rule of differentiation, you are urged to work Problems 4348 in Exercises 3.6. 36 Cauchy-Euler Equations | 145