In this section the central ideas of the theory of linear differential equations are introduced and illustrated concretely in the context of second-order equations. These key concepts include superposition of solutions (Theorem 1), existence and uniqueness of solutions (Theorem 2), linear independence, the Wronskian (Theorem 3), and general solutions (Theorem 4). This discussion of second-order equations serves as preparation for the treatment of nth order linear equations in Section 3.2. Although the concepts in this section may seem somewhat abstract to students, the problems set is quite tangible and largely computational.
In each of Problems 1-16 the verification that Yl and Y2 satisfy the given differential equation is a routine matter. As in Example 2, we then impose the given initial conditions on the general solution Y =ClYl + C2)l2. This yields two linear equations that determine the values of the constants Cl and C2.
1. Imposition of the initial conditions yeO) = 0, y'(O) = 5 on the general solution
y( x) = clex + c2e -x yields the two equations cl + c2 = 0, cl - c2 = 0 with solution cl = 5/2, c2 = - 5/2. Hence the desired particular solution is y(x) = 5(eX - e-X)/2.
2. Imposition of the initial conditions y(O) = -1, y'(O) = 15 on the general solution
y(x) = cle3x + c2e-3X yields the two equations cl + c2 = -1, 3cl - 3c2 = 15 with solution cl = 2, c2 = 3. Hence the desired particular solution is y(x) = 2e3x - 3e-3x.
3. Imposition of the initial conditions y(O) = 3, y'(O) = 8 on the general solution y(x) = cl cos2x + c2 sin2x yields the two equations cl = 3, 2c2 = 8 with solution cl = 3, c2 = 4. Hence the desired particular solution is y(x) = 3 cos 2x + 4 sin 2x.
4. Imposition of the initial conditions yeO) = 10, y'(O) = -10 on the general solution
y( x) = cl cos 5x + c2 sin 5x yields the two equations cl = 10, 5c2 = -10 with solution cl = 3, c2 = 4. Hence the desired particular solution is y(x) = 10 cos 5x - 2 sin 5x.
132
Chapter 3
5. Imposition ofthe initial conditions y(O) = 1, y'(O) = 0 on the general solution y(x) = c1ex + c2e2X yields the two equations c, + c2 = 1, c, + 2c2 = 0 with solution c1 = 2, c2 = -1. Hence the desired particular solution is y(x) = 2eX - e2x.
6. Imposition of the initial conditions y(O) = 7, y'(O) = -T on the general solution
y(x) = c,e2X + c2e-3X yields the two equations c, + c2 = 7, 2c, - 3c2 = -1 with solution c1 = 4, c2 = 3. Hence the desired particular solution is y(x) = 4e2x + 3e-3x.
7. Imposition ofthe initial conditions y(O) = - 2, y'(O) = 8 on the general solution y(x) = c1 + c2e-x yields the two equations c, + c2 = - 2, - c2 = 8 with solution c, = 6, c2 = - 8. Hence the desired particular solution is y(x) = 6 - 8e-x•
8. Imposition of the initial conditions y(O) = 4, y'(O) = - 2 on the general solution y(x) = c, + c2e3X yields the two equations c1 + c2 = 4, 3c2 = - 2 with solution
c, = 14/3, c2 = 2/3. Hence the desired particular solution is y(x) = (14 - 2e3X)/3.
9. Imposition of the initial conditions y(O) = 2, y'(O) = -T on the general solution y(x) = c1e-x + c2x e-x yields the two equations c1 = 2, - c1 + c2 = -1 with solution c1 = 2, c2 = 1. Hence the desired particular solution is y(x) = 2e-x + xe-x.
10. Imposition of the initial conditions y(O) = 3, y'(O) = 13 on the general solution y( x) = c,e5x + c2x e5x yields the two equations c1 = 3, 5c, + c2 = 13 with solution c, = 3, c2 = - 2. Hence the desired particular solution is y(x) = 3e5x - 2xe5x•
11. Imposition ofthe initial conditions y(O) = 0, y'(O) = 5 on the general solution
y(x) = c,ex cosx + c2ex sinx yields the two equations c, = 0, c, + c2 = 5 with solution c, = 0, c2 = 5. Hence the desired particular solution is y(x) = 5exsin x.
12. Imposition of the initial conditions y(O) = 2, y'(O) = 0 on the general solution
y(x) = c,e-3X cos2x + c2e-3x sin2x yields the two equations c, = 2, - 3c, + 2c2 = 5 with solution c, = 2, c2 = 3. Hence the desired particular solution is y(x)
e-3X(2 cos 2x + 3 sin 2x).
13. Imposition of the initial conditions y(l) = 3, y'(l) = 1 on the general solution y(x) = c,x + C2X2 yields the two equations c, + c2 = 3, c, + 2c2 = 1 with solution c, = 5, c2 = - 2. Hence the desired particular solution is y(x) = 5x - 2x2•
Section 3.1
133
14. Imposition of the initial conditions y(2) = 10, y'(2) = 15 on the general solution
y(x) = C1X2 + C2X-3 yields the two equations 4c1 + c2 18 = 10, 4c1 - 3c2 116 = 15 with solution c1 = 3, c2 = - 16. Hence the desired particular solution is y(x) = 3x2 - 16 I x3 .
15. Imposition of the initial conditions y(1) = 7, y'(1) = 2 on the general solution y(x) = c1x+c2xlnx yields the two equations c1=7, c1+c2=2withsolution c1 = 7, c2 = - 5. Hence the desired particular solution is y(x) = 7x - 5x In x.
16. Imposition of the initial conditions y(1) = 2, y'(I) = 3 on the general solution
y( x) = c1 cos(1n x) + c2 sin(ln x) yields the two equations c1 = 2, c2 = 3. Hence the desired particular solution is y(x) = 2 cos(ln x) + 3 sin(1n x).
17. If y=clx then y'+y2 = _clx2+c2Ix2 = c(c-l)lx2 -:f. ° unless either c==D or c = 1.
18. If y = cx3 then yy" = cx3• 6cx = 6C2X4 -:f. 6x4 unless c2 = 1.
20. Linearly dependent, because
21. Linearly independent, because x3 = + x2jxj if x> 0, whereas x3 = - x2jxj if x < o.
22. Linearly independent, because 1 + x = c(1 + jxj) would require that c = 1 with x = 0, but c = ° with x = -1. Thus there is no such constant c.
23. Linearly independent, because j(x) = +g(x) if x> 0, whereas j(x) = -g(x) if x < O.
24. Linearly dependent, because g(x) = 2j(x).
25. j(x) = eX sin x and g(x) = eXcos x are linearly independent, because j(x) = k g(x) would imply that sin x = k cos x, whereas sin x and cos x are linearly independent.
26. To see that j(x) and g(x) are linearly independent, assume that j(x) = c g(x), and then substitute both x = ° and x = Jd2.
27. Let Lfy] = y" + py' + qy. Then Lfyc] = ° and Lfyp] = f, so
134
Chapter 3
28. If y(x) = 1 + Cl cos X + c2sin x then y'(x) = -clsin x + C2COS X, so the initial conditions yeO) = y'(O) = -1 yield CI = -2, C2 = -1. Hence y = 1 - 2 cosx - sinx.
29. There is no contradiction because if the given differential equation is divided by x2 to get the form in Equation (8) in the text, then the resulting functions p(x) = -41x and q(x) = 61x2 are not continuous at x = 0.
30. (a) Y2 = x3 and Y2 = IX31 are linearly independent because x3 = C IX31 would require that C = 1 with x = 1, but C = - 1 with x = - 1.
(b) The fact that W(YI, Y2) = 0 everywhere does not contradict Theorem 3, because
when the given equation is written in the required form
y" - (3Ix)y' + (3Ix2)y = 0,
the coefficient functions p(x) = -31x and q(x) = 31x2 are not continuous at x = 0.
31. W(yl,Y2) = -2x vanishes at x = 0, whereas if YI and j-, were (linearly independent) solutions of an equation y" + py' + qy = ° with p and q both continuous on an open interval I containing x = 0, then Theorem 3 would imply that W,* 0 on /.
49. The solution curve with yeO) = 1, y'(O) = 6 is y(x) = 8e-x -7 e-2x. We find that y'(x)=O when x=ln(7/4) so e-x =417 and e-2x =16/49. It follows that
y(ln(7 / 4)) = 1617, so the high point on the curve is (In(7/ 4)), 1617) ~ (0.56,2.29), which looks consistent with Fig. 3.1.6.
50. The two solution curves with yeO) = a and yeO) = b (as well as y'(O) = 1) are
y = (2a + 1)e-X - (a + 1)e-2X, y = (2b+l)e-X -(b+1)e-2X.
136
Chapter 3
Subtraction and then division by a - b gives 2e-x = e-2X, so it follows that x = -In 2. Now substitution in either formula gives y = -2, so the cornmon point of intersection is (-In 2, -2).
51.
(a)
The substitution v = In x gives
, dy dydv ldy
y = - = -- =--
dx dv dx x dv
Then another differentiation using the chain rule gives
Substitution of these expressions for y' and y" into Eq. (21) in the text then yields immediately the desired Eq. (22):
d2y dy
a-+(b-a)-+cy = O.
dv2 dv
(b) If the roots fj and r2 of the characteristic equation ofEq. (22) are real and distinct, then a general solution of the original Euler equation is
52. The substitution v = In x yields the converted equation d? y I dv' - Y = 0 whose characteristic equation r2 -1 = 0 has roots fj = 1 and r2 = -1. Because e" = x, the corresponding general solution is
c y = cjeV + c2e-v = cjx +..2
x
53. The substitution v = In x yields the converted equation d? y I dv' + dy I dv -12 Y = 0 whose characteristic equation r2 + r -12 = 0 has roots fj = -4 and r2 = 3. Because e" = x, the corresponding general solution is
54. The substitution v = In x yields the converted equation 4 d2 y I dv' + 4 dy I dv - 3 y = 0 whose characteristic equation 4r2 + 4r - 3 = 0 has roots rj = -312 and r2 = 1/2. Because e" = x, the corresponding general solution is
Section 3.1
137
55. The substitution v = In x yields the converted equation d2 y 1 dv2 = 0 whose characteristic equation r2 = 0 has repeated roots lj, r2 = O. Because v = lnx, the corresponding general solution is
y = cj +C2V = cj +c2Inx.
56. The substitution v = In x yields the converted equation d2 y 1 dv' - 4 dy 1 dv + 4 y = 0 whose characteristic equation r2 - 4r + 4 = 0 has roots lj, r2 = 2. Because eV = x, the corresponding general solution is
2v 2 v 2 ( I)
y = cle + c2 v e = x cl + c2 n v .
SECTION 3.2
GENERAL SOLUTIONS OF LINEAR EQUATIONS
Students should check each of Theorems 1 through 4 in this section to see that, in the case
n = 2, it reduces to the corresponding theorem in Section 3.1. Similarly, the computational problems for this section largely parallel those for the previous section. By the end of Section 3.2 students should understand that, although we do not prove the existence-uniqueness theorem now, it provides the basis for everything we do with linear differential equations.
The linear combinations listed in Problems 1-6 were discovered "by inspection" - that is, by trial and error.
5. (1)(17) + (-34)(cos2x) + (17)(cos 2x) = 0, because 2 cos2x = 1 + cos 2x.
6. (-l)(eX) + (1)(cosh x) + (l)(sinh x) = 0, because cosh x = (eX + e-X)/2 and sinh x = (eX - e-X)/2.
138 Chapter 3
8.
1 X X2
W = 0 1 2x = 2 is nonzero everywhere.
0 0 2
eX e2x e3x
W X 2e2x 3e3x = 2e6x is never zero.
= e
eX 4e2x ge3x 7.
10. W = «' eX(x + 1)(x + 4) is nonzero for x> O.
11. W = x3 e2x is nonzero if x;t. O.
12. W = x-2[2 cos'(ln x) + 2 sirr'(ln x)] = 2x-2 is nonzero for x> O.
In each of Problems 13 - 20 we first form the general solution
then calculate y'(x) and y" (x) , and finally impose the given initial conditions to determine the values of the coefficients CJ, C2, C3.
13. Imposition of the initial conditions yeO) = 1, y'(O) = 2, y"(O) = 0 on the general solution y(x) = c1ex + c2e-x + c3e-2X yields the three equations
with solution c1 = 4/3, c2 = 0, c3 = -1/3. Hence the desired particular solution is given by y(x) = (4eX - e-2x)/3.
14. Imposition of the initial conditions y(O) = 0, y'(O) = 0, y"(O) = 3 on the general solution y(x) = c1ex + c2e2X + c3e3X yields the three equations
with solution c1 = 3/2, c2 = - 3, c3 = 3/2. Hence the desired particular solution is given by y(x) = (3eX - 6e2x + 3e3X)/2.
15. Imposition of the initial conditions yeO) = 2, y'(O) = 0, y"(O) = 0 on the general solution y(x) = c1ex + c2x e' + c3x2e3x yields the three equations
Section 3.2
139
with solution cj = 2, c2 = - 2, c3 = 1. Hence the desired particular solution is given by y(x) = (2 - 2x + x2)ex.
16. Imposition of the initial conditions j.(O) = 1, y'(O) = 4, y"(O) = 0 on the general solution
( ) x 2x 2x' ld h hr .
y x = cje + c2e + c3x e yie s t e t ee equations
with solution cj = -12, c2 = 13, c3 = -10. Hence the desired particular solution is given by y(x) = -12ex + 13e2x - 10xe2x.
17. Imposition of the initial conditions yeO) = 3, y'(O) = -1, y"(O) = 2 on the general solution y(x) = cj + c2 cos3x + c3 sin3x yields the three equations
with solution cj = 29/9, c2 = - 2/9, c3 = -1/3. Hence the desired particular solution is given by y(x) = (29 - 2 cos 3x - 3 sin 3x)/9.
18. Imposition of the initial conditions yeO) = 1, y'(O) = 0, y"(O) = 0 on the general solution y( x) = e' (cj + c2 cos X + c3 sin x) yields the three equations
with solution cj = 2, c2 = -1, c3 = -1. Hence the desired particular solution is given by y(x) = e\2 - cosx - sin x).
19. Imposition of the initial conditions y(1) = 6, y'(I) = 14, y"(I) = 22 on the general solution y(x) = cjx + C2X2 + C3X3 yields the three equations
with solution cj = 1, c2 = 2, c3 = 3. Hence the desired particular solution is given by y(x) = x + 2x2 + 3x3•
20. Imposition of the initial conditions y(l) = 1, y'(1) = 5, y"(1) = -lIon the general solution y(x) = cjx + C2X-2 + c3x-2lnx yields the three equations
140 Chapter 3
with solution c1 = 2, c2 = -1, c3 = 1. Hence the desired particular solution is given by y(x) = 2x - x-2 + x-21n x.
In each of Problems 21- 24 we first form the general solution
then calculate y'(x), and finally impose the given initial conditions to determine the values of the coefficients CI and C2.
21. Imposition of the initial conditions yeO) = 2, y'(O) = - 2 on the general solution y( x) = c1 cos X + c2 sin x + 3x yields the t~o equations ci = 2, c2 + 3 = - 2 with solution c1 = 2, c2 = - 5. Hence the desired particular solution is given by
y(x) = 2 cos x - 5 sin x + 3x.
22. Imposition of the initial conditions yeO) = 0, y'(O) = 10 on the general solution
y(x) = cle2x + c2e-2X - 3 yields the two equations ci + c2 - 3 = 0, 2cI - 2c2 = 10 with solution CI = 4, c2 = -1. Hence the desired particular solution is given by
y(x) = 4e2x - e-2x - 3.
23. Imposition of the initial conditions yeO) = 3, y'(O) = lIon the general solution
y(x) = c1e-x + c2e3X - 2 yields the two equations c1 + c2 - 2 = 3, - c1 + 3c2 = 11 with solution c1 = 1, c2 = 4. Hence the desired particular solution is given by
y(x) = e-x + 4e3x - 2.
24. Imposition ofthe initial conditions yeO) = 4, y'(O) = 8 on the general solution
y( x) = c1ex cos x + c.e' cos x + x + 1 yields the two equations ci + 1 = 4, ci + c2 + 1 = 8 with solution c1 = 3, c2 = 4. Hence the desired particular solution is given by
y(x) = eX(3 cos x + 4 sin x) + x + 1.
26.
(a)
Yl = 2 and Y2 = 3x
(b)
y = YI + Y2 = 2 + 3x
27. The equations
Section 3.2
141
(the latter two obtained by successive differentiation of the first one) evidently implyby substituting x = 0 - that Cl = C2 = C3 = O.
28. If you differentiate the equation Co + c,x + C2X2 + ... + c.x" = 0 repeatedly, n times in succession, the result is the system
C +2c x+·"+nc x"-' = 0
'2 n
(n -l)!c,,_, + n!cnx = 0
of n+ 1 equations in the n+ 1 coefficients co' C" c2' "', c", Upon substitution of
x = 0, the (k+l)st of these equations reduces to ksc, = 0, so it follows that all these coefficients must vanish.
29. If coerx + clxerx + ... + cwenerx = 0, then division by erx yields
Co + ClX + ... + c.x" = 0,
so the result of Problem 28 applies.
30. When the equation X2y" - 2xy' + 2y = 0 is rewritten in standard form
y" + (-2Ix)y' + (2Ii)y = 0,
the coefficient functions Pl(X) = -21x and P2(X) = 21i are not continuous at x = O. Thus the hypotheses of Theorem 3 are not satisfied.
31.
(a)
Substitution of x = a in the differential equation gives y"(a) = - p y'(a) - q(a).
(b) If j'(O) = 1 and y'(O) = 0, then the equation y" - 2y' - Sy = 0 implies that
y"(O) = 2y'(0) + Sy(O) = S.
32. Let the functions yl, Y2, ... ,Yn be chosen as indicated. Then evaluation at
x = a of the (k - l)st derivative of the equation ClYl + C2)!2 + ... C0ln = 0 yields Ck = O. Thus Cl = C2 = ... = c; = 0, so the functions are linearly independent.
33. This follows from the fact that
142 Chapter 3
1 1 1
abc = (b-a)(c-b)(c-a). a2 b2 c2
34. W(ji,/z, ···,In) = Vexp(riX), and neither V nor exp(rix) vanishes.
36. If Y = VYI then substitution of the derivatives
in the differential equation y" + PY' + qy = 0 gives
But the terms within brackets vanish because Yl is a solution, and this leaves the equation
Y1V" + (2y; + PYl)V' = 0 that we can solve by writing
v" Y'
=_2_1_p =>
v' Yl
In v' = - 2lny; - jp(x)dx + InC,
J -fp(x)dx
=> vex) = C e 2 dx+ K, Yl
With C = 1 and K = 0 this gives the second solution
37. When we substitute Y = vx3 in the given differential equation and simplify, we get the separable equation xv" + v' = 0 that we solve by writing
v" 1
=
=> In v' = -In x + In A,
v' x
, A
v =-
x
=> vex) = Alnx+B.
With A = 1 and B = 0 we get vex) = lnx and hence Y2(X) = x3lnx.
Section 3.2
143
38. When we substitute y = vx3 in the given differential equation and simplify, we get the separable equation xv" + 7v' = 0 that we solve by writing
V" 7
=
=> In v' = - 7ln x + In A,
v' X
I A
v =-
x7
A
vex) = --+B. 6x6
With A=-6 and B=O we get v(x)=l!x6 and hence Y2(x)=1/x3•
39. When we substitute Y = vexI2 in the given differential equation and simplify, we eventually get the simple equation v" = 0 with general solution vex) = Ax + B. With A=l and B=O we get v(x)=.x and hence Y2(x)=xex12•
40. When we substitute Y = vx in the given differential equation and simplify, we get the separable equation v" - v' = 0 that we solve by writing
v"
- = 1 => In v' = x + In A,
v'
v' = Aex => vex) = Aex + B.
With A = 1 and B = 0 we get vex) = e" and hence Y2(X) = xe",
41. When we substitute Y = vex in the given differential equation and simplify, we get the separable equation (1 + x )v" + X v' = 0 that we solve by writing
v" x 1
- = --- = -1+-- => lnv' = -x+ln(l+x)+lnA,
v' 1 + x 1 +x
v' = A(1 + x)e-X => vex) = A J(1 + x)e-X dx = - A(2 + x)e-X + B.
With A = -1 and B = 0 we get vex) = (2 + x)e " and hence Y2(X) = 2 + x.
42. When we substitute Y = v x in the given differential equation and simplify, we get the separable equation x (x2 -1 )v" = 2 v' that we solve by writing
144 Chapter 3
v" 2 2 1 1
= = --+-----
v' x( X2 - 1) x 1 + x 1 - x '
In v' = - 2lnx + In(1 + x) + In(l- x) + InA,
v' = A(l~X2) = A(:2 -1) => v(x) = A( - ~-x )+B.
With A=-1 and B=O we get v(x)=x+llx and hence Y2(X)=X2+1.
43. When we substitute y = v x in the given differential equation and simplify, we get the separable equation x(x2 -1)v" = (2 -4X2)V' that we solve by writing
v" 2 -4x2 2 1 1
= = -----+--
v' x( x2 - 1) x 1 + x 1 - x '
In v' = - 2lnx -In(1 + x) -In(l- x) + InA,
v' = A = A (_1 + 1 + 1 J
x2(1- x2) x2 2(1 + x) 2(1- x) ,
v(x) = A(-_!' + _!.In(1 + x) -_!.In(1- X)) + B.
x 2 2
With A = -1 and B = 0 we get
1 1 1
v( x) = - - -In(1 + x) + -In(l- x) =>
x 2 2
x 1+x Y2(X) = l--ln--. 2 I-x
44. When we substitute y = v x-l12 cosx in the given differential equation and simplify, we eventually get the separable equation (cosx)v" = 2(sinx)v' that we solve by writing
v" 2sinx
= =>
v' cosx
v' = Asec2 x =>
lnv' = -2Inlcosxl+lnA = ln sec/ x+lnA,
v(x) = A tan x + B.
With A = 1 and B = 0 we get v(x) = tanx and hence
Y2(X) = (tanx)(x-1/2 cosx) = x-l12 sinx.
Section 3.2
145
SECTION 3.3
HOMOGENEOUS EQUATIONS WITH CONSTANT COEFFICIENTS
This is a purely computational section devoted to the single most widely applicable type of higher order differential equations -linear ones with constant coefficients. In Problems 1-20, we write first the characteristic equation and its list of roots, then the corresponding general solution of the given differential equation. Explanatory comments are included only when the solution of the characteristic equation is not routine.
2. 2r2 - 3r = r(2r - 3) = 0; r = 0, 3/2; y(x) = CI + c2e3xI2
21. Imposition of the initial conditions yeO) = 7, y'(O) = lIon the general solution y(x) = c]ex + c2e3x yields the two equations c] + c2 = 7, c] + 3c2 = 11 with solution c] = S, C2 = 2. Hence the desired particular solution is y(x) = Sex + 2e3x.
22. Imposition of the initial conditions yeO) = 3, y'(O) = 4 on the general solution y(x) = e-xl3 [c] cos( xl -/3) + C2 sin( xl -/3) ] yields the two equations
c] = 3, - c] /3 + c2 / -/3 = 4 with solution c] = 3, c2 = s-/3. Hence the desired particular solution is y(x) = «:" [3cos( x / -/3) + s-/3 sin( xl -/3) 1
23. Imposition of the initial conditions yeO) = 3, y'(O) = 1 on the general solution
y( x) = e3x (c] cos 4x + c2 sin 4x) yields the two equations c] = 3, 3c] + 4c2 = 1 with solution c] = 3, c2 = - 2. Hence the desired particular solution is
y(x) = e3X(3 cos 4x - 2 sin 4x).
Section 3.3
147
24. Imposition of the initial conditions yeO) = 1, y'(O) = -1, y"(O) = 3 on the general solution y(x) = c1 + c2e2x + c3e-X12 yields the three equations
with solution c1 = - 7/2, c2 = 1/2, c3 = 4. Hence the desired particular solution is y(x) = (-7 + e2x + 8e-xI2)/2.
25. Imposition of the initial conditions yeO) = -1, y'(O) = 0, y"(O) = 1 on the general solution y(x) = c1 + c2x + c3e-2x13 yields the three equations
with solution c1 = -13/4, c2 = 3/2, c3 = 9/4. Hence the desired particular solution is y(x) = (-13+6x+ge-2x13)/4.
26. Imposition of the initial conditions yeO) = 1, y'(O) = -1, y"(O) = 3 on the general solution y( x) = c1 + c2e -5x + c3x e -5x yields the three equations
with solution c1 = 24/5, c2 = - 9/5, c3 = - 5. Hence the desired particular solution is y(x) = (24 - ge-5x - 25xe-5x)/5.
27. First we spot the root r = 1. Then long division of the polynomial r3 + 3r2 - 4 by r - 1 yields the quadratic factor r2 + 4r + 4 = (r + 2)2 with roots
r = -2, -2. Hence the general solution is y(x) = clex + c2e-2x + c3xe-2x.
28. First we spot the root r = 2. Then long division of the polynomial 2r3 - r2 - 5r - 2
by the factor r - 2 yields the quadratic factor 2r2 + 3r + 1 = (2r + 1)(r + 1) with roots r = -1, -1/2. Hence the general solution is y(x) = c1ix + C2e-x + C3e-x12.
29. First we spot the root r = -3. Then long division ofthe polynomial r3 + 27 by
r + 3 yields the quadratic factor r2 - 3r + 9 with roots r = 3 ( 1 ± i J3) 1 2. Hence the
general solution is y(x) = Cl e -3x + e3x12[ c2cos(3x J3 /2) + C3 sin(3x J3 12)].
30. First we spot the root r = -1. Then long division of the polynomial
148
Chapter 3
by r + 1 yields the cubic factor r3 - 2?- + 3r - 6. Next we spot the root r = 2, and another long division yields the quadratic factor r2 + 3 with roots r = ±i.J3. Hence the general solutionis y(x) = cle-x+c2e2x+c3cosx.J3 +c4sinx.J3.
31. The characteristic equation r3 + 3r2 + 4r - 8 = 0 has the evident root r = 1, and long division then yields the quadratic factor r2 + 4r + 8 = (r + 2i + 4 corresponding to the complex conjugate roots -2 ± 2 i. Hence the general solution is
32. The characteristic equation r4 + r3 - 3r2 - 5r - 2 = 0 has root r = 2 that is readily found by trial and error, and long division then yields the factorization
(r - 2)(r+ 1)3 = O.
Thus we obtain the general solution y(x) = cle2x + (C2 + C3X + c4X2)e-x.
33. Knowing that y = e3x is one solution, we divide the characteristic polynomial r3 + 3r2 - 54 by r - 3 and get the quadratic factor
r2 + 6r + 18 = (r + 3)2 + 9.
Hence the general solution is y(x) = cle3x + e-3x(c2COS 3x + cj sin 3x).
34. Knowing that y = e2x/3 is one solution, we divide the characteristic polynomial
3r3 - 2r2 + 12r - 8 by 3r - 2 and get the quadratic factor r2 + 4. Hence the general
solution is
y(X) = cle2x/3 + C2COS 2x + c3sin 2x.
35. The fact that y = cos 2x is one solution tells us that r2 + 4 is a factor of the characteristic polynomial
6r4 + 5r3 + 25r2 + 20r + 4.
Then long division yields the quadratic factor 6r2 + 5r + 1 = (3r + 1)(2r + 1) with roots r = - 1/2, -1/3. Hence the general solution is
36. The fact that y = e -x sin x is one solution tells us that (r + 1 i + 1 = r2 + 2r + 2 is a factor of the characteristic polynomial
Section 3.3
149
Then long division yields the linear factor 9r - 7. Hence the general solution is
37. The characteristic equation is r' - r3 = r'(r -1) = 0, so the general solution is
y( x) = A + Bx + Cx2 + D e", Imposition of the given initial conditions yields the equations
A+D = 18, B+D = 12, 2C+D = 13, D = 7
with solution A = 11, B = 5, C = 3, D = 7. Hence the desired particular solution is
38. Given that r = 5 is one characteristic root, we divide (r - 5) into the characteristic polynomial r3 - 5r2 + 1 OOr - 500 and get the remaining factor r2 + 100. Thus the general solution is
y(x) = Ae5x +Bcos10x+CsinlOx.
Imposition of the given initial conditions yields the equations
A+B = 0, 5A+10C = 10, 25A-IOOB = 250
with solution A = 2, B = - 2, C = O. Hence the desired particular solution is y(x) = 2e5x -2cos10x
39. (r - 2)3 = r3 - 6r2 + 12r - 8, so the differential equation is
y'll - 6 y" + 12 y' - 8 Y = O.
40. (r - 2)(r2 + 4) = r3 - 2r2 + 4r - 8, so the differential equation is
ylll - 2 y" + 4 y' - 8 Y = O.
41. (r2+4)(r2-4) = r4-16, so the differential equation is y(4)-16y = o.
42. (r2 + 4)3 = r6 + 12r4 + 48r2 + 64, so the differential equation is
44.
(a)
x = i, -2i
(b)
x = -i,3i
150
Chapter 3
45. The characteristic polynomial is the quadratic polynomial of Problem 44(b). Hence the general solution is
46. The characteristic polynomial is r2 - ir + 6 = (r + 2i)(r - 3i) so the general solution is
47. Thecharacteristicrootsare r = ±~-2+2iJ3 = ±(1+iJ3) so the general solutionis
48. The general solution is y(x) = Aex + Be'" + cef3x where a = (-1 + i J3 )/2 and p = (-1 - i J3 )/2. Imposition of the given initial conditions yields the equations
A+ B+ C = 1 A+ aB+ pC = 0 A+a2B+{J2C = 0
that we solve for A = B = C = 1/3. Thus the desired particular solution is given
by y(x) = t( e" + e(-I+i.J})xI2 + e(-I-.J})X/2), which (using Euler's relation) reduces to the
given real-valued solution.
49. The general solution is y = Aix + Be-x + C cos x + D sin x. Imposition of the given initial conditions yields the equations
A+B+C = 0
2A-B +D = 0 4A+B-C = 0 8A-B -D = 30
that we solve for A = 2, B = -5, C = 3, and D = -9. Thus
y(x) = 2e2x - 5e-x + 3 cos x - 9 sin x.
50. If x > 0 then the differential equation is y" + y = 0 with general solution y = A cosx + Bsinx .. But if x < 0 it is y" - y = 0 with general solution
y = C cosh x + D sin x. To satisfy the initial conditions Yl (0) = 1, y; (0) = 0 we choose
Section 3.3
151
A = c = 1 and B = D = O. But to satisfy the initial conditions Y2 (0) = 0, y; (0) = 1 we choose A = C = 0 and B = D = 1. The corresponding solutions are defined by
= { cos x if x ~ 0, cosh x if x s 0;
= {sinx if x z O, sinh x if x s o.
51. In the solution of Problem 51 in Section 3.1 we showed that the substitution v = lnx gives
, dy
y = - =
dx
1 dy x dv
and
A further differentiation using the chain rule gives
m d3y 2 dy 3 d2y 1 d3y
Y = dx3 = 7' dv -7' dv2 +7' dv3'
Substitution of these expressions for y', y", and y"' into the third-order Euler equation ax' y"' + bx' y" + CX y' + d y = 0 and collection of coefficients quickly yields the desired constant-coefficient equation
d3y d2y dy
a-+(b-3a)-+(c-b+2a)-+d y = O.
dv" dv' dv
In Problems 52 through 58 we list first the transformed constant-coefficient equation, then its characteristic equation and roots, and finally the corresponding general solution with v = In x
In this section we discuss four types of free motion of a mass on a spring - undamped, underdamped, critically damped, and overdamped. However, the undamped and underdamped cases - in which actual oscillations occur - are emphasized because they are both the most interesting and the most important cases for applications.
1. Frequency: (00 = .J k / m = .J16 / 4 = 2 rad / sec = lIJr Hz Period: P = 2Jr / (iJo = 2Jr /2 = Jr sec
2. Frequency (iJo =.Jk/m =.J48/0.75 =8 rad/sec=4/Jr Hz Period: P = 2Jr / (iJo = 2Jr / 8 = Jr / 4 sec
3. The spring constant is k = 15 N/0.20 m = 75 N/m. The solution of 3x" + 75x = 0 with x(O) = 0 and x'(O) = -10 is x(t) = -2 sin 5t. Thus the amplitude is 2 m; the
frequency is (iJo = .Jk/m = .J75/3 = 5 rad/sec = 2.5/Jr Hz; and the period is 2Jr/5 sec.
4. (a) With m = 1/4 kg and k = (9 N)/(0.25 m) = 36 N/m we find that ~ = 12 rad/sec. The solution of x" + 144x = 0 with x(O) = 1 and x'(O) = -5 is
(b) C = 13/12 :=:; 1.0833 m and T = 2nf12 :=:; 0.5236 sec.
5. The gravitational acceleration at distance R from the center of the earth is g = GMI R2.
According to Equation (6) in the text the (circular) frequency OJ of a pendulum is given
by 01 = gIL = GMIR2L, so its period is P = 2nfOJ = 2JrR-JLIGM.
6. If the pendulum in the clock executes n cycles per day (86400 sec) at Paris, then its period is PI = 86400ln sec. At the equatorial location it takes 24 hr 2 min 40 sec = 86560 sec for the same number of cycles, so its period there is P2 = 86560 In sec. Now let RI = 3956 mi be the Earth's "radius" at Paris, and R2 its "radius" at the equator. Then substitution in the equation pilp; = RIIR2 of Problem 5 (with LI = L2) yields
R2 = 3963.33 mi. Thus this (rather simplistic) calculation gives 7.33 mi as the thickness of the Earth's equatorial bulge.
7. The period equation P = 3960 -J100.l 0 = (3960 + x) Mo yields x :=:; 1.9795 mi :=:; 10,450 ft for the altitude of the mountain.
8. Let n be the number of cycles required for a correct clock with unknown pendulum length LI and period PI to register 24 hrs = 86400 sec, so nPI = 86400. The given
clock with length L2 = 30 in and period P2 loses 10 min = 600 sec per day, so
np2 = 87000. Then the formula of Problem 5 yields
so LI = (30)(86400 I 87000)2 :=:; 29.59 in.
10. The F = rna equation p:rr2hx" = p:rr2hg - m2xg simplifies to
x" + (gl ph)x = g.
The solution of this equation with x(O) = x'(O) = 0 is
x(t) = ph(1 - cos OJot)
where aJo = ~ g I ph. With the given numerical values of p, h, and g, the amplitude of oscillation is ph = 100 em and the period is P = 27r~phl g :=:; 2.01 sec.
154
Chapter 3
11. The fact that the buoy weighs 100 lb means that mg = 100 so m = 100/32 slugs.
The weight of water is 62.4lb/fe, so the F = ma equation of Problem 10 is
(l00/32)x" = 100 - 62.4nr2x.
It follows that the buoy's circular frequency OJ is given by
But the fact that the buoy's period is p = 2.5 sec means that OJ = 2Jl'12.5. Equating these two results yields r ~ 0.3173 ft ~ 3.8 in.
12.
(a)
Substitution of M; = (rl R)3 M in F, = -GMrmlr2 yields
(b) Because GMI R3 = gl R, the equation mr" = F, yields the differential equation
r" + (gl R)r = O.
(c) The solution of this equation with reO) = Rand r'(O) = 0 is ret) = Rcos OJot
where 0J0 = ~gl R. Hence, with g = 32.2 ft/sec' and R = (3960)(5280) ft, we find that the period of the particle's simple harmonic motion is
p = 27d0J0 = 2Jl'~RIg ~ 5063.10 sec ~ 84.38 min.
13. (a) The characteristic equation 10r2 + 9r + 2 = (5r + 2)(2r + 1) = 0 has roots
r = - 2/5, -1/2. When we impose the initial conditions x(O) = 0, x'(O) = 5 on the
general solution x(t) = cje-2tl5 + c2e-1/2 we get the particular solution
when t = 10 In( 5/4) ~ 2.23144. Hence the mass's farthest distance to the right is given by x(101n(5/4)) = 5121125 = 4.096.
14.
(a)
The characteristic equation 25r2 + lOr + 226 = (5r + 1)2 + 152 = 0 has roots
r = (-1 ± 15 i) 1 5 = -115 ± 3 i. When we impose the initial conditions x(0)=20~x'(0)=41 onthegeneralsolution x(t) = e-115(Acos3t+Bsin3t) we get A = 20, B = 15. The corresponding particular solution is given by
Section 3.4
155
x(t) = e-tf5(20cos3t+15sin3t) = 25e-tf5cos(3t-a) where a = tan " (3/ 4) :::: 0.6435.
(b) Thus the oscillations are "bounded" by the curves x = ±25e-tf5 and
the pseudoperiod of oscillation is T = 21C / 3 (because to = 3).
15. With damping The characteristic equation (1/ 2)r2 + 3r + 4 = 0 has roots r = - 2, - 4.
When we impose the initial conditions x(O) = 2, x'(O) = 0 on the general solution x(t) = cle-2t + c2e-4t we get the particular solution x(t) = 4e-2t - 2e-4t that describes overdamped motion.
Without damping The characteristic equation (1I2)r2 + 4 = 0 has roots r = ± 2i.fi. When we impose the initial conditions x(O) = 2, x'(O) = 0 on the general solution
u(t) = Acos(2.fi t) + BSin(2.fi t) we get the particular solution u(t) = 2cos(2.fi t). The graphs of x(t) and u(t) are shown in the following figure.
t
16. With damping The characteristic equation 3r2 + 30r + 63 = 0 has roots r = - 3, - 7.
When we impose the initial conditions x(O) = 2, x'(O) = 2 on the general solution x(t) = cle-3t + c2e-7t we get the particular solution x(t) = 4e-3t - ze" that describes overdamped motion.
Without damping The characteristic equation 3r2 + 63 = 0 has roots r = ± i.J2j. When we impose the initial conditions x(O) = 2, x'(O) = 2 on the general solution
u(t) = Acos(.J2j t) + B sin(.J2j t) we get the particular solution
The graphs of x(t) and u(t) are shown in the figure at the top of the next page.
156
Chapter 3
t
17. With damping The characteristic equation r2 +8r+ 16 = 0 has roots r = -4, -4.
When we impose the initial conditions x(O) = 5, x'(O) = -10 on the general solution x(t) = (cj + c2t )e-4t we get the particular solution x(t) = 5e-4t (2t + 1) that describes critically damped motion.
Without damping The characteristic equation r2 + 16 = 0 has roots r = ± 4i. When we impose the initial conditions x(O) = 5, x'(O) = -10 on the general solution
u(t) = Acos(4t) + Bsin(4t) we get the particular solution
The graphs of x(t) and u(t) are shown in the following figure.
t
18. With damping The characteristic equation 2r2 + 12r+50 = 0 has roots r = - 3± 4i.
When we impose the initial conditions x(O) = 0, x'(O) = - 8 on the general solution x(t) = e-3t (Acos4t+ Bsin4t) we get the particular solution
x(t) = - 2 e-3t sin 4t = 2 e-3t cos( 4t - 31l / 2) that describes underdamped motion.
Section 3.4
157
Without damping The characteristic equation 2r2 + 50 = 0 has roots r = ± 5i. When we impose the initial conditions x(O) = 0, x/CO) = - 8 on the general solution
u(t) = Acos(5t) + Bsin(5t) we get the particular solution
8 . 8 ( 3:r)
u(t)=--sm(5t)=-cos 5t-- .
5 5 2
The graphs of xU) and u(t) are shown in the following figure.
1
19. The characteristic equation 4r2 + 20r + 169 = 0 has roots r = - 512 ± 6 i. When we impose the initial conditions x(O) = 4, x/CO) = 16 on the general solution
x(t) = e-5t12 (Acos6t+ Bsin6t) we get the particular solution
x(t) = e-5t12 [ 4 cos 6t + Q sin 6t] 3
z _!_.J3i3 e-5t12 cos(6t - 0.8254) 3
that describes underdamped motion.
Without damping The characteristic equation 4r2 + 169 = 0 has roots r = ± 13i 1 2. When we impose the initial conditions x(O) = 4, x/CO) = 16 on the general solution uU) = Acos(13t 12) + B sin(13t 12) we get the particular solution
u(t) = 4cosC~t)+ ~~ sinC~t) z 1~ h33cosC: t-0.5517 ).
158
Chapter 3
20. With damping The characteristic equation 2r2 + 16r + 40 = 0 has roots r = - 4 ± 2 i.
When we impose the initial conditions x(O) = S, x'(O) = 4 on the general solution x(t) = e -41 (A cos 2t + B sin 2t) we get the particular solution
x(t) = e-41(S cos 2t + 12 sin 2t) ~ 13 e-4Icos(2t - 1.1760) that describes underdamped motion.
Without damping The characteristic equation 2r2 + 40 = 0 has roots r = ± 2J5 i. When we impose the initial conditions x(O) = s, x'(O) = 4 on the general solution
u(t) = A cos(2J5 t) + B sin(2J5 t) we get the particular solution
The graphs of x(t) and u(t) are shown in the following figure.
21. With damping The characteristic equation r2 + lOr + 12S = 0 has roots r = - S ± 10 i.
When we impose the initial conditions x(O) = 6, x'(O) = SO on the general solution x(t) = e-51 (Acos1 Ot + Bsin10t) we get the particular solution
x(t) = e-51(6 cos lOt + 8 sin lOt) ~ 10 e-5Icos(10t - 0.9273) that describes underdamped motion.
Without damping The characteristic equation r2 + 12S = 0 has roots r = ± sJ5 i. When we impose the initial conditions x(O) = 6, x'(O) = SO on the general solution
u(t) = A cos( sJ5 t) + B sin( sJ5 t) we get the particular solution
The graphs of x(t) and u(t) are shown in the figure at the top of the next page.
Section 3.4
159
22. (a) With m = 12/32 = 3/8 slug, c = 3 lb-sec/ft, and k = 24Ib/ft, the differential equation is equivalent to 3x" + 24x' + 192x = O. The characteristic equation
3r2 + 24r + 192 = 0 has roots r = - 4 ± 4.J3 i. When we impose the initial conditions
x(0)=1, x'(O)=0 on the general solution x(t) = e-4t (Acos4t.J3 + Bsin4t.J3) we get the particular solution
(b) The time-varying amplitude is 2/.J3 ::::: 1.15 ft; the frequency is 4.J3 ::::: 6.93
rad/sec; and the phase angle is Jt / 6 .
23.
(a)
With m = 100 slugs we get to = ..Jk/lOO. But we are given that
oi = (80 cyc1es/min)(21l)(l minl60 sec) = 8m3,
and equating the two values yields k ::::: 7018Ib/ft.
(b) With aJt = 21l(78/60) sec-l, Equation (21) in the text yields c ::::: 372.31
lb/(ftlsec). Hence p = c/2m ::::: 1.8615. Finally e !" = 0.01 gives t ::::: 2.47 sec.
30. In the underdamped case we have
x(t) = e-pt[A cos aJtt + B sin aJtt],
x'(t) = -pePt[A cos oxt + B sin aJtt] + e-pt[ -AaJtsin ox: + Baxcce mIt].
The conditions x(O) = xo, x'(O) = vo yield the equations A = Xo and -pA + BOlt = vo, whence B = (vo + pXo)/ Olt.
160
Chapter 3
31. The binomial series
(1 )a 1 a(a-1) 2 a(a-1)(a-2) 3
+ X = + ax + x + x + ...
2! 3!
converges if Ixi < 1. (See, for instance, Section 11.8 of Edwards and Penney, Calculus, 56h edition, Prentice Hall, 2002.) With a = 1/2 and x = - c2/4mk in Eq. (21) of Section 3.4 in this text, the binomial series gives
32. If x(t) = Ce-ptcos(OJlt - a) then
yields tan( OJI t - a) = -pi OJI.
33. If XI = X(tl) and X2 = X(t2) are two successive local maxima, then OJlt2 = OJltl + 27r so
Xl = C exp( -ptI) COS(OJItl - a),
X2 = C exp( -pt2) COS(OJlt2 - a) = C exp( =pta) COS(OJltl - a).
Hence xllx2 = exp[ -p(tl - t2)], and therefore
34. With tl = 0.34 and tz = 1.17 we first use the equation OJlt2 = OJltl + 27r from Problem 32 to calculate OJI = 27r/(0.83) I'::i 7.57 rad/sec. Next, with XI = 6.73 and X2 = 1.46, the result of Problem 33 yields
p = (1/0.83) In(6.7311.46) I'::i 1.84.
Then Equation (16) in this section gives
c = 2mp = 2(100/32)(1.84) I'::i l l.Sl lb-sec/ft,
and finally Equation (21) yields
Section 3.4
161
35. The characteristic equation r2 + 2r + 1 = 0 has roots r = -1, -1. When we impose the initial conditions x(O) = 0, x'(l) = 0 on the general solution x(t) = (c1 + cl)e-I we get the particular solution X1(f) = i«:
36. The characteristic equation r2 + 2r + (1-10-211) = 0 has roots r = -1 ± 10-11• When we impose the initial conditions x(O) = 0, x'(l) = 0 on the general solution
with solution c1 = 2,,-15", c2 = 2"-15". .This gives the particular solution
37. The characteristic equation r2 + 2r + (1 + 10-2") = 0 has roots r = -1 ± 10-11 i. When we impose the initial conditions x(O) = 0, x'(I) = 0 on the general solution
we get the equations c1 = 0, - c1 + 10-" c2 = 1 with solution c1 = 0, c2 = 10". This
gives the particular solution X3 (f) = 10" e -I sin(lO-" f).
38. lim x2 (f) = lim 10" e -I sinh(1 0-11 f) = t e -I • lim _si_nh_(;_l O_-_" t...:_) = t e -I and
11->00 11->00 11->00 10-" t
1· () li 10" -I • (10-11 ) -I li sin(1 0-11 t) -I
im X3 t = im e sm t = t e . im = t e ,
11->00 11->00 11->00 10-" t
using the fact that lim(sine)/ e = lim (sinh e)/ e = 0 (by L'Hopital's rule, for
0->0 0->0
instance).
162
Chapter 3
SECTION 3.5
NONHOMOGENEOUS EQUATIONS AND
THE METHOD OF UNDETERMINED COEFFICIENTS
The method of undetermined coefficients is based on "educated guessing". If we can guess correctly the form of a particular solution of a nonhomogeneous linear equation with constant coefficients, then we can determine the particular solution explicitly by substitution in the given differential equation. It is pointed out at the end of Section 3.5 that this simple approach is not always successful - in which case the method of variation of parameters is available if a complementary function is known. However, undetermined coefficients does tum out to work well with a surprisingly large number of the nonhomogeneous linear differential equations that arise in elementary scientific applications.
In each of Problems 1-20 we give first the form of the trial solution Ytria[, then the equations in the coefficients we get when we substitute Ytrial into the differential equation and collect like terms, and finally the resulting particular solution Yp.
15. This is something of a trick problem. We cannot solve the characteristic equation
r5 + 5r4 -1 = ° to find the complementary function, but we can see that it contains no constant term (why?). Hence the trial solution Ytrial = A leads immediately to the particular solution YP = -17.
19. First we note the duplication with the part c, + c2x of the complementary function (which corresponds to the factor r2 of the characteristic polynomial). Then:
20. First we note that the characteristic polynomial r3 - r has the zero r = I corresponding to the duplicating part eX of the complementary function. Then:
Ytria' = A+x(Bex); -A=7,3B=I; YP = -7+(1/3)xeX
In Problems 21-30 we list first the complementary function Ye, then the initially proposed trial function Yb and finally the actual trial function YP in which duplication with the complementary function has been eliminated.
21. Yc = eX(c,cosx+c2sinx); Yi = e' ( A cos x + B sin x)
Yi = (A + Bx + Cx2 ) cos 3x + ( D + Ex + Fx2 ) sin 3x
Yp = x {( A+ Bx +Cx2)cos3x+( D+ Ex + Fx2)sin3x ]
29. Yc = (c1 + c2x + C3X2 )ex + c4e2x + cse-2x; Yi = (A + Bx )eX + C e" + De-2x
Yp = x3 .(A+Bx)eX +x.(Ce2x)+x.(De-2X)
30. Yc = (c, +c2x)e-X +(c3 +c4x)eX
Yi = Yp = ( A + Bx + Cx2 ) cos X + ( D + Ex + Fx2 ) sin x
In Problems 31-40 we list first the complementary function Ye, the trial solution Ytr for the method of undetermined coefficients, and the corresponding general solution Yg = Ye + YP where YP results from determining the coefficients in Ytr so as to satisfy the given nonhomogeneous differential equation. Then we list the linear equations obtained by imposing the given initial conditions, and finally the resulting particular solution y(x).
41. The trial solution Ytr = A + Bx + Cx2 + Dx3 + EX4 + Fx5 leads to the equations
2A-B-2C-6D+24E = 0 -2B-2C-6D-24E+120F = 0
-2C-3D-12E-60F = 0
-2D-4E-20F = 0
-2E-5F = 0
-2F = 8
that are readily solve by back-substitution. The resulting particular solution is
y(x) = -255 - 450x + 30x2 + 20x3 + 10x4 - 4x5.
168
Chapter 3
42. The characteristic equation r4 - r3 - r' - r - 2 = 0 has roots r = -1, 2, ± i so the complementary function is Yc = c]e-x + c2e2x + c3 cosx + c4 sinx. We find that the coefficients satisfy the equations
Solution of this system gives finally the particular solution y = Yc + yp where YP is the particular solution of Problem 41 and
Yc = lOe-x + 35e2x + 210cosx + 390sinx.
43.
(a)
cos3x+isin3x = (cos x-i-r sin x)?
= cos' x+3icos2 xsin x +Scos xsin' x-isin3 x
When we equate real parts we get the equation
cos' X - 3 ( cos x) ( 1 - cos' x) = 4 cos' X - 3 cos x
and readily solve for cos' x = fcosx +tcos3x. The formula for sirr' x is derived similarly by equating imaginary parts in the first equation above.
(b) Upon substituting the trial solution Yp = A cos x + B sin x + C cos 3x + D sin 3x in the differential equation y" + 4 y = f cos x + t cos 3x, we find that A = 1/4, B = 0, C = -1/20, D = O. The resulting general solution is
y(x) = CI cos 2x + c2sin 2x + (1I4)cos x - (1I20)cos 3x.
44. We use the identity sinxsin3x = tcos2x -tcos4x, and hence substitute the trial solution yp = Acos2x + Bsin2x + C cos4x + Dsin4x in the differential equation y" + y' + y = tcos2x -tcos4x. We find that A = -3/26, B = liB, C = -14/482, D = 2/141. The resulting general solution is
y(x) = e-xI2(clCOSxJj/2+c2sinx.fil2)
+ (-3 cos 2x + 2 sin 2x)/26 + (-15 cos 4x + 4 sin 4x)/482.
45. We substitute
·42 sm x = (1 - cos 2x) 14
Section 3.5
169
= (I - 2 cos 2x + cos22x)/4 = (3 - 4 cos 2x + cos 4x)/8
on the right-hand side of the differential equation, and then substitute the trial solution Yp = Acos2x+Bsin2x+Ccos4x+Dsin4x+E. We find that A=-1I10,B=0,
C = -1156, D = 0, E = 1124. The resulting general solution is
Y = Cl cos 3x + c2sin 3x + 1/24 - (1110)cos 2x - (1I56)cos 4x.
46. By the formula for cos' x in Problem 43, the differential equation can be written as
y"+ Y = txcosx+txcos3x.
The complementary solution is Yc = Cl cos X + c2sin x, so we substitute the trial solution
Yp = x{(A+Bx)cosx+(C+Dx)sinx ]+[(E+Fx)cos3x+(G+Hx)sin3x J.
We find that A = 3116, B = C = O,D = 3/16,E = 0, F = -1I32,G = 3/128,H =0. Hence the general solution is given by Y = Yc + Yl + Y2 where
Yl = (3xcosx+3x2sinx)116 and Y2 = (3sin3x-4xcos3x)1128.
In Problems 47-49 we list the independent solutions Yl and Yz of the associated homogeneous equation, their Wronskian W = W(YI' Yz), the coefficient functions
U (x) = - f Yz(x) f(x) dx and U (x) = f Yl(X) f(x) dx
1 W(x) Z W(x)
in the particular solution Yp = U1Yl + uzYz ofEq. (32) in the text, and finally yp itself.
47. = e-2x -x W -3x
Yl , Y2 = e , = e
Ul = -(4/3)e3X, U2 = 2e2x,
YP = (2/3)eX
48. -2x 4x W = 6ix
Yl = e , Y2 = e ,
Ul = -xI2, U2 = _e-6x 112
,
YP = -(6x + l)e-2xI12
49. Yl = e2x _ 2x W = e4x
, Y2 - xe ,
= 2 U2 = 2x,
Ul -x,
YP = x2e2x 170
Chapter 3
50. The complementary function is YI = CI cosh 2x + c2sinh 2x, so the Wronskian is
51.
52.
53.
so when we solve Equations (31) simultaneously for u; and u;, integrate each and substitute in YP = YIUI + Y2U2, the result is
Y p = - (cosh 2x) ftc sinh 2x)( sinh 2x) dx + (sinh 2x) ftc cosh 2x)( sinh 2x) dx .
Using the identities 2 sinh2x = cosh 2x - 1 and 2 sinh x cosh x = sinh 2x, we evaluate the integrals and find that
YP = (4x cosh 2x - sinh 4x cosh 2x + cosh 4x sinh 2x)116, YP = (4x cosh 2x - sinh 2x)116.
YI = cos2x,
Y2 = sin 2x,
W = 2
Liberal use of trigonometric sum and product identities yields
U1 = (cosSx - Scosx)/20,
U1 = (sinSx-Ssinx)/20
YP = -(1/4)( cos 2x cos x - sin 2x sin x) + (1/20)( cos Sx cos 2x + sin Sx sin 2x) -(1/S)cos 3x (!)
YI = cos 3x,
Y2 = sin 3x,
W=3
Uj = -(6x-sin6x)/36, YP = -(x cos 3x)/6
Uj = -(1+cos6x)/36
YI = cos 3x,
Y2 = sin 3x,
W=3
U; = -(2/3)tan 3x,
U' = 2/3
2
YP = (2/9)[3x sin 3x + (cos 3x) In Icos 3xl ]
54.
W = 1
YI = cos x,
Y2 = smx,
u; = cos X csc2x
U' = -csc X
j ,
YP = -1 - (cos x) Inlcscx-cotxl
55.
Yl = cos 2x,
Y2 = sin 2x,
W=2
U; = -(1/2)sin2x sin 2x = -(1/4)(1 - cos 2x)sin 2x
Section 3.5
171
u; = (1I2)sin2x cos 2x = (1/4)(1 - cos 2x)cos 2x YP = (1 - x sin 2x)/8
56.
-2x YI = e ,
W = 4
UI = -(3x - 1 )e3x 136, YP = -eX(3x + 2)/9
U2 = -(x + 1)e-XI4
57. With YI = x, Y2 = x-I, and .f{x) = 72x3, Equations (31) in the text take the form
XU' + X -I U' = 0,
1 2
Upon multiplying the second equation by x and then adding, we readily solve first for
U' = 36x3 so
1 ,
- 9 4
UI - X
and then
so
Then it follows that
58. Here it is important to remember that - for variation of parameters - the differential equation must be written in standard form with leading coefficient 1. We therefore rewrite the given equation with complementary function Yc = CIX2 + c2x3 as
y" - (4Ix)y' + (6Ix2)y = X.
Thus .f{x) = x, and W = X4, so simultaneous solution of Equations (31) as in Problem 50 (followed by integration of u; and u;) yields
But we can drop the term sin x because it satisfies the associated homogeneous equation y" + Y = o.
Section 3.5
173
SECTION 3.6
FORCED OSCILLATIONS AND RESONANCE
1. Trial of x = A cos 2t yields the particular solution xp = 2 cos 2t. (Can you see that - because the differential equation contains no first-derivative term - there is no need to include a sin 2t term in the trial solution?) Hence the general solution is
x(t) = C! cos 3t + c2sin 3t + 2 cos 2t.
The initial conditions imply that Cl = -2 and C2 = 0, so x(t) = 2 cos 2t - 2 cos 3t. The following figure shows the graph of x(t).
27T
-3
3
3 IT
t
5lT
2. Trial of x = A sin 3t yields the particular solution xp = -sin 3t. Then we impose the initial conditions x(O) = x/CO) = 0 on the general solution
IT
xCt) = Cl cos 2t + c2sin 2t - sin 3t,
and find that x(t) = !sin 2t - sin 3t. The following figure shows the graph of x(t).
-2
174
2
27T
Chapter 3
3 zr
t
5 zr
3. First we apply the method of undetermined coefficients - with trial solution x = A cos 5t + B sin 5t - to find the particular solution
xp = 3 cos 5t + 4 sin 5t
= 5 [~cos5t+ ~ sin5t~ = 5cos(5t- 13)
where 13 = tan-1(4/3) :::::; 0.9273. Hence the general solution is
x(t) = Cl cos lOt + c2sin lOt + 3 cos 5t + 4 sin St.
The initial conditions x(O) = 375, x'(O) = 0 now yield CI = 372 and C2 = -2, so the part of the solution with frequency OJ = lOis
Xc = 372 cos lOt - 2 sin lOt
[ 372 2. l
= .J138388 .J cos lOt - .J sin lOti
138388 138388 'l
= .J138388cos(lOt-a)
where a = 21r - tan-I(1/186) :::::; 6.2778 is a fourth-quadrant angle. The following figure shows the graph of x(t).
375
-375
I 1T /5
t
7T
4. Noting that there is no first-derivative term, we try x = Acos4t and find the particular solution xp = lOcos 4t. Then imposition of the initial conditions on the general solution x( t) = c, cos 5t + c2 sin 5t + lOcos 4t yields
x(t) = (-10 cos 5t + 18 sin 5t) + 10 cos 4t
Section 3.6
175
= 2 (-5cos5t + 9sin5t) + lOcos4t
= 2M6(- b cos5t+ b sin5t)+ lOcos4t
,,106 ,,106
= 2M6 cos(5t - a)
where a = J'C - tan-I(9/5) :::::: 2.0779 is a second-quadrant angle. The following figure shows the graph of x(t).
30
I 27T I
1\ II
" A n
" n 1\
J j P
V ¥ \I U ~ " t
-30
5. Substitution of the trial solution x = C cos OJ{ gives C = Fol (k - mol). Then imposition of the initial conditions x(O) = xo' x/CO) = 0 on the general solution
gives the particular solution x(t) = (xo - C)cos ~t + C cos ox.
6. First, let's write the differential equation in the form x" + %' x = (Fo 1m) cos olaf, which is the same as Eq. (13) in the text, and therefore has the particular solution
xp = (Fo/2m%)tsinov given in Eq. (14). When we impose the initial conditions
x(O) = 0, x/CO) = Vo on the general solution
we find that c1 = 0, c2 = Vo I mo' The resulting resonance solution of our initial value problem is
xU)
2mvo + Fot .
= smmot.
2m%
176
Chapter 3
In Problems 7-10 we give first the trial solution xp involving undetermined coefficients A and B, then the equations that determine these coefficients, and finally the resulting steady periodic solution xsp. In each case the figure shows the graphs of xsp(t) and the adjusted forcing
function F;(t) = F(t)/mm.
7.
Xp = A cos 3t + B sin 3t;
-5A+12B = 10, 12A+5B = 0
50 120 . 10 (5 12.) 10
x (t) = --cos3t+-sm3t = - --cos3t+-sm3t = -cos(3t-a)
sp 169 169 13 13 13 13
a = JZ" - tan -I (12 / 5) z 1.9656 (2nd quadrant angle)
3
t
JT
-3
8.
Xp = Acos5t+Bsin5t;
-20A+15B = -4, 15A+20B = 0
16 12 . 4 (4 3.) 4
x (t) = -cos5t--sm5t = - -cos5t--sm5t = -cos(5t-a)
a == Jr+ tan-1(199 120) "'" 4.6122 (3rd quadrant angle)
0.1
Xsp
~ ~ JI ~ ~ I ~ n A
" " ~ h n ~ h b ~ ~
IV IV IV V IV ~ V IV IV IV 2JT
V \lFl \ ~ V ~ \I ~ V u t
-0.1
10. xp == AcosIOt+lOsin5t;
956 342 .
--cos10t- smlOt
10309 10309
-97A+30B == 8, 30A+97B =-6
2../257725 ( 478 10 171 . 10) 10 C761 (10 )
= -cos t- sm t == -YOlcos t-a
10309 ../257725 ../257725 793
a == Jr+ tan-1(171/ 478) "" 3.4851 (3rd quadrant angle)
sp
n 11 n ~ ~ ~ ~ ~ ~ "
h ~ b b b b b b h b
V V V rv rv IV' V V V V ~JT
~ ~ F 1 \j ~ " ~ ~ ~ ~ ~ x
1
t
-1
178
Chapter 3
Each solution in Problems 11-14 has two parts. For the first part, we give first the trial solution xp involving undetermined coefficients A and B, then the equations that determine these coefficients, and finally the resulting steady periodic solution xsp. For the second part, we give first the general solution x(t) involving the coefficients CI and C2 in the transient solution, then the equations that determine these coefficients, and finally the resulting transient solution xtr so that x(t) = Xtr(t) + xsp(t) satisfies the given initial conditions. For each problem, the
The figure at the top of the next page shows the graphs of x(t) and xsp (t).
Section 3.6
181
30 Xsp
t
-30
In Problems 15-18 we substitute x(t) = A(m)cosax+B(m)sinax into the differential equation mx" + ex' + kx = Fo cos ax with the given numerical values of m, e, k, and Fo' We give first the equations in A and B that result upon collection of coefficients of cos ax and sin ax, and then the values of A(m) and B(m) that we get by solving these equations. Finally,
C = .J A 2 + B2 gives the amplitude of the resulting forced oscillations as a function of the forcing frequency co. and we show the graph of the function C(m).
15. (2-m2)A+2mB = 2, -2mA+(2-m2)B = 0
A
B = 4m 4+m4
C( m) = 2/.J 4 + m4 begins with C(O) = 1 and steadily decreases as to increases. Hence there is no practical resonance frequency.
C
1
w
5
10
182
Chapter 3
16. (5-ai)A+4mB = 10, -4mA+(5-m2)B = 0
A = 10(5.-m2) 25 + 6m2 + m4 '
C(m) = 101 ,J25 + 6m2 + m4 begins with C(O) = 2 and steadily decreases as to increases. Hence there is no practical resonance frequency.
C
2
1
w
5
10
17. (45-al)A+6mB = 50, -6mA+(45-m2)B = 0
A = _50__;,( 4_5_- -;:-al...,:_) ---;-
2025 - 54m2 + m4 '
300m
B = ------------
2025 - 54m2 + m4
C(m) = 501"'/2025-54al +m4 so, to find its maximum value, we calculate the derivative
Hence the practical resonance frequency (where the derivative vanishes) is oi = m = 3Ji The graph of C( m) is shown at the top of the next page.
18. (650 - m2) A + 10m B = 100, - 10m A + (650 - m2) B = 0
100 ( 650 - m2)
A = ----~----~~~
422500 -1200m2 + m4 '
B = 1000m
422500-1200m2 + m4
C(m) = 100/ "'/422500-1200m2 + m4 so, to find its maximum value,
Section 3.6
183
c
1
w
10
20
we calculate the derivative
C/([O) = -200[0(-600+ [(2)
( 422500 -1200[02 + [04 )3/2 •
Hence the practical resonance frequency (where the derivative vanishes) is
[0 = J600 = 1OJ6.
c
0.4
w
25
50
19. m = 100/32 slugs and k = 1200 lb/ft, so the critical frequency is [00 = .Jk 1 m = .J384 rad/sec = .J384 1 21r ~ 3.12 Hz.
20. Let the machine have mass m. Then the force F = mg = 9.8m (the machine's weight) causes a displacement of x = 0.5 em = 11200 meters, so Hooke's law
184 Chapter 3
F = kx, that is, mg = k(l I 200)
gives the spring constant is k = 200mg (N/m). Hence the resonance frequency is
21. If e is the angular displacement from the vertical, then the (essentially horizontal) displacement of the mass is x = LB, so twice its total energy (KE + PE) is
Differentiation, substitution of e r:::, sin e, and simplification yields
B"+ (kim + gIL)B = 0
so
OJo = ~ kim + giL.
22. Let x denote the displacement of the mass from its equilibrium position, v = x' its velocity, and OJ = v I a the angular velocity of the pulley. Then conservation of energy yields
mv2/2 + /0112 + kx2/2 - mgx = C.
When we differentiate both sides with respect to t and simplify the result, we get the differential equation
(m + II a2)x" + kx = mg.
Hence OJ = ~ k I ( m + / I a' ).
23.
(a)
In ft-lb-sec units we have m = 1000 and k = 10000, so OJo =.JlO rad/sec
r:::, 0.50 Hz.
(b) We are given that OJ = 2:r 12.25 r:::, 2.79 rad/sec, and the equation
mx" + kx = F(t) simplifies to
x"+10x = (l/4)aisinOJt.
When we substitute x(t) = A sin ox we find that the amplitude is
A = OJ2/4(10-w2) r:::, 0.8854 ft r:::, 10.63 in.
Section 3.6
185
24. By the identity of Problem 43 in Section 3.5, the differential equation is
mx'+kx = Po (3cosmt+cos3mt)/4.
Hence resonance occurs when either m or 3m equals mo = -J kim, that is, when either m = mo or ca = mo 13.
25. Substitution of the trial solution x = A cos on + B sin tot in the differential equation, and then collection of coefficients as usual yields the equations
with coefficient determinant ~ = (k - moi' ) 2 + ( cm)2 and solution A = - ( coi ) Fo / ~, B = (k - mar: ) Po I~. Hence
where tan,8 = Po I Eo' The desired formula now results when we substitute the value of p defined above.
27. The derivative of C(m) = Fo/~(k-mm2r+(cm)2 isgivenby
C'(m) = _ mPo (c2 - 2km) + 2(mm)2 .
2 [ 2 2J3/2
( k - m(2) + ( eta )
186
Chapter 3
(a) Therefore, if e2 ~ 2km, it is clear from the numerator that C'(z») < 0 for all to,
so C( w) steadily decreases as ca increases.
(b) But if e2 < 2km, then the numerator (and hence C' (w) ) vanishes when
ia = wm = .J kim - e2 12m2 < .J kim = wo' Calculation then shows that
so it follows from the second-derivative test that C(wlll) is a local maximum value.
28. (a) The given differential equation corresponds to Equation (17) with Fa = mAw2• It therefore follows from Equation (21) that the amplitude of the steady periodic vibrations at frequency ca IS
C(w)
(b) Now we calculate
C'(w)
and we see that the numerator vanishes when
k ( 2mk) [k =
m Zmk=c' > ~-;;
29. We need only substitute Eo = acoi and Fa = ak in the result of Problem 26.
30. When we substitute the values ca = 2;rv I L, m = 800, k = 7 x 104, e = 3000 and
L = 10, a = 0.05 in the formula of Problem 29, simplify, and square, we get the function
25 (9;r2v2 + 122500)
Csq(v) = -----'-------'----::;-
16(16;r4v4 -64375;r2v2 + 76562500f
Section 3.6
187
giving the square of the amplitude C (in meters) as a function of the velocity v (in meters per second). Differentiation gives
Because the principal factor in the numerator is a quadratic in v2, it is easy to solve the equation Csq'(V) = 0 to find where the maximum amplitude occurs; we find that the
only positive solution is v ~ 14.36 mlsec ~ 32.12 mi/hr, The corresponding
amplitude ofthe car's vibrations is ~Csq(14.36) ~ 0.1364 m = 13.64 em.
SECTION 3.7
ELECTRICAL CIRCUITS
1. With E(t) == 0 we have the simple exponential equation 51'+ 251 = 0 whose solution with 1(0) = 4 is let) = 4e-5t•
2. With E(t) == 100 we have the simple linear equation 51'+ 251 = 100 whose solution with 1(0) = 0 is let) = 4(1 - e-5t).
3. Now the differential equation is 51'+ 251 = 100 cos 60t. Substitution of the trial solution
lp = A cos 60t + B sin 60t
yields
lp = 4(cos 60t + 12 sin 60t)1145.
The complementary function is Ie = ce-5t; the solution with 1(0) = 0 is
let) = 4(cos 60t + 12 sin 60t - e-5t)/145.
4. The solution of the initial value problem 21' + 401 = lOOe-lOt, 1(0) = 0 is let) = 5(e-lOt - e-20t). To find the maximum current we solve the equation
1'(t) = _50e-IOt +100e-20t =_50e-201(e201_2) = 0 for t = (ln2)/10. Then
lmax = 1((ln2)110) = 5/4.
5. The linear equation 1'+ 101 = 50 e -IOtcos 60t has integrating factor p = elOt. The resulting general solution is let) = e-10t [(5/6)sin60t+ cj.] To satisfy the initial condition 1(0) = 0, we take C = 0 and get let) = (5/6)e-10tsin 60t.
188 Chapter 3
6. Substitution of the trial solution I = A cos 60t + B sin 60t in the differential equation l' + 101 = 30cos60t + 40 sin 60t gives the equations lOA + 60B = 30, - 60A + lOB = 40 with solution A = -21/37, B = 22/37. The resulting steady periodic solution is
lsp(t) = (1/37)(-21 cos 60t + 22 sin 60t) = (51 ffi )cos(60t - a), where
a = 7£- tan-\22/21) ~ 2.3329 (2nd quadrant angle).
7. (a) The linear differential equation RQ'+ (l/C)Q = Eo has integrating factor p = eIIRC• The resulting solution with Q(O) = 0 is Q(t) = EoC(l - e-tIRC). Then let) = Q 'et) = (Eol R)e -tIRC.
(b) These solutions make it obvious that limQ(t) = EoC and liml(t) = O.
I-'>'IJ 1-+00
8. (a) The linear equation Q' + 5Q = lOe-51 has integrating factor p = e51• The resulting solution with Q(O) = 0 is Q(t) = 10te-5t, so let) = Q'(t) = 10(1 - 5t)e-5t.
(b) let) = 0 when t = liS, so Qmax= Q(1/5) = 2e-1•
9. Substitution of the trial solution Q = Acos120t + Bsin120t into the differential equation 200 Q' + 4000Q = 100 cos 120t yields the equations
4000A+24000B = 100, -24000A+4000B = 0
with solution A = 1/1480, B = 31740. The complementary function is Qc = ce-20I, and imposition of the initial condition Q(O) = 0 yields the solution
Q(t) = (cos 120t + 6 sin 120t - e-20t)11480. The current function is then
let) = Q'(t) = (36 cos 120t - 6 sin 120t + e-20t)174. Thus the steady-periodic current is
t; = ~(6cos120t-sin120t) 74
6J37( 6 1.) 3
= -- r,:;:; cos 120t - r,:;:; SIn l20t = r,:;:; cos(l20t - a)
74 ,,37 ,,37 ,,37
(with a = 27£ - tan -1 t), so the steady-state amplitude is 3 I m.
10. Substitution of the trial solution Q = A cos OJt + B sin OJt into the differential equation R Q' + (1 I C)Q = Eo cos OJt yields the equations
(UC)A+OJr B = Eo, -rOJA+(l/C)B = 0
Section 3.7
189
where f3 = tan " mRC (1st quadrant angle).
In Problems 11-16, we give first the trial solution Ip = Acoeox + Bsuuot, then the equations in A and B that we get upon substituting this trial solution into the RLC equation
L 1" + R I' + (1 I C) I = E' (t), and finally the resulting steady periodic solution.
In each of Problems 17-22, the first step is to substitute the given RLC parameters, the initial values 1(0) and Q(O), and the voltage E(t) into Eq. (16) and solve for the remaining initial value
/'(0) = ![ E(O) - R 1 (0) - (11 C) Q(O) J.
L
(*)
17. With 1(0) = 0 and Q(O) = 5, Equation (*) gives 1'(0) = -75. The solution of the RLC equation 21" + 161' + 501 = 0 with these initial conditions is I(t) = -25e-4tsin 3t.
18. Our differential equation to solve is
21" + 601' + 4001 = -1 OOe -to
We find the particular solution Ip = (-501171)e-t by substituting the trial solution A e-t; the general solution is
Section 3.7
191
The initial conditions are 1(0) = 0 and 1'(0) = 50, the latter found by substituting L = 2, R = 60, lie = 400, 1(0) = Q(O) = 0, and E(O) = 100 into Equation (*). Imposition of these initial values on the general solution above yields the equations ci + c2 - 50/171 = 0, -lOcI - 20c2 + 501171 = 50 with solution ci = 5019,
c2 = -100 I 19. This gives the solution
let) = (50/171)(1ge-IOt - 18e-20t - e-t).
19. Now our differential equation to solve is
21" + 601' + 4001 = -1000e-10t.
We find the particular solution Ip = -SOt e-10t by substituting the trial solution At e-lOt; the general solution is
The initial conditions are 1(0) = 0 and 1'(0) = -150, the latter found by substituting L = 2, R = 60, lie = 400, 1(0) = 0, Q(O) = 1, and E(O) = 100 into Equation (*). Imposition of these initial values on the general solution above yields the equations
ci + c2 = 0, -lOCI - 20c2 - 50 = -150 with solution c1 = -10, c2 = 10. Thus we
get the solution
let) = 10e-20t - lOe-IOt - 50te-IOt.
20. The differential equation 101" + 301' + 501 = 1 00 cos 2t has transient solution
and in Problem 11 we found the steady periodic solution
l,p(t) = .!.2.(cos2t+6sin2t). 37
When we impose the initial conditions 1(0) = 1'(0) = 0 on the general solution let) = Itr(t) + lsp(t), we get the equations
withsolution ci = -10/37, c2 = -270/37JU. Thefollowingfigureshowsthe graphs of let) and lsp(t).
192
Chapter 3
I
1
t
-5
21. The differential equation 101" + 201' + 1001 = -1000 sin 5t has transient solution
I tr (t) = e -I ( C1 cos 3t + c2 sin 3t),
and in Problem 13 we found the steady periodic solution
I. (t) = 20(2cos5t+3sin5t) .
. sp 13
When we impose the initial conditions 1(0) = 0, 1'(0) = -lOon the general solution I(t) = Itr(t)+ Isp(t), wegettheequations
C1 + 40/13 = 0, - c1 + 3 c2 + 300/13 = - 10
with solution c1 = -40/13, c2 = -470/39. The figure at the top of the next page shows the graphs of let) and Isp(t).
22. The differential equation 21" + 1001' + 2000001 = 66001Z'cOS 60m has transient solution
and in Problem 15 we found the steady periodic solution
(with the exact values of A and B given there). When we impose the initial conditions 1(0) = /'(0) = 0 on the general solution l(t) = Itr(t) + I,p(t), we find (with the aid of a computer algebra system) that
33JZ'(250-9;r2) 4 "" -0.157444,
250000 -17775JZ' + 324JZ'
llJZ'.Ji59(250+9JZ'2) ~ -0.026249
53(250000 -17775JZ'2 + 324JZ'4)
The following figure shows the graphs of let) and I,p(t).
I
I = Isp +I tr
0.2
t
-0.2
23. The LC equation L I" + (1/ C)I = 0 has general solution I (t) = c1 cos av + c2 sin u{l with critical frequency lUo = 11 JLC.
194
Chapter 3
24. We need only observe that the roots
-R±.JR2 -4LIC
r = -------------
2L
both necessarily have negative real parts.
25. According to Eq. (8) in the text, the amplitude of the steady periodic current is
Eo I ~ R2 + (OlL -1 I mC)2. Because the radicand in the denominator is a sum of squares, it is obvious that the denominator is least when OlL -1 I OlC = 0, that is, when
to = I/JLC.
SECTION 3.8
ENDPOINT PROBLEMS AND EIGENVALUES
The material on eigenvalues and endpoint problems in Section 3.8 can be considered optional at this point in a first course. It will not be needed until we discuss boundary value problems in the last three sections of Chapter 9 and in Chapter 10. However, after the concentration thus far on initial value problems, the inclusion of this section can give students a view of a new class of problems that have diverse and important applications (as illustrated by the subsection on the whirling string). If Section 3.8 is not covered at this point in the course, then it can be inserted just prior to Section 9.5.
1. If II, = 0 then y" = 0 implies that y(x) = A + Bx. The endpoint conditions
y'(O) = 0 and y(l) = 0 yield B = 0 and A = 0, respectively. Hence II, = 0 is not an eigenvalue.
If II, = el> 0, then the general solution of y" + ely = 0 IS
y(x) = A cos ax + B sin ax,
so
y'(x) = -Aasin ax + Bacos ax.
Then y'(O) = 0 yields B = 0, so y(x) = A cos ax. Next y(1) = 0 implies that
cos a = 0, so a is an odd multiple of 7rl2. Hence the positive eigenvalues are {(2n-1iJi'-14} withassociatedeigenfunctions {cos(2n-l)l1XI2} for n = 1,2,3,···.
2. If II, = 0 then y" = 0 implies that y(x) = A + Bx. The endpoint conditions
y'(O) = y'(7r) = 0 imply only that B = 0, so /4J = 0 is an eigenvalue with associated eigenfunction yo(x) = 1.
Section 3.7
195
If A = d> 0 , then the general solution of y" + dy = 0 IS
y(X) = A cos ax + B sin ax.
Then
y'(X) = -Aasin ax +Ba cos ax,
so y'(O) = 0 implies that B = O. Next, y'(Jr) = 0 implies that cot is an integral multiple of Jr. Hence the positive eigenvalues are {n2} with associated eigenfunctions {cos nx}, n = 1,2,3,···.
3. Much as in Problem 1 we see that A = 0 is not an eigenvalue. Suppose that A = d>o, so
y(X) = A cos ax + B sin ax.
Then the conditions y( - Jr) = y( Jr) = .0 yield
A cos aJr+ B sin arc = 0, A cos cot - B sin ax = O.
It follows that
A cos cot = 0 = B sin an:
Hence either A = 0 and B *- 0 with cot an even multiple of 7rl2, or A*-O and B = 0 with ast an odd multiple of 7rl2. Thus the eigenvalues are {n2/4} for n a positive integer, and the nth eigenfunction is ynCx) = cos(nx/2) if n is odd, Yn(x) = sin(nx/2) if n is even.
4. Just as in Problem 2, Ao = 0 is an eigenvalue with associated eigenfunction yo(x) = 1.
If A = d> 0 and
y(x) = A cos ax + B sin ax,
then the equations
y'( - Jr) = a( A sin aJr+ B cos aJr) = 0, y'( Jr) = a( -A sin an-+ B cos an-) = 0
yield A sin ast = B cos cot = O. If A = 0 and B *- 0, then cos cot = 0 so cot must be an odd multiple of 7rl2. If A*-O and B = 0, then sin ax = 0 so ax must be an even multiple of 7rl2. Therefore the positive eigenvalues are {Ii 14} with associated eigenfunctions Yn(x) = cos(nx/2) if the integer n is even, Yn(x) = sin(nx/2) if n is odd.
5. If A = d > 0 and
y(x) = A cos ax + B sin ax,
196 Chapter 3
y'(x) = -Aasin ax + Bacos ax
then the conditions y( -2) = y'(2) = 0 yield
A cos 2a - B sin 2a = 0, -A sin 2a+ B cos 2a = O.
It follows either that A = B and cos 2a = sin 2a, or that A cos 2a = -sin 2a. The former occurs if
-B and
2a = m'4, 5.7T14, 9.7T14, .. "
the latter if
2a = 3m'4, 7m'4, 1l.7T14,· . -.
Hence the nth eigenvalue is
for n = 1, 2, 3, .. " and the associated eigenfunction is
Yn(X) = cos anX + sin anX (n odd)
or
Yn(X) = cos anX - sin a.» (n even).
6. (a) If A = 0 and y(x) = A + Bx, then y'(O) = B = 0, so y(x) = A. But then y(l) + y'(l) = A = 0 also, so A = 0 is not an eigenvalue.
(b) If A = if > 0 and
y(x) = A cos ax + B sin ax,
then
y'(x) = a( -A sin ax + B cos ax),
so y'(O) = Ba = O. Hence B = 0 so y(x) = A cos ax. Then
y(l) + y'(l) = A(cos a - a sin a) = 0,
so a must be a positive root of the equation tan a = II a.
7. (a) If A = 0 and y(x) = A + Bx, then yeO) = A = 0, so y(x) = Bx. But then y(l) + y'(I) = 2B = 0, so A = B = 0 and A = 0 is not an eigenvalue.
(b) If A = if> 0 and y(x) = A cos ax + B sin ax, then yeO) = A = 0 so
y(x) = B sin ax. Hence
Section 3.8
197
y(1) + y'(l) = B(sin a+ acos a) = O.
so a must be a positive root of the equation tan a = -a, and hence the abscissa of a point of intersection of the lines y = tan z and y = - z. We see from the figure below
that an lies just to the right of the vertical line z = (2n -l)JZ' 12, and lies closer and closer to this line as n gets larger and larger.
y
10
z
-10
8. (a) If A = 0 and y(x) = A + Bx, then yeO) = A = 0, so y(x) = Bx. But then y(l) = y'(l) says only that B = B. Hence Ao = 0 is an eigenvalue with associated eigenfunction yo(x) = x.
(b) If A = /]2 > 0 and y(x) = Acos flx + Bsinflx, then yeO) = A = 0 so
y(x) = Bsinflx. Then y(l) = y'(l) says that Bsinfl = Bflcosfl, so fl must be a positive root of the equation tan fl = fl, and hence the abscissa of a point of intersection of the lines y = tan z and y = z. We see from the figure above that fln lies just to the left of the vertical line z = (2n + l)JZ' 12, and lies closer and closer to this line as n gets larger and larger.
9. If y" + AY = 0 and A = -d < 0, then
y(x) = AelXl: + Be -ox.
Then yeO) = A + B = 0, so B = -A and therefore
198
Chapter 3
Hence
y'(L) = Aa(eaL + e-aL) = 0.
But a* ° and eaL + e-aL > 0, so A = 0. Thus JL = -d is not an eigenvalue.
10. If JL = - a2 < 0, then the general solution of y" + JL y = ° is
y(x) = Acoshax + B sinh ax. Then y(o) = ° implies that A = 0, so y(x) = sinhax (or a nonzero multiple thereof). Next,
y(1) + y'(1) = sinhzr+ acosha = °
implies that tanh a = - a. But the graph of y = tanh a lies in the first and third quadrants, while the graph of y = - a lies in the second and fourth quadrants. It follows that the only solution of tanh a = - a is a= 0, and hence that our eigenvalue problem has no negative eigenvalues.
11. If JL = - a' < 0, then the general solution of y" + JL y = ° is
y(x) = Acoshax+ Bsinhax. Then y'(0) = ° implies that B = 0, so y(x) = coshax (or a nonzero multiple thereof). Next,
y(1)+y'(1) = cosh zr+zz sinhrz = °
implies that tanh a = - 1/ a. But the graph of y = tanh a lies in the first and third quadrants, while the graph of y = - 1/ a lies in the second and fourth quadrants. It follows that the only solution of tanh a = -1/ a is a= 0, and hence that our eigenvalue problem has no negative eigenvalues.
12. (a) If JL = ° and y(x) = A + Bx, then y( -1t) = y(Jr) means that
A + BJr = A - BJr, so B = ° and y(x) = A. But then y'( -1t) = y'(1t) implies nothing about A. Hence Ao = ° is an eigenvalue with Yo(x) = 1.
(b) If JL = -d < ° and
y(x) = Aeox + Be-OX,
then the conditions y( - Jr) = y(1t) and y'( -1t) = y'(1t) yield the equations
Addition of these equations yields 2AeaJl" = 2Ae -an: Since ea1! * e -ast because a* 0, it follows that A = 0. Similarly B = 0. Thus there are no negative eigenvalues.
Section 3.8
199
(c) If A = d > 0 and
y(x) = A cos ax + B sin ax,
then the endpoint conditions yield the equations
A cos ax + B sin ast = A cos an - B sin cot,
-A sin a7r+ B cos ax = A sin a7r+ B cos ax.
The first equation implies that B sin ax = 0, the second that A sin cot = O. If A and B are not both zero, then it follows that sin ast = 0, so a = n, an integer. In this case A and B are both arbitrary. Thus cos nx and sin nx are two different eigenfunctions associated with the single eigenvalue n2•
13.
(a)
With A = 1, the general solution of y" + 2y' + Y = 0 IS
y(x) = Ae-x + Bxe":
But then yeO) = A = 0 and y(1) = e-1(A + B) = O. Hence A = 1 is not an eigenvalue.
(b) If A < 1, then the equation y" + 2y' + /"'y = 0 has characteristic equation
r2 + 2r + A = O. This equation has the two distinct real roots (-2 ± J 4 - 4 A ) / 2; call
them r and s. Then the general solution is
and the conditions yeO) = y(l) = 0 yield the equations
A + B = 0, Aer + Be' = O.
If A, B * 0, then it follows that er = e', But r * s, so there is no eigenvalue A < 1.
(c) If A> 1 let A-I = d, so A = 1 + d. Then the characteristic equation
has roots -1 ± a i, so
y(x) = e-X(A cos ax + B sin ax).
Now yeO) = A = 0, so y(x) = Ae-xsin ax. Next, y(1) = Ae-1sin a = 0, so a = ntt with n an integer. Thus the nth positive eigenvalue is An = n2J?- + 1. Because A = d + 1, the eigenfunction associated with An is
200
Chapter 3
14. If A = 1 + if then we first impose the condition yeO) = 0 on the solution
y(x) = e-X(A cos ax + B sin ax)
found in Problem 13, and find that A = O. Hence
y(x) = Be-xsin ax,
y'(x) = B( - e-xsin ax + e-xacos ax),
so the condition y'(l) = 0 yields -sin a+ acos a = 0, that is, tan a = a.
15.
(a)
The endpoint conditions are
yeO) = y'(O) = y"(L) = y(3)(L) = o.
With these conditions, four successive integrations as in Example 5 yield the indicated shape function y(x).
(b) The maximum value Ymax of y(x) on the closed interval [0, L] must occur either at an interior point where y'(x) = 0 or at one of the endpoints x = 0 and x = L. Now
where k = w124EI, and the quadratic factor has no real zero. Hence x = 0 is the only zero of y'(x). But yeO) = 0, so it follows that Ymax = y(L).
16.
(a)
The endpoint conditions are
j'(D) = y'(O) = 0 and y(L) = y'(L) = O.
(b) The derivative
y'(x) = k(4x3 - 6Lx2 + 2L2x) = 2kx(2x - L)(x - L)
vanishes at x = 0, L12, L. Because yeO) = y(L) = 0, the argument of Problem 15(b) implies that Ymax = y(LI2).
17. If y(x) = kt x' - 2Lx3 + Ex) with k = w I 24EI, then
y'(x) = k(4x3 -6Lx2 +L3) = 0
has the solution x = LI2 that we can verify by inspection. Now long division of the cubic 4x3 - 6Lx2 + E by 2x - L yields the quadratic factor 2X2 - 2Lx - L2 whose
zeros (2L ± .J12L2)1 4 = (1 ± J3)LI2 both lie outside the interval [0, L]. Thus
x = Ll2 is, indeed, the only zero of y'(x) = 0 in this interval.
Section 3.8
201
18. (a)
The endpoint conditions are
yeO) = y'(O) = 0 and y(L) = y"(L) = O.
(b) The only zero of the derivative
y'(X) = 2kx(8x2 - 15Lx + 6L2)
interior to the interval [0, L] is
x.; = (15 - m )LI16,
and yeO) = y(L) = 0, so it follows by the argument of Problem 15(b) that Ymax = y(xm).