Matemáticas Avanzadas para Ingeniería Peter O'Neil 7ma Edición Sol PDF
Matemáticas Avanzadas para Ingeniería Peter O'Neil 7ma Edición Sol PDF
Matemáticas Avanzadas para Ingeniería Peter O'Neil 7ma Edición Sol PDF
www.librosysolucionarios.net
INSTRUCTOR'SSOLUTIONSMANUAL
TOACCOMPANY
ADVANCED
ENGINEERING
MATHEMATICS
SEVENTHEDITION
PETERV.O’NEIL
www.librosysolucionarios.net
Contents
iii
www.librosysolucionarios.net
iv CONTENTS
www.librosysolucionarios.net
v
8 Determinants 193
8.1 Definition of the Determinant 193
8.2 Evaluation of Determinants I 194
8.3 Evaluation of Determinants II 196
8.4 A Determinant Formula for A−1 198
8.5 Cramer’s Rule 199
8.6 The Matrix Tree Theorem 200
9 Eigenvalues and Diagonalization 203
9.1 Eigenvalues and Eigenvectors 203
9.2 Diagonalization 208
9.3 Some Special Matrices 214
10 Systems of Linear Differential Equations 223
10.1 Linear Systems 223
10.2 Solution of X = AX for Constant A 226
10.3 Solution of X = AX + G 231
10.4 Exponential Matrix Solutions 240
10.5 Applications and Illustrations of Techniques 243
10.6 Phase Portraits 253
11 Vector Differential Calculus 265
11.1 Vector Functions of One Variable 265
11.2 Velocity and Curvature 269
11.3 Vector Fields and Streamlines 273
11.4 The Gradient Field 275
11.5 Divergence and Curl 279
12 Vector Integral Calculus 283
12.1 Line Integrals 283
12.2 Green’s Theorem 285
12.3 An Extension of Green’s Theorem 289
12.4 Potential Theory 291
12.5 Surface Integrals 297
12.6 Applications of Surface Integrals 300
12.7 Lifting Green’s Theorem to R3 303
12.8 The Divergence Theorem of Gauss 304
12.9 The Integral Theorem of Stokes 306
12.10 Curvilinear Coordinates 309
www.librosysolucionarios.net
vi CONTENTS
www.librosysolucionarios.net
vii
www.librosysolucionarios.net
Chapter 1
First-Order Differential
Equations
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
4 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
14. The differential equation itself assumes that y = 0 and x = −1. Write
x dy 2y 2 + 1
= ,
y dx x+1
which separates as
1 1
dy = dx.
y(2y 2 + 1) x(x + 1)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 5
y2 1
dy = dx.
y+1 x
Write this as
1 1
y−1+ dy = dx.
1+y x
Integrate to obtain
1 2
y − y + ln |1 + y| = ln |x| + c.
2
Now use the initial condition y(3e2 ) = 2 to obtain
2 − 2 + ln(3) = ln(3) + 2 + c
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
y cos(3y) dy = 2x dx,
22. By Newton’s law of cooling the temperature function T (t) satisfies T (t) =
k(T −60), with k a constant of proportionality to be determined, and with
T (0) = 90 and T (10) = 88. This is based on the object being placed in
the environment at time zero. This differential equation is separable (as in
the text) and we solve it subject to T (0) = 90 to obtain T (t) = 60 + 30ekt .
Now
T (10) = 88 = 60 + 30e10k
gives us e10k = 14/15. Then
1 14
k= ln ≈ −6.899287(10−3 ).
10 15
Now 2
14
T (20) = 60 + 30 ≈ 86.13
15
degrees Fahrenheit. To reach 65 degrees, solve
t/10
14
65 = 60 + 30
15
to obtain
10 ln(1/6)
t= ≈ 259.7
ln(14/15)
minutes.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 7
23. Suppose the thermometer was removed from the house at time t = 0, and
let t > 0 denote the time in minutes since then. The house is kept at
70 degrees F. Let A denote the unknown outside ambient temperature,
which is assumed constant. The temperature of the thermometer at time
t is modeled by
together with the condition A(ln(2)) = e3 /2, since we must also find k.
Time is in weeks. Solve to obtain
t/ ln(2)
1
A(t) = e3
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
8 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
26. At any time t there will be A(t) = 12ekt gms, and A(4) = 9.1 requires
that e4k = 9.1/12, so
1 9.1
k = ln ≈ −0.06915805.
4 12
∗
The half-life is the time t∗ so that A(t∗ ) = 6, or ekt = 1/2. This gives
t∗ = − ln(2)/k ≈ 10.02 minutes.
27. Compute ∞
2x −(t2 +(x/t)2 )
I (x) = − e dt.
0 t
Let u = x/t to obtain
0
2
+u2 )
I (x) = 2 e−((x/u) du
∞
∞
2
+(x/u)2 )
= −2 e−(u du = −2I(x).
0
1
dI = −2 dx
I
28. (a) For water h feet deep in the cylindrical hot tub, V = 25πh, so
2
dh 5 √
25π = −0.6π 64h,
dt 16
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 9
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
r(t)
h(t)
18
h(t) - 18
18
h(t)
18
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 11
with h(0) = 36. Here h(t) is the height of the upper surface of the fluid
above the bottom of the sphere. This equation simplifies to
√
(36 h − h3/2 ) dh = −0.4 dt,
√
a separated equation with general solution h h(60 − h) = −t + k. Then
t = 0 when h = 36 gives us k = 5184. The tank runs empty when h = 0,
so t = 5184 seconds, about 86.4 minutes. This is the time it takes to drain
this spherical tank.
31. (a) Let r(t) be the radius of the exposed water surface and h(t) the depth
of the draining water at time t. Since cross sections of the cone are similar,
dh
πr2 = −kA 2gh,
dt
with h(0) = 9. From similar triangles (Figure 1.3), r/h = 4/9, so r =
(4/9)h. Substitute k = 0.6, g = 32 and A = π(1/12)2 and simplify the
resulting equation to obtain
dh
h3/2
= −27/160,
dt
with h(0) = 9. This separable equation has the general solution given
implicitly by
27
h5/2 = − t + k.
64
Since h(0) = 9, then k = 243 and the tank empties out when h = 0, so
64
t = 243 = 576
27
seconds, about 9 minutes, 36 seconds.
(b) This problem is modeled like part (a), except now the cone is inverted.
This changes the similar triangle proportionality (Figure 1.4) to
r 4
= .
9−h 9
Then r = (4/9)(9 − h). The separable differential equation becomes
(9 − h)2 27
√ dh = − ,
h 160
with h(0) = 9. This initial value problem has the solution
√ 2 27 1296
162 h − 12h3/2 + h5/2 = − t+ .
5 160 5
The tank runs dry at h = 0, which occurs when
160 1296
t= = 1536
27 5
seconds, about 25 minutes, 36 seconds.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
9
h
9
r
h
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 13
divided by
2
16 dh
π h2 = −kA 2gh
81 dt 2
yields
h − h0 (dh/dt)1
√ = = r,
h (dh/dt) 2
a known constant. We can therefore solve for h0 , the location of the hole
above the bottom of the tank.
33. Begin with the logistic equation
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
Here we assume that P (t) > 0 and a − bP (t) > 0. Write this equation as
P
ln = at + k,
a − bP
Then
P p0
= eat .
a − bP a − bp0
It is a straightforward algebraic manipulation to solve for P and obtain
ap0
P (t) = eat .
a − bp0 + bp0 eat
34. With a and b taking on the given values, and p0 = 3, 929, 214, the popula-
tion in 1790, we obtain the logistic model for the United States population
growth:
123, 141.5668
P (t) = e0.03134t .
0.03071576577 + 0.0006242342282e0.03134t
Table 1.1 shows compares the population figures given by P (t) with the
actual numbers, together with the percent error (positive if P (t) exceeds
the actual population, negative if P (t) is an underestimate).
An exponential model can also be constructed as Q(t) = Aekt . Then
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.1. TERMINOLOGY AND SEPARABLE EQUATIONS 15
1 5308483
k= ln ≈ 0.03008667012.
10 3929214
Thus the exponential model determined using these two data points (1790
and 1800) is
Q(t) = 3929214e0.03008667012t .
Population figures predicted by this model are also included in Table 1.1,
along with percentage errors. Notice that the logistic model remains quite
accurate until 1960, at which time the error increases dramatically for the
next three years. The exponential model becomes increasingly inaccurate
by 1870, after which the error rapidly becomes so large that it is not worth
computing further. Exponential models do not work well over time with
complex populations, such as fish in the ocean or countries throughout
the world.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
for x = 0.
R
2. e dx = ex is an integrating factor. Multiply the differential equation by
ex to obtain
1 2x
y ex + yex = (yex ) = e −1 .
2
Integrate to obtain
1 1
yex = e2x − x + c.
4 2
Then
1 1
y = ex − xe−x + ce−x .
4 2
R
3. e 2 dx = e2x is an integrating factor. Multiply the differential equation by
e2x to obtain
y e2x + 2y = (ye2x ) = xe2x .
Integrate to obtain
1 2x 1 2x
ye2x = xe2x dx = xe − e + c.
2 4
The general solution is
1 1
y= x − + ce−2x .
2 4
4. An integrating factor is
R
sec(x) dx
e = eln | sec(x)+tan(x)| = sec(x) + tan(x).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.2. LINEAR EQUATIONS 17
1 cos(x)
=
sec(x) + tan(x) 1 + sin(x)
to obtain
cos(x)
y = (x − cos(x) + k)
1 + sin(x)
x cos(x) − cos2 (x) + k cos(x)
= .
1 + sin(x)
R
−2 dx
5. An integrating factor is e = e−2x . Multiply the differential equation
by e−2x to obtain
Integrate to obtain
ye−2x = −8x2 e−2x dx = 4x2 e−2x + 4xe−2x + 2e−2x + c.
y = 4x2 + 4x + 2 + ce2x .
R
6. e 3 dx = e3x is an integrating factor. Multiply the differential equation by
e3x to obtain
y = e2x − 2 + ce−3x .
Now we need
y(0) = 1 − 2 + c = 2,
so c = 3. The initial value problem has the solution
y = e2x − 2 + 3e−3x .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
Integrate to obtain
(x − 2)y = x3 − 3x2 + c.
The general solution is
1
y= (x3 − 3x2 + c).
x−2
Now
y(3) = 27 − 27 + c = 4
so the initial value problem has the solution
x3 − 3x2 + 4
y= = x2 − x − 2.
x−2
(ye−x ) = 2e3x .
Integrate to obtain
2 3x
ye−x = e + c.
3
The general solution is
2 4x
y= e + cex .
3
Then
2
+c
y(0) = −3 =
3
so c = −11/3 and the initial value problem has the solution
2 4x 11 x
y= e − e
3 3
9. An integrating factor is
R 2
(2/(x+1)) dx
e = e2 ln |x+1| = eln((x+1) )
= (x + 1)2 .
Integrate to obtain
(x + 1)2 y = (x + 1)3 + c.
Then
c
y = (x + 1) + .
(x + 1)2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.2. LINEAR EQUATIONS 19
Now
y(0) = 5 = 1 + c
so c = 4 and the solution of the initial value problem is
4
y =x+1+ .
(x + 1)2
Integrate to obtain
27 41/9 9
yx5/9 = x + x23/9 + c.
41 23
Then
27 4 9
y= x + x2 + cx−5/9 .
41 23
We need
27 9
y(−1) = 4 = + − c,
41 23
so c = −2782/943. The solution is
27 4 9 2782 −5/9
y= x + x2 − x
41 23 943
11. Let (x, y) be a point on the curve. The tangent line at (x, y) must pass
through (0, 2x2 ), hence must have slope (y − 2x2 )/x. But this slope is y ,
so we have the differential equation
y − 2x2
y = .
x
This is the linear differential equation
1
y − y = −2x,
x
which has the general solution y = −2x2 + cx.
12. If A(t) is the amount of salt in the tank at time t ≥ 0, then
dA
= rate salt is added − rate salt is removed
dt
A(t)
=6−2 ,
50 + t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
20 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
The tank contains 100 gallons when t = 50 and A(50) = 176 pounds of
salt.
13. If A1 (t) and A2 (t) are the amounts of salt in tanks one and two, respec-
tively, at time t, we have
5 5A1 (t)
A1 (t) = − ; A1 (0) = 20
2 100
and
5A1 (t) 5A2 (t)
A2 (t) = − ; A2 (0) = 90.
100 150
Solve the first initial value problem to obtain
A1 (t) = 50 − 30e−t/20 .
2.5e−t/30 − 4.5e−t/20 = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.3. EXACT EQUATIONS 21
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
4.
∂M ∂N
= −2 sin(x + y) − 2x cos(x + y) =
∂y ∂x
so the equation is exact over the plane. Routine integrations yield the
potential function is ϕ(x, y) = 2x cos(x + y) and the general solution is
implicitly defined by 2x cos(x + y) = c.
ϕ(x, y) = ln |x| + xy + y 3 .
ϕ(x, y) = ln |x| + xy + y 3 = c
for x = 0.
x2
x3 + = c.
2y 2
x2 y 3 − 3xy − 3y 2 = c.
8. Compute
∂M
= 2 − 2y sec2 (xy 2 ) − 2xy 3 sec2 (xy 2 ) tan(xy 2 )
∂y
and
∂N
= 2 − 2y sec2 (xy 2 ) − 2xy 3 sec2 (xy 2 ) tan(xy 2 ).
∂x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.3. EXACT EQUATIONS 23
Since these partial derivatives are equal for all x and y for which the
functions are defined, the differential equation is exact for such x and
y. To find a potential function, we can start by integrating ∂ϕ/∂x =
2y − y 2 sec2 (xy 2 ) with respect to x to obtain
Now we need
∂ϕ
= 2x − 2xy sec2 (xy 2 )
∂y
= 2x − 2xy sec2 (xy 2 ) + α (y).
2xy − tan(xy 2 ) = c.
2(2) − tan(4) = c.
9. Since ∂M/∂y = 12y 3 = ∂N ∂x, the differential equation is exact for all x
and y. Straightforward integrations yield the potential function
ϕ(x, y) = 3xy 4 − x.
3xy 4 − x = c.
3(1)(24 ) − 1 = 47 = c.
The initial value problem has the unique solution implicitly defined by
3xy 4 − x = 47.
10. Compute
∂M 1 1 y
= ey/x − ey/x − 2 ey/x
∂y x x x
y ∂N
= − 2 ey/x = ,
x ∂x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
24 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
so the differential equation is exact for all x = 0 and all y. For a potential
function, begin with
∂ϕ
= ey/x
∂y
and integrate with respect to y to obtain
Then
∂ϕ y y
= 1 + ey/x − ey/x = ey/x − ey/x + β (x).
∂x x x
This requires that β (x) = 1 so choose β(x) = x. Then
ϕ(x, y) = xey/x + x.
xey/x + x = c.
e−5 + 1 = c.
xey/x + x = 1 + e−5 .
11. Compute
∂M ∂N
= −2x sin(2y − x) − 2 cos(2y − x) = ,
∂y ∂x
so the differential equation is exactly. For a potential function, integrate
∂ϕ
= −2x cos(2y − x)
∂y
with respect to y to get
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.3. EXACT EQUATIONS 25
−x sin(2y − x) = c.
Then we need
∂ϕ
= xey + α (y) = xey − 1.
∂y
Then α (y) = −1 and we can take α(y) = −y. Then
ϕ(x, y) = xey − y.
xey − y = c.
xey − y = 5.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
26 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
14. (a)
∂M ∂N
= 1 and = −1
∂y ∂x
so this differential equation is not exact over any rectangle in the plane.
(b) Multiply the differential equation by x−2 to obtain
yx−2 − x−1 y = 0.
−yx−1 = c.
y −1 − xy −2 y = 0.
xy −2 − x2 y −3 y = 0.
Now
∂M ∗∗∗ ∂N ∗∗∗
= −2xy −3 =
∂y ∂x
so this differential equation is exact. Integrate ∂ϕ/∂x = xy −2 with respect
to x to obtain
1
ϕ(x, y) = x2 y −2 + β(y).
2
Then
∂ϕ
= −x2 y −3 + β (y) = −x2 y −3
∂y
so choose β(y) = 0. The general solution in this case is given implicitly
by
x2 y −2 = c.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.3. EXACT EQUATIONS 27
This will be true for all x and y if we let b = 3/2, and then choose a so
that (b + 1)x = (a + 2)x, so b + 1 = a + 2. Therefore
1 3
a= and b = .
2 2
Multiply the original differential equation by µ(x, y) = x1/2 y 3/2 to obtain
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
28 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
Since x and y are independent, this equation can hold only if the coeffi-
cients of x and y are zero, giving us two equations for a and b:
Then
∂ϕ
= 3x2 y 2 − 6x3 y + β (y).
∂y
We may choose β(y) = 0, so ϕ(x, y) = x2 y 3 − 3x3 y 2 . The general solution
is implicitly defined by
x2 y 3 − 3x3 y 2 = c.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.4. HOMOGENEOUS, BERNOULLI AND RICCATI EQUATIONS 29
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
30 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
y ln |y| − x = cy.
6x3
y =4+ .
c − x3
xy − x2 − y 2 = c.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.4. HOMOGENEOUS, BERNOULLI AND RICCATI EQUATIONS 31
13. The equation is Riccati with one solution S(x) = ex . The general solution
is
2ex
y = 2x .
ce − 1
14. The equation is Bernoulli with α = 2 and general solution
2
y= .
3 + cx2
ah + bk = −c and dh + pk = −r.
This two by two system has a solution when the determinant of the coef-
ficients is nonzero: ap − bd = 0.
16. Here a = 0, b = 1, c = −3 and d = p = 1, r = −1. Solve
k = 3, h + k = 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
32 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
U ln |U | − 1 = −U ln |X| + KU,
(y − 3) ln |y − 3| − (x + 2) = K(y − 3).
dY 3X − Y
= .
dX X +Y
This homogeneous equation has general solution (in terms of x and y)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.5. ADDITIONAL APPLICATIONS 33
This has solution v(t) = 32(1 − e−t ) for 0 ≤ t ≤ 4. When the parachute
opens at t = 4, the skydiver has a velocity of v(4) = 32(1 − e−4 ) feet
per second. Velocity with the open parachute satisfies the initial value
problem
192 dv
= 192 − 3v 2 , v(4) = 32(1 − e−4 ) for t ≥ 4.
32 dt
This differential equation is separable and can be integrated using partial
fractions:
1 1
− dv = − 8t dt.
v+8 v−8
This yields
v+8 5 − 4e−4
ln = −8t + ln + 32.
v−8 3 − 4e−4
Solve for v(t) to obtain
8(1 + ke−8(t−4) )
v(t) = for t ≥ 4.
1 − ke−8(t−4)
We find using the initial condition that
3 − 4e−4
k= .
5 − 4e−4
Terminal velocity is limt→∞ v(t) = 8 feet per second. The distance fallen
is t
s(t) = v(ξ) dξ = 32(t − 1 + e−t )
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
34 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
for 0 ≤ t ≤ 4, while
2
s(t) = 32(3 + e−4 ) + 8(t − 4) + 2 ln(1 − ke−8(t−4) ) − 2 ln
5 − 4e−4
for t ≥ 4.
4. When fully submerged the buoyant force will be FB = (1)(2)(3)(62.5) =
375 pounds upward. The mass is m = 384/32 = 12 slugs. The velocity
v(t) of the sinking box satisfies
dv 1
12 = 384 − 375 − v; v(0) = 0.
dt 2
This linear problem has the solution
In t seconds the box has sunk s(t) = 18(t + 24e−t/24 − 24) feet. From v(t)
we find the terminal velocity
lim v(t) = 18
t→∞
feet per second. To answer the question about velocity when the box
reaches the bottom s = 100, we would normally solve s(t) = 100 and
substitute this t into the velocity. This would require a numerical solution,
which can be done. However, there is another approach we can also use.
Find t∗ so that v(t∗ ) = 10 feet per second, and calculate s(t∗ ) to see how
far the box has fallen. With this approach we solve 18(1 − e−t/24 ) = 10
to obtain t∗ = 24 ln(9/4) seconds. Now compute
feet. Therefore at the bottom s = 100, the box has not yet reached a
velocity of 10 feet per second.
5. If the box loses 32 pounds of material on impact with the bottom, then
m = 11 slugs. Now
dv 1
11 = −352 + 375 − v; v(0) = 0
dt 2
in which we have taken up as the positive direction. This gives us
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.5. ADDITIONAL APPLICATIONS 35
dv gr
v =− ,
dr R
with the condition v(R) = 0. Integrate to obtain
gr2
v 2 = gR − .
R
√
Put r = 0 to get the speed at the center of the Earth. This is v =
√ gR =
24 ≈ 4.9 miles per second.
7. Let θ be the angle the chord makes with the vertical. Then
dv
m = mg cos(θ); v(0) = 0.
dt
where s is the length of the chord. By the law of cosines, the length of
this chord satisfies
Therefore
R
t=2 ,
g
and this is independent of θ.
10i1 + 15(i1 − i2 ) = 10
15(i2 − i1 ) + 30i2 = 0
so
1 1
i1 = amp and i2 = amp.
2 6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
36 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
q + 2q = 32(10−5 ),
a linear equation with solution q(t) = 16(10−5 )(1 − e−2t ). The capacitor
voltage is
1
EC = q = 80(1 − e−2t ).
C
The voltage reaches 76 volts when t = (1/2) ln(20), which is approximately
1.498 seconds after the switch is closed. Calculate the current at this time
by
1
ln(20)i = q (ln(20)/2) = 32(10−5 )e− ln(20) = 16 micro amps.
2
10. The loop currents satisfy
From the upper node between loops 1 and 2, we conclude that i1 (0+) =
i2 (0+). Therefore
3
i1 (0+) = i2 (0+) = amps.
20
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.5. ADDITIONAL APPLICATIONS 37
E t/RC
(qet/RC ) = e
R
so
q(t) = EC + ke−t/RC .
so
q0 − EC
t = RC ln .
0.1EC
dy 4x
= .
dx 3
Orthogonal trajectories satisfy
dy 3
=−
dx 4x
and are given by
3
y = − ln |x| + c.
4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
38 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
y 2 (ln(y 2 ) − 1) = c − 2x2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.5. ADDITIONAL APPLICATIONS 39
or
2(A − x)y = (1 + (y )2 )1/2 ,
subject to y(0) = y (0) = 0. Let u = y to obtain the separable equation
1 1
√ du = dx.
1+u 2 2(A − x)
This has the solution
1
ln(u + 1 + u2 ) = − ln(A − x) + c.
2
Using y (0) = u(0) = 0 gives us
√
A
u+ 1 + u2 = √ ,
A−x
or, equivalently,
√
A
y + (1 + (y )2 ) =√ ; y(0) = 0.
A−x
From the equation for y , we obtain
1 + (y )2 = 2(A − x)y ,
so √
A
y + 2(A − x)y = √ ; y(0) = y (0) = 0
A−x
for x < A. Let w = y to obtain the linear first order equation
√
1 A
w + w= .
2(A − x) 2(A − x)3/2
√
An integrating factor is 1/ A − x and we can write
√
d w A
√ = .
dx A−x 2(A − x)2
The solution, subject to w(0) = 0, is
A 1 1 √ dy
w(x) = √ √ − √ A−x= .
2 A−x 2 A dx
Integrate one last time to obtain
√ √ 1 2
y(x) = − A A − x + √ (A − x)1/2 + A,
3 A 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
40 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
On the other hand, the tangent direction must be from (f (θ), θ) to (f (θ), θ+
π/2), so
f (θ) + f (θ) = 0
√
with f (0) = a/ 2. Then
a
r = f (θ) = √ e−θ
2
is the polar coordinate equation of the pursuit curve.
(b) The distance traveled by each bug is
∞
D= (r )2 + r2 dθ
0
2 2 1/2
∞
a −a
= √ e−θ + √ e−θ dθ
0 2 2
∞
=a e−θ dθ = a.
0
√
(c) Since r = f (θ) = ae−θ / 2 > 0 for all θ, no bug reaches its quarry.
The distance between pursuer and quarry is ae−θ .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.5. ADDITIONAL APPLICATIONS 41
20. (a) Assume the disk rotates counterclockwise with angular velocity ω ra-
dians per second and the bug steps on the rotating disk at point (a, 0).
By the chain rule,
dr dr dθ
= ,
dt dθ dt
so
dr v
=− .
dθ ω
Then
θv
r =c− , r(0) = a
ω
gives us
θv
r(θ) = a − .
ω
This is a spiral.
(b) To reach the center, solve r = 0 = a − θv/ω to get θ = aω/v radians,
or θ = aω/2πv revolutions.
(c) The distance traveled is
aω/v
s= r2 + (r )2 dθ
0
aω/v 2
vθ v 2
= a− + dθ.
0 ω ω
21. Let x(t) denote the length of chain hanging down from the table at time t,
and note that once the chain starts moving, all 24 feet move with velocity
v. The motion is modeled by
24ρ dv 3ρ dv
ρx = = v ,
g dt 4 dx
4 2
v2 = (x − 36).
3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
42 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
√
When the end leaves the table, x = 24 so v = 12 5 ≈ 26.84 feet per
second. The time is
24 24 √
1 3
tf = dx = √ dx
6 v(x) 6 2 x2 − 36
√
3 √
= ln(6 + 35) ≈ 2.15
2
seconds.
22. The force pulling the chain off the table is due to the four feet of chain
hanging between the table and the floor. Let x(t) denote the distance the
free end of the chain on the table has moved. The motion is modeled by
d ρ
4ρ = (22 − x) v ; v = 0 when x = 0.
dt g
Rewrite this as
dv
128 + v 2 = (22 − x)v
,
dx
a separable differential equation which we solve to get
1
ln(128 + v 2 )
c − ln |22 − x| =
2
√
Since v = 0 when x = 0, then c = ln(176 2). The end of the chain leaves
the table when x = 18, so at this time
√
v = 3744 ≈ 61.19 feet per second.
∂f 8x
= −2y − 2
∂y y
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.6. EXISTENCE AND UNIQUENESS QUESTIONS 43
4. Both f (x, y) = cos(exy ) and ∂f /∂y = −xexy sin(exy ) are continuous over
the entire plane.
5. By taking |y | = y , we get y = 2y and the initial value problem has the
solution y(x) = y0 e2(x−x0 ) . However, if we take |y | = −y , then the initial
value problem has the solution y(x) = y0 e−2(x−x0 ) .
In this problem we have |y | = 2y = f (x, y), so we actually have y = ±2y,
and f (x, y) = ±2y. This is not even a function, so the terms of Theorem
1.2 do not apply and the theorem offers no conclusion.
6. (a) Since both f (x, y) = 2−y and ∂f /∂y = −1 are continuous everywhere,
the initial value problem has a unique solution. In this case the solution
is easy to find: y = 2 − e−x . This is the answer to (b).
(c)
x
y0 = 1, y1 = 1 + dt = 1 + x,
0
x
x2
y2 = 1 + (1 − t) dt = 1 + x − ,
0 2
x 2
t x2 x3
y3 = 1 + 1−t+ dt = 1 + x − + ,
0 2 2 3!
x
t2 t3 x2 x3 x4
y4 = 1 + 1−t+ − dt = 1 + x − + − ,
0 2 3! 2 3! 4!
x
x2 x3 x4 x5
y5 = 1 + y4 (t) dt = 1 + x − + − + ,
0 2 3! 4! 5!
x 2 3 4
x x x x5 x6
y6 = 1 + y5 (t) dt = 1 + x − + − + − .
0 2 3! 4! 5! 6!
(d)
−x x2 x3 x4 xn
2−e =2− 1+x− + − + · · · + (−1)n
2 3! 4! n!
2 3 n
x x x
=1+x− + − · · · + (−1)n+1 + ···
2! 3! n!
Since
n
xk
2 − e−x = 2 − lim (−1)k = lim yn (x),
n→∞ k! n→∞
k=0
the Picard iterates converge to the unique solution of the initial value
problem.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
44 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
x2 x3 xn
yn (x) = 3 + 7x + 7 + 7 + ··· + 7 .
2 3! n!
Note that
n
xk
yn (x) = −4 + 7
k!
k=0
and that
∞
xk
lim yn (x) = −4 + 7 = −4 + 7ex .
n→∞ k!
k=0
8. (a) Both f (x, y) = 2x2 and ∂f /∂y = 0 are continuous everywhere, so the
initial value problem has a unique solution.
(b) The solution is
2 3 7
y= x + .
3 3
(c) x
2 3 7
y0 = 3, y1 = 3 + 2t2 dt = x + .
1 3 3
Because f (x, y) is independent of y, yn (x) = y1 (x) for all n.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
1.6. EXISTENCE AND UNIQUENESS QUESTIONS 45
The nth partial sum Tn of this Taylor series does not agree with the nth
Picard iterate yn (x). However,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
46 CHAPTER 1. FIRST-ORDER DIFFERENTIAL EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 2
Linear Second-Order
Equations
2. The general solution is y(x) = c1 e4x + c2 e−4x . For the initial conditions,
compute
y(0) = c1 + c2 = 12 and y (0) = 4c1 − 4c2 = 3.
Solve these algebraic equations to obtain c1 = 51/8 and c2 = 45/8. The
solution of the initial value problem is
51 4x 45 −4x
y(x) = e + e .
8 8
3. The general solution is y(x) = c1 e−2x + c2 e−x . For the initial conditions,
we have
y(0) = c1 + c2 = −3 and y (0) = −2c1 − c2 = −1.
Solve these to obtain c1 = 4, c2 = −7. The solution of the initial value
problem is
y(x) = 4e−2x − 7e−x .
47
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
48 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
11. For conclusion (1), begin with the hint to the problem to write
Multiply the first equation by y2 and the second by −y1 and add the
resulting equations to obtain
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.1. THEORY OF THE LINEAR SECOND-ORDER EQUATION 49
Since W = y1 y2 − y2 y1 , then
W = y1 y2 − y1 y2 ,
so
W + pW = y1 y2 − y1 y2 + p(y1 y2 − y1 y2 ) = 0.
Therefore the Wronskian satisfies theR linear differential equation W +
pW = 0. This has integrating factor e p(x) dx and can be written
R
W e p(x) dx = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
50 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
Then
y (x ) y2 (x0 )
W (x0 ) = 1 0 = 0.
0 0
Therefore y1 and y2 are linearly dependent.
15. Suppose ϕ (x0 ) = 0. Then ϕ is the unique solution of the initial value
problem
y + py + qy = 0; y(x0 ) = y (x0 ) = 0
on I. But the functions that is identically zero on I is also a solution of
this problem. Therefore ϕ(x) = 0 for all x in I.
y = c1 ex cos(3x) + c2 ex sin(3x).
y = c1 e−3x + c2 xe−3x .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.2. THE CONSTANT COEFFICIENT CASE 51
y(x) = c1 ex + c2 e−3x .
Now we need
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
52 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
14. y = e2x (3 − x)
15.
1 3(x−2)
y= [9e + 5e−4(x−2) ]
7
16. √
6 x
√6x √
y= e e − e− 6x
4
18. √
√ 5(x−2)/7 23
y = −4(5 − 23)e sin (x − 2)
2
19. √ √
(x+2)/2 15 5 15
y=e cos (x + 2) + √ sin (x + 2)
2 15 2
20. √ √
y = ae(−1+ 5)x/2
+ be(−1− 5)x/2
,
where √
(9 + 7 5) −2+√5
a= √ e
2 5
and √
(7 5 − 9) −2−√5
b= √ e
2 5
(c) In general,
lim y (x) = eαx (c1 + c2 ) = y(x).
→0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.2. THE CONSTANT COEFFICIENT CASE 53
(b) We obtain
1 αx
y = ψ (x) = e (d − ac + c)ex + (ac − d + c)e−x .
2
lim ψ (x) =
→0
1 αx
e lim (d − ac + c)xex − (ac − d + c)xe−x + ce( x + ce−x )
2 →0
= eαx (c + (d − ac)x) = ψ(x).
1 1
λ1 = (−a + a2 − 4b), λ2 = (−a − a2 − 4b).
2 2
y = e−ax/2 (c1 + c2 x) → 0 as x → ∞,
because a > 0.
If a2 > 4b, then a2 − 4b < a2 and λ1 and λ2 are both negative, so
y = c1 eλ1 x + c2 eλ2 x → 0 as x → ∞.
Finally, if a2 < 4b, then the general solution has the form
24. We will use the fact that, for any positive integer n,
Now suppose a is real and split the exponential series into two series, one
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
54 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
for even values of the summation index, and the other for odd values:
∞
1 n n
eia = i a
n=0
n!
∞
∞
1 2n 2n 1
= i a + i2n+1 a2n+1
n=0
(2n)! n=0
(2n + 1)!
∞ ∞
(−1)n 2n (−1)n 2n+1
= a + ia
n=0
2n! n=0
n!
∞ ∞
(−1)n 2n (−1)n 2n+1
= a +i a
n=0
n! n=0
(2n + 1)!
= cos(a) + i sin(a).
and
y1 (x)f (x)
u2 (x) = dx = cos(x) tan(x) dx
W (x)
= sin(x) dx = − cos(x).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.3. THE NONHOMOGENEOUS EQUATION 55
y(x) = c1 e3x + c2 ex
3 1
− cos(x + 3) + sin(x + 3)
10 10
1 1
+ cos(x + 3) − sin(x + 3).
2 2
This can be written
y(x) = c1 e3x + c2 ex
1 2
+ cos(x + 3) − sin(x + 3).
5 5
For Problems 3 through 6 we will omit some of the details and give an outline
of the solution.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
56 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
4. y1 = e3x and y2 = e−x , with Wronskian −4e−2x . With f (x) = 2 sin2 (x) =
1 − cos(2x), obtain u1 and u2 to write the general solution
1 7 4
y(x) = c1 e3x + c2 e−x − + cos(2x) + sin(2x).
3 65 65
yp (x) = Ax2 + Bx + C.
Then
2A − B − 2C = 5,
−2A − 2B = 0,
−2A = 2.
y = c1 e2x + c2 e−x − x2 + x − 4.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.3. THE NONHOMOGENEOUS EQUATION 57
10. y1 = e2x cos(x) and y2 = e2x sin(x). With f (x) = 21e2x , try yp (x) = Ae2x
to obtain
y = e2x [c1 cos(x) + c2 sin(x)] + 21e2x .
11. y1 = e2x and y2 = e4x . With f (x) = 3ex , try yp (x) = Aex , noting that
ex is not a solution of the associated homogeneous equation. Obtain the
general solution
y = c1 e2x + c2 e4x + ex .
1
y = e−3x [c1 + c2 x] + sin(3x).
2
Although the general solution does not contain a cos(3x) term, this does
not automatically follow and in general both the sine and cosine term
must be included in our attempt at yp (x).
14. y1 = 1 and y2 = e−4x . With f (x) = 8x2 + 2e3x , try yp (x) = Ax2 + Bx +
C + De3x , since e3x is not a solution of the homogeneous equation. This
gives us the general solution
2 1 1 2
y = c1 + c2 e−4x − x3 − x2 − x − e3x .
3 2 4 3
15. y1 = e2x cos(3x) and y2 = e2x sin(3x). Since neither e2x nor e3x is a
solution of the homogeneous equation, try yp (x) = Ae2x + Be3x to obtain
the general solution
1 1
y = e2x [c1 cos(3x) + c2 sin(3x)] + e2x − e3x .
3 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
58 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
In Problems 17 through 24, we first find the general solution of the differential
equation, then solve for the constants to satisfy the initial conditions. Problems
17 through 22 are well suited to the method of undetermined coefficients, while
Problems 23 and 24 can be solved fairly directly by variation of parameters.
7 1
y = c1 e2x + c2 e−2x − xe2x − x.
4 4
Now
7
y(0) = c1 + c2 = 1 and y (0) = 2c1 − 2c2 −
= 3.
4
Then c1 = 7/4 and c2 = −3/4. The solution of the initial value problem
is
7 3 7 1
y = − e2x − e−2x − xe2x − x.
4 4 4 4
18. Two independent solutions of the homogeneous equation are y1 = 1 and
y2 = e−4x . For a particular solution we might try A + B cos(x) + C sin(x),
but A is a solution of the homogeneous equation, so try yp (x) = Ax +
B cos(x) + C sin(x). The general solution is
Now
y(0) = c1 + c2 − 2 = 3 and y (0) = −4c2 + 8 + 2 = 2.
These lead to the solution of the initial value problem:
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.3. THE NONHOMOGENEOUS EQUATION 59
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
60 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
0
0 2 4 6 8 10
t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.4. SPRING MOTION 61
0
0 1 2 3 4 5
t
5. The solution is
A
y(t) = √ e−2t sinh( (2)t)
2
and is graphed for A = 1, 3, 6, 10, −4 and −7 in Figure 2.5.
6. The solution is y(t) = Ae−2t (1 + 2t) and is graphed for A = 1, 3, 6,
10, −4, −7 in Figure 2.6.
7. The solution is y(t) = Ate−2t , graphed for A = 1, 3, 6, 10, −4 and −7 in
Figure 2.7.
8. The solution is
A −t
e [2 cos(2t) + sin(2t)],
y(t) =
2
graphed in Figure 2.8 for A = 1, 3, 6, 10, −4 and −7.
9. The solution is
A −t
e sin(2t)
y(t) =
2
and is graphed for A = 1, 3, 6, 10, −4 and −7 in Figure 2.9.
10. From Newton’s second law of motion,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
62 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
0
0 1 2 3 4 5 6
t
-1
0
0 0.5 1 1.5 2 2.5 3
t
-4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.4. SPRING MOTION 63
1.5
0.5
0
0 1 2 3 4 5 6 7
t
-0.5
-1
0
0 0.5 1 1.5 2 2.5 3 3.5
t
-4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
64 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
1.5
0.5
0
0 1 2 3 4
t
-0.5
-1
0
0 1 2 3 4
t
-4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.4. SPRING MOTION 65
0
0 1 2 3 4
t
-1
A + B = y0 and − α(A + B) + βB = v0 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
66 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
2.5
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1 1.2
t
with α > 0 and A and B determined by the initial conditions. From the
linear factor, the bob can pass through the equilibrium at most once, and
will do this for some t > 0 if and only if B = 0 and AB < 0. Now note
that y0 = A and v0 = y (0) = −αA + B. Thus to ensure that the bob
never passes through equilibrium we need AB > 0, which becomes the
condition (v0 + αy0 )y0 > 0. No condition on y0 alone can ensure this.
We would also need to specify v0 > −αy0 , and this will ensure that the
critically damped bob never passes through the equilibrium point.
13. For underdamped motion, the solution has the appearance
y(t) = e−ct/2m [c1 cos( 4km − c2 t/2m) + c2 sin( 4km − c2 t/2m)]
having frequency √
4km − c2
ω= .
2m
Thus increasing c decreases the frequency of the the motion, and decreas-
ing c increases the frequency.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.4. SPRING MOTION 67
y(0)
y(t∗ ) = (2m + c)e−2m/(2m+c) .
c
y + 6y + 2y = 4 cos(3t)
is
√ √
y(t) = e−3t [c1 cosh( 7t) + c2 sinh( 7t)]
28 72
− cos(3t) + sin(3t).
373 373
2266 6582
c1 = and c2 = √ .
373 373 7
Now the solution is
1 −3t √ 6582 √
ya (t) = [e [2266 cosh( 7t)+ √ sinh( 7t)]−28 cos(3t)+72 sin(3t)].
373 7
1 −3t √ 2106
√
yb (t) = [e [29 cosh( 7t) + √
7
sinh( 7t)] − 28 cos(3t) + 72 sin(3t)].
373
y + 4y + 4y = 4 cos(3t)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
68 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
0
0 4 8 12 16
t
0
0 1 2 3 4 5
t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.5. EULER’S EQUATION 69
is
20 48
y(t) = e−2t [c1 + c2 t] − cos(3t) + sin(3t).
169 169
(a) The initial conditions y(0) = 6, y (0) = 0 give us the unique solution
1 −2t
ya (t) = [e [1034 + 1924t] − 20 cos(3t) + 48 sin(3t)].
169
(b) The initial conditions y(0) = 0, y (0) = 6 give us the unique solution
1 −2t
yb (t) = [e [20 + 910t] − 20 cos(3t) + 48 sin(3t)].
169
These solutions are graphed in Figure 2.12.
17. The general solution of the underdamped problem
y (t) + y + 3y = 4 cos(3t)
is
√ √
−t/2 11t 11t 24 12
y(t) = e c1 cos + c2 sin − cos(3t) + sin(3t).
2 2 45 45
(a) The initial conditions y(0) = 6, y (0) = 0 yield the unique solution
√ √
1 −t/2 11t 74 11t
ya (t) = e 98 cos + √ sin − 8 cos(3t) + 4 sin(3t) .
15 2 11 2
(b) The initial conditions y(0) = 0, y (0) = 6 yield the unique solution
√ √
1 −t/2 11t 164 11t
yb (t) = e 8 cos + √ sin − 8 cos(3t) + 4 sin(3t) .
15 2 11 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
70 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
0
0 2 4 6 8 10 12
t
-2
Y + 2Y + Y = 0,
Then
1
y(x) = c1 x−1 + c2 x−1 ln(x) = (c1 + c2 ln(x)).
x
3. Solve
Y + 4Y = 0
to obtain
Y (t) = c1 cos(2t) + c2 sin(2t).
Then
y(x) = c1 cos(2 ln(x)) + c2 sin(2 ln(x)).
4. y(x) = c1 x2 + c2 x−2
5. y(x) = c1 x4 + c2 x−4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
2.5. EULER’S EQUATION 71
y(x) = c1 x3 + c2 x−7 .
We need
Solve for c1 and c2 to obtain the solution of the initial value problem
7 x 3 3 x −7
y(x) = +
10 2 10 2
y(x) = −3 + 2x2
λ2 + (a − 1)λ + b = 0,
r2 + (a − 1)r + b = 0.
This equation for r is the same as the quadratic equation for λ, so its roots
are r1 = λ1 and r2 = λ2 . Therefore both the transformation method,
and direct substitution of y = xr into Euler’s equation, lead to the same
solutions.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
72 CHAPTER 2. LINEAR SECOND-ORDER EQUATIONS
x2 y + xy + y = 0
for x < 0. We know that, for x > 0, this Euler equation transforms to
Y + Y = 0,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 3
3(s2 − 4)
F (s) =
(s2 + 4)2
3. From entries (2) and (6) and the linearity of the transform,
14 7
H(s) = 2
− 2
s s + 49
5. From entries (4) and (6) and the linearity of the transform,
−10 3
K(s) = + 2
(s + 4)3 s +9
7
6. From entry (12) of the table, r(t) = 3 sinh(3t).
7. From (7) of the table, q(t) = cos(8t).
8. From entries (6) and (7),
√ √
g(t) = √5 sin( 12t) − 4 cos( 8t).
12
73
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
74 CHAPTER 3. THE LAPLACE TRANSFORM
For each R, let N be the largest integer so that (N + 1)T ≤ R and use
the additivity of the integral to write
R N
(n+1)T R
e−st f (t) dt = e−st f (t) dt + e−st f (t) dt.
0 n=0 nT (N +1)T
since f (u + nT ) = f (u).
13. By the results of Problems 11 and 12,
∞ (n+1)T
L[f ](s) = e−st f (t) dt
n=0 nT
∞ T
= e−snT e−st f (t) dt
n=0 0
∞
T
−snT
= e e−st f (t) dt,
n=0 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.1. DEFINITION AND NOTATION 75
5
= .
s(1 + e−3s )
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
76 CHAPTER 3. THE LAPLACE TRANSFORM
Now
25 10
5 −5s
e−st f (t) dt = 5e−st dt = e (1 − e−5s ).
0 5 s
Then
5e−5s (1 − e−5s )
L[f ](s) = .
s(1 − e−25s )
to obtain
1 1 − 6se−6s − e−6s
L[f ](s) = .
3s2 1 − e−6s
Compute
2π/ω π/ω
f (t)e−st dt = E sin(ωt)e−st dt
0 0
Eω
= (1 + e−πs/ω ).
s2 + ω2
Then
Eω 1 + e−πs/ω
L[f ](s) = 2
s + ω 2 1 − e−2πs/ω
Eω 1
= 2 2
.
s + ω 1 − e−πs/ω
20.
3t/2 for 0 < t < 2,
f (t) =
0 for 2 ≤ 2 ≤ 8.
Here T = 8 and
8
1
L[f ](s) = e−st f (t) dt
1 − e−8s 0
3 1 − 2se−2s − e−2s
= 2 .
2s 1 − e−8s
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.2. SOLUTION OF INITIAL VALUE PROBLEMS 77
21. We have
h for 0 < t ≤ a,
f (t) =
0 for a < t ≤ 2a
and T = 2a. Now
2a a
h
e−st f (t) dt = he−st dt = (1 − e−as ).
0 0 s
Then
h 1 − e−as h 1
L[f ](s) = = .
s 1 − e−2as s 1 + e−as
22. T = 2a and
ht/a for 0 ≤ t < a,
f (t) = h
− a (t − 2a) for a < t ≤ 2a.
Now
2a
h a −st h 2a
e−st f (t) dt = te dt − (t − 2a)e−st dt
0 a 0 a a
h
= 2 (1 − e−as )2 .
as
Then
h (1 − e−as )2
L[f ](s) =
as2 1 − e−2as
h 1 − e−as
= 2 .
as 1 + e−as
This can also be written in terms of the hyperbolic tangent function:
h as
L[f ](s) = 2 tanh .
as 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
78 CHAPTER 3. THE LAPLACE TRANSFORM
or
1 1 −3s + 1
Y (s) = −3 = .
s+4 s s(s + 4)
Use a partial fractions decomposition to write this as
13 1 1 1
Y (s) = − + .
4 s+4 4 s
5
sY (s) − y(0) − 9Y (s) = .
s
Substitute y(0) = 5 and solve for Y to obtain
1 1
Y (s) =
s − 9 s2 + 5
406 1 1 1 1 1
= − − ,
81 s − 9 9 s2 81 s
For the last part of this equation we have again used a partial fractions
decomposition. Finally, take the inverse transform of Y (s) to obtain the
solution
406 9t 1 1
y(t) = e − t−
81 9 81
3. Take the transform of the differential equation, insert the initial condition
and solve for Y to obtain
1 s
Y (s) =
s + 4 s2 + 1
4 1 1 4s + 1
=− + .
17 s + 4 17 s2 + 1
The solution is
4 −4t 4 1
y(t) = − e + cos(t) + sin(t)
17 17 17
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.2. SOLUTION OF INITIAL VALUE PROBLEMS 79
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
80 CHAPTER 3. THE LAPLACE TRANSFORM
The solution is
13 2t 22 2t 3 4
y(t) = − te + e + cos(t) − sin(t)
5 25 25 25
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.3. SHIFTING AND THE HEAVISIDE FUNCTION 81
12. Begin with the definition of the Laplace transform, applied to f (t) and
integrate by parts, then use the fact that
to obtain:
∞
L[f (t)](s) = e−st f (t) dt
0
∞
∞
= e−st f (t) 0 − −se−st f (t) dt
0
= −f (0) + sL[f (t)](s)
= −f (0) + s(sF (s) − f (0))
= s2 F (s) − sf (0) − f (0).
[f (t)]t→t−a .
L[f (t)]s→s−a .
1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
82 CHAPTER 3. THE LAPLACE TRANSFORM
3. Write
Then
1 s 1
L[f (t)](s) = (1 − e−7s ) + 2 cos(7)e−7s − 2 sin(7)e−7s
s s +1 s +1
4.
1 s
L[f ](s) = −
s2 s→s+4 s2 + 1 s→s+4
1 s+4
= − .
(s + 4)2 (s + 4)2 + 1
5. Write
Then
1 11 4
L[f (t)](s) = 2
− e−3s − 2 e−3s .
s s s
6. Write
f (t) = [2(t − π) + 2π + sin(t − π)][1 − H(t − π)],
to obtain
2 1 2 2π −πs 1
L[f (t)](s) = 2
− 2 − 2 e−πs − e − 2 e−πs .
s s +1 s s s +1
Then
2 17 17 8
L[f (t)](s) = − + 2 + 3 e−2s .
s3 s s s
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.3. SHIFTING AND THE HEAVISIDE FUNCTION 83
9. First write
to obtain
s 2 s 1
L[f ](s) = + − − e−2πs
s2 + 1 s s2 + 1 s2 + 1
10. Write
so
4 4 e−3 −3s
L[f (t)](s) = − + e−s + e .
s s s+1
11. Since
s2 − 9
L[t cos(3t)](s) = ,
(s2 + 9)2
we obtain the transform of te−t cos(3t) by replacing s with s + 1:
(s + 1)2 − 9
L[f ](s) = .
((s + 1)2 + 9)2
1 s
− ,
s s2 − 1
so the transform of et (1 − cosh(t)) is obtained by replacing s with s − 1 to
obtain
1 s−1
L[f ](s) = −
s − 1 (s − 1)2 − 1
1 s−1
= − .
s − 1 s(s − 2)
Then
1 2 1 1
F (s) = − + − e−16s
s2 s s s2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
84 CHAPTER 3. THE LAPLACE TRANSFORM
14. Write
f (t) = (1 − cos(2t)) − (1 − cos(2t))H(t − 3π).
Then
1 s
F (s) = ( − 2 )(1 − e−3πs ).
s s +4
15. Replace s with s + 5 in the transform of t4 + 2t2 + t to obtain
24 4 1
F (s) = + + .
(s + 5)5 (s + 5)3 (s + 5)2
16. Write
1
F (s) = .
(s + 2)2 + 8
√ √
This is the transform of (1/2 2) sin(2 2t) with s replaced by s+2. There-
fore
1 √
f (t) = √ e−2t sin(2 2t).
2 2
17. Write
1
F (s) = ,
(s − 2)2 + 1
which we recognize as the transform of sin(t) with s replaced by s − 2.
Therefore
f (t) = e−2t sin(t).
18.
−1 e−5s −1 1
L (t) = L
s3 s3 t→t−5
1
= (t − 5)2 H(t − 5).
2
21. Since
1
F (s) = ,
(s + 3)2 − 2
then
1 √
f (t) = √ sinh( 2t)e−3t .
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.3. SHIFTING AND THE HEAVISIDE FUNCTION 85
22. Since
s−4
F (s) = ,
(s − 4)2 − 6
then √
f (t) = e4t cosh( 6t).
23. Write
(s + 3) − 1
F (s) =
(s + 3)2 − 8
to obtain
√ 1 √
f (t) = e−3t cosh(2 2t) − √ e−3t sinh(2 2t).
2 2
24. Put a = 1 and F (s) = 1/(s − 5) in the second shifting theorem. Then
f (t) = e5t in this formula, yielding
−1 1 −s
L e (t) = e5(t−1) H(t − 1).
s−5
25. Write
1 1 1 1 s
= −
s(s2 + 16) 16 s 16 s2 + 16
to obtain
1
f (t) = (1 − cos(4(t − 21)))H(t − 21).
16
26. By the first shifting theorem
t
L e−2t e2w cos(3w) dw = F (s + 2),
0
where
t
F (s) = L e2w cos(3w) dw .
0
Now t
d
e2w cos(3w) dw = e2t cos(3t).
dt 0
By the operational rule for the Laplace transform, applied in the case of
a first derivative, we can write
t
d
L e cos(3w) dw = L e2t cos(3t)
2w
dt 0
t
= sL e2w cos(3w) dw = sF (s).
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
86 CHAPTER 3. THE LAPLACE TRANSFORM
Therefore
1 2t 1 s−2
F (s) = L e cos(3t) = .
s s (s − 2)2 + 9
Therefore
t
−2t 2w s
L e e cos(3w) dw = .
0 (s + 2)(s2 + 9)
transforms to
3 −4s
(s2 + 4)Y (s) =
e + s.
s
Then
3 1 s s
Y (s) = − e−4s + 2 .
4 s s2 + 4 s +4
Inverting this gives the solution
3
y(t) = cos(2t) + (1 − cos(2(t − 4)))H(t − 4).
4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.3. SHIFTING AND THE HEAVISIDE FUNCTION 87
1 1 9 1 43 1
Y (s) = + − −
4s 4s2 4s−2 4 s − 2)2
1 1 1 1
+ − + e−3s .
2s 2 s − 2 (s − 2)2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
88 CHAPTER 3. THE LAPLACE TRANSFORM
10
0
0 2 4 6 8 10 12
t
33. Assume that the switch is held in position for B seconds, then switched
to position A and left there. The charge q on the capacitor is modeled by
the initial value problem
250, 000q + 106 q = 10H(t − 5); q(0) = C, E(0) = 5(10−6 ).
Transform this problem and solve for Q(s) to obtain
5(10−6 ) −6 1 1
Q(s) = + 10 − e−5s .
s+4 s s+4
Invert this to obtain
q(t)
Eout = = 106 q(t) = 5e−4t + 10(1 − e−4(t−5) )H(t − 5).
C
This output function is graphed in Figure 3.1.
34. The current is modeled by the initial value problem
Li + Ri = 2H(t − 5); i(0) = 0.
Transform this problem and solve for I(s) to obtain
2 2 1 1
I(s) = e−5s = − e−5s .
s(Ls + 4) R s s + R/L
Invert this to obtain the current function
2
i(t) = (1 − e−R(t−5)/L )H(t − 5).
R
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.3. SHIFTING AND THE HEAVISIDE FUNCTION 89
t
0 2 4 6 8 10 12 14
0
-1E8
-2E8
-3E8
-4E8
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
90 CHAPTER 3. THE LAPLACE TRANSFORM
3.4 Convolution
1. Let
1 1
F (s) = and G(s) = 2 .
s2 + 4 s −4
Then
1 1
L−1 [F (s)] = sin(2t) and L−1 [G(s)] = sinh(2t).
2 2
−1 1 1
L
s2 + 4 s2 − 4
1
= sin(2t) ∗ sinh(2t)
4
1 t
= sin(2(t − τ )) sinh(2τ ) dτ
4 0
1 t
= [sin(2(t − τ )) cosh(2τ ) + cos(2(t − τ )) sinh(2τ )]0
16
1
= [sinh(2t) − sin(2t)].
16
2. Choose
s e−2s
F (s) = and G(s) = .
s2 + 16 s
Then
e−2s 1
L−1 = sin(4(t − 2))H(t − 2).
s2 + 16 4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.4. CONVOLUTION 91
If b2 = a2 ,
s 1 sin(at)
L−1 = cos(at) ∗
(s2 + a2 ) (s2 + a2 ) a
1 t
= cos(a(t − τ )) sin(aτ ) dτ
a 0
t
1
= (sin(at) + sin(2aτ − at)) dτ
2a 0
t
1 cos(a(2τ − t))
= τ sin(at) −
2a 2a 0
t sin(at)
= .
2a
4. First write
s 1 s
= .
(s − 3)(s2 + 5) s−3 s2 + 5
Then
s √
L−1 2
= e3t ∗ cos( 5t)
(s − 3)(s + 5)
t √
= cos( 5τ )e3(t−τ ) dτ
0
t √
= e3t cos( 5τ )e−3τ dτ
0
t
e−3τ √ √ √
= e3t − 3 cos( 5τ ) + 5 sin( 5τ )
14 0
√
3 3t 3 √ 5 √
= e − cos( 5t) + sin( 5t).
14 14 14
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
92 CHAPTER 3. THE LAPLACE TRANSFORM
Then
1 1
L−1 = 2 [1 − cos(at)] ∗ sin(at)
s(s2 + a2 )2 a
t
1
= 3 [1 − cos(a(t − τ ))] sin(aτ ) dτ
a 0
t
1 cos(aτ ) τ sin(at) cos(2aτ − at)
= 3 − − +
a a 2 4a 0
1 t
= 4 [1 − cos(at)] − 3 sin(at).
a 2a
6.
1 1 t3
L−1 4
= e5t ∗
s (s − 5) 6
t
1 t 5(t−τ ) 3 1
= e τ dτ = e5t τ 3 e−5τ dτ
6 0 6 0
1 5t 1 1 1 1
= e − t3 − t2 − t− .
625 30 50 125 625
7.
1 e−4s
L−1 = e−2t ∗ H(t − 4)
(s + 2) s
t
1
= e−2(t−τ ) = (1 − e−2(t−4) )
4 2
8.
√
−1 2 1 2 sin( 5t)
L =t ∗ √
s2 (s2 + 5) 5
t √
1
=√ τ 2 sin( 5(t − τ )) dτ
5 0
1 2 2 √
= t− + cos( 5t).
5 25 25
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.4. CONVOLUTION 93
F (s) s
Y (s) = +
(s + 6)(s + 4) (s + 6)(s + 4)
1 1 1 3 2
= F (s) − + − .
2 s+4 s+6 s+6 s+4
Then
1 −4t 1
y(t) = e ∗ f (t) − e−6t ∗ f (t) + 3e−6t − 2e−4t .
2 2
In Problems 11 - 16, we give the solution of the initial value problem without
the details of taking the transform of the differential equation.
11. We obtain
1 6t 1
y(t) = e ∗ f (t) − e2t ∗ f (t) + 2e6t − 5e2t .
4 4
12.
1 5t 1 1 3
y(t) = e ∗ f (t) − e−t ∗ f (t) + e5t + e−t
6 6 2 2
13.
1 1
y(t) = sin(3t) ∗ f (t) − cos(3t) + sin(3t)
3 3
14.
1 4
y(t) = sinh(kt) ∗ f (t) − 2 cosh(kt) − sinh(kt)
k k
15.
1 2t 1 1 1 1 4
y(t) = e ∗ f (t) + e−2t ∗ f (t) − et ∗ f (t) − e2t − e−2t + et
4 12 3 4 12 3
16.
√ √
1 3t 1 −3t 2 √2t 2 −√2t
y(t) = e ∗ f (t) − e ∗ f (t) − e ∗ f (t) − e ∗ f (t)
42 42 28 28
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
94 CHAPTER 3. THE LAPLACE TRANSFORM
1 −2t 3
f (t) = e − .
2 2
18. The equation is f (t) = −t + f (t) ∗ sin(t). Take the transform to obtain
(s2 + 1) 1 1
F (s) = − = − 2 − 4.
s4 s s
Then
1
f (t) = −t − t3 .
6
19. The equation is f (t) = e−t + f (t) ∗ 1. Transform this and solve for F (s)
to obtain
s 1 1 1
F (s) = = + .
(s + 1)(s − 1) 2 s+1 s−1
Now invert to obtain
1 −t 1 t
f (t) = e + e = cosh(t).
2 2
20. Write the equation as f (t) = −1 + t − 2f (t) ∗ sin(t). Take the transform
and solve for F (s) to obtain
(1 − s)(s2 + 1)
F (s) =
s2 (s2 + 3)
1 1 2 s 2 1
= 2− − + .
3s 3s 3 s2 + 3 3 s2 + 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.4. CONVOLUTION 95
3(s2 + 4) 3 3
F (s) = = + 2 .
s(s2 − s + 4) s s −s+4
Transform this and solve the resulting equation for F (s) to obtain
s(s + 2)
F (s) =
(s + 1)(s2 + 1)
1 3 s 1 1
=− + + .
2(s + 1) 2 s2 + 1 2 s2 + 1
23. We want r(t) if f (t) = A = constant and m(t) = e−kt . Begin with
A 1
F (s) = and M (s) = .
s s+k
Then
A A
s − s+k Ak
R(s) = s = .
s+k s2
Therefore r(t) = Akt. This function has a straight line graph, shown in
Figure 3.3 for A = 3, k = 1/5.
A B 1
F (s) = + 2 and M (t) = .
s s s+k
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
96 CHAPTER 3. THE LAPLACE TRANSFORM
0
0 2 4 6 8 10
t
Then
A B A
s + s2 − s+k
R(s) = s
s+k
Ak + B
= .
s2 + Bk
s3
Then
1
r(t) = (Ak + B)t + Bkt2 .
2
This is graphed in Figure 3.4 for A = 2, B = 1, k = 1/5.
25. Now f (t) = A + Bt + Ct2 and m(t) = e−kt , so
A B 2C 1
F (s) = + 2 + 3 and M (s) = .
s s s s+k
Then, by a routine algebraic calculation,
A B 2C 1
s + s2 + s3 − A s+k
R(s) = s
s+k
Ak + B 2C + Bk 2Ck
= + + 4 .
s2 s3 s
Then
1 1
r(t) = (Ak + B)t + Bk + C t2 + Ckt3 .
2 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.4. CONVOLUTION 97
20
15
10
0
0 2 4 6 8 10
t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
98 CHAPTER 3. THE LAPLACE TRANSFORM
400
300
200
100
0
0 2 4 6 8 10
t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.5. IMPULSES AND THE DIRAC DELTA FUNCTION 99
Then
1 1 1 1
Y (s) = 3 − e−2s − 4 − e−5s .
s+2 s+3 s+2 s+3
2.
4 2(t−3)
y(t) = e sin(3(t − 3))H(t − 3)
3
3.
y(t) = 6(e−2t − e−t + te−t )
4.
y(t) = 3 cos(4t) + 3 sin(4(t − 5π/8))H(t − 5π/8)
5.
ϕ(t) = (B + 9)e−2t − (B + 6)e−3t ; ϕ(0) = 3, ϕ (0) = B
The Dirac delta function δ(t−t0 ) applied at time t0 imparts a unit velocity
to the unit mass.
6. The motion is modeled by the initial value problem
my + ky = 0; y(0) = 0, y (0) = v0 .
We find that
mv0
Y (s) = ,
ms2 + k
so
m k
y(t) = v0 sin t .
k m
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
100 CHAPTER 3. THE LAPLACE TRANSFORM
1 1
y + 9y = δ(t); y(0) = y (0) = 0.
16 4
Transform to obtain
4
Y (s) = ,
s2 + 144
so
1
y(t) = sin(12t).
3
The initial velocity is y (0) = 4 feet per second. The frequency is 6/π
hertz and the amplitude is 1/3 feet, or 4 inches.
By the mean value theorem for integrals, there is some t between a and
a + such that
a+
f (t) dt = f (t ).
a
Then ∞
f (t)δ (t − a) dt = f (t ).
0
1
,
sX − 2sY =
s
sX − X + Y = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.6. SOLUTION OF SYSTEMS 101
x(t) = −t − 2 + 2et/2 ,
y(t) = −t − 1 + et/2 .
The solution is
1 4 4 2 2
x(t) = t + − e3t/4 and y(t) = − + e3t/4 .
3 9 9 3 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
102 CHAPTER 3. THE LAPLACE TRANSFORM
Then
2s2 1 1 1 s 1
X(s) = = − + ,
(s2 + 4)(s2 − 3s) 2 s − 1 2 s2 + 4 s2 + 4
−s2 1 1 1 s 1 1
Y (s) = 2 =− + − .
(s + 4)(s2 − 3s) 4 s − 1 4 s2 + 4 2 s2 + 4
The solution is
1 t 1 1
x(t) = e − cos(2t) + sin(2t),
2 2 2
1 t 1 1
y(t) = − e + cos(2t) − sin(2t).
4 4 4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.6. SOLUTION OF SYSTEMS 103
The solution is
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
104 CHAPTER 3. THE LAPLACE TRANSFORM
6
(s − 1)X + 2sY = 0 and 4sX + (3s + 1)Y = − .
s
Solve for X and Y to get
−12 6 30
X(s) = =− + ,
5s2 + 2s s 5s + 2
6(s − 1) 21 1 3 105 1
Y (s) = 2
=− + 2+ .
s(5s + 2s) 2 s s 2 5s + 2
21 21
x(t) = −6 + 6e−2t/5 , y(t) = − + 3t + e−2t/5 .
2 2
1 + s − s2 1 1 1 1 1 1
Y1 (s) = =− 2 − + + ,
s2 (s2 − 1) s s 2s−1 2s+1
s+1 1 1 1 1
Y2 (s) = − 3 = − 2 − ,
2s 2s 2 s3
2
−2s + 1 1 1 1 1 1 1
Y3 (s) = 2 2 =− 2 − + .
3s (s − 1) 3s 6s−1 6s+1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.6. SOLUTION OF SYSTEMS 105
Simplify these equations and solve take the Laplace transform to obtain
s+1 1 2e−4s
I1 (s) = − 2
5(2s + 1) s s + 4
1 1 1 2 2 s 9
= − − − 2 + 2 e−4s ,
5 s 2s + 1 85 2s + 1 s + 4 s + 4
1 2se−4s
I2 (s) = 1− 2
5(2s + 1) s +4
1 2 2 s 8
= + − − e−4s .
5(2s + 1) 85 2s + 1 s2 + 4 s2 + 4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
106 CHAPTER 3. THE LAPLACE TRANSFORM
2(s2 + 5) 5 1 s 8 s
X1 (s) = = − − ,
s(x4
+ 13s2 + 36) 18s 10 s2 + 4 45 s2 + 9
s−4 1 1 s 4 s
X2 (s) = =− + − .
s(s2 + 13s2 + 36) 9s 5 s2 + 4 45 s2 + 9
From the inverse Laplace transform we obtain the solution for the dis-
placement functions
5 1 8
x1 (t) = − cos(2t) − cos(3t),
18 10 45
1 1 4
x2 (t) = − cos(2t) + cos(3t).
9 5 45
Then
s2 + 5
X1 (s) = (1 − e−2s ),
s(s4
+ 13s2 + 36)
2
X2 (s) = (1 − e−2s ).
s(s4 + 13s2 + 36)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.6. SOLUTION OF SYSTEMS 107
Transform the system and solve for Y1 (s) and Y2 (s) to obtain
ms2 + k2
Y1 (s) = Y2 (s),
k2
Aωk2
Y2 (s) = .
(s2 + ω 2 )(M ms4 + mc1 s3 + (mk1 + mk2 + M k2 )s2 + k2 c1 s + k1 k2 )
(b) If ω = k2 /m then
s2 + ω 2
Y1 (s) = Y2 (s)
ω2
Aω
= .
M s4 + c1 s3 + (k1 + k2 + M ω 2 )s2 + ω 2 c1 s + k1 ω 2
my1 = k(y2 − y1 ),
m2 y2 = k(y1 − y2 ),
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
108 CHAPTER 3. THE LAPLACE TRANSFORM
Apply the transform to the system and solve for Y1 (s) and Y2 (s) to obtain
kd
Y1 (s) = ,
m1 +m2
m1 s s2 + (m 1 m2 )
k
d(m1 s2 + k)
Y2 (s) = .
m1 +m2
m1 s s2 + (m 1 m2 )
k
The quadratic factor in the denominator shows that the motion has fre-
quency
m1 + m2
ω= k,
m1 m2
with initial conditions i1 (0) = i2 (0) = 0. Transform the system and solve
for I1 (s) and I2 (s) to obtain
5(30s + 20)e−5s
I1 (s) =
s(600s2 + 700s + 100)
1 1 27 1
= − − e−5s ,
s 10(s + 1) 5 6s + 1
50e−5s
I2 (s) = 2
s(600s + 700s + 100)
1 1 18 1
= + − e−5s .
2s 10(s + 1) 5 6s + 1
19. As in the solution of Problem 18, except with E(t) = 5δ(t − 1), the trans-
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.6. SOLUTION OF SYSTEMS 109
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
110 CHAPTER 3. THE LAPLACE TRANSFORM
21. Using the notation of the solution of Problem 20, we can write the system
6 3
x1 = − x1 + x2 ,
200 100
4 4
x2 = x1 − x2 + 5δ(t − 3),
200 200
with initial conditions
Then
100000s + 5500 + 1500e−3s
X1 (s) =
10000s2 + 700s + 6
50 900 300 150
= + + − e−3s ,
50s + 3 100s + 1 100s + 1 50s + 3
50000s + 3500 + (50000s + 1500)e−3s
X2 (s) =
10000s2 + 700s + 6
50 600 150 200
=− + + + e−3s .
50s + 3 100s + 1 50s + 3 100s + 1
Invert these equations to obtain the solution
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
3.7. POLYNOMIAL COEFFICIENTS 111
z(u) = ce−2u − 1
or
y(t) = −1 + ce−2/t .
This problem can also be solved as a first order linear differential equation,
after dividing it by t2 .
2. Transform the initial value problem to obtain
d
s2 Y − sy(0) − y (0) − 4 (sY − y(0)) − 4Y = 0.
ds
Upon taking the derivative and inserting the initial values, we obtain
4sY + (8 − s2 )Y = 7,
In Problems 3 through 9, the details of the solution are like those of Problems
1 and 2, and only the solution is given.
3. y(t) = 7t2
4. y(t) = −4t
5. y(t) = ct2 e−t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
112 CHAPTER 3. THE LAPLACE TRANSFORM
6. y(t) = 3t2
7. y(t) = 4
8. y(t) = 10t
9. y(t) = 3t2 /2
3s2 − 4s + 4 e−s
Y (s) = + K
s2 (s − 1) s2 (s − 1)
3 4 1 1 1
= − 2 +K − − 2 e−s .
s−1 s s−1 s s
K is arbitrary and can be given any real value. This illustrates a bi-
furcation in the solution. At t = 1, the solution splits off and travels
along different curves, depending on the choice of K. Notice that the ex-
istence/uniqueness theorem for solutions of this differential equation does
not apply at t = 1, which is a singular point.
11. When we wrote factorials and inverted terms of the form 1/s2n+k in the
binomial expansion used in the derivation, we assumed that n is a non-
negative integer.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 4
Series Solutions
2. Write
∞
∞
y − x3 y = nxn xn−1 − an xn+3
n=1 n=0
∞
= a1 + 2a2 x + 3a3 x2 + (nan − an−4 )xn−1 = 4.
n=4
113
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
114 CHAPTER 4. SERIES SOLUTIONS
3. Write
∞
∞
∞
y + (1 − x2 )y = nan xn−1 + an xn − an xn+2
n=1 n=0 n=0
∞
= (a1 + a0 ) + (2a2 + a1 )x + (nan + an−1 − an−3 )xn−1
n=3
= x.
4. Begin with
∞
∞
∞
y + 2y + xy = n(n − 1)an xn−2 + 2nan xn−1 + an xn+1
n=2 n=1 n=0
= (2a2 + 2a1 ) + (3 · 2a3 + 2 · 2a2 + a0 )x
∞
+ (n(n − 1)an + 2(n − 1)an−1 + an−3 )xn−2 = 0.
n=4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
4.1. POWER SERIES SOLUTIONS 115
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
116 CHAPTER 4. SERIES SOLUTIONS
7. Write
∞
y − x2 y + 2y = n(n − 1)an xn−2
n=2
∞
∞
− nan xn+1 + 2an xn = (2a2 + 2a0 ) + (6a3 + 2a1 )x
n=1 n=0
∞
+ (n(n − 1)an − (n − 3)an−3 + 2an−2 )xn−2 = x.
n=4
Then a0 and a1 are arbitrary, a2 = −a0 , 6a3 + 2a1 = 1, and, for the
recurrence relation,
(n − 3)an−3 − 2an−2
an = for n = 4, 5, · · · .
n(n − 1)
The general solution has the form
1 1 1
y(x) = a0 1 − x2 + x4 − x5 − x6 + · · ·
6 10 90
1 1 1 7 6
+ a1 x − x3 + x4 + x5 − x + ···
3 12 30 180
1 3 1 5 1 6 1 7 1 8
+ x − x + x + x − x + ··· .
6 6 60 1260 480
Here a0 = y(0) and a1 = y (0). The third series in the solution represents
a particular solution obtained from the recurrence by putting a0 = a1 = 0.
8. Write
∞
∞
y + xy = nan xn−1 + an xn+1
n=1 n=0
∞
∞
= a1 + (2na2n + a2n−2 )x2n−1 + ((2n + 1)a2n+1 + a2n−1 )x2n
n=0 n=1
∞
(−1)n x2n
= .
n=0
(2n)!
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
4.1. POWER SERIES SOLUTIONS 117
9. We have
∞
y + (1 − x)y + 2y = n(n − 1)an xn−2
n=2
∞
∞
∞
+ nan xn−1 − nan xn + 2 2an xn
n=1 n=1 n=0
∞
= (2a2 + a1 + 2a0 ) + (n(n − 1)an + (n − 1)an−1 − (n − 4)an−2 )xn−2
n=3
= 1 − x2 ,
Then a0 and a1 are arbitrary, 2a2 + a1 + 2a0 = 1, 6a3 + 2a2 + a1 = 0,
12a4 + 3a3 = −1, and
−(n − 1)an−1 + (n − 4)an−2
an =
n(n − 1)
for n = 5, 6, · · · . The general solution is
2 1 3 1 4 1 5
y(x) = a0 1 − x + x − x + x − · · ·
3 12 30
1 1 1 1 1 6 1 7
+ a1 x − x2 + x2 − x3 − x4 − x + x + ··· ,
2 2 6 24 360 2520
where a0 = y(0) and a1 = y (0). The last series is a particular solution of
the nonhomogeneous equation.
10. We have
∞
∞
y + xy = n(n − 1)an xn−2 + nan xn
n=2 n=1
∞
= 2a2 + (n(n − 1)an + (n − 2)an−2 )xn−2
n=3
∞
1
=− xn−2 .
n=3
(n − 2)!
Then a0 and a1 are arbitrary, a2 = 0, and
−(n − 2)an−2 − 1/(n − 2)!
an =
n(n − 1)
for n = 3, 4, · · · . The solution is
1 3 3 5 15 7 105 9
y(x) = a0 + a1 x − x + x − x + x + ···
3! 5! 7! 9!
1 1 2 3 11 19
+ − x3 − x4 + x5 + x6 − x7 + x8 + · · · .
3! 4! 5! 6! 7! 8!
Here a0 = y(0) and a1 = y (0).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
118 CHAPTER 4. SERIES SOLUTIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
4.2. FROBENIUS SOLUTIONS 119
A second solution is
∞
1
y2 (x) = (1 − x) ln(x) + 3x − xn .
n=2
n(n − 1)n!
For the remaining problems of this section we give the essential elements
of the solution without all of the details of the calculations.
2(n + r − 2)
cn = cn−1
(n + r)(n + r − 1)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
120 CHAPTER 4. SERIES SOLUTIONS
(−1)n
3/2 2n
y1 (x) = x 1+ x
n=1
2n n!(5 · 7 · 9 · · · (2n + 3))
and ∞
(−1)n+1
−3/2 3n
y2 (x) = x 1+ x .
n=1
2n+1 n!(3) · · · (2n − 3)
and
∞
y2 (x) = c∗n xn .
n=0
Substitute these into the differential equation in turn to obtain
∞
(−1)n
1/2 n
y1 (x) = x 1+ x
n=1
2n n!(3 · 5 · 7 · · · (2n + 1))
1/2 1 1 2 1 3 1 4
=x 1− x+ x − x + x + ···
6 120 5040 362880
and
∞
(−1)n
y2 (x) = 1 + n n!(1 · 3 · 5 · · · (2n − 1))
xn
n=1
2
1 1 1 3 1
= 1 − x + x2 − x + x4 − · · · .
2 24 720 40320
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
4.2. FROBENIUS SOLUTIONS 121
Upon substituting these into the differential equation, we obtain the simple
solutions
y1 (x) = x1/2 , y2 (x) = x−1/2 .
We could also have observed that the differential equation in this problem
is an Euler equation.
1 4 1 1
y1 (x) = x2 + x + x6 + x8 + · · · = x sinh(x)
3! 5! 7!
and
1 3 1 1
y2 (x) = x − x2 + x − x4 + x4 − · · · = xe−x .
2! 3! 4!
∞
(−1)n 2n+1 n+2 2 1 1 1 6
y1 (x) = x = x2 − x3 + x4 − x5 + x − ··· .
n=0
n!(n + 2)! 3 6 45 540
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
122 CHAPTER 4. SERIES SOLUTIONS
with c∗0 = 1 for simplicity, we obtain c∗1 = 2, k = −2, c∗2 arbitrary (we take
this coefficient to be zero), and
1 (−1)n 2n+1
c∗n = − 2c∗n−1 +
n(n − 2) n((n − 2)!)2
and
∞
y2 (x) = y1 (x) ln(x) + c∗n xn .
n=1
and ∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 5
Approximation of Solutions
123
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
124 CHAPTER 5. APPROXIMATION OF SOLUTIONS
y(x)0
-4 -2 0 2 4
x
-2
-4
y(x)0
-3 -2 -1 0 1 2 3
x
-1
-2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
5.2. EULER’S METHOD 125
y(x)0
-2 -1 0 1 2
x
-2
-4
y(x)0
-4 -2 0 2 4
x
-2
-4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
126 CHAPTER 5. APPROXIMATION OF SOLUTIONS
y(x)0
-3 -2 -1 0 1 2 3
x
-2
-4
y(x)0
-4 -2 0 2 4
x
-2
-4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
5.2. EULER’S METHOD 127
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
128 CHAPTER 5. APPROXIMATION OF SOLUTIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
5.3. TAYLOR AND MODIFIED EULER METHODS 129
xk Euler approx. yk
0.0 4
0.05 3.20
0.10 2.895
0.15 2.3324875
0.20 1.99078478
0.25 1.802625739
0.30 1.52652761
0.35 1.531089704
0.40 1.431377920
0.45 1.348435782
0.50 1.280454395
6. We do not have the exact solution in closed form for this problem. Table
5.6 lists approximate solution values.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
130 CHAPTER 5. APPROXIMATION OF SOLUTIONS
xk Runge-Kutta approximation
0.0 2
0.2 2.162573
0.4 2.27782433
0.6 2.34197299
0.8 2.35937518
1.0 2.33748836
1.2 2.28390814
1.4 2.20518645
1.6 2.10658823
1.8 1.99221666
2.0 1.86523474
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
5.3. TAYLOR AND MODIFIED EULER METHODS 131
xk Runge-Kutta approximation
0.0 1
0.2 1.26465161
0.4 1.45389723
0.6 1.58483705
0.8 1.67216598
1.0 1.72743772
1.2 1.75944359
1.4 1.77479969
1.6 1.77846513
1.8 1.77414403
2.0 1.76458702
xk Runge-Kutta approx.
3.0 2
3.2 0.927007472
3.4 0.281610758
3.6 0.0797232508
3.8 0.0218981447
4.0 5.92711033E-03
4.2 1.60552109E-03
4.4 4.42481887E-04
4.6 1.26188752E-04
4.8 3.78589406E-05
5.0 1.21355052E-05
6. Table 5.12 lists approximate values for Problem 6, using RK4. An exact
solution for comparison is not available for this problem.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
132 CHAPTER 5. APPROXIMATION OF SOLUTIONS
xk Runge-Kutta approx.
0.78 1
0.98 1.12897598
1.18 1.1538066
1.38 1.12922007
1.58 1.08698233
1.78 1.04366376
1.98 1.00569426
2.18 0.974171102
2.38 0.948175352
2.58 0.926448449
2.78 0.907969373
2.98 0.891968617
3.18 0.877955391
3.38 0.86554404
3.58 0.85445476
3.78 0.844473215
3.98 0.835431462
4.18 0.827195488
4.38 0.819656709
4.58 0.812725995
4.78 0.806329358
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 6
2. F + G √
= i + 4j − 3k, F − G = i − 4j − 3k, 2F = 2i − 6k, 3G = 12j,
F = 10
3. F + G =
√3i − k, F − G = i − 10j + k, 2F = 4i − 10j, 3G = 3i + 15j − 3k,
F = 29
√ √ √
4. F+G = ( 2+8)i+j−4k, √ F−G = ( 2−8)i+j−8k, 2F = 2 2i+2j−12k,
3G = 24i + 6k, F = 41
In each of Problems 6 through 9, we first use the given points to find a vector
from the first point to the second. This vector may or may not be a unit vector,
but at least it is in the right direction. Divide this vector by its length to obtain
a unit vector in the direction from the first to the second point. If this vector
is then multiplied by a positive scalar α, then we have a vector of length α in
the direction from the first point to the second. We include these details for
Problem 6 and give just the answer for Problems 7 - 9.
6. −5i + j − 2k is a vector from the first point to the second. Divide this
vector by its length to obtain a unit vector, then multiply by 5 to obtain
a vector of length 5 in the direction from (0, 1, 4) to (−5, 2, 2):
5
√ (−5i + j − 2k).
30
133
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
134 CHAPTER 6. VECTORS AND VECTOR SPACES
7.
9
√ (−5i − 4j + 2k)
45
8.
12
√ (10i − 3j − 4k)
125
9.
4
(−4i + 7j + 4k)
9
In each of Problems 10 through 15, we follow the procedure of the text
to find parametric equations of a line through the given points. Details are
provided for Problem 10 only. However, it must be understood that any line in
three space can be described by infinitely many different parametric equations.
For example, if in Example 6.1 we replace t by 2t, we obtain slightly different
looking parametric equations of the same line, since 2t takes on all real values
as t does.
10. Let L be the line containing these points. A vector from (1, 0, 4) to (2, 1, 1)
is M = i + j − 3k. A vector from (1, 0, 4) to (x, y, z) on L is (x − 1)i +
yj + (z − 4)k. These two vectors are parallel, so for some scalar t,
Then
x − 1 = t, y = t, z = 4 − 3t.
Parametric equations of L are
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.2. THE DOT PRODUCT 135
1. F · G = 2 and
F·G 2
cos(θ) = =√ .
F G 14
The vectors are not orthogonal.
√
2. F · G = 8, cos(θ) = 8/ 82, not orthogonal
√ √
3. F · G = −23, cos(θ) = −23/ 29 41, not orthogonal
√ √
4. F · G = −63, cos(θ) = −63/ 75 74, not orthogonal
In Problems 7 - 12, if the given point is (x0 , y0 , z0 ) and the normal vector is
N = ai + bj + ck, then the equation of the plane is
3(x + 1) − (y − 1) + 4(z − 2) = 0.
3x − y + 4z = 4.
8. x − 2y = −1
9. 4x − 3y + 2z = 25
10. −3x + 2y = 1
11. 7x + 6y − 5z = −26
12. 4x + 3y + z = −6
For each of Problems 13, 14 and 15, the projection of v onto u is calculated
as
u·v
u.
u 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
136 CHAPTER 6. VECTORS AND VECTOR SPACES
13.
−9
proju v = u
14
14.
11
− u
30
15.
1
u
62
2. F × G = i + 12j + 6k
3. F × G = −8i − 12j − 5k
4. F × G = 112k
In Problems 5 through 9, the three points are used to find two vectors in
the plane that is wanted. Their cross product produces a normal vector to this
plane, and then, knowing a point on the plane and a normal vector, we can find
an equation of the plane, as in Section 6.1. This procedure produces N = O
exactly when the three points are collinear and do not define a unique plane.
The details of this procedure are included only for Problem 5.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.4. THE VECTOR SPACE RN 137
6. The points are not collinear and the plane containing them has equation
x + 2y + 6z = 12.
7. The points are not collinear and the plane containing them has equation
2x − 11y + z = 0.
8. The points are not collinear and the plane containing them has equation
5x + 16y − 2z = −4.
9. The points are not collinear and the plane containing them has equation
29x + 37y − 12z = 30.
For Problems 10, 11 and 12, recall that the vector ai + bj + ck is normal to
the plane ax + by + cz = d. Any nonzero scalar multiple of this normal vector
is also a normal vector.
10. N = 8i − j + k
11. N = i − j + 2k
12. N = i − 3j + 2k
13. The area of a parallelogram in which two incident sides have an angle of
θ between them is the product of the lengths of the sides times the cosine
of θ. If the sides are along the vectors F and G, drawn from a common
point, then this area is
F G cos(θ),
and this is exactly F × G .
(G × H) · F = N · F = N F cos(θ)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
138 CHAPTER 6. VECTORS AND VECTOR SPACES
h F
2. If
α(2i) + β(3j) + γ(5i − 12k) + δ(i + j + k) = 0
then
2α + 5γ + δ = 0,
3β + δ = 0,
−12γ + δ = 0.
2 < 1, 2, −3, 1 > + < 4, 0, 0, 2 > − < 6, 4, −6, 4 > = < 0, 0, 0, 0 > .
6. Suppose
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.4. THE VECTOR SPACE RN 139
Then
−3β − 2γ + δ = 0,
α + 2β + 2γ + δ = 0,
α + 4β + 34γ − 6δ = 0,
α + 4β + 2γ + 2δ = 0.
α = β = γ = δ = 0.
The only linear combination of the given vectors that equals the zero
vector is the trivial linear combination (all coefficients zero). Therefore
the vectors are linearly independent.
7. The vectors are linearly dependent, since
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
140 CHAPTER 6. VECTORS AND VECTOR SPACES
21. First,
(X − Y) · (X + Y) = X · X + X · Y − Y · X − Y · Y
= X 2 − Y 2 = 0,
so X − Y is orthogonal to X + Y.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.4. THE VECTOR SPACE RN 141
22. Let
k
Y =X− (X · Vj )Vj .
j=1
Then
0 ≤ Y 2 = Y · Y
⎛ ⎞ ⎛ ⎞
k k
= ⎝X − (X · Vj )Vj ⎠ · ⎝X − (X · Vj )Vj ⎠
j=1 j=1
k
= X · X − 2X · (X · Vj )Vj
j=1
⎛ ⎞ ⎛ ⎞
k
k
+ ⎝ (X · Vj )Vj ⎠ · ⎝ (X · Vj )⎠ Vj
j=1 j=1
k
= X · X − 2X · (X · Vj )Vj
j=1
k
k
+ (X · Vj )(X · Vr )Vj · Vr .
j=1 r=1
k
0 ≤ X 2 −2 (X · Vj )2
j=1
k
+ (X · Vj )2
j=1
k
= X 2 − (X · Vj )2 .
j=1
Therefore
k
(X · Vj )2 ≤ X 2 .
j=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
142 CHAPTER 6. VECTORS AND VECTOR SPACES
We have
X 2 = X · X
⎛ ⎞ ⎛ ⎞
n n
= ⎝ (X · Vj )Vj ⎠ · ⎝ (X · Vj )Vj ⎠
j=1 j=1
n
n
= (X · Vj )(X · Vr )Vj · Vr
j=1 r=1
n
= (X · Vj )2 .
j=1
0 = 0V1 + · · · + 0Vk ,
Vm = c1 V1 + · · · + cm−1 Vm−1 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.5. ORTHOGONALIZATION 143
form a basis for Rn . If not, there is some vector in Rn that is not a linear
combination of u1 , · · · , uk , v1 , v2 , and we repeat the argument. After
n − k applications of this argument, we reach a linearly independent set
of n vectors
u1 , · · · , uk , v1 , · · · , vn−k
spanning Rn , and these form a basis for Rn .
6.5 Orthogonalization
The arithmetic of carrying out the Gram-Schmidt process can be tedious and
computations are most easily carried out using a software package such as
MAPLE.
In each problem, the given vectors are denoted X1 , · · · , Xk in the given
order.
2. Let V1 = X1 and
11
V2 = X 1 + X1 =< 0, 4/5, 2/5, 0 > .
5
3. Let V1 = X1 , then
−7
V2 = X1 − X1 =< 0, 4/3, 13/6, 29/6 > .
6
Finally,
3 43/2
V3 = X 3 − V 1 − V2
6 179/6
1 129
= X3 − V1 − V2
2 179
1
= < 0, 7, −11, 3 > .
179
4. V1 = X1 ,
5
V2 = X 2 − X1
26
1
= < 109, 0, −41, 0, 58 >,
26
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
144 CHAPTER 6. VECTORS AND VECTOR SPACES
17 331/26
V3 = X 3 − X1 − V2
26 651/26
17 331
= X3 − x1 − V2
26 651
1
= < −962, 0, −1406, 0, 814 > .
651
5. V1 = X1 ,
5
V2 = X2 − X1
7
1
= < 0, 0, −1, −19, 40 >,
9
2 17
V3 = X3 + V1 + V2
9 9
1
= < 0, 218, −341, 279, 62 >,
218
6 13 435
V4 = X1 + V2 − V3
9 3 1179
1
= < 0, 248, 88, −24, −32 > .
393
6. V1 = X1 ,
1
V2 = X2 − X1
10
1
= < 21, −8, −60, −31, −18, 0 >,
10
3 163/10
V3 = X3 − X1 − V2
10 269/10
1
= < −423, −300, 489, −759, 132, 0 >,
269
−15 13/2 4455/269
V4 = X4 − X1 − V2 − V3
10 269/10 4095/269
1
= < 337, −145, 250, 29, −9, 0 > .
91
7. V1 = X1 ,
3
V 2 = X2 + X 1
2
1
= < 0, 0, −3, 3, 0, 0 > .
2
8. V1 = X1 , V2 = X2 because X2 and X1 are orthogonal. Finally,
4 4
V3 = X3 + V1 + V2 = < 0, −8/3, 0, −8/3, 0, 16/3 > .
12 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.6. ORTHOGONAL COMPLEMENTS AND PROJECTIONS 145
and
u⊥ = u − uS =< 0, 0, 1, 7 > .
2.
2 1
uS = V1 + V2 =< 0, 0, 0, 1, 0 >,
5 5
3.
7
uS = V1 + V2 − 3v3 =< 9/2, −1/2, 0, 5/2, −13/2 >,
2
4.
31
uS = −3V1 + V2 =< −86/39, 148/39, 62/13, 31/39 >,
39
5.
1
uS = 3V1 + V2 =< 3, 1/2, 3, 1/2, 3, 0, 0 >,
2
S = (S ⊥ )⊥ .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
146 CHAPTER 6. VECTORS AND VECTOR SPACES
v1 , · · · , vk , u1 , · · · , ur .
v1 , · · · , vk , u1 , · · · , ur
9. Let
V1 =< 2, 1, −1, 0, 0 >, V2 =< −1, 2, 0, 1, 0 > and V3 =< 0, 1, 1, −2, 0 > .
These form an orthogonal basis for S. With u =< 4, 3, −3, 4, 7 >, compute
7 4
uS = V 1 + V2 − V 3
3 3
=< 11/3, 3, −11/3, 11/3, 0 > .
10. Let
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.7. THE FUNCTION SPACE C[A, B] 147
V1 (x) = X1 (x) = ex .
Next choose
X2 · X 1
V2 (x) = X2 (x) − X1
X1 · X1
1
1 dx x
= e−x −
10 e
e 2x dx
0
2
= e−x − 2 ex .
e −1
It is routine to check that indeed V1 and V2 are orthogonal, since
1
V 1 · V2 = V1 (x)V2 (x) dx = 0.
0
cos(x) · sin(x)
V2 (x) = cos(x) − sin(x) = cos(x),
sin(x) · sin(x)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
148 CHAPTER 6. VECTORS AND VECTOR SPACES
The formulas for the orthogonal basis functions is the same, but now the
dot product that appears in the coefficients is different. Choose V1 (x) =
X1 (x) = 1, and then
X 2 · X1
V2 (x) = cos(πx/2) − X1
X 1 · X1
4
= cos(πx/2) + 2 .
π
Finally,
X 3 · X1 X 3 · X2
V3 (x) = X3 − X1 − X2
X 1 · X1 X2 · X2
2 π(16 − π 2 ) 4
= sin(πx/2) − − cos(πx/2) + .
π π 4 − 32 π2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.7. THE FUNCTION SPACE C[A, B] 149
10
0
0 0.5 1 1.5 2 2.5 3
x
π2 4(−1)k
c0 = and ck =
3 k2
for k = 1, 2, 3, 4. Then
π2 1 1
fS (s) = − 4 cos(x) + cos(2x) − cos(3x) + cos(4x).
3 2 4
Figure 6.2 compares a graph of f (x) and fS (x). It happens that these
graphs are fairly close, but in applications f (x) is probably not approxi-
mated closely enough by fS (x) for reliable calculations. The point, how-
ever, is that fS (x) is the function in C[0, π] nearest to the subspace S
spanned by the five given functions, in the sense of distance in this func-
tion space. If we wanted a better numerical approximation (graphs closer
together), we could change S and include more functions cos(kx). This is
the idea of a Fourier cosine expansion, treated in Chapter Fourteen.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
150 CHAPTER 6. VECTORS AND VECTOR SPACES
10
0
0 0.5 1 1.5 2 2.5 3
x
have
5
fS (x) = ck sin(kx)
k=1
where
π
x2 sin(kx) dx
ck =
0 π 2
0
sin (kx) dx
2 −2 + 2(−1)k − k 2 π 2 (−1)k
= .
π k3
Figure 6.3 shows graphs of f and fS . It is clear that fS does not approx-
imate f very well in the numerical sense that f (x) and fS (x) are close,
within some small error tolerance. However, it remains true that fS is the
function in C[0, π] closest to S. If we want a better numerical approxima-
tion of f (x) by a sum of multiples of functions sin(kx), we must choose k
larger. Later we will see this as one idea behind Fourier sine series.
7. We want the function fS in S that is closest (in the distance defined on
this function space) to f (x) = x(2 − x), where S is the subspace spanned
by the orthogonal functions 1, cos(kπx/2 and sin(kπx/2 for k = 1, 2, 3.
This orthogonal projection has the form
3
fS (x) = c0 + (ck cos(kπx/2) + dk sin(kπx/2)).
k=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
6.7. THE FUNCTION SPACE C[A, B] 151
x
-2 -1 0 1 2
0
-2
-4
-6
-8
16(−1)k+1 8(−1)k+1
ck = 2 2
and dk = .
π k πk
Figure 6.4 shows graphs of f and fS .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
152 CHAPTER 6. VECTORS AND VECTOR SPACES
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 7
7.1 Matrices
1. ⎛ ⎞
14 −2 6
2A − 3B = ⎝ 10 −5 −6⎠
−26 −43 −8
2. ⎛⎞
19 2
⎜ 6 −2⎟
−5A + 3B = ⎜
⎝−28 38 ⎠
⎟
−27 35
3.
2 + 2x − x2 −12x + (1 − x)(x + ex + 2 cos(x))
A2 + 2AB =
4 + 2x + 2ex + 2xex −22 − 2x + e2x + 2ex cos(x)
4.
−3A − 4B = (18)
This is a 1 × 1 matrix, which we think of as just the number 18. Here the
matrix structure serves no purpose, since there are no row and column
locations to distinguish between.
5.
−36 0 68 196 20
4A + 8B =
128 −40 −36 −8 72
6.
−17 18 −40 8 27 10
A3 − B2 = − =
6 1 −5 −39 11 40
153
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
154 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
7. ⎛ ⎞
−10 −34 −16 −30 −14
AB = ⎝ 10 −2 −11 −8 −45⎠ ; BA is not defined.
−5 1 15 61 −63
8.
−16 0 12 −32
AB = ; BA =
17 28 −14 0
9. ⎛ ⎞
3 −18 −6 −42 66
⎜−2 12 4 28 −44 ⎟
⎜ ⎟
⎜
AB = (115); BA = ⎜−6 36 12 84 −132⎟⎟
⎝0 0 0 0 0 ⎠
4 −24 8 −56 88
10. ⎛ ⎞
48 1 1 −58
⎜ −96 220 ⎟
AB = ⎜
2 2 ⎟ ; BA = 76 152
⎝−288 −22 −22 −68⎠ 50 136
−16 6 6 184
11.
410 36 −56 227
AB is not defined; BA =
17 253 40 −1
12.
−22 30 −10 −4
AB = ; BA is not defined.
−42 45 30 6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.1. MATRICES 155
Then B = C, but
12 6
AB = CA = .
6 3
Compute
⎛ ⎞ ⎛ ⎞
2 7 7 4 4 14 17 17 18 18
⎜7 8 9 9 9⎟ ⎜17 34 33 26 26⎟
⎜ ⎟ ⎜ ⎟
A3 = ⎜
⎜7 9 8 9 9⎟ 4 ⎜
⎟ and A = ⎜17 33 34 26 26⎟⎟.
⎝4 9 9 6 7 ⎠ ⎝18 26 26 25 24⎠
4 9 9 7 6 18 26 26 24 25
Compute
⎛ ⎞ ⎛ ⎞
3 2 1 2 1 19 18 11 14 11
⎜2 3 1 2 1⎟ ⎜18 19 11 14 11⎟
⎜ ⎟ ⎜ ⎟
A2 = ⎜
⎜1 1 3 0 3⎟ 4 ⎜
⎟ and A = ⎜11 11 20 4 20⎟⎟.
⎝2 2 0 2 0 ⎠ ⎝14 14 4 12 4⎠
1 1 3 0 3 11 11 20 4 20
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
156 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
Then
⎛ ⎞ ⎛ ⎞
4 2 3 3 2 10 10 11 11 10
⎜2 3 2 2 3⎟ ⎜10 6 10 10 6⎟
⎜ ⎟ 3 ⎜ ⎟
A2 = ⎜
⎜3 2 4 3 2⎟ ⎜
⎟ , A = ⎜11 10 10 11 10⎟⎟,
⎝3 2 3 4 2⎠ ⎝11 10 11 10 10⎠
2 3 2 2 3 10 6 10 10 6
and ⎛ ⎞
42 32 41 41 32
⎜32 30 32 32 30⎟
⎜ ⎟
A4 = ⎜
⎜41 32 42 41 32⎟
⎟.
⎝41 32 41 42 32⎠
32 30 32 32 30
The number of v4 − v5 walks of length 2 is 2, the number of v2 − v3 walks
of length 3 is 10, the number of v1 − v2 walks of length 4 is 32, and the
number of v4 − v5 walks of length 4 is 32.
26. (a) The i, i element of A2 is the number of vi − vi walks of length 2 in
the graph. Each such walk has the form vi − vj − vi , for some j = i,
hence corresponds to a vertex vj adjacent to vi in the graph. Therefore
Aii counts the number of vertices adjacent to vi .
(b) The i, i element of A3 is the number of walks vi − vi walks of length 3
in G. Any such walk has the form vi − vj − vk − vi , for some j = k, and
neither j nor k equal to i. These three vertices therefore form the vertices
of a triangle in the graph. However, each such triangle is counted twice in
the i, i element of A3 , because this triangle actually represents two vi − vi
walks, namely vi − vj − vk − vi and (going the other way), vi − vk − vj − vi .
Therefore
A3ii = 2(number of triangles in G).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.2. ELEMENTARY ROW OPERATIONS 157
28. We can reason as in Problem 27, except, in stringing out the rows of an
n × m matrix with complex entries, we can string out all the nm real
parts of the entries, followed by the nm complex parts of the entries.
This matches the set of all n × m complex matrices with R2nm , with
the operations of addition and scalar multiplication corresponding as in
Problem 27. We conclude that this vector space of complex matrices has
dimension 2nm.
As an example, the 2 × 3 complex matrix
2 − i 4 6 + 7i
1 − i 2i 3 − 4i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
158 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
√
3. Thus form ⎛ ⎞
1 √0 0
Ω = ⎝0 3 0⎠ .
0 0 1
As a check, observe that
⎛ ⎞
−2 √1 4√ 2
√
ΩA = ⎝ 0 3 16 3 3 3⎠
1 −2 4 8
0 0 0 1
Then ⎛⎞
3 −6
⎜1 1⎟
ΩA = ⎜ ⎟
⎝14 4 ⎠ ,
0 5
and this is the matrix obtained by performing the given row operation on
A.
3. ⎛ ⎞⎛ ⎞⎛ √ ⎞
5 0 0 0 1 0 1 0 13
Ω = ⎝0 1 0⎠ ⎝1 0 0⎠ ⎝ 0 1 0 ⎠
0 0 1 0 0 1 0 0 1
⎛ ⎞
0 5 √0
= ⎝1 0 13⎠
0 0 1
and ⎛ ⎞
40 √ 5√ −15
√
ΩA = ⎝−2 + 2 13 14 + 9 13 6 + 5 13⎠
2 9 5
4. ⎛ ⎞⎛ ⎞ ⎛ ⎞
1 0 0 1 0 0 1 0 0
Ω = ⎝−1 1 0⎠ ⎝0 0 1⎠ = ⎝−1 0 1⎠
0 0 1 0 1 0 0 1 0
and ⎛ ⎞
−4 6 −3
ΩA = ⎝ 5 −3 3 ⎠
12 4 −4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.2. ELEMENTARY ROW OPERATIONS 159
5. √
0 1 1 0 1 3 0 15
√
Ω= =
1 0 0 15 0 1 1 3
and
30 √ 120√
ΩA =
−3 + 2 3 15 + 8 3
6. ⎛ ⎞⎛ ⎞⎛ ⎞⎛ ⎞
1 0 0 1 0 0 √1 0 0 1 0 0
Ω = ⎝0 1 0⎠ ⎝0 1 0⎠ ⎝ 3 1 0⎠ ⎝0 1 0⎠
0 1 1 0 0 4 0 0 1 1 0 1
⎛ ⎞
√1 0 0
=⎝ 3√ 1 0⎠
4+ 3 1 4
and ⎛ ⎞
3√ −4√ 5√ 9 √
ΩA = ⎝ 2 + 3 √3 1 − 4 √3 3 + 5 √3 −6 + 9√ 3⎠
18 + 3 3 37 − 4 3 31 + 5 3 54 + 9 3
7. ⎛ ⎞⎛ ⎞⎛ ⎞ ⎛ ⎞
1 0 0 1 0 0 1 0 0 1 0 0
Ω = ⎝0 0 1⎠ ⎝14 1 0⎠ ⎝0 1 0⎠ = ⎝ 0 0 4⎠
0 1 0 0 0 1 0 0 4 14 1 0
and ⎛ ⎞
−1 0 3 0
ΩA = ⎝−36 28 −20 28⎠
−13 3 44 9
8. ⎛ ⎞⎛ ⎞⎛ ⎞⎛ ⎞
1 0 0 0 0 1 1 0 0 1 0 0
Ω = ⎝0 1 0⎠ ⎝0 1 0⎠ ⎝0 1 0⎠ ⎝0 0 1⎠
0 0 5 1 0 0 0 3 1 0 1 0
⎛ ⎞
0 1 3
= ⎝0 0 1⎠
5 0 0
and ⎛ ⎞
28 50 2
ΩA = ⎝ 9 15 0⎠
0 −45 70
In these and later problems, it is sometimes useful to use the delta notation,
defined by
1 if i = j,
δij =
0 if i = j.
For example, In is the n × n matrix whose i, j− element is δij .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
160 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
For i = s,
n
n
(EA)sj = esk akj = δtk akj = atj = bsj .
k=1 k=1
And for i = t,
n
n
(EA)tj = eik akj = δsk akj = asj = btj
k=1 k=1
for j = 1, 2, · · · , m. Therefore EA = B.
10. Let A be n × m. Since B and E are formed, respectively, by multiplying
row s of A and In by α, then, for i = s, bij = aij and, eij = δij , while for
i = s, bsj = αasj and esj = αδsj .
Now consider the i, j− element of EA. For i = s,
n
n
(EA)ij = eik akj = αδik akj = aij = bij ,
k=1 k=1
while
n
n
(EA)sj = esk akj = αδsk akj = bsj
k=1 k=1
for j = 1, 2, · · · , m. Therefore EA = B.
11. Let A be n × m. Now B and E are obtained, respectively, from A and In
by adding α times row s to row t. Then, for i = t, bij = aij and eij = δij ,
while for i = t, btj = atj + αasj and etj = δtj + αδsj .
Now consider the i, j− element of EA. For i = t,
n
n
(EA)ij = eik akj = δik akj = aij
k=1 k=1
while, for i = t,
n
n
(EA)tj = etk akj = (δtk + αδsj )akj
k=1 k=1
= atj + αasj = bsj .
Therefore EA = B.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.3. REDUCED ROW ECHELON FORM 161
2. We can reduce A by first adding row two to row one of I2 , then multiplying
row one (of the new matrix) by 1/3. Thus proceed:
1 0 1 −1 1/3 −1/3
I2 = → → = Ω.
0 1 0 1 0 1
This yields
1 0 1/3 4/3
AR = .
0 1 0 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
162 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
5. ⎛ ⎞ ⎛ ⎞
0 0 1 −3 1 0
⎜0 0 0 1⎟ ⎜0 1⎟
Ω=⎜
⎝1
⎟,A = ⎜ ⎟
0 −6 17 ⎠ R ⎝0 0⎠
0 1 0 0 0 0
6.
0 1 1 1
Ω= , AR =
1 −2 0 0
7. ⎛ ⎞ ⎛ ⎞
−8 −2 38 1 0 0
1 ⎝
Ω= 37 43 −7⎠ , AR = ⎝0 1 0⎠ = I3
270
19 −29 11 0 0 1
8.
−1/3 0 1 −4/3 −4/3
Ω= , AR =
0 1 0 0 0
9.
0 1 1 0 0 0
Ω= , AR =
1/2 1/2 0 1 3/2 1/2
10. ⎛ ⎞ ⎛ ⎞
0 0 1 1 0 0 −3/4
1⎝
Ω= 4 −4 −8⎠ , AR = ⎝0 1 0 3 ⎠
4
−4 8 8 0 0 1 0
11. ⎛ ⎞ ⎛ ⎞
0 1/2 −1 1 0 0
Ω=⎝ 0 0 1 ⎠ , A R = ⎝0 1 0⎠ = I3
−1/7 2/7 −3/7 0 0 1
12. ⎛ ⎞ ⎛ ⎞
0 0 1 0 1
⎜0 1 3 0⎟ ⎜ ⎟
⎜
Ω=⎝ ⎟ , AR = ⎜0⎟
1 0 −6 0⎠ ⎝0⎠
0 0 −1 1 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.4. ROW AND COLUMN SPACES 163
These are linearly independent as vectors in R3 and form a basis for the
row space of A.
The columns of A are
−4 1 3
C1 = , C2 = C3 = .
2 2 0
3 3
C3 = − C1 + C2 .
5 5
Therefore C1 and C2 form a basis for the column space, which also has
dimension 2.
Note that we can actually read the row and column space dimensions from
the reduced matrix, since the rank of A is the number of nonzero rows of
AR , and this rank is equal to both the row and column ranks.
In addition, as an example, we looked at the row and column vectors
explicitly in this solution, but this is not necessary if all we want is the
rank of the matrix. For this, either the row rank or the column rank is
sufficient, since these numbers must be equal.
2. ⎛ ⎞
1 0 7
A R = ⎝0 1 3⎠ .
0 0 0
Therefore the rank of A equals 2, and this is also the row rank and the
column rank. The first two rows of A are independent in R3 , hence form
a basis for the row space, and the first two columns are also independent
in R3 and form a basis for the column space.
3. ⎛ ⎞
1 0
A R = ⎝0 1⎠ ,
0 0
so A has rank 2. The first two rows and the two columns of A are bases
for the row and column spaces, respectively.
4. ⎛ ⎞
1 0 0 1/6 1/6
A R = ⎝0 1 0 1/6 1/6⎠ ,
0 0 0 0 0
so A has rank 2. The second row is 2 times the first row of A, but the
first and third rows are independent and form a basis for the row space in
R5 . The first two columns form a basis for the column space.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
164 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
5.
1 0 −1/4 1/2
AR = ,
0 1 −5/4 1/2
so A has dimension 2. The two rows of A form a basis for the row space
in R4 and the first two columns form a basis for the column space in R2 .
7. ⎛ ⎞
1 0 0
⎜0 1 0⎟
AR = ⎜
⎝0
⎟,
0 1⎠
0 0 0
so A has rank 3. The first, second and fourth rows are linearly independent
and form a basis for the row space in R3 . All three columns are linearly
independent and form a basis for the column space in R4 .
8. ⎛ ⎞
0 1 0
AR = ⎝0 0 1⎠ ,
0 0 0
so A has rank 2. The first two rows span the row space in R3 and columns
two and three span the column space in R3 .
9. We find that AR = I3 , so A has rank 3. The row space has all the rows
for a basis and the column space has all the columns.
10. ⎛ ⎞
1 0 0
⎜0 0 1⎟
AR = ⎜
⎝0
⎟,
0 0⎠
0 0 0
so A has rank 2. Rows one and three form a basis for the row space in
R3 , and columns one and three form a basis for the column space in R4 .
11. AR = I3 , so A has rank 3. All of the rows form a basis for the row space
and all of the columns form a basis for the column space.
12. ⎛ ⎞
1 0 0 0
A R = ⎝0 1 0 −13/2⎠ ,
0 0 1 −7
so A has rank 3. The rows form a basis for the row space in R4 and the
first three columns for a basis for the column space in R3 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.5. LINEAR HOMOGENEOUS SYSTEMS 165
13. ⎛ ⎞
1 0 −11
AR = ⎝0 1 −3 ⎠ ,
0 0 0
so A has rank 2. The first two rows are linearly independent and form a
basis for the row space in R3 , and the first two columns form a basis for
the column space in R3 .
14.
1 −2/3 −1/3 −1/3 0
AR = ,
0 0 0 0 0
so A has rank 1. Either row forms a basis for the row space in R5 , and
any of the first four columns forms a basis for the column space in R2 .
15. Use the fact that, for any matrix, the rank, row rank and column rank are
the same. Since the rows of A are the columns of At , then
rank of A = row rank of A
column rank of At = rank of At .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
166 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.5. LINEAR HOMOGENEOUS SYSTEMS 167
These two column vectors form a basis for the solution space of AX = O,
which has dimension 4 − 2 = 2.
5. The coefficient matrix
⎛ ⎞
1 −1 3 −1 4
⎜2 −2 1 1 0⎟
A=⎜
⎝1
⎟
0 −2 0 1⎠
0 0 1 1 −1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
168 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.5. LINEAR HOMOGENEOUS SYSTEMS 169
The solution space is clearly one-dimensional. We can also see this dimen-
sion from m − rank(A) = 4 − 3 = 1.
10. The coefficient matrix
⎛ ⎞
4 −3 0 1 1 −3
⎜0 2 0 4 −1 −6⎟
⎜
A=⎝ ⎟
3 −2 0 0 4 −1⎠
2 1 −3 4 0 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
170 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.5. LINEAR HOMOGENEOUS SYSTEMS 171
13. Yes. All that is required is that m − rank(A) > 0, so that the solution
space has something in it. As a specific example, consider the system
AX = O, with ⎛ ⎞
1 0 3
A = ⎝0 1 −1⎠ .
3 0 9
This is a homogeneous system with three equations in three unknowns.
We find that ⎛ ⎞
1 0 3
AR = ⎝0 1 −1⎠ ,
0 0 0
so A has rank 2. The solution space has dimension 3 − 2 = 1, hence has
nonzero vectors in it. The general solution is
⎛ ⎞
3
X = α ⎝1⎠ .
1
a1 C1 + a2 C2 + · · · + am Cm = O,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
172 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
This yields a linear combination of the columns equal to the zero vector,
but with not all coefficients zero, so the columns are linearly dependent.
Conversely, if the columns are linearly dependent, then there are numbers
a1 , · · · , am , not all zero, such that
a1 C1 + a2 C2 + · · · + am cm = O,
15. (a) Let R1 , · · · , Rn be the rows of A. These vectors span R, the row space
of the matrix. Now, X is in the solution space if and only if X = O, and
this is true exactly when Rj · X = 0 for j = 1, · · · , n, which in turn is true
if and only if X is orthogonal to each row of A. But this is equivalent to
X being orthogonal to every linear combination of the rows of A, hence
to every vector in the row space of A. Therefore the solution space of A
is the orthogonal complement of the row space, or
R⊥ = S(A).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.6. NONHOMOGENEOUS SYSTEMS 173
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
174 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.6. NONHOMOGENEOUS SYSTEMS 175
0 1
with α arbitrary.
7. The augmented matrix is
⎛ ⎞
..
⎜8 −4 0 0 10 . 1⎟
⎜ . ⎟
⎜0 1 0 1 −1 .. 2⎟ .
⎝ ⎠
.
0 0 1 −3 2 .. 0
(The x1 column has been omitted since x1 does not appear in the equa-
tions). The reduced form of this matrix is
⎛ ⎞
..
⎜1 0 0 1/2 3/4 . 9/8⎟
⎜ .. ⎟
⎜0 1 0 1 −1 . 2 ⎟ .
⎝ ⎠
..
0 0 1 −3 2 . 0
.
Since rank(A) = rank([A..B]), this system has solutions, which we read as
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
9/8 −1/2 −3/4
⎜ 2 ⎟ ⎜ −1 ⎟ ⎜ 1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
X = ⎜ 0 ⎟ + α⎜ 3 ⎟ + β⎜
⎜ ⎟ ⎜ ⎟ ⎟
⎜ −2 ⎟ ,
⎝ 0 ⎠ ⎝ 1 ⎠ ⎝ 0 ⎠
0 0 1
in which α and β are arbitrary.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
176 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
.
Now rank(A) = rank([A..B]) = number of unknowns = 3, so the system
has the unique solution ⎛ ⎞
3/4
X = ⎝−1/12⎠ .
1/6
.
Note that rank(A) = rank(A..B]), so there are solutions. We read from
the augmented matrix that the general solution has the form
X=
⎛ ⎞
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞ −1 ⎛ ⎞
−29/7 −1 1 −3/14 ⎜0⎟ 1/14
⎜ 0 ⎟ ⎜1⎟ ⎜0⎟ ⎜ 0 ⎟ ⎜ ⎟ ⎜ 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜0⎟ ⎜ ⎟
⎜ 1/7 ⎟ ⎜0⎟ ⎜0⎟ ⎜ 3/14 ⎟ ⎜ ⎟ ⎜−1/14⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜0⎟ ⎜ ⎟
⎜ 0 ⎟ + α ⎜ 0 ⎟ + β ⎜1⎟ + γ ⎜ 0 ⎟ + δ ⎜ ⎟ + ⎜ 0 ⎟ ,
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜0⎟ ⎜ ⎟
⎜ 0 ⎟ ⎜0⎟ ⎜0⎟ ⎜ 1 ⎟ ⎜ ⎟ ⎜ 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜0⎟ ⎜ ⎟
⎝ 0 ⎠ ⎝0⎠ ⎝0⎠ ⎝ 0 ⎠ ⎜ ⎟ ⎝ 0 ⎠
⎝1⎠
0 0 0 0 1
0
with α, β, γ, δ and arbitrary.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.6. NONHOMOGENEOUS SYSTEMS 177
0 1
in which α is arbitrary.
12. The augmented coefficient matrix is
⎛ ⎞
..
⎜ −4 5 −6 . 2 ⎟
⎜ .. ⎟
⎜ 2 −6 1 . −5 ⎟,
⎝ ⎠
..
−6 16 −11 . 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
178 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.7. MATRIX INVERSES 179
.
Now A and [A..B] have the same rank, so the system has solutions which
we read from the reduced augmented matrix
⎛ ⎞ ⎛ ⎞
−15/23 −21/33
⎜−25/23⎟ ⎜ 11/23 ⎟
X=⎜ ⎟ ⎜ ⎟
⎝ 40/23 ⎠ + α ⎝ 171/23 ⎠ .
0 1
15. Write ⎛⎞
a1
⎜ a2 ⎟
⎜ ⎟
X=⎜ . ⎟
⎝ .. ⎠
am
and let C1 , · · · , Cn be the columns of A. Now AX = B if and only if
a1 C1 + a2 C2 + · · · + am Cm = B.
This means that the system has a solution X if and only if X is a linear
combination of the columns of A, hence is in the column space of A.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
180 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
3.
1 −2 2
A−1 =
12 1 5
4.
1 4 0
A−1 = −
4 −4 −1
5.
1 3 −2
A−1 =
12 −3 6
6. ⎛ ⎞
64 −4 49
1 ⎝
A−1 = −8 4 −7⎠
56
0 0 14
7. ⎛ ⎞
−6 11 2
1 ⎝
A −1
= 3 10 −1⎠
31
1 −7 10
9. ⎛ ⎞
6 −6 0
1 ⎝
A−1 =− −3 −9 2⎠
12
3 −3 −2
11. ⎛ ⎞⎛ ⎞ ⎛ ⎞
−1 −1 8 4 1 −23
⎜ 14 ⎟ ⎜ ⎟ ⎜ ⎟
X = A−1 B =
1 ⎜−9 2 −5 ⎟ ⎜ 2 ⎟ = 1 ⎜−75⎟
11 ⎝ 2 2 −5 3 ⎠ ⎝ 0 ⎠ 11 −9 ⎠
⎝
3 3 −2 −1 −5 14
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.8. LEAST SQUARES VECTORS AND DATA FITTING 181
12. ⎛ ⎞⎛ ⎞ ⎛ ⎞
5 5 5 4 9
1 ⎝−10 1
X = A−1 B = 34 23⎠ ⎝0⎠ = ⎝15⎠
55 11
5 6 17 5 13
13.
X = A−1 B
⎛ ⎞⎛ ⎞ ⎛ ⎞
−11 −12 −9 −4 22
1 1
= − ⎝ −3 −16 −5⎠ ⎝ 5 ⎠ = ⎝27⎠
28 7
−8 −24 −4 8 30
14. ⎛ ⎞⎛ ⎞ ⎛ ⎞
4 4 0 4 −4
1 ⎝7 1
X = A−1 B = −6 39⎠ ⎝−5⎠ = ⎝ 58 ⎠
52 52
1 14 13 0 −66
15. ⎛ ⎞⎛ ⎞ ⎛ ⎞
5 −15 −15 0 −21
1 1
X = A−1 B = − ⎝−10 15 10 ⎠ ⎝ 0 ⎠ = ⎝ 14 ⎠
25 5
−5 10 0 −7 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
182 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
3. Compute
37 12 1 −3
At A = and (At A)−1 = .
12 4 −3 37/4
Next,
−26
At B = .
−8
Then
−2
X∗ = (At A)−1 (At B) = .
4
4. We find that ⎛ ⎞
5 −5 −12
At A = ⎝ −5 10 13 ⎠
−12 13 29
and this is a singular matrix. Thus obtain values of X∗ as solutions of
AX∗ = BS , where BS is the orthogonal projection of B onto the column
space S of A. The first two columns of A are linearly independent and
form a basis for R2 , so S = R2 . Since B is in R2 , then BS = B. Therefore
solve the nonhomogeneous system
AX∗ = B
to obtain ⎛⎞ ⎛ ⎞
7/3 −2
X∗ = α ⎝ 1 ⎠ + ⎝ 0 ⎠ ,
5/3 −1
in which α is an arbitrary constant.
6. Form ⎛ ⎞
1 1
⎜−2 3⎟
⎜ ⎟
A=⎜
⎜0 −1⎟⎟,
⎝2 2⎠
−3 7
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.8. LEAST SQUARES VECTORS AND DATA FITTING 183
so
1 −2 0 2 −3
At = .
1 3 −1 2 7
Then
t 18 −22 t −1 16/167 11/334
AA= and (A A) = .
−22 64 11/334 9/334
Next, compute
6
At B = .
6
Then
−63/167
X∗ = (At A)−1 (At B) = .
−6/167
7. We have ⎛ ⎞ ⎛ ⎞
1 1 3.8
⎜1 3⎟ ⎜11.7⎟
⎜ ⎟ ⎜ ⎟
A=⎜
⎜1 5⎟ and B = ⎜
⎟ ⎟
⎜20.6⎟ .
⎝1 7⎠ ⎝26.5⎠
1 9 35.2
Compute
5 25 33/40 −1/8
At A = and (At A)−1 = .
25 165 −1/8 1/40
Further,
97.80000
At B = .
644.20000
Then
0.1599
X∗ = (At A)−1 (At B) = .
3.8799
The line has the equation y = a + bx, with a = 3.8799 and b = 0.1599.
8. We have ⎛ ⎞ ⎛ ⎞
1 −5 21.2
⎜1 −3⎟ ⎜ 13.6 ⎟
⎜ ⎟ ⎜ ⎟
⎜1 −2⎟ ⎜ 10.7 ⎟
⎜ ⎟ ⎜ ⎟
A=⎜ 0⎟ ⎜ ⎟
⎜1 ⎟ and B = ⎜ 4.2 ⎟ .
⎜1 ⎟ ⎜ ⎟
⎜ 1⎟ ⎜ 2.4 ⎟
⎝1 3 ⎠ ⎝ −3.7 ⎠
1 6 −14.2
Then
7 0 1/7 0
At A = and (At A)−1 = .
0 84 0 1/84
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
184 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
Next,
34.20000
At B = .
−262.10000
Finally, compute
60.97750
X∗ = (At A)−1 (At B) = .
−10.82750
Next, compute
−9.79999
At B = .
227
Then
−9.266855
X∗ = .
4.167394
The equation of the line is y = a + bx, with a = 4.167394 and b =
−9.266855.
10. We have ⎛ ⎞ ⎛ ⎞
1 −3 −7.4
⎜1 −1⎟ ⎜−4.2⎟
⎜ ⎟ ⎜ ⎟
⎜1 0⎟⎟ ⎜−3.7⎟
⎜ ⎜ ⎟
A=⎜ 2⎟ ⎜ ⎟.
⎜1 ⎟ and B = ⎜−1.9 ⎟
⎜1 4⎟ ⎜ 0.3 ⎟
⎜ ⎟ ⎜ ⎟
⎝1 7⎠ ⎝ 2.8 ⎠
1 11 7.2
Then
7 20 1/5 −1/50
At A = and (At A)−1 = .
20 200 −1/50 7/1000
Finally, compute
−6.89999
At B = .
122.59999
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.9. LU FACTORIZATION 185
Then
−5.364589
X∗ = .
2.298867
The equation of the line is y = a + bx, with a = 2.298867 and b =
−5.364589.
7.9 LU Factorization
1. Given A, first produce U. Proceed
⎛ ⎞
2 4 −6
A = ⎝ 8 2 1 ⎠ → add −4 row one to row two, 2 row one to row three
−4 4 10
⎛ ⎞
2 4 −6
→ ⎝0 −14 25 ⎠
0 12 −2
⎛ ⎞
2 4 −6
→ add 6/7 row two to row three → ⎝0 −14 25 ⎠ .
0 0 136/7
This is U: ⎛ ⎞
2 4 −6
U = ⎝0 −14 25 ⎠ .
0 0 136/7
Now use the boldface entries in the formation of U to obtain L. Start
with ⎛ ⎞
2 0 0
D = ⎝ 8 −14 0 ⎠.
−4 12 136/7
Here we have listed the boldface elements from the formation of U, with
zeros above, to form a lower triangular matrix. This is not yet L. In
D, divide each column by the reciprocal of the diagonal element of that
column to obtain ⎛ ⎞
1 0 0
L=⎝ 4 1 0⎠ .
−2 −6/7 1
It is routine to check that LU = A.
2. Proceed
⎛ ⎞
1 5 2
A = ⎝3 −4 2 ⎠ → −3 times row one to row two, subtract row one from row three
1 4 10
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
186 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
⎛ ⎞
1 5 2
⎝0 −14 4⎠ → add 1/19 row two to row three
0 −1 8
⎛ ⎞
1 5 2
⎝0 −19 −4 ⎠ .
0 0 156/19
Then ⎛ ⎞
1 5 2
U = ⎝0 −19 −4 ⎠ .
0 0 156/19
To form L, begin with
⎛ ⎞
1 0 0
D = ⎝3 −19 0 ⎠
1 −1 156/19
Then LU = A.
For Problems 3, 4 and 5 the same algorithm is used and we give only the
matrices L and U.
3. ⎛ ⎞ ⎛ ⎞
−2 1 12 1 0 0
U = ⎝ 0 −5 13 ⎠ , L = ⎝−1 1 0⎠
0 0 119/5 −1 −3/5 1
4. ⎛ ⎞ ⎛ ⎞
1 7 2 −1 1 0 0
U = ⎝0 −16 −4 9 ⎠,L = ⎝ 3 1 0⎠
0 0 25/2 7/8 −3 −7/8 1
5.
⎛ ⎞ ⎛ ⎞
1 4 2 −1 4 1 0 0 0
⎜0 −5 2 0 0 ⎟ ⎜1 1 0 0⎟⎟
U=⎜ ⎟,L = ⎜
⎝0 0 88/5 4 6 ⎠ ⎝−2 −14/5 1 0⎠
0 0 0 195/22 −691/44 4 14/5 −63/88 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.9. LU FACTORIZATION 187
6. This problem has a little twist to it. It is the only problem in which the
number of rows exceeds the number of columns. If the algorithm is carried
out starting with A, a difficulty occurs.
However, we can still write the LU − decomposition of A by working with
At , which is 3 × 4. The strategy is to find upper and lower triangular
matrices U and L so that
At = LU.
Then
A = Ut Lt .
But the transpose of an upper triangular matrix is lower triangular, and
the transpose of a lower triangular matrix is upper triangular, so this is
the decomposition we want for A.
Thus start with ⎛ ⎞
4 2 −3 0
A = ⎝−8
t
24 2 1 ⎠.
2 −2 14 −5
Applying the algorithm to this matrix, we find that
⎛ ⎞ ⎛ ⎞
4 2 −3 0 1 0 0
U = ⎝0 28 −4 1 ⎠ , L = ⎝ −2 1 0⎠ .
0 0 211/14 −137/28 1/2 −3/28 0
Problems 7 - 12 are small in the sense that the matrices are low-dimensional
and the entries are integers. In such cases it would be just as efficient to solve
the system AX = B directly. The LU − factorization method only reveals
computational efficiencies when the systems are large. However, these problems
are intended to promote familiarity with the method.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
188 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
8. The LU − decomposition of A is
2 1 1 3 1 0
U= ,L = .
0 7/2 11/2 1/2 1/2 1
Solve LY = B to obtain
2
Y= .
3
Now solve UX = Y to obtain
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
10 16 −8
⎜1⎟ ⎜ 0 ⎟ ⎜0⎟
X = α⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝ 0 ⎠ + β ⎝ 1 ⎠ + ⎝ 0 ⎠.
−7 −11 6
9. We find that
⎛ ⎞ ⎛ ⎞
−1 1 1 6 1 0 0
U=⎝ 0 3 2 16 ⎠ , L = ⎝−2 1 0⎠ .
0 0 17/3 52/3 −1 −1/3 1
Solve LY = B to obtain ⎛
⎞
2
Y=⎝ 5 ⎠
29/3
and then solve UX = Y for
⎛ ⎞ ⎛ ⎞
1 0
⎜ 28/3 ⎟ ⎜−5/3⎟
X = α⎜ ⎟ ⎜ ⎟
⎝ 26/3 ⎠ + ⎝−1/3⎠ .
−17/6 2/3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.9. LU FACTORIZATION 189
10. Obtain ⎛ ⎞ ⎛ ⎞
7 2 −4 1 0 0
U = ⎝0 20/7 44/7⎠ , L = ⎝−3/7 1 0⎠ .
0 0 16 4/7 1 1
Solve LY = B to obtain ⎛
⎞
7
Y = ⎝−1⎠ .
3
Solve UX = Y to obtain
⎛ ⎞
93/154
X = ⎝ 89/88 ⎠ .
−3/16
11. Obtain
⎛ ⎞ ⎛ ⎞
6 1 −1 3 1 0 0 0
⎜0 4/3 5/3 3 ⎟ ⎜ 2/3 1 0 0⎟
⎟.
U=⎜ ⎟,L = ⎜
⎝0 0 13/4 13/4⎠ ⎝−2/3 5/4 1 0⎠
0 0 0 5 1/3 −1 4/13 1
Solve LY = B: ⎛ ⎞
4
⎜ 28/3 ⎟
Y=⎜
⎝ −7 ⎠
⎟
93/13
The solution of LY = B is
⎛ ⎞
0
Y = ⎝ −4 ⎠ .
−10/3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
190 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
Solve UX = Y:
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
46 −58 19 −25 47/3
⎜−35⎟ ⎜87/2⎟ ⎜−14⎟ ⎜37/2⎟ ⎜−73/6⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ 1 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
X = α⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ 0 ⎟ + β ⎜ 1 ⎟ + γ ⎜ 0 ⎟ + γ ⎜ 0 ⎟ + ⎜ 0 ⎟.
⎜ 0 ⎟ ⎜ 0 ⎟ ⎜ 1 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝ 0 ⎠ ⎝ 0 ⎠ ⎝ 0 ⎠ ⎝ 1 ⎠ ⎝ 0 ⎠
−6 15/2 −2 7/2 −5/2
T (1, 0, 0) =< 3, 1, 0 >, T (0, 1, 0) =< 0, −1, 0 >, T (0, 0, 1) =< 0, 0, 2 >
so ⎛⎞
3 0 0
AT = ⎝1 −1 0⎠ .
0 0 2
Because AT has rank 3, T is one-to-one and onto and the dimension of
the null space is 3 − 3 = 0 (contains only the zero vector).
2. T is linear and
1 −1 0 0
AT = .
0 0 1 −1
T is not one-to-one (all vectors < α, α, β, β > map to < 0, 0, 0, 0 >). T is
onto. Since AT has rank 2, the dimension of the null space is 4 − 2 = 2.
4. T is linear and ⎛ ⎞
0 0 0 0 1
⎜0 0 0 1 0⎟
⎜ ⎟
AT = ⎜
⎜ 1 −1 0 0 0⎟⎟.
⎝1 0 −1 0 0⎠
−1 −3 0 0 1
T is one-to-one and onto and the null space has dimension 5 − 5 = 0, since
AT has rank 5.
5. T is linear and ⎛ ⎞
1 0 0 −1 0
AT = ⎝0 1 −1 0 0⎠ .
0 0 0 1 1
T is not one-to-one, but is onto. The null space has dimension 5 − 3 = 2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
7.10. LINEAR TRANSFORMATIONS 191
6. T is linear and
1 1 4 −8
AT = .
−1 1 −1 0
T is onto but not one-to-one. The null space has dimension 4 − 2 = 2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
192 CHAPTER 7. MATRICES AND SYSTEMS OF LINEAR EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 8
Determinants
2. In the 2 × 2 case,
a11 (1/α)a12
B=
αa21 a22
so
1
|B| = a11 a22 − (α)(a12 − a21 )
α
= a11 a22 − a12 a21 = |A|.
In the 3 × 3 case,
⎛ ⎞
a11 (1/α)a12 (1/α2 )a13
B = ⎝ αa21 a22 (1/α)a23 ⎠
α2 a31 (1/α)a32 a33
193
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
194 CHAPTER 8. DETERMINANTS
If n is odd, then
|A| = −|A|,
and this implies that |A| = 0.
4. From the definition of determinant,
|In | = σ(p)(In )1p(1) (In )2p(2) · · · (In )np(n) .
p
Now (In )ij = 0 if i = j, so the only way a term of this sum can be nonzero
is if each factor
(In )jp(j) = 0
in this term. But this can occur only if p(j) = j for j = 1, · · · , n, so
p(1) = 1, p(2) = 2, · · · , p(n) = n.
This means that p must be the identity permutation that leaves each j
unchanged for j = 1, 2, · · · , n. But, if p is the identity permutation, than
σ(p) = 1. Further, each Ijj = 1. Therefore In has determine 1 · 1 · · · 1 = 1.
5. If p is a permutation of 1, 2, · · · , n, then it is impossible for each p(j) ≥ j
or for each p(j) ≤ j for j = 1, 2, · · · , n, unless each p(j) = j and p is the
identity permutation. Thus, the only (possibly) nonzero term in the sum
defining the determinant is
|A| = σ(p)A1p(1) A2p(2) · · · Anp(n)
= A11 A22 · · · Ann .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
8.2. EVALUATION OF DETERMINANTS I 195
2. Add 3 times row two to row one, then add row two to row three to obtain
2
−3 7 44 0 10
44 10
14 1 1 = 14 2+2
1 1 = (−1) (1) = 254
−13 −1 5 1 1 6
0 6
3. Add column two to column one, then 3 times column two to column three:
−4 5 6 1 5 21
−2 3 5 = 1 3 14 = (−1)3+2 (−2) 1 21 = −14
1 14
2 −2 6 0 −2 0
4. Add 2 times row three to row one and 2 times row three to row two to
obtain
2 −5 8 28
−5 0
28 −5
4
3 8 = 30 3 0 = (−1)3+3 (−4) = −936
30 3
13 0 −4 13 0 −4
5. Add 2 times column three to column one and then add column three to
column two to obtain
17 −2 5 27 3
5
27 3
1 12 3+3
0 = 1 12 0 = (−1) (−7) = −2, 247
1 12
14 7 −7 0 0 −7
6. Add column one to column two, then 3 times column one to column three,
then 2 times column one to column four to obtain
−3 3
9 6 −3 0 0 0
−1 18 8
1 −2 15 6 1 −1 18 8
= = (−1)1+1 (−3) 8 22 19
7 1 1 5 7 8 22 19
3 5 7
2 1 −1 3 2 3 5 7
−1 18 8
= −3 8 22 19 .
3 5 7
Now we have reduced the problem of evaluating a 4×4 determinant to one
of evaluating a 3 × 3 determinant. In this 3 × 3 determinant, add 18 times
column one to column two, then 8 times column one to column three to
obtain
−1 18 8 −1 0 0
8 22 19 = 8 166 83 = (−1) 166 83 = −249
59 31
3 5 7 3 59 31
Putting the two steps together,
−3 3 9 6
1 −2 15 6
= (−3)(−249) = 747.
7 1 1 5
2 1 −1 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
196 CHAPTER 8. DETERMINANTS
7. −122
8. 293
9. 72
2. Use row operations to reduce column one, then expand by this column:
1 1 6 1 1 16
2 −2 1 = 0 −4 −11 = −4 −11 = −4 −11 = 12
−4 −14 0 −3
3 −1 4 0 −4 −14
3. 3
4. 124
5. −773
6. 3, 775
7. −152
8. 4, 882
9. 1, 693
10. 3, 372
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
8.3. EVALUATION OF DETERMINANTS II 197
11.
1
α α2 1 α α2
1
β β 2 = 0 β−α β 2 − α2
1 γ γ 2 0 γ−α γ 2 − α2
1 α α2 1 α α2
= 0 β − α (β α)(β + α) = (β − α)(γ − α) 0 1
− β + α
0 γ − α (γ − α)(γ + α) 0 1 γ + α
1 β + α
= (β − α)(γ − α) = (β − α)(γ − α)(γ − β).
1 γ + α
12. Add columns two, three and four to column one and factor (α + β + γ + δ)
out of column one to obtain
a b c d 1 b c d
b c d a
= (a + b + c + d) 1 c d a .
c d a b 1 d a b
d a b c 1 a b c
Now add
(−1)row two + row three − row four
to row one and factor out (b − a + d − c) from the new row one to obtain
0 1 −1 1
1 c d a
(a + b + c + d)(b − a + d − c) .
1 d a b
1 a b c
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
198 CHAPTER 8. DETERMINANTS
2.
−1 1 4 0
A =
12 −1 3
3.
1 −4 1
A−1 =
5 1 1
4.
1 −3 −5
A−1 =
29 7 2
5. ⎛ ⎞
5 3 1
1 ⎝
A−1 = −8 −24 24⎠
32
−2 −14 6
6. ⎛ ⎞
−10 −10 0
1 ⎝
A−1 = −11 −95 36⎠
120
3 15 12
7. ⎛ ⎞
−1 25 −21
1 ⎝
A−1 = −8 −3 6 ⎠
29
−1 −4 8
8. ⎛ ⎞
9 35 5
1 ⎝
A−1 = 0 119 0 ⎠
119
−4 77 11
9. ⎛ ⎞
210 −42 42 0
1 ⎜⎜ 899 −124 223 −135⎟
⎟
A−1 =
378 ⎝ 275 −64 109 −27 ⎠
−601 122 −131 81
10. ⎛ ⎞
−52 131 −62 54
1 ⎜
⎜ 208 −132 248 −216⎟
⎟
A−1 =
784 −496 360 −320 304 ⎠
⎝
−212 127 −102 190
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
8.5. CRAMER’S RULE 199
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
200 CHAPTER 8. DETERMINANTS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
8.6. THE MATRIX TREE THEOREM 201
5. ⎛ ⎞
3 −1 0 0 −1 −1
⎜−1 3 −1 0 −1 0 ⎟
⎜ ⎟
T=⎜ ⎟
⎜ 0 −1 4 −1 −1 −1⎟
⎝0 0 −1 2 −1 0 ⎠
−1 0 −1 0 0 2
and each cofactor is equal to 61.
6. The tree matrix for the complete graph Kn is
⎛ ⎞
n−1 −1 −1 ··· −1
⎜ −1 n − 1 −1 ··· −1 ⎟
⎜ ⎟
⎜ ⎟
T = ⎜ −1 −1 n−1 ··· −1 ⎟.
⎜ .. .. .. .. ⎟
⎝ . . . ··· . ⎠
−1 −1 −1 ··· n−1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
202 CHAPTER 8. DETERMINANTS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 9
Eigenvalues and
Diagonalization
2.
pA (λ) = λ2 − 2λ − 8,
0 6
λ1 = 4, V1 = , λ2 = −2, V2 = .
4 −1
The Gershgorin circle is of radius 1 about (4, 0). The other Gershgorin
”circle” has radius 0 and so is not really a circle. We may think of this as
a degenerate circle containing the eigenvalue −2 in its interior.
3.
pA (λ) = λ2 + 3λ − 10,
7 0
λ1 = −5, V1 = , λ2 = 2, V2 = .
−1 1
The Gershgorin circle has radius 1 and center (2, 0).
203
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
204 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
4.
pA (λ) = λ2 − 10λ + 18,
√ 2√ √ 2√
λ1 = 5 + 7, V1 = , λ2 = 5 − 7, V2 = .
1− 7 1+ 7
The Gershgorin circles have radius 2, center (6, 0) and radius 3, center
(4, 0).
5. pA (λ) = λ2 − 3λ + 14,
√
√ −1 + 47i
λ1 = (3 + 14i)/2, V1 =
4
√
√ −1 − 14i
λ2 = (3 − 14i)/2, V2 = .
4
The Gershgorin circles have radius 6, center (1, 0) and radius 2, center
(2, 0).
6.
pA (λ) = λ2 ,
with roots λ1 = λ2 = 0. The only eigenvectors are nonzero scalar multiples
of
1
V1 = .
0
The Gershgorin circle has radius 1, center the origin.
7.
pA (λ) = λ3 − 5λ2 + 6λ,
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0 2 0
λ1 = 0, V1 = ⎝1⎠ , λ2 = 2, V2 = ⎝1⎠ , λ3 = 3, V3 = ⎝2⎠ .
0 0 3
The Gershgorin circle has radius 3, center the origin.
8.
pA (λ) = (λ + 1)(λ2 − λ − 7),
⎛ ⎞ ⎛ ⎞
0 √ 2√
λ1 = 1, V1 = ⎝0⎠ , λ2 = (1 + 29)/2, V2 = ⎝5 + 29⎠ ,
1 0
⎛ ⎞
√ 2√
λ3 = (1 − 29)/2, V3 = ⎝5 − 29⎠ .
0
The Gershgorin circles have radius 1, center (−2, 0), and radius 1, center
(3, 0).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.1. EIGENVALUES AND EIGENVECTORS 205
9.
pA (λ) = λ2 (λ + 3),
⎛ ⎞ ⎛ ⎞
1 1
λ1 = −3, V1 = ⎝0⎠ , λ2 = λ3 = 0, V2 = ⎝0⎠ .
0 3
There is only one independent eigenvector associated with eigenvalue 0.
The Gershgorin circle has radius 2, center (−3, 0).
10.
pA (λ) = λ3 + 2λ,
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0 √ 1 √ 1
λ1 = 0, V1 = ⎝1⎠ , λ2 = 2i, V2 = ⎝ −1 √
⎠ , λ3 = − 2i, V3 = ⎝ −1 ⎠ .
√
0 −2 2i 2i
The Gershgorin circles have center (0, 0) and radii 1 and 2.
11.
pA (λ) = (λ + 14)(λ − 2)2 ,
⎛ ⎞
−16
λ1 = −14, V1 = ⎝ 0 ⎠
1
⎛ ⎞
0
λ2 = λ3 = 2, V2 = ⎝0⎠ ,
1
with only one independent eigenvector associated with the multiple eigen-
value λ2 . The Gershgorin circles have radius 1, center (−14, 0) and radius
3, center (2, 0).
12.
pA (λ) = (λ − 3)(λ2 + λ − 42),
⎛
⎞ ⎛ ⎞ ⎛ ⎞
0 30 0
λ1 = 6, V1 = ⎝ 1 ⎠ , λ2 = 3, V3 = ⎝−2⎠ , λ3 = −7, V3 = ⎝8⎠ .
−1 5 5
The Gershgorin circles have radius 9, center (−2, 0), and radius 5, center
(1, 0).
13.
pA (λ) = λ(λ2 − 8λ + 7),
⎛⎞ ⎛ ⎞ ⎛ ⎞
14 6 0
λ1 = 0, V1 ⎝ 7 ⎠ , λ2 = 1, V2 = ⎝0⎠ , λ3 = 7, V3 = ⎝0⎠
10 5 1
The Gershgorin circles have radius 2, center (1, 0) and radius 5, center
(7, 0).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
206 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
14.
pA = λ2 (λ2 + 2λ − 1),
λ1 = λ2 = 0,
with two associated independent eigenvectors
⎛ ⎞ ⎛ ⎞
1 0
⎜2⎟ ⎜0⎟
V1 = ⎝ ⎠ , V 2 = ⎝ ⎟
⎜ ⎟ ⎜ .
0 1⎠
−1 0
0 0
The Gershgorin circles have radius 1, center (−2, 0) and radius 2, center
(0, 0).
15.
pA (λ) = (λ − 1)(λ − 2)(λ2 + λ − 13),
⎛ ⎞ ⎛ ⎞
−2 0
⎜−11⎟ ⎜0⎟
λ1 = 1, V1 = ⎜ ⎟ ⎜ ⎟
⎝ 0 ⎠ , λ2 = 2, V2 = ⎝1⎠ ,
1 0
⎛√ ⎞ ⎛ √ ⎞
√ 53 − 7 √ − 53 − 7
−1 + 53 ⎜ ⎟ ⎜ ⎟
λ3 = , V3 = ⎜
0 ⎟ , λ4 = −1 − 53 , V4 = ⎜ 0 ⎟.
2 ⎝ 0 ⎠ 2 ⎝ 0 ⎠
2 2
The Gershgorin circles have radius 2, center (−4, 0) and radius 1 and
center (3, 0).
16.
pA (λ) = λ2 (λ − 1)(λ − 5),
⎛ ⎞ ⎛ ⎞
1 1
⎜−4⎟ ⎜0⎟
λ1 = 1, V1 = ⎜ ⎟ ⎜ ⎟
⎝ 0 ⎠ , λ2 = 5, V2 = ⎝0⎠ ,
0 0
⎛ ⎞
0
⎜0⎟
λ3 = λ4 = 0, V3 = ⎜ ⎟
⎝1⎠ .
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.1. EIGENVALUES AND EIGENVECTORS 207
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
208 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
√
(λ − 2)(λ2 − 2λ − 2), so eigenvalues are λ1 = 2, λ2 = 1 +
22. pA (λ) = √ 3,
λ3 = 1 − 3. Corresponding eigenvectors are
⎛ ⎞ ⎛ √ ⎞ ⎛ √ ⎞
0 −1 + 3 −1 − 3
V1 = ⎝−1⎠ , V2 = ⎝ 1 ⎠ and V3 = ⎝ 1 ⎠.
1 1 1
9.2 Diagonalization
1.
pA (λ) = λ2 − 3λ + 4
√
is the√characteristic polynomial, with roots λ1 = (3 + 7i)/2 and λ2 =
(3 − 7i)/2. Corresponding eigenvectors are
√ √
−3 + 7i −3 − 7i
V1 = and V2 = .
8 8
The matrix √ √
−3 + 7i −3 − 7i
P=
8 8
diagonalizes A and
√
(3 + 7i)/2 0
√
P−1 AP = .
0 (3 − 7i)/2
2.
pA (λ) = λ2 − 8λ + 12,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.2. DIAGONALIZATION 209
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
210 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
Then ⎛ ⎞
0 0 0
P−1 AP = ⎝0 5 0 ⎠.
0 0 −2
6.
pA (λ) = λ(λ − 3λ − 2),
√ √
so the eigenvalues are λ1 = 0, λ2 = (3 + 17)/2 and λ3 = (3 − 17)/2.
Corresponding eigenvectors are
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
−2 0 0
V1 = ⎝−3⎠ , V2 = ⎝ 4√ ⎠ and V3 = ⎝ 4√ ⎠ .
1 3 + 17 3 − 17
Let ⎛ ⎞
−2 0 0
P = ⎝−3 4√ 4√ ⎠ .
1 3 + 17 3 − 17
Then ⎛ ⎞
0 √0 0
P−1 AP = ⎝0 (3 + 17)/2 0
√
⎠.
0 0 (3 − 17)/2
7.
pA (λ) = (λ + 2)2 (λ − 1),
so eigenvalues and corresponding eigenvectors are
⎛ ⎞ ⎛ ⎞
0 −3
λ1 = 1, V1 = ⎝1⎠ , λ2 = λ3 = −2, V2 = ⎝ 1 ⎠ .
0 0
Let ⎛ ⎞
1 0 0
P = ⎝0 1 1⎠
0 i −i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.2. DIAGONALIZATION 211
and then ⎛ ⎞
2 0 0
P−1 AP = ⎝0 2 + i 0 ⎠.
0 0 2−i
9.
pA (λ) = (λ − 1)(λ − 4)(λ2 + 5λ + 5),
√
√ and eigenvectors are λ1 = 1, λ2 = 4, λ3 = (−5 + 5)/2 and
so eigenvalues
λ4 = (−5 − 5)/2. Corresponding eigenvectors are
⎛ ⎞ ⎛ ⎞
1 0
⎜0⎟ ⎜1⎟
V1 = ⎜ ⎟ ⎜ ⎟
⎝0⎠ , V2 = ⎝0⎠ ,
0 0
⎛ ⎞ ⎛ ⎞
0
√ 0
√
⎜(2 − 3 5)/41⎟ ⎜(2 + 3 5)/41⎟
V3 = ⎜ √ ⎟ ⎜ √ ⎟
⎝ (−1 + 5)/2 ⎠ and V4 = ⎝ (−1 − 5)/2 ⎠ .
1 1
Let ⎛ ⎞
1 0 0
√ 0
√
⎜0 1 (2 − 3 √5)/41 (2 + 3 √5)/41⎟
P=⎜
⎝0
⎟.
0 (−1 + 5)/2 (−1 − 5)/2 ⎠
0 0 1 1
Then ⎛ ⎞
1 0 0 0
⎜0 4 0√ 0 ⎟
−1 ⎜
P AP = ⎝ ⎟.
0 0 (−5 + 5)/2 0√ ⎠
0 0 0 (−5 − 5)/2
10.
pA (λ) = (λ + 2)4 ,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
212 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
Ak = (PDP−1 )k
= (PDP−1 )(PDP−1 ) · · · (PDP−1 )
= PDk P−1 ,
Compute √
1 7 − √73 6
P−1 = √ .
12 73 −7 − 73 −6
Then √
16 ((1 + 73)/2)16 √0
A =P P−1
0 ((1 − 73)/2)16
6(216 ) − 316 3(216 ) − 317
= .
−217 + 2(316 ) −216 + 6(316 )
13.
pA (λ) = λ2 + 6λ + 5,
so the eigenvalues are −1 and −5. Form P using corresponding eigenvec-
tors as columns:
4 0
P= .
1 1
Then
A18 = PAP−1
4 0 1 0 1/4 0
=
1 1 0 518 −1/4 1
1 0
= .
(1 − 518 )/4 518
√ √
14. A has eigenvalues −3 + 10 and −3 − 10. Form P using corresponding
eigenvectors as columns:
3√ 3√
P= .
1 − 10 1 + 10
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.2. DIAGONALIZATION 213
Then √ √
(10 + √10)/60 −√ 10/20
P−1 = .
(10 − 10)/60 10/20
Then √ 31
(−3 + 10) 0√
A31 = P P−1
0 (−3 − 10)31
a b
= ,
c d
where
1
√ √ √ √
a= √ (1 + 10)(−3 + 10)31 + ( 10 − 1)(−3 − 10)31 ,
2 10
3
√ √
b= √ (−3 + 10)31 + (3 + 10)31 ,
2 10
1
√ √ √ √
c= √ (−1 + 10)(−3 + 10)31 + ( 10 + 1)(3 + 10)31
2 10
3
√ √
d= √ (−3 + 10)31 + (3 + 10)31 .
2 10
√ √
15. Eigenvalues of A are λ1 = 2 and λ2 = − 2, with corresponding eigen-
vectors √ √
2 − 2
V1 = , V2 = .
1 1
Let √ √
2 − 2
P= .
1 1
We find that √
P−1 = √2/4 1/2 .
− 2/4 1/2
Then
√ √ √ 43 √
43 2 − 2 ( 2) √0 √2/4 1/2
A =
1 1 0 (− 2)43 − 2/4 1/2
0 222
= .
221 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
214 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
for j = 1, 2, · · · , n. Then
pA2 (λ) = (A − λj In )(A + λj In )
= pA ( λj )pA (− λj ) = O
for j = 1, 2, · · · , n. Then
pA ( λj ) = 0 or pA (− λj ) = 0.
But this means that λj or − λj is an eigenvalue of A with associated
eigenvector Xj . This implies that A has n linearly independent eigenvec-
tors and is therefore diagonalizable.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.3. SOME SPECIAL MATRICES 215
5. Eigenvectors are
⎛ ⎞ ⎛ √ ⎞ ⎛ √ ⎞
0 1+ 2 1− 2
V1 = ⎝0⎠ , V2 = ⎝ 1 ⎠ , and V3 = ⎝ 1 ⎠ .
1 0 0
6. Eigenvectors are
⎛ ⎞ ⎛ √ ⎞ ⎛ √ ⎞
0 −1 + 3 −1 − 3
V1 = ⎝−1⎠ , V2 = ⎝ 1 ⎠ and V3 = ⎝ 1 ⎠.
1 1 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
216 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
7.
pA (λ) = λ(λ2 − 5λ − 4)
√ √
and A has eigenvalues 0, (5 + 41)/2 and 5 − 41)/2, with corresponding
eigenvectors
⎛ ⎞ ⎛ √ ⎞ ⎛ √ ⎞
0 5 + 41 5 − 41
V1 = ⎝1⎠ , V2 = ⎝ 0 ⎠ , V3 = ⎝ 0 ⎠.
0 4 4
8.
pA (λ) = λ(λ2 − 2λ − 16)
√ √
so 0, 1 + 17 and 1 − 17 are eigenvalues. We find the corresponding
eigenvectors
⎛ ⎞ ⎛ √ ⎞ ⎛ √ ⎞
0 1 + 17 1 − 17
V1 = ⎝0⎠ , V2 = ⎝ −4 ⎠ , V3 = ⎝ −4 ⎠ .
1 0 0
Then
⎛ √ √ √ √ ⎞
0 (1 + 17)/
34 + 2 17 (1 − 17)/
34 − 2 17
⎜ √ √ ⎟
Q = ⎝0 −4/ 34 + 2 17 −4/ 34 − 2 17 ⎠.
1 0 0
9.
pA (λ) = λ(λ2 − λ − 4)
√ √
and the eigenvalues are 0, (1 + 17)/2, (1 − 17)/2. Corresponding
eigenvectors are
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
1 0√ 0√
⎝0⎠ , V2 ⎝−1 − 17⎠ , V3 = ⎝−1 − 17⎠ .
0 4 4
Then
⎛ ⎞
1 0 0
⎜ √ √ √ √ ⎟
Q = ⎝0 (−1 − 17)/ 34 + 2 17 (−1 + 17)/ 34 + 2 17⎠ .
√ √
0 4/ 34 + 2 17 4/ 34 + 2 17
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.3. SOME SPECIAL MATRICES 217
10.
pA (λ) = (λ − 1)(λ2 − λ − 10)
√ √
and the eigenvalues are 1, (1 + 41)/2 and (1 − 41)/2. Corresponding
eigenvectors are
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
1 6√ 6√
V1 = ⎝ 0 ⎠ , V2 = ⎝−1 + 41⎠ , V3 = ⎝−1 − 41⎠ .
−3 2 2
Then
⎛ √ √ √ ⎞
1/ 10 6/ 82 −
2 41 √ 6/ 82 +
2 41 √ ⎟
⎜ √ √
Q=⎝ 0 (−1 + 41)/ 82 − 2 41 (−1 − 41)/ 82 − 2 41⎠ .
√ √ √
−2/ 10 2/ 82 − 2 41 2/ 82 − 2 41
11.
pA (λ) = λ2 (λ2 − 2λ − 3)
so the eigenvalues are 0, 0, −1 and 3. Corresponding eigenvectors are
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
1 0 0 0
⎜ 0⎟ ⎜0⎟ ⎜1⎟ ⎜−1⎟
V 1 = ⎝ ⎠ , V 2 = ⎝ ⎠ , V3 = ⎝ ⎠ , V 4 = ⎝ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ .
0 0 1 1⎠
0 1 0 0
Then ⎛ ⎞
1 0 0√ 0√
⎜0 0 1/ 2 −1/√ 2⎟
Q=⎜ √
⎝0 0 1/ 2
⎟.
1/ 2 ⎠
0 1 0 0
12.
pA = λ(λ − 5)(λ2 − 1)
and the eigenvalues are 0, 5, 1 and −1. Corresponding eigenvectors are
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0 1 0 0
⎜ 0⎟ ⎜0⎟ ⎜1⎟ ⎜1⎟
V 1 = ⎝ ⎠ , V 2 = ⎝ ⎠ , V4 = ⎝ ⎠ , V 4 = ⎝ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ .
0 0 −1 1⎠
1 0 0 0
Then ⎛ ⎞
0 1 0√ 0√
⎜0 0 1/ 2 1/ 2⎟
Q=⎜ √ √ ⎟
⎝0 0 −1/ 2 1/ 2⎠ .
1 0 0 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
218 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
pA (λ) = (λ − 2)2 .
pA (λ) = (λ + 1)2
St = −S.
ps (λ) = λ(λ2 + 3)
√ √
so the eigenvalues are 0, 3i and − 3i, with corresponding eigenvectors
⎛ ⎞⎛ ⎞ ⎛ ⎞
2 √1 √1
⎝ 0 ⎠ ⎝ 3i ⎠ , and ⎝ − 3i ⎠ .
1 + i, −1 − i −1 − i
Let P have these eigenvectors as columns (in the given order). Then
⎛ ⎞
0 √0 0
P−1 SP = ⎝0 3i √0 ⎠.
0 0 − 3i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.3. SOME SPECIAL MATRICES 219
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
220 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
y12 + 6y22 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
9.3. SOME SPECIAL MATRICES 221
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
222 CHAPTER 9. EIGENVALUES AND DIAGONALIZATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 10
Systems of Linear
Differential Equations
Therefore these solutions are linearly independent. We can form the fun-
damental matrix using these solutions as columns:
2t
−e 3e6t
Ω(t) = .
e2t e6t
In terms of the fundamental matrix, the general solution of the system is
X(t) = Ω(t)C, where
c
C= 1 .
c2
To satisfy the initial condition x1 (0) = 0, x2 (0) = 4, solve for C in X(0) =
Ω(0)C, which is
0 −1 3 c1
= .
4 1 1 c2
223
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
224 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Then
0 −1 0
= Ω (0)
4 4
1 1 −3 0 3
=− = .
4 −1 1 4 1
The solution of the initial value problem is
3 −3e2t + 3e6t
X(t) = Ω(t) = .
1 3e3t + e6t
A fundamental matrix is
e4t cos(t) e4t sin(t)
Ω(t) = .
2e4t (cos(t) − sin(t)) 2e4t (cos(t) − sin(t))
Note that |Ω(0)| = 2 = 0. The general solution is X(t) = Ω(t)C. For the
solution of the initial value problem, choose
−1 1 2 0 −2 −2
C = Ω (0)X(0) = = .
2 −2 1 1 5/2
A fundamental matrix is
√ √
4e(1+2 3)t √ 4e(1−2 3)t √
Ω(t) = √ (1+2 3)t √ .
(−1 + 3)e (−1 − 3)e(1−2 3)t
√
Notice that |Ω(0)| = −8 3 = 0. The general solution is X(t) = Ω(t)C
For the initial value problem, choose
√
1 −1 −√ 3 −4 2
C = Ω−1 (0)X(0) = − √
8 3 1− 3 4 2
√
1 3+5 3
= √ .
12 3 − 5 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.1. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS 225
The unique solution of the initial value problem is (after some manipula-
tion),
t √ √ √
2e cosh(2 √3t) + (10/√ 3)et sinh(2√ 3t)
X(t) = Ω(t)C = .
2et cosh(2 3t) − (1/ 3)et sinh(2 3t)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
226 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
A fundamental matrix is
7e3t 0
Ω(t) = .
5e3t e−4t
For Problems 2 through 5, we give the solution in the form X(t) = Ω(t)C,
with Ω(t) a fundamental matrix. Note that we can read the eigenvalues and
corresponding eigenvectors of the coefficient matrix from the fundamental ma-
trix.
2.
2et e6t c1 2c1 et + c2 e6t
X(t) = Ω(t)C = =
−3et e6t c2 −3c1 et + c2 e6t
3.
1 e2t c1 c1 + c2 e2t
X(t) = Ω(t)C = =
−1 e2t c2 −c1 + c2 e2t
4.
⎛ ⎞⎛ ⎞ ⎛ ⎞
et e−t e2t c1 c1 et + c2 e−t + c3 e2t
X(t) = Ω(t)C = ⎝et e−t 2e 2t ⎠ ⎝ ⎠
c2 = ⎝c1 e + c2 e + 2c3 e ⎠
t −t 2t
5. ⎛ ⎞⎛ ⎞
1 2e3t −e−4t c1
X(t) = Ω(t)C = ⎝ 6 3e3t 2e−4t ⎠ ⎝c2 ⎠
−13 −2e3t e−4t c3
⎛ 3t −4t
⎞
c1 + 2c2 e − c3 e
= ⎝ 6c1 + 3c2 e3t + 2c3 e−4t ⎠
−13c1 − 2c2 e3t + c3 e−4t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.2. SOLUTION OF X = AX FOR CONSTANT A 227
6.
t −1 t
2e e−t 2 1 7 4e + 3e−t
X(t) = =
et e−t 1 1 5 2et + 3e−t
7.
−1
2e4t e−t 2 1 1 6e4t − 5e−3t
X(t) = =
−3e4t 2e−3t −3 2 −19 −9e4t − 10e−3t
8.
−3t −1
2e −5e4t 2 −5 −3 1 54e−3t − 75e4t
X(t) = =
e−3t e4t 1 1 6 7 27e−3t + 15e4t
9.
⎛ ⎞⎛ ⎞−1 ⎛ ⎞
0 e2t 3e3t 0 1 3 1
⎝
X(t) = 1 e2t e3t ⎠ ⎝1 1 1⎠ ⎝5⎠
1 0 e3t 1 0 1 1
⎛ ⎞
4e2t − 3e3t
= ⎝2 + 4e2t − e3t ⎠
2 − e3t
10.
⎛ t ⎞⎛ ⎞−1 ⎛ ⎞
e e−t −e−3t 1 1 −1 1
X(t) = ⎝et 3e−t 3e−3t ⎠ ⎝1 3 3 ⎠ ⎝7⎠
et 3e−t −e−3t 1 3 −1 3
⎛ ⎞
et + e−t − e−3t
= ⎝et + 3e−t + 3e−3t ⎠
et + 3e−t − e−3t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
228 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
13. Eigenvalues of A are roots of λ2 −2λ+2, and are λ1 = 1+i and λ2 = 1−i,
with corresponding eigenvectors
5 5
and .
2−i 2+i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.2. SOLUTION OF X = AX FOR CONSTANT A 229
Divide by e2t . Further, AE1 = 2E1 , so two terms in the last equation
cancel. This leaves
E1 + 2E2 = AE2 .
The unknown here is
e
E2 = 1 .
e2
This system of equations reduces to
2e1 = 2e1
1 + 2e2 = 5e1 + 2e2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
230 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
In Problems 17 through 21, we omit some of the details given in the solution
of Problem 16.
A fundamental matrix is
e3t 2te3t
Ω(t) = .
0 e3t
A fundamental matrix is
⎛ ⎞
et 5tet 0
Ω(t) = ⎝ 0 et 0⎠.
4tet (10t + 8t)et
2
et
A fundamental matrix is
⎛ ⎞
e2t 3e5t 27te5t
Ω(t) = ⎝ 0 3e5t (3 + 27t)e5t ⎠ .
0 −e5t (2 − 9t)e5t
20. The coefficient matrix has a multiplicity 2 eigenvalue i, with single eigen-
vector ⎛ ⎞
i
⎜−1⎟
⎜ ⎟
⎝ −i ⎠
1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.3. SOLUTION OF X = AX + G 231
A fundamental matrix is
⎛ ⎞
cos(t) cos(t) + t sin(t) sin(t) sin(t) − t cos(t)
⎜ − sin(t) t cos(t) cos(t) t sin(t) ⎟
⎜
Ω(t) = ⎝ ⎟.
− cos(t) cos(t) − t sin(t) − sin(t) sin(t) + t cos(t) ⎠
sin(t) −t cos(t) − 2 sin(t) − cos(t) −t sin(t) + 2 cos(t)
10.3 Solution of X = AX + G
For a linear system of diffeential equations, a fundamental matrix is not unique,
and different fundamental matrices may be derived using different methods.
The general solution can be written using any fundamental matrix.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
232 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Compute
1 − 2t −2t
Ω−1 (t) = e−3t .
2t 1 + 2t
Now compute a particular solution of the given nonhomogeneous system
as
1 − 2t −2t −3et
u(t) = Ω−1 (t)G(t) dt = e−3t dt
2t 1 + 2t e3t
−2t
6te − 3e−2t − 2t −3te−2t − t2
= dt = .
−6te−2t + 1 + 2t (3/2)(1 + 2t)e−2t + t + t2
The general solution is
A fundamental matrix is
2 1 + 2t
Ω(t) = .
1 t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.3. SOLUTION OF X = AX + G 233
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
234 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
In Problems 6 through 9, where initial values are given, the method is to find
the general solution of the system and then solve for the constants to satisfy the
initial values. For these four problems only the solution is given.
6.
−1 + e2t
X(t) =
−5t + (3 + 5t)e2t
7.
(−1 − 14t)et
X(t) =
(3 − 14t)et
8. ⎛ ⎞
13t − (8 + 12t + 3t2 )e2t
X(t) = ⎝ 4et + (7 + 2t)e2t ⎠
−et − e2t
9. ⎛ ⎞
(6 + 12t + (1/2)t2 )e−2t
X(t) = ⎝ (2 + 12t + (1/2)t2 )e−2t ⎠
(3 + 38t + 66t2 + (13/6)t3 )e−2t
For the remaining problems, the solution is expressed in the form X(t) =
PZ(t), where P is a matrix having eigenvectors of A as columns, A(t) is the
solution of the uncoupled system Z = DZ + P−1 G, and D is a diagonal matrix
having the eigenvalues of A on its main diagonal.
We find that
1 4 −1
P−1 = .
3 −1 1
Then
−1 0
P−1 AP =
0 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.3. SOLUTION OF X = AX + G 235
and Z satisfies
−1 0 1 4 −1 0
Z = Z+ .
0 2 3 −1 1 10 cos(t)
Then
c1 et + c2 e−2t − 3 cos(t) − sin(t)
X(t) = PZ = .
c1 et + 4c2 e−2t − 7 cos(t) + sin(t)
Then
1 1 −1
P−1 =
4 1 3
and the uncoupled system for Z is
2 0 1 1 −1 8
Z = Z+ .
0 6 4 1 3 4e3t
Then
3c1 e2t + c2 e6t − 4e3t − 10/3
X(t) = PZ(t) = .
−c1 e2t + c2 e6t + 2/3
12. Here
1 1
A= ,
1 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
236 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Then
1 1 −1
P−1 =
2 1 1
and the uncoupled system for Z is
3t
0 0 1 1 −1 6e
Z = Z+ .
0 2 2 1 1 4
Then
c1 + 2c2 e2t − 1 − 2t + 4e3t
X(t) = PZ(t) = .
−c1 + c2 e2t − 1 + 2t + 2e3t
Then
1 1 −5
P−1 =
6 1 1
and the system for Z is
1 0 1 1 −5 −4 cos(3t)
Z = Z+ .
0 7 6 1 1 8
Then
X(t) = PZ(t) =
t 7t
c1 e + 5c2 e + (68/145) cos(3t) − (54/145) sin(3t) + 40/7
.
−c1 et + c2 e7t + (2/145) cos(3t) + (24/145) sin(3t) − 48/7
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.3. SOLUTION OF X = AX + G 237
We find that
1 −3i 1+i
P−1 = .
6 3i 1−i
Z satisfies
2t
3i 0 1 −3i 1+i 3e
Z = Z+ .
0 −3i 6 3i 1−i e2t
Write
1 1
(c1 + c2 i) and d2 = (c1 − c2 i)
d1 =
2 2
with c1 and c2 real, to obtain
(1/2)(c1 cos(3t) − c2 sin(3t)) + (1/2)(c2 cos(3t) + c1 sin(3t))i + ((2 − i)/6)e2t
Z(t) = .
(1/2)(c1 cos(3t) − c2 sin(t)) − (1/2)(c2 cos(3t) + c1 sin(3t))i + ((2 + i)/6)e2t
Then
X(t) = PZ(t) =
c1 (cos(3t) − sin(3t)) − c2 (sin(3t) + cos(3t)) + e2t
.
3c1 cos(3t) − 3c2 sin(3t) + 2e2t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
238 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Then
2 + 4(1 + t)e2t
X(t) = PZ(t) = .
−2 + 2(1 + 2t)e2t
16. With
1 −2
A=
−1 2
we find the eigenvalues 0 and 3 and, from corresponding eigenvectors,
2 −1
P= .
1 1
Then
1 1 1
P−1 = .
3 −1 2
The uncoupled system is
0 0 1/3 1/3 2t
Z = Z+ ,
0 3 −1/3 2/3 5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.3. SOLUTION OF X = AX + G 239
18. The eigenvalues of the coefficient matrix are 1, 1, −3. Form P having
independent eigenvectors as columns:
⎛ ⎞
1 −1 1
P = ⎝1 0 3⎠ .
0 1 1
We find that ⎛ ⎞
3 −2 3
P−1 =⎝ 1 −1 2 ⎠.
−1 1 −1
With X = PZ we obtain the uncoupled system
⎛ ⎞ ⎛ ⎞⎛ ⎞
1 0 0 3 −2 3 −3e−3t
Z = ⎝0 1 0 ⎠ Z + ⎝ 1 −1 2 ⎠⎝ t ⎠.
0 0 −3 −1 1 −1 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
240 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.4. EXPONENTIAL MATRIX SOLUTIONS 241
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
242 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Then
∞
1 −1
eBt = (P AP)n t
n=0
n!
∞
1 −1 n
= P A Pt
n=0
n!
∞
1
−1 n
=P ( A t P
n=0
n!
= P−1 eAt P.
9. First deal with the matrix A of Problem 1. The eigenvalues, with corre-
sponding eigenvectors, are
1 − 2i 1 + 2i
2i, , −2i, .
5 5
The matrix
1 − 2i 1 + 2i
P=
5 5
diagonalizes A, so
−1 2i 0
P AP = D = .
0 −2i
Now,
e2it 0
eDt = .
0 e−2it
Further, we find that
−1 (1/4)i 1/10 − i/20
P = .
−(1/4)i 1/10 + i/20
Then
eAt = PeDt P−1
2it
1 − 2i 1 + 2i e 0 (1/4)i 1/10 − i/20
=
5 5 0 e−2it −(1/4)i 1/10 + i/20
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.5. APPLICATIONS AND ILLUSTRATIONS OF TECHNIQUES 243
(1/2 + i/4)e2it + (1/2 − i/4)e−2it (i/4)(e−2it − e2it )
= .
(5i/4)(e2it − e−2it ) (1/2 − i/4)e2it + (1/2 + i/4)e−2it
This appears to be different from the solution obtained using MAPLE.
However, recall that
If these are substituted into the exponential matrix we have just found,
we obtain the exponential matrix produced by MAPLE.
Now turn to the matrix of Problem 2. Eigenvalues and eigenvectors are
−1 1
−3, , 0, .
1 2
Let
−1 1
P= .
1 2
Then
−3 0
P−1 AP = D = .
0 0
Now
e−3t 0
eDt = .
0 1
Then
1 1 + 2e−3t 1 − e−3t
PeDt P−1 = −3t .
3 2 − 2e−3t 2+e
Because only real quantities were involved in the computation, we obtain
the same result as that returned by MAPLE.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
244 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
volts.
2. Denote the amount (in pounds) of salt in tank j at time t by xj (t). Then
x x 1
1 2
x1 = −16 + 12 + (4),
x 200 300
x 4
1 2
x2 = 12 − 18 .
200 300
Together with the given initial conditions, we now have the initial value
problem
4 2
x1 = − x1 + x2 + 1,
50 50
3 3
x2 = x1 − x2 ,
50 50
x1 (0) = 200, x2 (0) = 150.
4 1
x1 = − x1 + x2 + 1,
50 50
1 4
x2 = x1 − x2 + 2,
50 50
x1 (0) = 40, x2 (0) = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.5. APPLICATIONS AND ILLUSTRATIONS OF TECHNIQUES 245
From the first equation and the derivative of the third, we have
1 −1 i1 −50 0 i1 5
= + ,
1 0 i2 0 −20 i2 0
1 1
i1 (t) = − e−10t sin(30t),
10 15
1 −2t
i2 (t) = e (3 cos(30t) − sin(30t)).
30
5. Designate down as positive, y1 (t) the position of the upper weight relative
to the equilibrium position of this weight, and y2 (t) the position of the
lower weight relative to the equilibrium position of the lower weight. Then
x1 = x3 ,
x2 = x4 ,
x3 = −22x1 + 6x2 ,
x4 = 6x1 − 6x2 ,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
246 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
6 −6 0 0
√
with eigenvalues ±2i and ±2 6i. One eigenvector associated with 2i is
⎛ ⎞
1
⎜3⎟
⎜ ⎟
⎝2i⎠
6i
√
and an eigenvector associated with 2 6i is
⎛ ⎞
3
⎜ −1 ⎟
⎜ √ ⎟.
⎝ 6 6i ⎠
√
−2 6i
Use these to write the general solution of the system of differential equa-
tions in terms of real functions:
⎛ √ √ ⎞
c1 cos(2t) + c2 sin(2t) + 3c3 cos(2 √6t) + 3c4 sin(2√6t)
⎜ 3c1 cos(2t) + 3c2 sin(2t)√− c3 cos(2√6t) − c4√sin(2 6t)√ ⎟
X(t) = ⎜ ⎟
⎝2c2 cos(2t) − 2c1 sin(2t) + 6 6c4 cos(2 6t) − 6 6c3 sin(2 6t)⎠ .
√ √ √ √
6c2 cos(2t) − 6c1 sin(2t) − 2 6c4 cos(2 6t) + 2 6c3 sin(2 6t)
Substitute the initial conditions and recall that y1 = x1 and y2 = x2 to
obtain
2 3 √
y1 (t) = cos(2t) + cos(2 6t),
5 5
6 1 √
y2 (t) = cos(2t) − cos(2 6t).
5 5
6. Using the same assignment of variables as in the solution to Problem 5,
we have
y1 = −22y1 + 6y2 ,
y2 = 6y1 − 6y2 + 4 sin(3t),
y1 (0) = y2 (0) = y1 (0) = y2 (0) = 0.
Proceeding as in the solution to Problem 5, we obtain
√
9 3 6 √ 8
y1 (t) = sin(2t) + sin(2 6t) − sin(3t),
25 150
√ 25
27 6 √ 52
y2 (t) = sin(2t) − sin(2 6t) − sin(3t).
25 150 75
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.5. APPLICATIONS AND ILLUSTRATIONS OF TECHNIQUES 247
7. Consider the direction to the right as positive and let y1 be the displace-
ment of the left weight from the equilibrium position and y2 the displace-
ment of the right weight from its equilibrium position. The spring/mass
system is modeled by the initial value problem
2y1 = −8y1 + 5(y2 − y1 ),
2y2 = −5(y2 − y1 ) − 8y2 ,
y1 (0) = 1, y2 (0) = −1, y1 (0) = y2 (0) = 0.
As we have done before, let
x1 = y1 , x2 = y2 , x3 = y1 , x4 = y2 .
This gives us the first-order system
x1 = x3 ,
x2 = x4 ,
13 5
x3 = − x1 + x2 ,
2 2
5 13
x4 = x1 − x2 .
2 2
x1 (0) = 1, x2 (0) = −1, x3 (0) = x4 (0) = 0.
The matrix of this system has eigenvalues ±2i and ±3i. Eigenvectors
corresponding to 2i and one for 3i are, respectively,
⎛ ⎞ ⎛ ⎞
1 1
⎜1⎟ ⎜ ⎟
⎜ ⎟ and ⎜ −1 ⎟ .
⎝2i⎠ ⎝ 3i ⎠
2i −3i
Using these, write the general solution of the system:
⎛ ⎞
c1 cos(2t) + c2 sin(2t) + c3 cos(3t) + c4 sin(4t)
⎜ c1 cos(2t) + c2 sin(2t) − c3 cos(3t) − c4 sin(3t) ⎟
X(t) = ⎜ ⎟
⎝2c2 cos(2t) − 2c1 sin(2t) + 3c4 cos(3t) − 3c3 sin(3t)⎠ .
6c2 cos(2t) − 6c1 sin(2t) − 3c4 cos(3t) + 3c3 sin(3t)
Upon using the initial conditions and setting y1 = x1 and y2 = x2 we
obtain the solution for the displacement functions:
y1 (t) = cos(3t),
y2 (t) = − cos(3t).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
248 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Use the derivative of the first equation together with the third equation
to obtain the system
1 0 i1 −25 −25 i1 0
= + ,
0 2 i2 −8 0 i2 1
with the initial conditions given previously. Upon multiplying this system
on the left by
−1
1 0
,
0 2
which is the matrix
1 0
,
0 1/2
we obtain the system
i1 −25 25 i1 0
= + ,
i2 −4 0 i2 1/2
with the given initial conditions. This system has the solution
5 −20t 5 −5t 1
i1 (t) = e − e + ,
24 24 8
1 −20t 1 −5t 1
i2 (t) = e − e + .
24 6 8
9. From Kirchhoff’s laws,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.5. APPLICATIONS AND ILLUSTRATIONS OF TECHNIQUES 249
We obtain
i1 0 −20 0
= − ,
i2 5 −20 1/2
with the given initial conditions. The coefficient matrix of this system has
repeated eigenvalue −10, and only one independent eigenvector
2
,
1
We need to integrate
0 −(1/2)(1 + 10t)e10t
Ω−1 G(t) = Ω−1 = .
−1/2 e10t
Then
1/10
Ω(t)u(t) =
0
is a particular solution. The general solution of the nonhomogeneous
system is
1
i1 (t) = 2c1 e−10t + c2 (1 + 10t)e−10t + ,
10
i2 (t) = c1 e−10t + c2 te−10t .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
250 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
Upon inserting the initial conditions, we obtain the solution for the current
functions:
1
i1 (t) = − 2te−10t ,
10
1
i2 (t) = − t e−10t
10
amperes.
10. The circuit can be modeled using any three of the following six equations:
Using the derivative of the first equation, the second equation, and the
derivative of the third equation, we obtain the system
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞
2 0 0 i1 −5 5 0 i1 0
⎝0 2 −2⎠ ⎝i2 ⎠ = ⎝−4 −5 0⎠ ⎝i2 ⎠ + ⎝9⎠ ,
4 5 5 i3 0 0 0 i3 0
with the above initial conditions. The coefficient matrix of this system
has eigenvalues 0, −5/2, −9/4, with corresponding eigenvectors
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0 5 10
⎝0⎠ , ⎝ 0 ⎠ , ⎝ 1 ⎠ .
1 −4 −9
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.5. APPLICATIONS AND ILLUSTRATIONS OF TECHNIQUES 251
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
252 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
√
x1 (t) = 26 e−t (c1 − 2c2 ) cos(2t) + (2c1 + c2 ) sin(2t)
+ e−2t [(2c3 − c4 ) cos(t) + (c3 + 2c4 ) sin(t)] ,
x2 (t) = 5e−t [2c1 cos(2t) + 2c2 sin(2t)]
+ e−2t [c3 cos(t) + c4 sin(t)] ,
√
x3 (t) = 26e−t [(3c1 + 4c2 ) cos(2t)] + (4c1 + 3c2 ) sin(2t),
+ e−2t [(−3c3 + 4c4 ) cos(t) + (−4c3 − 3c4 ) sin(t)] ,
x4 (t) = 5e−t [(−2c1 + 4c2 ) cos(2t) + (−4c1 − 2c2 ) sin(2t)]
+ e−2t [(−2c3 + c4 ) cos(t) + (−c3 − 2c4 ) sin(t)] .
We also find the following solution of the nonhomogeneous system:
√ √
9 26 3 26
x1 (t) = − cos(t) + sin(t),
40 40
9 9
x2 (t) = − cos(t) + sin(t),
8
√ 8 √
3 26 9 26
x3 (t) = cos(t) + sin(t),
40 40
9 9
x4 (t) = cos(t) + sin(t).
8 8
Add this particular solution to the general solution of the associated ho-
mogeneous equation and then insert the initial conditions to solve for the
constants, obtaining
6 3 21 3
c1 = , c2 = , c3 = , c4 = − .
100 100 200 200
Upon inserting these constants and putting y1 = x1 and y2 = x2 we obtain
the displacement functions for the weights:
√
3 26 −t
y1 (t) = 2e sin(2t)
40
−2t
+e (3 cos(t) + sin(t)) − 3 cos(t) + sin(t)
3 −t
y2 (t) = e (8 cos(2t) + 4 sin(2t))
40
+e−2t (7 cos(t) − sin(t)) − 15 cos(t) + 15 sin(t) .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.6. PHASE PORTRAITS 253
y 0
-6 -4 -2 0 2 4 6
x
-2
-4
-6
2. The eigenvalues of A are −3, 4 and the origin is a saddle point. The
general solution is
−c1 e−3t + (4/3)c2 e4t
X(t) = .
c1 e−3t + c2 e4t
4. The eigenvalues are 3, 2 and the origin is a nodal source. The general
solution is
7c1 e3t + c2 e2t
X(t) = 3t 2t .
6c1 e + c2 e
A phase portrait consists of straight lines emanating from the origin.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
254 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
10
-10 0 10 20
0
-10
-20
y 0
-15 -10 -5 0 5 10 15
x
-2
-4
-6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.6. PHASE PORTRAITS 255
600
400
200
y 0
-800 -400 0 400 800
x
-200
-400
-600
5. The eigenvalues are 4 ± 5i and the origin is a spiral point. The general
solution is
(3c1 − 5c2 )e4t sin(5t) + (5c1 + 3c2 )e4t cos(5t)
X=
2c1 e4t sin(5t) + 2c2 e4t cos(5t)
Figure 10.4 is a phase portrait.
6. The eigenvalues are −3, −5 and the origin is a nodal sink. A phase portrait
consists of straight lines moving into the origin. The general solution is
7c1 e−3t + c2 e−5t
X(t) = .
5c1 e−3t + c2 e−5t
7. The eigenvalues are 3, 3 and the origin is an improper node. The general
solution is
c1 e3t + c2 te3t
X=
(c1 + c2 )e3t + c2 te3t
√
8. The eigenvalues of the coefficient matrix are ± 31i, so the origin is a
center, with periodic closed orbits enclosing (0, 0). A phase portrait for
this system is shown in Figure 10.5.
The general solution is
√ √ √ √
(3c1 − 31) sin( √31t) + ( 31c1 +√3c2 ) cos( 31t)
X(t) =
8c1 sin( 31t) + 8c2 cos( 31t)
√
9. The eigenvalues are −2 ± 3i, so the origin is a spiral point. The general
solution is √ √
c1 e−2t cos(√ 3t) − c2 e−2t sin( √3t)
X=
c1 e−2t sin( 3t) + 3c2 e−2t cos( 3t)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
256 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
y 0
-1.5 -1 -0.5 0 0.5 1 1.5
x
-1
-2
400
200
y 0
-400 -200 0 200 400
x
-200
-400
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.6. PHASE PORTRAITS 257
y3
10. The eigenvalues are −13, −13 and the origin is an improper node. The
general solution is
(c1 + c2 t)e−13t
X(t) = .
(c1 + c2 t − (1/7)c2 )e−13t
11. Let H be the constant of proportionality for the outside agent that at
any time removes members of both species at a rate proportional to their
population at that time. Coupling this term with a predator/prey model,
we have the system.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
258 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
3
y
0 1 2 3 4 5
x
3
y
0
0 1 2 3 4
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.6. PHASE PORTRAITS 259
30
y 20
10
0
0 0.4 0.8 1.2
x
10
0
8 12 16 20 24
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
260 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
16
12
y 8
0
0 2 4 6 8 10
x
10
0
0 10 20 30 40 50 60 70
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.6. PHASE PORTRAITS 261
50
45
40
35
y
30
25
20
15
0 5 10 15 20
x
25
20
15
y
10
0 5 10 15 20
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
262 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
25
20
15
y
10
0 5 10 15 20
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
10.6. PHASE PORTRAITS 263
25
20
15
y
10
0
4 8 12 16 20
x
25
20
15
y
10
0 5 10 15 20
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
264 CHAPTER 10. SYSTEMS OF LINEAR DIFFERENTIAL EQUATIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 11
1. First, applying the ”product rule” for a scalar function times a vector
function,
so
2. First,
265
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
266 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
so
(F(t) · G(t)) = 1 − 6t cos(t) + 3t2 sin(t).
To carry out the cross product and then differentiate, first compute
i j k
F(t) × G(t) = t 1 4
1 − cos(t) t
= (t + 4 cos(t))i + (4 − t2 )j − (t cos(t) + 1)k.
Then
4.
5.
6.
(F(t) · G(t)) = sin(t) + t cos(t) + 4 + 5t4
7.
(F(t) × G(t)) = tet (2 + t)(j − k)
8.
(F(t) · G(t)) = −16 cos2 (t) + 16 sin2 (t)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.1. VECTOR FUNCTIONS OF ONE VARIABLE 267
9.
F(t) = sin(t)i + cos(t)j + 45tk, 0 ≤ t ≤ 2π
is a position vector for the curve C. Then
Then
s
t3 = √ − 1
3
so 1/3
s
t= √ −1 .
3
Set 1/3
s
G(s) = f (T (s)) = √ −1 (i + j + k).
3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
268 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
Then
1
G (s) = √ (i + j + k)
3
and this is a unit tangent vector.
11 .
F(t) = t2 (2i + 3j + 4k)
is a position vector, for 1 ≤ t ≤ 3, and
12. Suppose F(t)×F (t) = O for all t. Then either F(t) = O, or F (t) = O, or
both vectors are nonzero and parallel, for all t. In the first case the particle
sits at the origin for all time. In the second case there is no motion and
the particle is at rest. In the last case, if the position and tangent vectors
are always parallel, then the velocity vector is always directed along the
path of motion and the motion is in a straight line.
We could also argue as follows in the last case. If F and F are parallel
for all t, then for some number c,
This forces
x(t) = x0 ect , y(t) = y0 ect , z(t) = z0 ect ,
where F(0) = x0 i+y0 j+z0 k. These are parametric equations of a straight
line.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.2. VELOCITY AND CURVATURE 269
In this way we do not have to compute s and attempt to write vectors in terms
of s, which is often quite awkward or even impossible in terms of elementary
functions. We could also compute
dT/dt
N(t) = ,
dT/dt
T (t)
κ(t) =
F (t)
In Problems 1 - 10, we will provide full details just for the first problem.
The methodology is the same for the other problems.
1. The velocity is
v(t) = F (t) = 3i + 2tk,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
270 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
√
the speed is v(t) = 9 + 4t2 , acceleration is
a(t) = F (t) = 2k,
and a unit tangent is
1 1
T(t) = F (t) = √ (3i + 2tk).
F (t) 9 + 4t2
The curvature is
T (t) 6
κ(t) = = ,
F (t) (9 + 4t2 )3/2
in which we have omitted routine differentiations. The normal and tan-
gential components of the acceleration are given by
dv 4t
aT = =√
dt 9 + 4t2
and
6
aN = a 2 −a2T = √ .
9 + 4t2
2.
v(t) = (sin(t) + t cos(t))i + (cos(t) − t sin(t))j,
a(t) = (2 cos(t) − t sin(t))i − (2 sin(t) + t cos(t))j,
1
T(t) = √ v,
1 + t2
v(t) = 1 + t2 ,
t 2 + t2
aT = √ , aN =
1 + t2 1 + t2
2 + t2
κ=
(1 + t2 )3/2
3.
v(t) = 2i − 2j + k, v = 3,
1
T = (2i − 2j + k)
3
a T = aN = κ = 0
4. √
v(t) = et (sin(t) + cos(t))i + et (cos(t) − sin(t))k, v = 2et ,
a(t) = 2et (cos(t)i − sin(t)k),
1
T(t) = √ ((sin(t) + cos(t))i + (cos(t) − sin(t))k),
2
√
aT = 2et = aN
1
κ = √ e−t
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.2. VELOCITY AND CURVATURE 271
5.
v(t) = −3e−t (i + j − 2k), a(t) = 3e−t (i + j − 2k),
√ 1
v(t) = 3 6e−t , T(t) = √ (−i − j + 2k),
6
√ −t
aT = −3 6e , aN = 0, κ = 0
6.
cosh(2t) cosh(2t)
1
κ=
2(cosh(2t))3/2
Here we have used the hyperbolic identity
8.
1 1
v(t) = (i − j + 2k), a(t) = − 2 (i − j + 2k)
t t
√
6 1
v= , T(t) = √ (i − j + 2k),
t 6
√
6
aT = − 2 , aN = 0, κ = 0
t
9.
v(t) = 2t(αi + βj + γk),
a(t) = 2(αi + βj + γk),
v(t) = 2|t| α2 + β 2 + γ 2 ,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
272 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
1
T(t) =
(αi + βj + γk),
α + β2 + γ2
2
aN = 0, κ = 0,
and
aT = 2(sgn(t)) α2 + β 2 + γ 2 ,
where
1 if t > 0,
sgn(t) =
−1 if t < 0.
10.
v(t) = (3 cos(t) − 3t sin(t))j − (3 sin(t) + 3t cos(t))k,
v(t) = 3 1 + t2 ,
1
T(t) = √ ((cos(t) − t sin(t))j − (sin(t) + t cos(t))k)
1 + t2
3t (3t2 + 6)2
aT = √ , aN = √ ,
1 + t2 1 + t2
(3t2 + 6)2
κ=
9(1 + t2 )3/2
If
F(s) = f (s)i + g(s)j + h(s)k,
then
f (s) = g (s) = h (s) = 0
which means that f (s) = a + bs, g(s) = c + ds and h(s) = p + qs for some
constants a, b, c, d, p, h. Then F is the position vector for a straight line.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.3. VECTOR FIELDS AND STREAMLINES 273
Then
T (t) = − cos(t)i − sin(t)j,
a unit vector. Finally,
T (t) 1
κ=
= .
F (t) r
vT × F = vT(aT T + aN N)
= vaT T × T + vaN T × N
= vaN T × N = v(v 2 κ)T × N.
F × F = v 3 κ.
Finally, v = F , so
F(t) × F(t)
κ= .
F(t) 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
274 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
1
1= , 1 = e2−c2 .
2 − c1
1
x = x, y = , z = ex−2 .
x−1
y = −2x + c1 , z = x + c2 .
3. Streamlines satisfy
dy dz
x dx = = .
ex −1
Integration xex dx = dy to obtain y = xex − ex + c1 . Integrate x dx = −dz
to obtain x2 = −2z + c2 . Using x as parameter, streamlines are given by
1
y = xex − ex + c1 , z = (c2 − x2 ).
2
For the streamline passing through (2, 0, 4), we need
1
e2 + c1 = 0 and 4 = (c2 − 4).
2
1
x = x, y = xex − ex − e2 , z = (12 − x2 ).
2
4. Streamlines satisfy
dx dy
= , dz = 0.
cos(y) sin(x)
Integrate sin(x) dx = cos(y) dy and dz = 0 to obtain
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.4. THE GRADIENT FIELD 275
6. Streamlines satisfy
dx dy dz
2
=− = 3.
3x y z
Integrate the equations
1 3 1 1
dx = − dy and dy = − 3 dz
x2 y y z
to obtain
1
= −3 ln |y| + c1 and 2 ln |y| + c2 = z −2 .
x
For the streamline passing through (2, 1, 6), we need c1 = 1/2 and c2 =
1/36. Using y as the parameter, this streamline can be written
2 6
x= , y = y, z =
.
1 + 6 ln(y) 1 + 72 ln(y)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
276 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
and
∇ϕ(1, 1, 1) = i + j + k.
√
√ value of Du ϕ(1, 1, 1) is ∇ϕ(1, 1, 1) =
The maximum 3. The minimum
value is − 3.
2.
3.
∇ϕ(x, y, z) = (2y + ez )i + 2xj + xez k
∇ϕ(−2, 1, 6) = (2 + e6 )i − 4j − 2e6 k
The maximum value of Du ϕ(−2, 1, 6) is
4.
∇ϕ(x, y, z) = −yz sin(xyz)i − xz sin(xyz)j − xy sin(xyz)k,
π π
∇ϕ(−1, 1, π/2) = i − j − k
2 2
The maximum value of Du (−1, 1, π/2) is
π2
∇ϕ(−1, 1, π/2) = 1+ .
2
The minimum value is the negative of this maximum value.
5.
∇ϕ(x, y, z) = 2y sinh(2xy)i + 2x sinh(2xy)j − cosh(z)k,
∇ϕ(0, 1, 1) = − cosh(1)k,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.4. THE GRADIENT FIELD 277
6.
1
∇ϕ(x, y, z) =
[xi + yj + zk] ,
x2 + y 2 + z 2
1
∇ϕ(2, 2, 2) = √ (i + j + k),
3
max Du = ∇ϕ(2, 2, 2) = 1,
min Du = − ∇ϕ(2, 2, 2) = −1
7.
Du ϕ(x, y, z) = ∇ϕ(x, y, z) · u
1
= ((8y 2 − z)i + 16xyj − xk) · √ (i + j + k)
3
1 2
= √ (8y − z + 16xy − x)
3
8 .
Du ϕ(x, y, z) = ∇ϕ(x, y, z) · u
1
= (− sin(x − y)i + sin(x − y)j + ez k) · √ (i − j + 2k)
6
1 z
= √ (−2 sin(x − y) + 2e )
6
9.
Du ϕ(x, y, z) = ∇ϕ(x, y, z) · u
1
= (2xyz 3 i + x2 y 3 j + 3x2 yz 2 k) · √ (2j + k)
5
1
= √ (2x2 z 3 + 3x2 yz 2 )
5
10.
Du ϕ(x, y, z) = ∇ϕ(x, y, z) · u
1
= ((z + y)i + (z + x)j + (y + x)k) · √ (i − 4k)
17
1
= √ (z − 3y − 4x)
17
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
278 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
√
The tangent plane to the surface at (1, 1, 2) has equation
√ √
2(x − 1) + 2(y − 1) + 2 2(z − 2) = 0,
or √
x+y+ 2z = 4.
The normal line to the surface at this point has parametric equations
√
x = y = 1 + 2t, z = 2(1 + 2t) for − ∞ < t < ∞.
−2x + y − z = 1
N = ∇(x2 − y 2 − z 2 )|(1,1,0) = 2i − 2j
The tangent plane at (1, 1, 0) has equation y = x and the normal line has
parametric equations
The tangent plane has the equation x = 1 and the normal line has para-
metric equations
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.5. DIVERGENCE AND CURL 279
The tangent plane has equation x + y + 2z = 4 and the normal line has
parametric equations
17. Since ∇ϕ(x, y, z) = i + k for all (x, y, z), the normal to the level surface
ϕ(x, y, z) = K is the constant vector N = i + k, so the surface must be the
plane x + z = K. The streamlines of the vector field ∇ϕ(x, y, z) = i + k
are solutions of
dx = dz, dy = 0.
Integrate to obtain
x = z + c 1 , y = c2 .
Using t as parameter,
These streamlines are lines in 3− space which are orthogonal to the surface
x + z = K.
∇ · (∇ × F) = 0
2.
∇ · F = xz cosh(xyz),
∇ × F = −xy cosh(xyz)i + yz cosh(xyz)k
∂ ∂
∇ · (∇ × F) = (−xy cosh(xyz)) + (yz cosh(xyz))
∂x ∂z
= cosh(xyz)(−y + y) + sinh(xyz)(−xy 2 z + zy 2 z) = 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
280 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
3.
∇ · F = 2y + xey + 2
∇ × F = (ey − 2x)k
∂ y
∇ · (∇ × F) = (e − 2x) = 0
∂z
4.
∇·F=1+1+2=4
∇×F=O
∇ · (∇ × F) = 0
5.
∇ · F = cosh(x) + xz sinh(xyz) − 1
∇ × F = (−1 − xy sinh(xyz))i − j + yz sinh(xyz)k
∂ ∂
∇·∇×F= (−1 − xy sinh(xyz)) + (yz sinh(xyz))
∂x ∂y
= (−y + y) sinh(xyz) + cosh(xyz)(−xy 2 z + xy 2 z) = 0
6.
∇ · F = cosh(x − z) + 2 + 1 = cosh(x − z) + 3
∇ × F = −2yi − cosh(x − z)j
∂ ∂
∇ · (∇ × F) = (−2y) + (− cosh(x − z)) = 0
∂x ∂y
7.
∇ϕ = i − j + 4zk
i
j k
∇ × (∇ϕ) = ∂/∂x ∂/∂y ∂/∂z = O
1 −1 4z
8.
∇ϕ = (18yz + ex )i + 18xzj + 18xyk
∇ × (∇ϕ) = (18x − 18x)i + (18y − 18y)j + (18z − 18z)k
9.
∇ϕ = −6x2 yz 2 i − 2x3 z 2 j − 4x3 yzk
and
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
11.5. DIVERGENCE AND CURL 281
10.
∇ϕ = z cos(xz)i + x cos(xz)k
i j k
∇ × (∇ϕ) = ∂x ∂ ∂
∂y
∂
∂z
z cos(xz) 0 x cos(xz)
11.
12.
∇ϕ = ex+y+z (i + j + k)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
282 CHAPTER 11. VECTOR DIFFERENTIAL CALCULUS
Next,
i
j k
∇ × (ϕF) = ∂/∂x ∂/∂y ∂/∂z
ϕf ϕg ϕh
∂ ∂
= (ϕh) − (ϕg) i
∂y ∂z
∂ ∂
+ (ϕf ) − (ϕh) j
∂z ∂x
∂ ∂
+ (ϕg) − (ϕf ) k
∂x ∂y
∂ϕ ∂ϕ ∂ϕ ∂ϕ
= h− g i+ f− h j
∂y ∂z ∂z ∂x
∂ϕ ∂ϕ
+ g− f k
∂x ∂y
∂h ∂g ∂f ∂h ∂g ∂f
+ϕ − i+ − j+ − k
∂y ∂z ∂z ∂x ∂x ∂y
= ∇ϕ × F + ϕ(∇ × F).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 12
2.
1
2
−4x dx + y dy − yz dz = (−4(−t2 )(−2t) + 02 − 0) dt
C 0
1
= −8t3 dt = −2.
0
3.
2
(x + y) ds = (t + t) 1 + 1 + 4t2 dt
C 0
2 √
2 1 26 2
= 2t 2 + 4t2 dt = (2 + 4t2 )3/2 =
0 6 0 3
283
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
284 CHAPTER 12. VECTOR INTEGRAL CALCULUS
5.
3
F · dR = (cos(t)i + t2 j + tk) · (i − 2tj + 0k) dt
C 0
3
81
= (cos(t) − 2t3 ) dt = sin(3) − .
0 2
6. √
2
2
√ 28 6
4xy ds = 4t 6 dt = .
C 1 3
7. Parametrize C as x = 2 cos(t), y = 2 sin(t), z = 0 for 0 ≤ t ≤ 2π. Then
2π
F · dR = (2 cos(t)i + 2 sin(t)j) · (−2 sin(t)i + 2 cos(t)j) dt
C 0
2π
= (−4 cos(t) sin(t) + 4 sin(t) cos(t)) dt = 0.
0
9.
9 √
−xyz dz = −z z dz
C 4
9
2 422
= − z 5/2 =−
5 4 5
10. 3
xz dy = t(−4t2 ) dt = −80
C 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.2. GREEN’S THEOREM 285
Because the density function and the position of the wire are symmetric
in the first octant, we will have x = y = z. Compute
3
2
x= √ xδ(x, y, z) ds
27 3 0
3 √
2
= √ t(3t) 3 dt = 2.
27 3 0
13. Take F(x) = f (x)i and R(t) = ti, for a ≤ t ≤ b. The graph of the curve
defined by this position vector is [a, b], and
b
F · dR = f (x) dx.
C a
2.
work = F · dR
C
= (ex − y + x cosh(x)) dx + (y 3/2 + x) dy
C
∂ 3/2 ∂ x
= (y + x) − (e − y + x cosh(x)) dA
∂x ∂y
D
= 2 dA = 2(area of D) = 2(62 π) = 72π.
D
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
286 CHAPTER 12. VECTOR INTEGRAL CALCULUS
3.
work = (− cosh(4x4 ) + xy) dx + (e−y + x) dy
C
∂ −y ∂
(e + x) − (− cosh(4x4 ) + xy) dA
D ∂x ∂y
3 7
= (1 − x) dA = (1 − x) dy dx
D 1 1
3
= 6(1 − x) dx = −12.
1
4.
F · dR
C
∂ ∂
= (−x) − (2y) dA
∂x ∂y
D
= (−3) dA = −3(area of D) = −3(16π) = −48π.
D
5.
F · dR
C
∂ ∂ 2
= (−2xy) − (x ) dA
D ∂x ∂y
6 (22−2y)/5
= (−2y) dA = −2y dx dy
D 1 (y+4)/5
6
2y
= (3y − 18) dy = −40.
1 5
6.
∂ ∂
F · dR = (x − y) − (x + y) dA
C ∂x ∂y
D
= 0 dA = 0.
D
7.
∂
F · dR = (8xy 2 ) = 8y 2 dA.
C D ∂x D
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.2. GREEN’S THEOREM 287
8.
∂ ∂ 2
F · dR = (cos(2y) − e3y + 4x) − (x − y)
C ∂x ∂y
D
= 5 dA = 125.
D
9.
∂ x ∂ x
F · dR = (e sin(y)) − (e cos(y))
C ∂x ∂y
D
= (−ex sin(y) + ex sin(y)) dA = 0.
D
10.
∂ ∂ 2
F · dR = (−xy 2 ) − (x y)
C ∂x ∂y
π/2 2
= (−y 2 − x2 ) dA = (r2 )r dr dθ
D 0 0
2
π
= −r3 dr = −2π.
2 0
11.
∂ 2 cos(y) ∂
F · dR = (xy − e )− (xy)
C D ∂x ∂y
3 5−5x/3
= (y 2 − x) dA = (y 2 − x) dy dx
D 0 0
3 3
1 5x 5x
= 5− dx − x 5− dx
0 3 3 0 3
95
= .
4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
288 CHAPTER 12. VECTOR INTEGRAL CALCULUS
14. Assume that C is a join of two curves in two ways. First, C has an upper
piece y = p(x) and a lower piece y = q(x) for a ≤ x ≤ b, so D consists of
all (x, y) with
a ≤ x ≤ b, q(x) ≤ y ≤ p(x).
Second, C also has a right piece y = β(x) and a left piece y = α(x) for
c ≤ y ≤ d, so, looking left to right instead of bottom to top, D can also
be described as consisting of all (x, y) with
c ≤ y ≤ d, α(y) ≤ x ≤ β(y).
Now use both of these descriptions in turn as follows. Using the second
description of C (look at C from left to right),
d c
g(x, y) dy = g(β(y), y) dy + g(α(y), y) dy.
C c d
Therefore
∂g
g(x, y) dy = dA.
C D ∂x
This is ”half” of the conclusion of Green’s theorem. For the rest, use the
first description of C. Now, looking from bottom to top, we have (keeping
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.3. AN EXTENSION OF GREEN’S THEOREM 289
and
b p(x)
∂f
dA = f (x, p(x)) − f (x, q(x)) dA.
D ∂y a q(x)
Then
∂f
f (x, y) dx = − dA.
C D ∂y
Upon adding these two equations, we obtain
∂g ∂f
f (x, y) dx + g(x, y) dx = − dA.
C D ∂x ∂y
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
290 CHAPTER 12. VECTOR INTEGRAL CALCULUS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.4. POTENTIAL THEORY 291
since the partial derivatives in the integral are equal and therefore cancel.
If C does enclose the origin, use the extension of Green’s theorem. If K
is a circle of radius r enclosing the origin and enclosed by C, we obtain
F · dR = F · dR
C K
2π
r cos(θ)
= + 2r cos(θ) (−r sin(θ)) dθ
0 r
2π
r sin(θ)
+ − 3r2 sin2 (θ) r cos(θ) dθ
0 r
2π
= −r2 (2 cos(θ) sin(θ) + 3 sin2 (θ) cos(θ)) dθ
0
2π
= −r2 (sin2 (θ) + sin3 (θ) 0
= 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
292 CHAPTER 12. VECTOR INTEGRAL CALCULUS
Then
∂ϕ
= 6x + xexy + k (y) = 6x + xexy .
∂y
Then k (y) = 0 and we may choose k(y) = 0. A potential function is
ϕ(x, y) = 6xy + exy .
3. Since
∂ ∂
(16x) = 0 = (2 − y 2 ),
∂y ∂x
then F is conservative. Integrate ∂ϕ/∂x = 16x with respect to x to obtain
Then
∂ϕ
= 2 − y 2 = k (y)
∂y
so we may choose k(y) = 2y − y 3 /3 to obtain the potential function
1
ϕ(x, y) = 8x2 + 2y − y 3 .
3
4. Since
∂ ∂
(2xy cos(x2 )) = 2x cos(x2 ) = (sin(x2 )),
∂y ∂x
then F is conservative. Integrate
∂ϕ
= 2xy cos(x2 )
∂x
to obtain
ϕ(x, y) = y sin(x2 ) + k(y).
Then
∂ϕ
= sin(x2 ) = sin(x2 ) + k (y),
∂y
and we may choose k(y) = 0. A potential function is
ϕ(x, y) = y sin(x2 ).
5. Since
∂ 2x 4xy ∂ 2y
=− = ,
∂y x2 + y 2 (x2 + y 2 )2 ∂x x2 + y 2
we know that F is conservative on any region not containing the origin.
A potential function ϕ(x, y) must satisfy
∂ϕ 2x ∂ϕ 2y
= 2 and = 2 .
∂x x + y2 ∂y x + y2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.4. POTENTIAL THEORY 293
Then we need
∂ϕ 2x 2x
= 2 + c (x) = 2 .
∂x x + y2 x + y2
Then c (x) = 0 and we may choose c(x) = 0, yielding the potential func-
tion
ϕ(x, y) = ln(x2 + y 2 ).
k(y, z) = −y 2 + c(z).
Next we need
∂ϕ ∂k
= z sin(x) + 1 = z sin(x) +
∂y ∂y
so
∂k
= 1.
∂y
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
294 CHAPTER 12. VECTOR INTEGRAL CALCULUS
k(x, y) = y + c(z).
Thus far
ϕ(x, y, z) = yz sin(x) + y + c(z).
Finally, we need
∂ϕ
= y sin(x) + c = y sin(x).
∂z
We may choose c(x) = 0, yielding the potential function
ϕ(x, y, z) = yz sin(x) + y.
9. We find that
∇ × F = (−z 2 − xy)i + yzk = O
so F is not conservative and there is no potential function.
10. Compute
ϕ(x, y) = x3 (y 2 − 4y).
Then
F · dR = ϕ(2, 3) − ϕ(−1, 1) = −24 − 3 = −27.
C
ϕ(x, y) = x2 y − ln |y|
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.4. POTENTIAL THEORY 295
16. The easiest way to find a potential function for F is to integrate ∂ϕ/∂y =
x cos(xz) with respect to y and obtain ϕ(x, y, z) = xy cos(xz) + k(x, z).
Now observe that ϕ(x, y, z) = xy cos(xz) is a potential function for F if
we choose k(x, z) = 0. Then
F · dR = ϕ(1, 1, 7) − ϕ(1, 0, π) = cos(7).
C
21. Let C be a smooth path of motion given by R(t) = x(t)i + y(t)j + z(t)k
and let L be the kinetic energy plus the potential energy. Then
m m
L(t) = R (t) 2 −ϕ(x(t), y(t), z(t)) = R (t)·R (t)−ϕ(x(t), y(t), z(t)).
2 2
Then
dL m ∂ϕ ∂ϕ ∂ϕ
= (2R (t) · R (t)) − x (t) − y (t) − z (t)
dt 2 ∂x ∂y ∂z
= (mR (t) · R (t)) − ∇ϕ · R (t)
= (mR (t) − ∇ϕ) · R (t).
But ∇ϕ is the force acting on the particle, so by Newton’s second law,
mR = ∇ϕ, and therefore dL/dt = 0. Therefore L(t) is a constant of the
motion.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
296 CHAPTER 12. VECTOR INTEGRAL CALCULUS
This is a function because its value depends only on (x, y) and not on the
path in D from P to (x, y). We claim that ∇ϕ = F.
To show this, we will first show that
∂ϕ
= f (x, y).
∂x
Choose Δx small enough that (x + Δx, y) is in D. Now
This is a line integral over over a horizontal line segment from (x, y) to
(x + Δx, y), with y fixed on this segment. Parametrize this segment by
ξ = x + tΔx, η = y for 0 ≤ t ≤ 1.
On this segment,
dξ = (Δx) dt and dη = 0.
Then 1
ϕ(x + Δx, y) − ϕ(x, y) = Δx f (x + tΔx, y) dt.
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.5. SURFACE INTEGRALS 297
Then 1
ϕ(x + Δx, y) − ϕ(x, y)
= f (x + tΔx, y) dt.
Δx 0
By the mean value theorem for integrals, there is some t0 in (0, 1) such
that 1
f (x + tΔx, y) dt = f (x + t0 Δx, y).
0
Therefore
ϕ(x + Δx, y) − ϕ(x, y)
= f (x + t0 Δx, y).
Δx
As Δx → 0, x + t0 Δx → x and, by continuity, f (x + t0 Δx, y) → f (x, y).
Therefore,
A similar argument, using a vertical path from (x, y) to (x, y + Δy) shows
that
∂ϕ
= g(x, y).
∂y
and
√
x dσ = 3 2x dA
Σ D
√
√ 5/2 10−4y
3 2 5/2
=3 2 x dx dy = (10 − 4y)2 dy
0 0 2 0
√
5/2
2 √
=− (10 − 4y)3 = 125 2.
8
0
√ √
2. On the surface, z = x so dσ = 12 + 12 + 02 = 2 dA and
√
2
y dσ = 2y 2 dA
Σ D
√
√ 2 4 2 128 2
= 2 y dx dy = .
0 0 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
298 CHAPTER 12. VECTOR INTEGRAL CALCULUS
3. On Σ,
dσ = 12 + (2x)2 + (2y)2 dA = 1 + 4(x2 + y 2 ) dA.
5. On the surface, z 2 = x2 + y 2 , so
∂z ∂z
2z = 2x and 2z z = 2y.
∂x ∂y
Then
∂z x ∂z y
= and = .
∂x z ∂y z
Then
x2 y2 √
dσ = 1+ 2
+ 2 dA = 2 dA.
z z
Then
√
z dσ = 2 x2 + y 2 dA
Σ D
√ π/2 4
28π √
= 2 r2 dr dθ = 2.
0 2 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.5. SURFACE INTEGRALS 299
√ √
6. On Σ, dσ = 1 + 1 + 1 dσ = 3 dσ, and z = x + y, so
√
xyz dσ = 3 xy(x + y) dA
Σ D
√
√ 1 1
2 2 3
= 3 (x y + xy ) dy dx = .
0 0 3
√
7. On Σ, dσ = 1 + 4x2 dA, so
y dσ = y 1 + 4x2 dA
Σ D
2 3 2 √ √
9 9
= y 1 + 4x2 dy dx = 1 + 4x2 dx = (ln(4 + 17) + 4 17).
0 0 2 0 8
8. On the surface, dσ = 1 + 4(x2 + y 2 ) dA, so
x2 dσ = x2 1 + 4(x2 + y 2 ) dA
Σ D
2π 2
= (r2 cos2 (θ) 1 + 4r2 r dr dθ
0 0
2π 2
= cos2 (θ) dθ r3 1 + 4r2 dr.
0 0
2
For the first integral, use the identity cos (θ) = (1 + cos(2θ))/2. For the
second integral, use the substitution u = 1 + 4r2 , so r2 = (u − 1)/4 and
r dr = (1/8)du. These yield
17
1 2π 1
x2 dσ = (1 + cos(2θ))dθ (u3/2 − u1/2 ) du
Σ 2 0 32 1
π √
= (782 17 + 2).
240
√
9. On this surface, dσ = 3 dA and z = x − y so
√
z dσ = 3(x − y) dA
Σ D
√ 1 5 √
= 3 (x − y) dy dx = −10 3.
0 0
10. On the surface, dσ = 1 + 4y 2 dA and z = 1 + y 2 , so
xyz dσ = xy(1 + y 2 ) 1 + 4y 2 dA
Σ D
1 1
1 1
= xy(1 + y 2 ) 1 + 4y 2 dy dx = y(1 + y 2 ) 1 + 4y 2 dy.
0 0 2 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
300 CHAPTER 12. VECTOR INTEGRAL CALCULUS
For the last integral let u = 1+4y 2 , y dy = (1/8)du, and 1+y 2 = (u+3)/4
to obtain
5
1 1 3/2 1/2 1 √ 3
xyz dσ = (u + 3u ) du = 5 5− .
Σ 2 32 1 16 5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.6. APPLICATIONS OF SURFACE INTEGRALS 301
Then
1
z= z dσ
m Σ
√ √
2K 2π 3 2 18Kπ 2
= r dr dθ = = 2.
m 0 0 m
The center of mass is (0, 0, 2).
4. On Σ, dσ = 1 + 4(x2 + y 2 ) dA. Σ projects onto the x, y− plane to give
the quarter annulus D consisting of points (x, y) with x ≥ 0, y ≥ 0 and
1 ≤ x2 + y 2 ≤ 9. The mass is
xy
m= dσ
Σ 1 + 4(x2 + y 2 )
π/2 3
= xy dA = r3 cos(θ) sin(θ) dr dθ
D 0 1
π/2 4 3
1 2 r
= sin (θ) = 10.
2 0 4 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
302 CHAPTER 12. VECTOR INTEGRAL CALCULUS
Finally,
1 1
z= zδ(x, y, z) dσ = (16 − x2 − y 2 )xy dA
m Σ 10 D
π/2 3
1 331
= (16 − r2 )r3 sin(θ) cos(θ) dr dθ = .
10 0 1 48
Finally,
1 1
z= zδ(x, y, z) dσ = (6 − x2 − y 2 )(1 + 4x2 + 4y 2 ) dA
m Σ m D
√
2π 6
1 162π 27
= (6 − r2 )(1 + 4r2 )r dr dθ = = .
m 0 0 m 13
6. By symmetry, x = y = z. On Σ,
dσ = 1 + (x/z)2 + (y/z)2 dA = (1/z) dA.
The mass is
4πK
m= K dσ = K(area of )Σ = = Kπ.
Σ 4
Finally,
1 K 1 1
z= Kz dσ = z(1/z) dA = (area of )D = .
m Σ Kπ D π 4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.7. LIFTING GREEN’S THEOREM TO R3 303
On Σ, z = 8 − x − 2y, so
1
F · n = √ (2x + 4y − 8).
6
√ √
Further, dσ = 1 + 4 + 1 dA = 6 dA. Therefore, the flux of F across Σ
is
4 8−2y
128
F·n dσ = (2x+4y −8) dA = (2x+4y −8) dx dy = .
Σ D 0 0 3
1
n= (xi + yj + zk).
2
Then
1 2
F·n= (x z − yz).
2
Now, dσ = 1 + (−x/z)2 + (−y/z)2 dA. Therefore, the flux of F across
Σ is
F · n dσ = (x2 − y) dA.
Σ D
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
304 CHAPTER 12. VECTOR INTEGRAL CALCULUS
2. ∇ · F = 4 − 6 = −2, so compute
∇ · F dV = −2(volume of V ) = −2π(22 )(2) = −16π.
M
3. Since ∇ · F = 0, ∇ · F dV = 0.
M
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.8. THE DIVERGENCE THEOREM OF GAUSS 305
5. since ∇ · F = 4, compute
8π
∇ · F dV = 4 dV = 4(volume of M ) = .
3
M M
6. ∇ · F = 2 + x, so
3 2 4 4
∇ · F dV = (2 + x) dx dy dz = 3(2 + x)2 0
= 96.
0 0 0
M
Next,
√
2 √ 1 1 1
r2 (cos(θ) + sin(θ))( 2 − r) + r(2 − r2 ) d r = (cos(θ)+sin(θ))+ .
0 2 3 2
Finally,
2π
1 1
(cos(θ) + sin(θ)) + dθ = π.
0 3 2
Therefore
∇ · F dV = 2π.
M
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
306 CHAPTER 12. VECTOR INTEGRAL CALCULUS
But ∇ · ∇ × F = 0, so
(∇ × F) · n dσ = 0.
Σ
R = 2 cos(t)i + 2 sin(t)j + 0k
so
F·dR = (−16 cos2 (t) sin2 (t)−16 cos2 (t) sin2 (t)) dt = −32 cos2 (t) sin2 (t) dt.
Then 2π
F · dR = −32 cos2 (t) sin2 (t) dt = −8π.
C 0
If we use the surface integral, then evaluate Σ
(∇ × F) · n dσ. Compute
∇ × F = −(x2 + y 2 )k.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.9. THE INTEGRAL THEOREM OF STOKES 307
Further, a normal to Σ is
∇(x2 + y 2 + z 2 ) = 2xi + 2yj + 2zk.
A unit normal vector is
1
n= (xi + yj + zk).
2
Finally, dσ = 1 + (x/z)2 + (y/z)2 dA = (2/z) dA. Then
2π 2
(∇ × F) · n dσ = − (x2 + y 2 ) dA = − r3 dr dθ = −8π.
Σ D 0 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
308 CHAPTER 12. VECTOR INTEGRAL CALCULUS
5. Notice that the boundary curve C is piecewise smooth and must be parametrized
in three smoothcurves.
This is not difficult but is tedious. We therefore
try to compute Σ
(∇ × F) · n dσ. First,
∇ × F = (x − y)i − yj − xk.
And
1
n = √ (2i + 4j + k).
21
√
Finally, dσ = 21 dA. Then
2 4−2y
32
(∇ × F) · ndσ = (x − 6y) dA = (x − 6y) dx dy = − .
Σ D 0 0 3
6. The circulation is C F · dR. Take Σ to be the disk 0 ≤ x2 + y 2 ≤ 1, with
boundary C parametrized by x = cos(t), y = sin(t), z = 0 for 0 ≤ t ≤ 2π.
The proper unit normal to Σ is n = k. Now
so
(∇ × F) · n = 2xy + 1.
Further, dσ = dA. Then
F · dR = (∇ × F) · n dσ
C
Σ
= (2xy + 1) dA = area of A = π
D
since D
2xy dA = 0.
7. Compute
∇ × F = −i − j − k.
A normal to the surface is
N = i + 4j + k
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.10. CURVILINEAR COORDINATES 309
Now
6
(∇ × F) · n = − √
18
and √
dσ = N dx dy = 18
so
−6 √
(∇ × F) · ndσ = √ 18 dx dy
D D 18
= −6( area of D) = −6(18) = −108.
u1 = ur = r, u2 = uθ = θ, u3 = uz = z.
From Example 12.31, we know that, for cylindrical coordinates, the scale
factors are
h1 = hr = 1, h2 = hθ = r, h3 = hz = 1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
310 CHAPTER 12. VECTOR INTEGRAL CALCULUS
Given g(r, θ, z), we can compute the gradient and Laplacian in cylindrical
coordinates as
∂g 1 ∂g ∂g
∇g = ur + uθ + uz
∂r r ∂θ ∂z
and
1 ∂ ∂g ∂ 1 ∂g ∂ ∂g
∇2 g = r + + r
r ∂r ∂r ∂θ r ∂θ ∂z ∂z
1 ∂g ∂ 2 g 1 ∂2g ∂2g
= + 2 + 2 2 + 2.
r ∂r ∂r r ∂θ ∂z
Given a vector field F(r, θ, z) in cylindrical coordinates, write
F = f1 u1 + f2 u2 + f3 u3 .
The divergence is given by
1 ∂ ∂ ∂
∇·F= (f1 r) + (f2 ) + (rf1 )
r ∂r ∂θ ∂z
1 ∂f1 ∂f2 ∂f3
= f1 + r + +r .
r ∂r ∂θ ∂r
Finally, the curl is given by
ur
uθ uz
∇ × F = ∂/∂r ∂/∂θ ∂/∂z .
f1 rf2 f3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
12.10. CURVILINEAR COORDINATES 311
and
hz = 1.
If g(u, v, w) is a scalar-valued function, then
1 ∂g 1 ∂g ∂g
∇g = uu + uv + uz ,
h1 ∂u h2 ∂v ∂z
where
a
uu = (sinh(u) cos(v)i + cosh(u) sin(v)j),
h1
a
uu = (− cosh(u) sin(v)i + sinh(u) cos(v)j),
h2
uz = k.
The divergence and curl of a vector field F(u, v, z), with component func-
tions f1 , f2 , f3 , are
1 ∂ ∂ ∂g
∇·F= 2 (f1 h1 ) + (f2 h2 ) + ,
h1 ∂u ∂v ∂z
and
1 ∂f3 ∂f2
∇×F= − uu
h1 ∂v ∂z
∂f1 1 ∂f3
+ − uv
∂z h1 ∂u
1 ∂ ∂
+ 2 (f2 h2 ) − (f1 h1 ) uz .
h1 ∂u ∂v
a sinh(v) a sin(u)
x= ,y = , z = z.
cosh(v) − cos(u) cosh(v) − cos(u)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
312 CHAPTER 12. VECTOR INTEGRAL CALCULUS
and
h1 uu h2 uv uz
(cosh(v) − cos(u))2
∇×F= ∂/∂u ∂/∂v ∂/∂z .
a2
hu f1 hv f2 f3
4. Parabolic cylindrical coordinates are given by
1 2
x = uv, y = (u − v 2 ), z = z.
2
We find that the scaling factors are
hu = hv = u2 + v 2 , hz = 1.
If g(u, v, z) is a scalar field, then
1 ∂g 1 ∂g ∂g
∇g = √ uu + √ uv + uz
u2 + v ∂u
2 u + v ∂v
2 2 ∂z
and
2 1 ∂2g ∂2g ∂ 2 2∂g
∇ g= 2 + + (u + v ) .
u + v2 ∂u2 ∂v 2 ∂z ∂z
And, if F(u, v, z) is a vector field, then
1 ∂ 2
∇·F= 2 2
( u + v 2 f1 )
u +v ∂u
∂ 2 2
∂ 2 2
+ ( u + v f2 ) + ((u + v )f3 ) .
∂v ∂z
and √ √
u2 + v 2 u u u2 + v 2 uv uz
1
∂/∂u ∂/∂z .
∇×F= 2
u + v2 √ √ ∂/∂v
u2 + v 2 f1 u2 + v 2 f2 f3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 13
Fourier Series
313
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
314 CHAPTER 13. FOURIER SERIES
2.5
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
2.5
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.2. THE FOURIER SERIES OF A FUNCTION 315
0.5
0
-1 -0.5 0 0.5 1
x
-0.5
-1
This Fourier series has just one term, 4 itself. Of course, this converges to
4 on [−3, 3].
2. The Fourier series of f (x) = −x on [−1, 1] has the form
∞
1
a0 + (an cos(nπx) + bn sin(nπx)).
2 n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
316 CHAPTER 13. FOURIER SERIES
10
-1 -0.5 0 0.5 1
x
and, for n = 1, 2, · · · ,,
1
2 sinh(π) (−1)n
an = 2 cosh(πx) cos(nπx) dx = .
0 π 1 + n2
For Problems 4 through 10, we give the Fourier series, analyze its conver-
gence, and show a graph of one of its partial sums.
4. The series is
∞
8 1
cos((2n − 1)πx/2),
π 2 n=1 (2n − 1)2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.2. THE FOURIER SERIES OF A FUNCTION 317
0.5
x
-2 -1 0 1 2
0
-0.5
-1
converging to ⎧
⎪
⎨−4 for −π < x < 0,
4 for 0 < x < 4,
⎪
⎩
0 for 0, π, −π.
Figure 13.6 shows a graph of this function and the twentieth partial sum
of its Fourier series.
6. Since f (x) is odd and periodic of period π, then f (x) = sin(2x) is
its own Fourier series (the Fourier series has just one term, the function
itself).
7. The Fourier series is
∞
13 n 16 4
+ (−1) cos(nπx/2) + sin(nπx/2) ,
3 n=1
(nπ)2 nπ
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
318 CHAPTER 13. FOURIER SERIES
x
-3 -2 -1 0 1 2 3
0
-2
-4
-2 -1 0 1 2
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.2. THE FOURIER SERIES OF A FUNCTION 319
25
20
15
10
0
-4 -2 0 2 4
x
where
25
an = (11(−1)n − 1)
(nπ)2
and
5 25
bn = (1 − 21(−1)n ) + ((−1)n − 10).
nπ (nπ)3
The series converges to
⎧
⎪
⎪ −x for −5 < x < 0,
⎪
⎨1 + x2 for 0 < x < 5,
⎪
⎪ 1/2 for x = 0,
⎪
⎩
31/2 for x = ±5.
A graph of f (x) and the twenty-fifth partial sum of this Fourier series is
shown in Figure 13.8.
9. The Fourier series of f (x) on [−π, π] is
∞
3 2 1
+ sin((2n − 1)x).
2 π n=1 2n − 1
This converges to
⎧
⎪
⎨1 for −π < x < 0,
2 for 0 < x < π,
⎪
⎩
3/2 for x = 0, π, and −π.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
320 CHAPTER 13. FOURIER SERIES
1.8
1.6
1.4
1.2
-3 -2 -1 0 1 2 3
x
Figure 13.9 shows the function and the thirtieth partial sum of this Fourier
series.
10. The Fourier series is
∞
2 4 (−1)n
− sin(x) − cos(nx),
π π n=1 4n2 − 1
sin(3) ∞
(−1)n+1
nπx
+ 6 sin(3) cos ,
3 n=1
n2 π 2 − 9 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.2. THE FOURIER SERIES OF A FUNCTION 321
1.5
0.5
0
-3 -2 -1 0 1 2 3
x
Figure 13.10: Fourth partial sum of the Fourier series in Problem 10.
0.5
0
-3 -2 -1 0 1 2 3
x
-0.5
-1
Figure 13.11: Fifth partial sum of the Fourier series in Problem 11.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
322 CHAPTER 13. FOURIER SERIES
1.5
0.5
0
-1 -0.5 0 0.5 1
x
Figure 13.12: Twentieth partial sum of the Fourier series in Problem 12.
converging to
⎧
⎪
⎨1 − x for −1 < x < 0,
1/2 for x = 0,
⎪
⎩
1 for x = −1, 1.
Figure 13.12 shows this function and the twentieth partial sum of its
Fourier series on [−1, 1].
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.2. THE FOURIER SERIES OF A FUNCTION 323
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
324 CHAPTER 13. FOURIER SERIES
0
0 0.5 1 1.5 2 2.5 3
x
Figure 13.13: Partial sums of the sine series in Problem 1, Section 13.3.
converging to ⎧
⎪
⎨1 for 0 ≤ x < 1,
0 for x = 1,
⎪
⎩
−1 for 1 < x ≤ 2.
Figure 13.14 shows a graph of the function and the tenth and twentieth
partial sums of this expansion.
The sine series is
∞
21
(1 + (−1)n − 2 cos(nπ/2)) sin(nπx/2),
π n=1 n
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 325
0.5
0
0 0.5 1 1.5 2
x
-0.5
-1
Figure 13.14: Partial sums of the cosine series in Problem 2, Section 13.3.
converging to ⎧
⎪
⎨1 for 0 < x < 1,
0 for x = 0, 1, 2,
⎪
⎩
−1 for 1 < x < 2.
Figure 13.15 is a graph of f (x) and the tenth and sixty-fifth partial sums
of this sine expansion.
3. The cosine series is
∞
1 2 (−1)n (2n − 1)
cos(x) − cos((2n − 1)x/2),
2 π n=1 (2n − 3)(2n + 1)
converging to ⎧
⎪
⎪0 for 0 ≤ x < π,
⎪
⎨−1/2 for x = π,
⎪
⎪cos(x) for π < x < 2π,
⎪
⎩
0 for x = 2π.
Figure 13.16 shows a graph of the function and the fifteenth partial sum
of this cosine expansion.
The sine series is
∞
2 2n
− sin(x/2) − 2 − 4)π
((−1)n + cos(nπ/2)) sin(nx/2),
3π n=3
(n
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
326 CHAPTER 13. FOURIER SERIES
0.5
0
0 0.5 1 1.5 2
x
-0.5
-1
Figure 13.15: Partial sums of the sine series in Problem 2, Section 13.3.
0.5
x
0 1 2 3 4 5 6
0
-0.5
-1
Figure 13.16: Partial sum of the cosine series in Problem 3, Section 13.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 327
0.5
0
0 1 2 3 4 5 6
x
-0.5
-1
Figure 13.17: Partial sum of the sine series in Problem 3, Section 13.3.
converging to ⎧
⎪
⎪ 0 for 0 ≤ x < π,
⎪
⎨−1/2 for x = π,
⎪cos(x) for
⎪ π < x < 2π,
⎪
⎩
0 for x = 2π.
Figure 13.17 is a graph of the function and the fortieth partial sum of this
sine series.
4. The cosine series is
∞
8 1
1− cos((2n − 1)πx),
π n=1 (2n − 1)2
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
328 CHAPTER 13. FOURIER SERIES
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1
x
Figure 13.18: Fifth partial sum of the cosine expansion in Problem 4, Section
13.3.
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1
x
Figure 13.19: Fiftieth partial sum of the sine expansion in Problem 4, Section
13.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 329
0
0 0.5 1 1.5 2
x
Figure 13.20: Tenth partial sum of the cosine expansion in Problem 5, Section
13.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
330 CHAPTER 13. FOURIER SERIES
0
0 0.5 1 1.5 2
x
Figure 13.21: Fiftieth partial sum of the sine expansion in Problem 5, Section
13.3.
0.9
0.8
0.7
0.6
0.5
0.4
Figure 13.22: Tenth partial sum of the cosine expansion in Problem 6, Section
13.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 331
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
Figure 13.23: Sixtieth partial sum of the sine expansion in Problem 6, Section
13.3.
converging to ⎧
⎪
⎨x for 0 ≤ x < 2,
1 for x = 2
⎪
⎩
2 − x for 2 < x ≤ 3.
Figure 13.25 compares f (x) with the fortieth partial sum of this cosine
series.
The sine expansion is
∞
12 4 2 n
sin(2nπ/3) − cos(2nπ/3) + (−1) sin(nπx/3),
n=1
n2 π 2 nπ nπ
converging to ⎧
⎪
⎪x for 0 ≤ x < 2,
⎪
⎨1 for x = 2,
⎪2 − x for 2 < x < 3,
⎪
⎪
⎩
0 for x = 3.
Figure 13.24 shows the fifty-fifth partial sum of this sine series.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
332 CHAPTER 13. FOURIER SERIES
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
-0.5
-1
Figure 13.24: Fortieth partial sum of the cosine expansion in Problem 7, Section
13.3.
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
-0.5
-1
Figure 13.25: Fifty-fifth partial sum of the sine expansion in Problem 7, Section
13.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 333
0.5
0
0 1 2 3 4 5
x
-0.5
-1
Figure 13.26: Sixtieth partial sum of the cosine expansion in Problem 8, Section
13.3.
∞
1 41
− + cos(nπ/5) sin(2nπ/5) cos(nπx/5),
5 π n=1 n
converging to
⎧
⎪
⎪1 for 0 ≤ x < 1,
⎪
⎪
⎪
⎪ x = 1,
⎨1/2 for
0 for 1 < x < 3,
⎪
⎪
⎪
⎪−1/2 for x = 3,
⎪
⎪
⎩−1 for 3 < x < 5.
Figure 13.26 shows the sixtieth partial sum of this cosine expansion.
∞
4 1
(1 + (−1)n − 2 cos(nπ/5) cos(2nπ/5)) sin(nπx/5),
π n=1 2n
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
334 CHAPTER 13. FOURIER SERIES
0.5
x
0 1 2 3 4 5
0
-0.5
-1
Figure 13.27: Sixty-fifth partial sum of the sine expansion in Problem 8, Section
13.3.
converging to
⎧
⎪
⎪1 for 0 < x < 1,
⎪
⎪
⎪
⎪ for x = 1,
⎨1/2
0 for 1 < x < 3 or x = 0 or x = 5,
⎪
⎪
⎪−1/2 for x = 3,
⎪
⎪
⎪
⎩−1 for 3 < x < 5.
Figure 13.27 shows the sixty-fifth partial sum of this sine expansion.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 335
0.8
0.6
0.4
0.2
0
0 1 2 3 4
x
Figure 13.28: Tenth partial sum of the cosine expansion in Problem 9, Section
13.3.
converging to 1−x2 for 0 < x < 2 and to 0 for x = 0 and for x = 2. Figure
13.31 compares the thirtieth partial sum of this series with the function.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
336 CHAPTER 13. FOURIER SERIES
1.2
0.8
0.6
0.4
0.2
0
0 1 2 3 4
x
Figure 13.29: Twentieth partial sum of the sine expansion in Problem 9, Section
13.3.
x
0 0.5 1 1.5 2
0
-2
-4
-6
Figure 13.30: Tenth partial sum of the cosine expansion in Problem 10, Section
13.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.3. SINE AND COSINE SERIES 337
x
0 0.5 1 1.5 2
0
-2
-4
-6
Figure 13.31: Thirtieth partial sum of the sine expansion in Problem 10, Section
13.3.
12. Write
f (x) + f (−x)
fe (x) =
2
and
f (x) − f (−x)
fo (x) = .
2
Then fe (x) = fe (−x), so fe is even. And fo (−x) = −f (x), so fo is an odd
function. Further,
f (x) = fe (x) + fo (x).
13. Suppose f is both even and odd on [−L, L]. Then, for any x in this
interval,
f (x) = f (−x) = −f (x)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
338 CHAPTER 13. FOURIER SERIES
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.4. INTEGRATION AND DIFFERENTIATION OF FOURIER SERIES339
both converge. As with any convergent series, the general term has limit
zero as n → ∞, so
lim a2n = lim b2n = 0.
n→∞ n→∞
lim an = lim bn = 0.
n→∞ n→∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
340 CHAPTER 13. FOURIER SERIES
converge. Using the inequality of the preceding line in the comparison test
for positive series, we conclude that
∞
(|an | + |bn |)
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.5. PHASE ANGLE FORM 341
2.
g(t + p/α) = f (α(t + p/α)) = f (αt + p) = f (αt) = g(t),
and
t + αp
h(t + αp) = f = f (t/α + p) = f (t/α) = h(t).
α
3.
f (t + p + h) − f (t + p)
f (t + p) = lim
h→0 h
f (t + h) − f (t)
= lim = f (t).
h→0 h
4. Expanding f in a Fourier series on [0, 2] yields the series
∞
21
1− sin(nπx).
π n=1 n
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
342 CHAPTER 13. FOURIER SERIES
19
∞
1 1
nπx
+ 2 d n cos + δ n ,
8 π n=1 n2 2
where
dn = 8 + 5n2 π 2 − 12nπ sin(3nπ/2) + 4(n2 π 2 − 2) cos(3nπ/2)
and
nπ/2 + nπ cos(3nπ/2) − sin(3nπ/2)
δn = arctan .
nπ sin(3nπ/2) + cos(3nπ/2) − 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.5. PHASE ANGLE FORM 343
8
∞
n
π
cos 2nπx − .
π n=1 4n2 − 1 2
9. We can write
x for 0 ≤ x < 1,
f (x) =
x − 2 for 1 < x ≤ 2,
and f (x + 2) = f (x), so f has period 2. The Fourier series is
∞
2 (−1)n+1
sin(nπx).
π n=1 n
21
∞
π
cos nπx + (−1)n+1 .
π n=1 n 2
11. Write ⎧
⎪
⎨1 for 0 ≤ x < 1,
f (x) = 2 for 1 < x < 3,
⎪
⎩
1 for 3 < x < 4,
with f (x + 4) = f (x). The Fourier series of this function is
∞
3 2 (−1)n (2n − 1)πx
+ cos .
2 π n=1 2n − 1 2
3
∞
2 1
πx π
+ cos (2n − 1) + (1 − (−1)n ) .
2 π n=1 2n − 1 2 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
344 CHAPTER 13. FOURIER SERIES
12. Write
k for 0 < x < 1,
f (x) =
0 for 1 < x < 2,
and, for n = 0,
3
1 3
dn = 2xe−2nπit/3 dt = i.
3 0 nπ
The complex Fourier series expansion of f (x) is
∞
3i 1 2nπix/3
3+ e
π n
n=−∞,n=0
3 for x = 0 or x = 3,
=
2x for 0 < x < 3.
This converges to
2 for x = 0 or x = 2,
x2 for 0 < x < 2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.6. COMPLEX FOURIER SERIES 345
This converges to
⎧
⎪
⎨1/2 for x = 0 or x = 1 or x = 4,
0 for 0 < x < 1,
⎪
⎩
1 for 1 < x < 4.
This converges to
−2 for x = 0 or x = 6,
1−x for 0 < x < 6.
converging to ⎧
⎪
⎨1/2 for x = 0, 2, 4,
−1 for 0 < x < 2,
⎪
⎩
2 for 2 < x < 4.
Points of the frequency spectrum are
nπ 3
(0, 1/2), , .
2 (2n − 1)π
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
346 CHAPTER 13. FOURIER SERIES
converging to
(1 − e−5 )/2 for x = 0 or x = 5,
e−x for 0 < x < 5.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 347
0.5
0
-2 -1 0 1 2
t
-0.5
-1
Figure 13.32: Fifth partial sum and Cesáro sum in Problem 1, Section 13.7.
Figures 13.32, 13.33 and 13.34 compare f (t), SN (t) and σN (t) for N =
5, 10, 25, respectively. Notice that the Cesáro sums have the effect of
smoothing the Gibbs effect seem at 0 and the ends of the interval.
2. The N th partial sum of the Fourier series of f has the form
N
13 nπt nπt
SN (t) = + an sin + bn cos ,
8 n=1
2 2
where
2
an = sin(nπ/2)(−4 − nπ 2 ) + nπ cos(nπ/2) + 5nπ(−1)n
n3 π 3
and
2
bn = − −nπ sin(nπ/2) + cos(nπ/2)(−4 − n2 π 2 ) + 4(−1)n .
n3 π 3
Form σN (t) by inserting a factor of 1 − n/N into SN (t). Figures 13.35,
13.36 and 13.37 compare the N th partial sums and Cesáro sums and the
function for N = 5, 10, 25, respectively.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
348 CHAPTER 13. FOURIER SERIES
0.5
0
-2 -1 0 1 2
t
-0.5
-1
Figure 13.33: Tenth partial sum and Cesáro sum in Problem 1, Section 13.7.
0.5
0
-2 -1 0 1 2
t
-0.5
-1
Figure 13.34: Twenty-fifth partial sum and Cesáro sum in Problem 1, Section
13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 349
0
-2 -1 0 1 2
t
Figure 13.35: Fifth partial sum and Cesáro sum in Problem 2, Section 13.7.
0
-2 -1 0 1 2
t
Figure 13.36: Tenth partial sum and Cesáro sum in Problem 2, Section 13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
350 CHAPTER 13. FOURIER SERIES
0
-2 -1 0 1 2
t
Figure 13.37: Twenty-fifth partial sum and Cesáro sum in Problem 2, Section
13.7.
3. We find that
N
2
SN (t) = [cos(nπ/2) − (−1)n ] sin(nπt)
n=1
nπ
and
N
n 2
σN (t) = 1− [cos(nπ/2) − (−1)n ] sin(nπt)
n=1
N nπ
Figures 13.38, 13.39 and 13.40 compare the fifth, tenth and twenty-fifth
partial sums of these sums with f (t).
1
SN (t) = sin(3)
6
N
−1 n nπt n nπt
+ 3 sin(3)(−1) cos + nπ(−1 + (−1) cos(3)) sin
n=1
n2 π 2 − 9 3 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 351
0.5
0
-1 -0.5 0 0.5 1
t
-0.5
-1
Figure 13.38: Fifth partial sum and Cesáro sum in Problem 3, Section 13.7.
0.5
t
-1 -0.5 0 0.5 1
0
-0.5
-1
Figure 13.39: Tenth partial sum and Cesáro sum in Problem 3, Section 13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
352 CHAPTER 13. FOURIER SERIES
0.5
0
-1 -0.5 0 0.5 1
t
-0.5
-1
Figure 13.40: Twenty-fifth partial sum and Cesáro sum in Problem 3, Section
13.7.
and
1
σN (t) = sin(3)
6
N
n −1 n nπt
+ 1 − 3 sin(3)(−1) cos
n=1
N n2 π 2 − 9 3
nπt
+nπ(−1 + (−1)n cos(3)) sin
3
Figures 13.41, 13.42 and 13.43 compare the fifth, tenth and twenty-fifth
partial sums of these sums with f (t).
5. We find the partial sums
N
17 1 − (−1)n 5 − 6(−1)n
SN (t) = + cos(nπt) + sin(nπt) ,
4 n=1
n2 π 2 nπ
and
17
σN (t) =
4
N
17 n 1 − (−1)n 5 − 6(−1)n
+ + 1− cos(nπt) + sin(nπt) .
4 n=1
N n2 π 2 nπ
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 353
0.5
0
-3 -2 -1 0 1 2 3
t
-0.5
-1
Figure 13.41: Fifth partial sum and Cesáro sum in Problem 4, Section 13.7.
0.5
0
-3 -2 -1 0 1 2 3
t
-0.5
-1
Figure 13.42: Tenth partial sum and Cesáro sum in Problem 4, Section 13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
354 CHAPTER 13. FOURIER SERIES
0.5
0
-3 -2 -1 0 1 2 3
t
-0.5
-1
Figure 13.43: Twenty-fifth partial sum and Cesáro sum in Problem 4, Section
13.7.
Figures 13.44, 13.45 and 13.46 compare the fifth, tenth and twenty-fifth
partial sums of these sums with f (t).
6. The partial sums of the Fourier series are
N
2 n nπt
SN (t) = (1 − (−1) ) sin .
n=1
nπ t
The Cesáro, Hamming and Gaussian filtered N partial sums are, respec-
tively,
2
n
N
n nπt
σN (t) = 1− (1 − (−1) ) sin ,
n=1
nπ N t
N
2 nπt
HN (t) = (0.54 + 0.46 cos(πn/N ))(1 − (−1)n ) sin ,
n=1
nπ t
N
2 −n2 π2 /N 2 nπt
GN (t) = e (1 − (−1)n ) sin ,
n=1
nπ t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 355
0
-1 -0.5 0 0.5 1
t
Figure 13.44: Fifth partial sum and Cesáro sum in Problem 5, Section 13.7.
0
-1 -0.5 0 0.5 1
t
Figure 13.45: Tenth partial sum and Cesáro sum in Problem 5, Section 13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
356 CHAPTER 13. FOURIER SERIES
0
-1 -0.5 0 0.5 1
t
Figure 13.46: Twenty-fifth partial sum and Cesáro sum in Problem 5, Section
13.7.
0.5
t
-2 -1 0 1 2
0
-0.5
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 357
0.5
0
-2 -1 0 1 2
t
-0.5
-1
Figure 13.48: Fourier, Cesáro, Hamming, and Gauss partial sums, N = 10, in
Problem 6, Section 13.7.
0.5
t
-2 -1 0 1 2
0
-0.5
-1
Figure 13.49: Fourier, Cesáro, Hamming, and Gauss partial sums, N = 25, in
Problem 6, Section 13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
358 CHAPTER 13. FOURIER SERIES
0
-2 -1 0 1 2
t
-1
-2
N
2 nπt
SN (t) = 1 + (1 − 3(−1)n ) sin ,
n=1
nπ 2
2
n
N
nπt
σN (t) = 1 + 1− (1 − 3(−1)n ) sin ,
n=1
nπ N 2
N
2 nπt
HN (t) = 1 + (0.54 + 0.46 cos(πn/N ))(1 − 3(−1)n ) sin ,
n=1
nπ 2
N
2 −n2 π2 /N 2 nπt
GN (t) = 1 + e (1 − 3(−1)n ) sin .
n=1
nπ 2
Graphs of these partial sums are given for N = 5, 10, 25, respectively, in
Figures 13.50, 13.51 and 13.52.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
13.7. FILTERING OF SIGNALS 359
0
-2 -1 0 1 2
t
-1
-2
Figure 13.51: Fourier, Cesáro, Hamming, and Gauss partial sums, N = 10, in
Problem 7, Section 13.7.
0
-2 -1 0 1 2
t
-1
-2
Figure 13.52: Fourier, Cesáro, Hamming, and Gauss partial sums, N = 25, in
Problem 7, Section 13.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
360 CHAPTER 13. FOURIER SERIES
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 14
2.
∞ 10
|f (x)| dx = k dx = 20k,
−∞ −10
361
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
362 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
This converges to
⎧
⎪
⎨k for −10 < x < 10,
0 for |x| > 10,
⎪
⎩
k/2 for x = −1 and for x = 10.
∞
3. Certainly −∞ |f (x)| dx converges, and each Aω = 0 because f is an odd
function. Compute
1 π 2
Bω = f (t) sin(ωt) dt = (1 − cos(πω)).
π −π πω
This converges to
⎧
⎪
⎪−1/2 for x = −π,
⎪
⎪
⎪
⎪
⎨−1 for −π < x < 0,
0 for x = 0 and for |x| > π,
⎪
⎪
⎪
⎪1 for 0 < x < π,
⎪
⎪
⎩1/2 for x = π.
∞
4. Certainly −∞
|f (x)| dx converges. Compute
0 4
1 1
Aω = sin(t) cos(ωt) dt + cos(t) cos(ωt) dt
π −4 π 0
1
= (1 + sin(4(ω − 1)) − cos(4(ω − 1)))
2π(ω − 1)
1
− (1 − sin(4(ω + 1)) − cos(4(ω + 1)))
2π(ω + 1)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.1. THE FOURIER INTEGRAL 363
if |ω| = 1, and
1
Bω = (1 − cos(4(ω − 1)) + sin(4(ω − 1)))
2π(ω − 1)
1
+ (1 − cos(4(ω + 1))) − sin(4(ω + 1))
2π(ω + 1)
if |ω| = 1. We also take A1 = limω→1 Aω , with a similar assignment if
ω = −1. B1 and B−1 are treated the same way.
The Fourier integral representation of f (x) is
∞
(Aω cos(ωx) + Bω sin(ωx)) dω.
0
This converges to
⎧
⎪
⎪ 1/2 for x = 0,
⎪
⎨cos(4)/2 for x = 4,
⎪
⎪ − sin(4)/2 for x = −4,
⎪
⎩
f (x) for −4 < x < 0 and for 0 < x < 4.
∞
5. Clearly −∞
|f (x)| dx converges. Since f (x) is even, Bω = 0. Compute
1 100 2 2 100 2
Aω = t cos(ωt) dt = t cos(ωt) dt
π −100 π 0
100
2 t2 cos(ωt) 2t cos(ωt) 2 sin(ωt)
= + −
π ω ω2 ω3 0
20000 sin(100ω) 4 sin(100ω) 400 sin(100ω)
= − + .
πω πω 3 πω 2
The Fourier integral representation of f (x) is
∞
400 cos(100ω) 20000ω 2 − 4
+ sin(100ω) cos(ωx) dω.
0 πω 2 πω 3
This converges to
⎧
⎪ 2
⎨x for −100 < x < 100,
0 for |x| > 100,
⎪
⎩
5000 for x = 100 and for x = −100.
∞
6. −∞
|f (x)| dx converges. Compute
1 2π
Aω = |t| cos(ωt) dt
π −π
cos(πω) + πω sin(πω) + 2 cos2 (πω) − 3 + 4πω sin(πω) cos(πω)
=
πω 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
364 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
and
2π
1
Bω = |t| sin(ωt) dt
π −π
− sin(πω) + πω cos(πω) + 2 sin(πω) cos(πω) − 4πω cos2 (πω) + 2πω
= .
πω 2
The Fourier integral representation of f (x) is
∞
(Aω cos(ωx) + Bω sin(ωx)) dω.
0
This converges to
⎧
⎪
⎪|x| for −π < x < 2π,
⎪
⎨0 for x < −π and for x > 2π,
⎪
⎪π/2 for x = −π,
⎪
⎩
π for x = 2π.
∞
7. Certainly −∞
|f (t)| dt converges. Compute
π
1 4 cos(πω)(cos2 (πω) − 1)
Aω = sin(t) cos(ωt) dt =
π −3π π(ω 2 − 1)
and π
1 4 sin(πω) cos2 (πω)
Bω = sin(t) sin(ωt) dt = − .
π −3π π(ω 2 − 1)
The Fourier integral representation is
∞
(Aω cos(ωx) + Bω sin(ωx)) dω.
0
This converges to
where
1 1 1 1 5
Aω = cos(ωt) dt + cos(ωt) dt
π −5 2 π 1
sin(ω)
= 24 cos4 (ω) − 18 cos2 (ω) + 1
πω
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.1. THE FOURIER INTEGRAL 365
and
1 5
1 1
Bω = sin(ωt) dt + sin(ωt) dt
π
−5 2 1
2 cos(ω)
=− 4 cos4 (ω) − 5 cos2 (ω) + 1 .
πω
Interchange the order of integration and use the fact that f (t) cos(ω(t−x))
is an even function of ω to write this integral representation as
∞ ∞
1
f (t) cos(ω(t − x)) dω dt.
2π −∞ −∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
366 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.2. FOURIER COSINE AND SINE INTEGRALS 367
converging to ⎧
⎪
⎪1 for 0 < x < 1,
⎪
⎪
⎪
⎪ x = 1,
⎨3/2 for
2 for 1 < x < 4,
⎪
⎪
⎪
⎪1 for x = 0 and for x = 4,
⎪
⎪
⎩0 for x > 4.
The sine integral is
∞
2
(1 + cos(ω) − 2 cos(4ω)) sin(ωx) dω,
0 πω
converging to ⎧
⎪
⎪1 for 0 < x < 1,
⎪
⎪
⎪
⎪
⎨3/2 for x = 1,
2 for 1 < x < 4,
⎪
⎪
⎪
⎪1 for x = 4,
⎪
⎪
⎩0 for x = 0 and for x > 4.
converging to
⎧
⎪
⎪ cosh(x) for 0 < x < 5,
⎪
⎨cosh(5)/2 for x = 5,
⎪
⎪ 0 for x > 5
⎪
⎩
1 for x = 0.
The sine integral is
∞
2
2 + 1)
(5ω sinh(5) − ω cosh(5) cos(5ω) + ω) sin(ωx) dω,
0 π(ω
converging to
⎧
⎪
⎨cosh(x) for 0 < x < 5,
cosh(5)/2 for x = 5,
⎪
⎩
0 for x > 5 and for x = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
368 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
converging to
⎧
⎪
⎪1 + 2x for 0 < x < π,
⎪
⎪
⎪
⎪ x = π,
⎨(3 + 2π)/2 for
2 for π < x < 3π,
⎪
⎪
⎪1
⎪ for x = 3π and for x = 0,
⎪
⎪
⎩0 for x > 3π.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.2. FOURIER COSINE AND SINE INTEGRALS 369
converging to
Both converge to e−2x cos(x) for x > 0, while the sine integral converges
to 0 at 0, and the cosine integral converges to 1 at 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
370 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
11. From the Laplace integrals and the convergence theorem, we can write
2k ∞ 1
e−kx = cos(ωx) for x ≥ 0
π 0 k2 + ω2
and
2 ∞ ω
e−kx = sin(ωx) dω for x > 0.
π 0 k2 + ω2
Put k = 1 and interchange the symbols x and ω to obtain
∞
πe−ω 1
Aω = = cos(ωx) dx
2 0 1 + x2
and ∞
πe−ω x
Bω = = sin(ωx) dx.
2 0 1 + x2
From these it follows that the Fourier cosine integral representation of
1/(1 + x2 ) is
∞
1
C(x) = e−ω cos(ωx) dω = for x ≥ 0
0 1 + x2
and the Fourier sine integral for x/(1 + x2 ) is
∞
x
S(x) = e−ω sin(ωx) dω = for x > 0.
0 1 + x2
By direct computation, we also have S(0) = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.3. THE FOURIER TRANSFORM 371
1.4
1.2
0.8
0.6
0.4
0.2
0
-20 -10 0 10 20
w
2.5
1.5
0.5
0
-10 -5 0 5 10
w
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
372 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
40
30
20
10
0
-10 -5 0 5 10
w
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.3. THE FOURIER TRANSFORM 373
0
-10 -5 0 5 10
w
Then
4e−k/4
|fˆ(ω)|(ω) = √ .
1 + 16ω 2
The amplitude spectrum is shown in Figure 14.5 for k = 4.
6.
2
fˆ(ω) = 3 (k 2 ω 2 sin(kω) + 2kω cos(kω) − 2 sin(kω)).
ω
Figure14.6 shows the amplitude spectrum for k = 2 and ω > 0.
7.
fˆ(ω) = πe−|ω| .
The amplitude spectrum is shown in Figure 14.7.
8. Write
f (t) = 3e−6 H(t − 2)e−3(t−2)
so
e−2iω
fˆ(ω) = 3e−6 .
3 + iω
We can also write
e−2(3+iω)
fˆ(ω) = .
3 + iω
Then
3e−6
|fˆ(ω)| = .
9 + ω2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
374 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
1.4
1.2
0.8
0.6
0.4
0.2
-10 -5 0 5 10
w
60
50
40
30
20
10
0
1 2 3 4 5
w
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.3. THE FOURIER TRANSFORM 375
2.5
1.5
0.5
0
-4 -2 0 2 4
w
0.0008
0.0007
0.0006
0.0005
0.0004
0.0003
0.0002
-6 -4 -2 0 2 4 6
w
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
376 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
1.4
1.2
0.8
0.6
0.4
-8 -4 0 4 8
w
so
e−3(2+iω)
fˆ(ω) = .
2 + iω
Then
e−6
|fˆ(ω)| =
.
4 + ω2
The amplitude spectrum has the same appearance as that of Problem 8.
11.
2 −4it −8t2
f (t) = 18 e e
π
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.3. THE FOURIER TRANSFORM 377
12. Write
e−4(ω−5)i
fˆ(ω) = ,
3 + (ω − 5)i
so −4iω
e
f (t) = e5it fˆ−1 = e5it H(t − 4)e−3(t−4) .
3 + iω
13. Write
e2(ω−3)i
fˆ(ω) = ,
5 + (ω − 3)i
so
3it ˆ−1 e2iω
f (t) = e f
5 + iω
= e3it H(t + 2)e−5(t+2) = H(t + 2)e−(10+(5−3i)t) .
14. Write
10 sin(3ω) 10 sin(3(ω + π))
fˆ(ω) = =− ,
ω+π ω+π
so
−πit ˆ−1 2 sin(3ω)
f (t) = −5e f = 5e−πit (H(t + 3) − H(t − 3)).
ω
15. Write
1 + iω 2 1
fˆ(ω) = = − .
(3 + iω)(2 + iω) 3 + iω 2 + iω
Then
f (t) = H(t)(2e−3t − e−2t ).
16.
1
fˆ−1 = H(t)e−t ∗ H(t)e−2t
(1 + iω)(2 + iω)
∞
= H(τ )e−τ H(t − τ )e−2(t−τ )
−∞
t
= H(t)e−2t eτ dτ = H(t)e−2t (et − 1)
0
= H(t)(e−t − e−2t ).
17.
1
fˆ−1 = H(t)e−t ∗ H(t)e−t
(1 + iω)2
∞
= H(τ )e−τ H(t − τ )e−(t−τ ) dτ
−∞
t
= H(t)e−t dτ = H(t)te−t .
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
378 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
18.
sin(3ω) 1
fˆ−1 = (H(t + 3) − H(t − 3)) ∗ H(t)e−2t
(2 + iω)ω 2
1 ∞
= (H(t + 3) − H(t − 3))H(t − τ )e−2(t−τ ) dτ
2 −∞
t t
1 −2t 2τ 2τ
= e H(t + 3) e dτ − H(t − 3) e dτ
2 −3 3
1 1
= (1 − e−2(t+3) )H(t + 3) − (1 − e−2(t−3) )H(t − 3)
4 4
19. Compute
∞ ∞ ∞
1 1
|f (t)|2 dt = fˆ(ω)fˆ(ω) dω = |fˆ(ω)|2 dω.
−∞ 2π −∞ 2π −∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.3. THE FOURIER TRANSFORM 379
22. Let ŷ(ω) = fˆ[y(t)](ω) and transform the differential equation to obtain
Then
e−3iω
ŷ(ω) =
+ 6iω + 5 −ω 2
e−3iω 1 e−3iω e−3iω
= = − .
(1 + iω)(5 + iω) 4 1 + iω 5 + iω
23. Compute
5
fˆwin (ω) = t2 e−iωt dt
−5
2
= (25ω 2 sin(5ω) + 10ω cos(5ω) − 2 sin(5ω)).
ω3
Since w(t) = 1 and the support of g is [−5, 5], then tC = 0. For the RMS
bandwidth of the window function,
5 1/2
−5
t2 dt 10
wRMS = 2 5 =√ .
−5
dt 3
24. Compute
4π
fˆwin (ω) = cos(at)e−iωt dt
−4π
2
= (ω sin(4πω) cos(4aπ) − a cos(4πω) sin(4aπ)).
ω2 − a2
Since w(t) is constant on [−4π, 4π], tC = 0. We also have
4π 1/2
−4π
t2 dt 8π
wRMS = 2 4π =√ .
−4π
dt 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
380 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
25. Compute
4
1
fˆwin (ω) = e−t e−iωt dt = (1 − e−4(1+iω) )
0 1 + iω
1
= (1 − e−4 (cos(4ω) − i sin(4ω))(1 − iω)
1 + ω2
1 − e−4 cos(4ω) + e−4 sin(4ω)
=
1 + ω2
−4
e sin(4ω) + (e−4 cos(4ω) − 1)ω
+i .
1 + ω2
We also have 4
t dt
tC = 04 =2
0
dt
and 4 1/2
0
(t − 2)2 dt 4
wRMS = 2 4 =√ .
0
dt 3
26. Compute
1
fˆwin (ω) = et sin(πt)e−iωt dt
−1
1
= sin(πt)e(1−iω)t dt
−1
π 2 sinh(1)(1 + π 2 ) − 2ω 2 cos(ω) + cosh(1)ω sin(ω)
=
(1 + (π + ω)2 )(1 + (π − ω)2 )
π sinh(1)(2ω 2 sin(ω) − (2 + 2π 2 ) sin(ω)) + cosh(1)ω cos(ω)
+i .
(1 + (π + ω)2 )(1 + (π − ω)2 )
27. First,
2
fˆwin (ω) = (t + 2)2 e−iωt dt
−2
4
= 3
(4ω 2 − 1) sin(2ω) + 2ω cos(2ω)
ω
8i
+ 2 (2ω cos(2ω) − sin(2ω)) .
ω
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.4. FOURIER COSINE AND SINE TRANSFORMS 381
28. We have
5π
fˆwin (ω) = e−iωt dt
3π
1 5πiω
=− (e − e3πiω )
iω
2eπiω e4πiω − e−4πiω
=− = −2eπiω sin(4πω)
ω 2i
= −2 cos(πω) sin(4πω) − 2i sin(πω) sin(4πω).
Finally,
5π
t dt
tC = 3π5π = 4π
3π
dt
and 5π 1/2
3π
(t − 4π)2 dt 2π
wRMS = 2 5π =√ .
3π
dt 3
1. ∞
1
fˆC (ω) = e−t cos(ωt) dt =
0 1 + ω2
∞
ω
fˆS (ω) = e−t sin(ωt) dt =
0 1 + ω2
2.
a2 − ω 2
fˆC (ω) =
(a2 + ω 2 )2
2aω
fˆS (ω) =
(a2 + ω 2 )2
3.
1 sin(K(ω + 1)) sin(K(ω − 1))
fˆC (ω) = + for ω = ±1
2 ω+1 ω−1
K 1
fˆC (1) = fˆC (−1) = + sin(2K)
2 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
382 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
ω 1 cos((ω + 1)K) cos((ω − 1)K)
fˆS (ω) = − + for ω = ±1
ω2 − 1 2 ω+1 ω−1
1 1
fˆS (1) = (1 − cos(2K)), fˆS (−1) = − (1 − cos(2K))
4 4
4.
1
fˆC (ω) = (2 sin(Kω) − sin(2Kω))
ω
1
fˆS (ω) = (1 − 2 cos(Kω) + cos(2Kω))
ω
5.
1 1 1
fˆC (ω) = +
2 1 + (ω + 1)2 1 + (ω − 1)2
1 ω+1 ω−1
fˆS (ω) = +
2 1 + (ω + 1)2 1 + (ω − 1)2
6.
1
fˆC (ω) = (cosh(2K) cos(2Kω) − cosh(K) cos(Kω))
1 + ω2
1
+ (ω sinh(2K) sin(2Kω) − ω sinh(K) sin(Kω))
1 + ω2
and
1
fˆS (ω) = (cosh(2K) cosh(2Kω) − cosh(K) sin(Kω))
1 + ω2
1
+ (−ω sinh(2K) cos(2Kω) + ω sinh(K) cos(Kω)) .
1 + ω2
7. Suppose for each L > 0, f (4) (t) is piecewise continuous on [0, L], f (3) (t) is
continuous, and, as t → ∞, f (3) (t) → 0, f (t) → 0 and f (t) → 0. Then
we can integrate by parts four times to obtain
∞
(4)
FS [f (t)](ω) = f (4) (t) sin(ωt) dt
0
∞
= f (3) (t) sin(ωt) − ωf (t) cos(ωt) − ω 2 f (t) sin(ωt) + ω 3 cos(ωt)f (t)
0
∞
4
+ω sin(ωt)f (t) dt
0
= ω fˆS (ω) − ω 3 f (0) + ωf (0).
4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.5. THE DISCRETE FOURIER TRANSFORM 383
∞
FC [f (4) (t)](ω) = f (4) (t) cos(ωt) dt
0
∞
= f (3) (t) cos(ωt) + ωf (t) sin(ωt) − ω 2 f (t) cos(ωt) − ω 3 f (t) sin(ωt)
0
∞
4
+ω f (t) cos(ωt) dt
0
= ω 4 fˆC (ω) + ω 2 f (0) − f (3) (0).
5
D[u](k) = u(j)e−πkji/3
j=0
for k = −4, −3, −2, −1, 0, 1, 2, 3, 4. For Problems 1 through 6, these values were
computed using MAPLE and rounded to the five decimal places.
1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
384 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
2.
3.
4.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.5. THE DISCRETE FOURIER TRANSFORM 385
5.
6.
For Problems 7 through 12, the N − point inverse discrete Fourier transform
of the sequence [Uj ]N −1
j=0 is the sequence computed by
N −1
1
uj = Uk e2πijk/N .
N
k=0
Values were computed using MAPLE to nine decimal places, with results recorded
below to six places.
5
1
uj = (1 + i)k e2πijk/6 .
6
k=0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
386 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
We obtain
u0 ≈ −1.333333 + 0.166667i,
u1 ≈ −0.427030 + 0.549038i,
u2 ≈ −0.016346 + 0.561004i,
u3 ≈ 0.333333 + 0.5000000i,
u4 ≈ 0.849679 + 0.272329i,
u5 ≈ 1.593696 − 2.049038i.
8. Here, N = 6 and
4
1 −k 2nπijk/5
uj = (i )e .
5
k=0
We obtain
u( 0) ≈ 0.200000,
u1 ≈ 0.731375 − 0.531375i,
u2 ≈ −0.096262 + 0.296261i,
u3 ≈ 0.049047 + 0.150953i,
u4 ≈ 0.115838 + 0.084162.
9. N = 7 and
6
1 −ik 2nπijk/7
uj = (e )e .
7
k=0
u0 ≈ 0.103479 + 0.014751i,
u1 ≈ 0.933313 − 0.296094,
u2 ≈ −0.094163 + 0.088785i,
u3 ≈ −0.023947 + 0.062482i,
u4 ≈ 0.004307 + 0.051899i,
u5 ≈ 0.025788 + 0.043852i,
u6 ≈ 0.051222 + 0.034325i.
10. N = 5 and
4
1 2 2πijk/5
uj = (k )e .
5
k=0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.5. THE DISCRETE FOURIER TRANSFORM 387
11. N = 5 and
4
1
uj = (cos(k))e2πijk/5 .
5
k=0
Approximate values are
u0 ≈ −0.103896,
u1 ≈ 0.420513 + 0.294562i,
u2 ≈ 0.131434 + 0.031205i,
u3 ≈ 0.131434 − 0.031205i,
u4 ≈ 0.420513 − 0.294562i.
12. N = 6 and
5
1
uj = ln(k + 1)e2πijk/6 .
6
k=0
Approximate values are
u0 ≈ 1.096542,
u1 ≈ −0.249644 − 0.232302i,
u2 ≈ −0.201697 − 0.084840i,
u3 ≈ −0.193858,
u4 ≈ −0.201697 + 0.084840i,
u5 ≈ −0.249644 + 0.232302i.
for k = −3, −2, −1, 0, 1, 2, 3. These values were computed using MAPLE to nine
decimal places and are given below rounded to six places.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
388 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
k dk fk
-3 −0.005177 + 0.075984i 0.000346 + 0.075849i
-2 −0.011816 + 0.115622i −0.006293 + 0.115532i
-1 −0.051259 + 0.250780i −0.045737 + 0.250753i
0 0.454649 0.460171
1 −0.051259 − 0.250798i −0.045737 − 0.250753i
2 −0.011816 − 0.115622i −0.006293 − 0.115532i
3 −0.005177 − 0.075984i 0.000346 − 0.075849i
k dk fk
-3 0.007825 + 0.049165i 0.11551 + 0.049074i
-2 0.017079 + 0.071538i 0.020805 + 0.071478i
-1 0.058802 + 0.123155i 0.062528 + 0.123125i
0 0.316738 0.320464
1 0.058802 − 0.123155i 0.062528 − 0.123125i
2 0.017079 − 0.071538i 0.020804 − 0.071478i
3 0.007825 − 0.049165i 0.011551 − 0.049074i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.6. SAMPLED FOURIER SERIES 389
k dk fk
-3 0.005629 − 0.053051i 0.001733 − 0.052956i
-2 0.012665 − 0.078577i 0.008769 − 0.079514i
-1 0.050661 − 0.159155i 0.046765 − 0.159123i
0 0.333333 0.329437
1 0.050661 + 0.159155i 0.046765 + 0.159123i
2 0.012665 + 0.079577i 0.008769 + 0.079514i
3 0.005629 + 0.053052i 0.001733 + 0.052956i
k dk fk
-3 247.246215 − 515.579355i 201.215105 − 514.436038i
-2 452.586443 − 626.547636i 406.555000 − 625.785580i
-1 894.543813 − 612.101891i 848.512176 − 611.720909i
0 1304.231619 1258.199915
1 894.543813 + 612.1018911i 848.512177 + 611.720909i
2 452.586443 + 626.547636i 406.555000 + 625.785580i
3 247.246215 + 515.579355i
127
jp
Un = f e−2πijn/128 .
j=0
128
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
390 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
Then, with
Vn = Un for n = 0, 1, · · · , 10, 118, 119, · · · , 127
and
Vn = 0 for 11 ≤ n ≤ 117,
we obtain the DFT approximation
127
1
w= Vk e2πikt0 /p .
128
k=0
The nonzero values if Un (to six decimal places) are recorded below for each
problem, followed by the DFT approximation w and the difference
Finally,
var ≈ |S10 (1/8) − w| ≈ 0.034520.
Values of Un are given in Table 14.5.
2. Compute
1
1 2nπ + (2n2 π 2 − 1)i
d0 = , dn = t2 e−2πint dt = .
3 0 n3 π 3
The complex Fourier series is
∞
1 2nπ + (2n2 π 2 − 1)i 2nπit
+ e .
3 n3 π 3
n=−∞,n=0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.6. SAMPLED FOURIER SERIES 391
U0 255
U1 −1 + 40.735481i U118 −1 − 3.992224i
U2 −1 + 20.355468i U119 −1 − 4.453202i
U3 −1 + 13.556669i U120 −1 − 5.027339i
U4 −1 + 10.153170i U121 −1 − 5.763142i
U5 −1 + 8.107786i U122 −1 − 6.741452i
U6 −1 + 6.74152i U123 −1 − 8.107786i
U7 −1 + 5.763142i U124 −1 − 10.153170i
U8 −1 + 5.027339i U125 −1 − 13.556670i
U9 −1 + 4.453202i U126 −1 − 20.355468i
U10 −1 + 3.992224i U127 −1 − 40.735484i
U0 42.167969
U1 5.985858 + 20.367742i U118 −0.433837 − 1.996112i
U2 1.122442 + 10.177734i U119 −0.418629 − 2.226601i
U3 0.221810 + 6.778335i U120 −0.397367 − 2.513670i
U4 −0.093411 + 5.076585i U121 −0.366352 − 2.881571i
U5 −0.239312 + 4.053893i U122 −0.318566 − 3.370726i
U6 −0.318566 + 3.370726i U123 −0.239313 − 4.053893i
U7 −0.366352 + 2.881571i U124 −0.093411 − 5.076585i
U8 −0.397367 + 2.513670i U125 0.221810 − 6.678335i
U9 −0.418629 + 2.226601i U126 1.122442 − 10.177734i
U10 −0.433837 + 1.996112i U127 5.985857 − 20.367742i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
392 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
U0 58.901925
U1 −5.854287 − 32.096339i U118 0.647851 + 2.829713i
U2 −0.805518 − 14.788044i U119 0.633992 + 3.157208i
U3 0.044274 − 9.708611i U120 0.614603 + 3.565443i
U4 0.336014 − 7.235154i U121 0.586989 + 4.089267i
U5 0.470070 − 5.764387i U122 0.542633 + 4.787014i
U6 0.542633 − 4.787014i U123 0.470070 + 5.764387i
U7 0.586299 − 4.089267 U124 0.336014 + 7.235154i
U8 −3.565442i0.614603 U125 0.044274 + 9.708611i
U9 0.63391 − 3.157208i U126 −0.805518 + 14.788044i
U10 0.647851 − 2.829712i U127 −5.854287 + 32.096339i
Then
S10 (1/2) ≈ 0.2504564.
For the DFT approximation, compute From these we obtain
w ≈ 0.246560 + 0.156250(10−9 )i
and
var ≈ 0.003896.
Approximate values for Un are listed in Table 14.6.
3. Compute
w ≈ 1.042757 − 0.267410(10−9 )i
and
var ≈ 0.024403.
Approximate values for Un are given in Table 14.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.6. SAMPLED FOURIER SERIES 393
U0 31.907298
U1 9.553181 − 14.227057i U118 0.620232 + 1.951481i
U2 3.383468 − 9.071385i U119 0.649821 + 2.174758i
U3 1.847063 − 6.366930i U120 0.691097 + 2.451901i
U4 1.269469 − 4.860090i U121 0.751110 + 2.805358i
U5 0.994545 − 3.915837i U122 0.843127 + 3.271884i
U6 0.843127 − 3.271884i U123 0.994545 + 3.915837i
U7 0.741110 − 2.805358i U124 1.269469 + 4.860090i
U8 0.691097 − 2.451901i U125 1.847063 + 6.366930i
U9 0.649821 − 2.174658i U126 3.383468 + 9.071385i
U10 0.620232 − 1.951481i U127 9.553181 + 14.227057i
4. Compute
(e − 1)(1 + 2nπi)
d0 = e − 1, dn = .
1 + 4n2 π 2
The complex Fourier series of f (t) is
∞
e−1 (e − 1)(1 + 2nπi) 2nπit
e .
+ 1 + 4n2 π 2
n=−∞,n=0
Then
S10 (1/4) ≈ 0.827534 − 0.9(10−10 )i.
For the DFT approximation, compute From these obtain the DFT ap-
proximation
w ≈ 0.810504 − 0.954242(10−11 )i
with
var ≈ 0.017031.
Table 14.8 gives the approximate values for Un .
5. Compute
1 3nπ + (2n2 π 2 − 3)i
d0 =
, dn = .
4 4n3 π 3
The complex Fourier series is
∞
1 3nπ + (2n2 π 2 − 3)i 2nπit
+ e .
4 4n3 π 3
n=−∞,n=0
Then
S10 (1/4) ≈ −0.000729.
For the DFT calculations, we need
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
394 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
U0 31.501953
U1 9.228787 + 17.271595i U118 −0.400755 − 1.993017i
U2 1.933662 + 9.790716i U119 −0.377943 − 2.222355i
U3 0.582715 + 6.663663i U120 −0.346050 − 2.507623i
U4 0.109884 + 5.028208i U121 −0.299528 − 2.872545i
U5 −0.108968 + 4.029124i U122 −0.227849 − 3.356393i
U6 −0.227849 + 3.356393i U123 −0.108968 − 4.029124i
U7 −0.299528 + 2.872544i U124 0.109884 − 5.028208i
U8 −0.346050 + 2.507623i U125 0.582715 − 6.663663i
U9 −0.377943 + 2.222355i U126 1.933662 − 9.790715i
U10 −0.400755 + 1.993017i U127 9.228787 − 17.271595i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
14.7. DFT APPROXIMATION OF THE FOURIER TRANSFORM 395
U0 60.953531
U1 −94.509581 − 26.479226i U118 1.064899 + 6.040217i
U2 −11.274203 − 30.060278i U119 −.061000 + 6.737012i
U3 −3.690170 − 20.379819i U120 0.815252 + 7.604519i
U4 −1.331661 − 15.319442i U121 0.601640 + 8.715653i
U5 −0.286124 − 12.249522i U122 0.270130 − 10.191613i
U6 0.270130 − 10.191613i U123 −0.286124 + 12.249522i
U7 0.601640 − 8.715653i U124 −1.331661 + 15.319442i
U8 0.815252 − 7.604519i U125 −3.690170 + 20.379819i
U9 −0.961000 − 6.737012i U126 −11.274203 + 30.060278i
U10 1.064899 − 6.040217i U127 −0.945096 + 26.479226i
Then
1
fˆ(4) = (1 − i).
8
The DFT approximation to fˆ(4) with L = 3 and N = 512 is
511
3π 3πj
f e−3πij/64 ≈ 0.143860 − 0.124549i.
256 j=0 256
4 − ω2 4ω
fˆ(ω) = − 2 i.
(ω 2 + 4)2 (ω + 4)2
Then
fˆ(12) ≈ −0.006392 − 0.002191i.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
396 CHAPTER 14. THE FOURIER INTEGRAL AND TRANSFORMS
To compare this with the exact value, write the decimal expansion
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 15
Eigenfunction Expansions
y = a + bx
y = c1 eαx + c2 e−αx .
Now y(0) = c1 + c2 = 0, so
Next,
y (L) = 2c1 α cosh(αL) = 0.
But cosh(αL) > 0, and α > 0, so c1 = 0 and the problem has only the
trivial solution for λ > 0. This problem has no positive eigenvalue.
Case 3. λ < 0, say λ = −α2 , with α > 0. Now the differential equation
has the general solution
y = c1 cos(αx) + c2 sin(αx).
397
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
398 CHAPTER 15. EIGENFUNCTION EXPANSIONS
Now y(0) = c1 = 0, so
y = c2 sin(αx).
Next,
y (L) = c2 α cos(αL) = 0.
To have a nontrivial solution we want to be able to choose c2 = 0, so we
must have cos(αL) = 0. Then αL is a positive zero of the cosine function,
(2n − 1)π
αL = ,
2
in which n can be any positive integer. Then
√ (2n − 1)π
α= λ= .
2L
Since λ = α2 , the eigenvalues of this problem, indexed by n, are
2
(2n − 1)π
λn =
2L
for n = 1, 2, 3, · · · . Corresponding to each such eigenvalue we have the
eigenfunction
(2n − 1)π
ϕn (x) = sin x .
2L
Of course, any nonzero constant multiple of this eigenfunction is also an
eigenfunction.
Problems 2, 3 and 5 are solved by an analysis similar to that just done for
Problem 2 and also in Example 15.2. We therefore omit the details for the
solutions of these problems. Problems 6 through 10 are more involved and more
details are provided.
λ0 = 0, λn = n2 for n = 1, 2, · · · .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.1. GENERAL EIGENFUNCTION EXPANSIONS 399
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
400 CHAPTER 15. EIGENFUNCTION EXPANSIONS
8. The problem is regular on [0, 1]. The differential equation has character-
istic equation
r2 + 2r + (1 + λ) = 0,
with roots √
r = −1 ± λi.
The general solution is
√ √
y(x) = c1 ex cos( λx) + c2 e−x sin( λx).
so
λ n = n2 π 2
is an eigenvalue for each positive integer n. Corresponding eigenfunctions
are
ϕn (x) = e−x sin(nπx)
or any nonzero constant multiples of this function.
9. The problem is regular on [0, π]. The differential equation can be written
y + 2y + λy = 0
Consider cases on λ.
Case 1: 1 − λ = a2 > 0. The general solution is
Now
y(0) = c1 + c2 = 0
so c2 = −c1 . Next
y(π) = c1 e−π eaπ − e−π e−aπ .
eaπ − e−aπ = 0,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.1. GENERAL EIGENFUNCTION EXPANSIONS 401
10. The problem is regular on [0, 8]. Similar to the solution of Problem 9,
eigenvalues are
λ n = 8 + n2 π 2
and eigenfunctions are
ϕn (x) = e3x sin(nπx).
11. The eigenfunctions are ϕn (x) = sin(nπx/L). The coefficients in the eigen-
function expansion are
2 L 2
cn = (1 − ξ) sin(nπξ/L) dξ = (1 + (−1)n (L − 1))
L 0 nπ
for n = 1, 2, · · · . The expansion is
∞
2
1−x= (1 + (−1)n (L − 1)) sin(nπx/L)
n=1
nπ
for 0 < x < L. The fortieth partial sum of this series is compared to the
function in Figure 15.1 for L = 1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
402 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.5
x
0 0.5 1 1.5 2 2.5 3
0
-0.5
-1
-1.5
-2
for 0 < x < π. Figure 15.2 shows the thirtieth partial sum of this expan-
sion and the function itself.
13. From Problem 3 the eigenfunctions are
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.1. GENERAL EIGENFUNCTION EXPANSIONS 403
2.5
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
404 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.5
0
0 1 2 3 4
x
-0.5
-1
4 ((−1)n − 1)
cos(nx) = sin(2x)
π n2 − 4
n=1,n=2
for 0 < x < π. Figure 15.4 shows a graph of sin(2x) compared to the
thirtieth partial sum of this expansion.
15. The eigenfunctions are
ϕ0 (x) = 1, ϕn (x) = an cos(nx/3) + bn sin(nx/3) for n = 1, 2, · · · .
The coefficients in the eigenfunction expansion x2 on [−3π, 3π] are
3π
1
c0 = ξ 2 dξ = 3π 2 ,
6π −3π
3π
1 36
an = ξ 2 cos(nξ/3) dξ = 2 (−1)n for n = 1, 2, · · · ,
3π −3π n
and 3π
1
bn = ξ 2 sin(nξ/3) dξ = 0 for n = 1, 2, · · · .
3π −3π
The expansion is
∞
(−1)n
3π 2 + 36 cos(nx/3) = x2
n=1
n2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.1. GENERAL EIGENFUNCTION EXPANSIONS 405
0.5
x
0 0.5 1 1.5 2 2.5 3
0
-0.5
-1
for −3π < x < 3π. Figure 15.5 is a graph of this function compared to
the tenth partial sum of this expansion.
16. The eigenfunctions are ϕn (x) = e−x sin(nπx) for n = 1, 2, · · · . Notice that
the Sturm-Liouville form of the differential equation is
cn e−x sin(nπx)
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
406 CHAPTER 15. EIGENFUNCTION EXPANSIONS
80
60
40
20
0
-8 -4 0 4 8
x
1.2
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.1. GENERAL EIGENFUNCTION EXPANSIONS 407
1.2
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
Figure 15.6 shows a graph of f (x) compared to the thirtieth partial sum
of this expansion. This partial sum is not a very good fit to the function.
Figure 15.7 shows a graph of the function and the ninetieth partial sum,
a better fit. For improved accuracy we would have to take more terms in
the partial sum.
17. Normalized eigenfunctions for Problem 3 are obtained by dividing each
eigenfunction by its length, whose square is the dot product of this eigen-
function with itself.
4
(2n − 1)πx
cos2 dx = 2.
0 8
Therefore the normalized eigenfunctions are
1 (2n − 1)πx
ϕn (x) = √ cos ,
2 8
for n = 1, 2, 3, · · · . Now calculate the dot product
4
1 (2n − 1)πx
ϕn · f = √ x(4 − x cos dx
2 0 8
256 4(−1)n + (2n − 1)π
= −√ .
2 (2n − 1)3 π 3
Since 4
512
f ·f = x2 (4 − x)2 dx = ,
0 15
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
408 CHAPTER 15. EIGENFUNCTION EXPANSIONS
Further,
π
1
f ·f = e−2x dx = (1 − e−2π ) = e−π sinh(π).
0 2
Now Bessel’s inequality gives us
∞
2λ2n 2
−π
√ 1 + e cos( λ n π)
n=1
(1 + λn )2 (π + 2 cos2 ( λn π))
≤ e−π sinh(π).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.2. LEGENDRE POLYNOMIALS 409
1. For this problem, use the recurrence relation to write Pn+1 (x) in terms of
Pn (x) and Pn−1 (x):
2n + 1 n
Pn+1 (x) = xPn (x) + Pn−1 (x)
n+1 n+1
1
P6 (x) = (231x6 − 315x4 + 105x2 − 5)
16
1
P7 (x) = (429x7 − 693x5 + 315x3 − 35x)
16
1
P8 (x) = (6435x8 − 12012x6 + 6930x4 − 1260x2 + 35).
128
2. Using the Rodrigues formula, we obtain the following, which can be checked
with the polynomials listed previously:
1 d2
P2 (x) = ((x2 − 1)2 )
22 1! dx2
1 d2 4 1
= 2
(x − 2x2 + 1) = (3x2 − 1),
8 dx 2
1 d3 2 3
P3 (x) = ((x − 1) )
23 3! dx3
1 d3 6 1
= (x − 3x4 + 3x2 − 1) = (5x3 − 3x),
48 dx3 2
1 d4 2 4
P4 (x) = ((x − 1) )
24 4! dx4
1 d4 8 1
= (x − 4x6 + 6x4 − 4x2 + 1) = (35x4 − 30x2 + 3),
384 dx4 8
1 d5 2 5
P5 (x) = ((x − 1) )
25 5! dx5
1 d5 10
= (1 − 5x8 + 10x6 − 5x2 + 1)
3840 dx5
1
= (63x5 − 70x3 + 15x).
8
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
410 CHAPTER 15. EIGENFUNCTION EXPANSIONS
P0 (x) = (−1)0 x0 = 1,
(−1)0 2!
P1 (x) = x = x,
2
4! 2 (−1)2! 0 1
P2 (x) = (−1)0 x + x = (3x2 − 1);
2!2! 22 2
(−1)0 6! 3 (−1)4! 1
P3 (x) = 3 x + 3 x = (5x3 − 3x),
2 3!3! 2 2! 2
(−1)0 8! 4 6! 2 4! 1
P4 (x) = 4 x − 4 x + 4 = (35x4 − 30x2 + 3),
2 4!4! 2 3!2! 2 2!2! 8
(−1)0 10! 5 8! 3 6! 1
P5 (x) = 5
x − 5 x + 5 x = (63x5 − 70x3 + 15x).
2 5!5! 2 4!3! 2 2!3! 8
4. Attempt to find a second solution of the form Qn (x) = z(x)Pn (x). Sub-
stitute this into Legendre’s equation to obtain, after some manipulation,
z[(1−x2 )Pn −2xPn +n(n+1)Pn ]+z (1−x2 )Pn +z [2(1−x2 )Pn −2xPn ] = 0.
ln |z | + ln |1 − x2 | + 2 ln |Pn | = c.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.2. LEGENDRE POLYNOMIALS 411
Finally,
1
Q2 (x) = 2(3x2 − 1) dx
(3x2 − 1)2 (1 − x2 )
1 1 1 1 1
= (3x2 − 1) − + √ + √ dx
4 x + 1 x − 1 (x + 1/ 3)2 (x − 1/ 3)2
1 1+x 3
= (3x2 − 1) ln − x.
4 1−x 2
R2 = r2 + d2 − 2rd cos(θ)
so
R2 r r2
2
= 1 − 2 cos(θ) + 2 .
d d d
Then
1 1d 1 1
ϕ(x, y, z) = = = .
R dR d 1 − 2 r cos(θ) + r2
d d2
This concludes part (a). For (b), suppose r/d < 1. By comparing the
result of (a) with the generating function for Legendre polynomials (with
x = cos(θ) and t = r/d), we have
∞ n
1
r
ϕ(r) = Pn (cos(θ)) ,
d n=0 dn
which is equivalent to
∞
1
ϕ(r) = P (cos(θ))rn .
n+1 n
n=0
d
R2 d d2
2
= 1 − 2 cos(θ) + 2 .
r r r
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
412 CHAPTER 15. EIGENFUNCTION EXPANSIONS
Then
r 1
= .
R d
1 − 2 r cos(θ) + d2
r2
This is equivalent to
∞
1
n
ϕ(r) = d Pn (cos(θ))r−n .
r n=0
∞
1
√ = Pn (x)tn .
1 − 2xt + t2 n=0
Then
∞
2 1 1
√ = n
Pn .
3 n=0 2 2
7. One way to derive these results is to use the expression for Pn (x) given in
Problem 3. First,
2n + 1 1
= n+ = n,
2 2
so
n
(4n + 2 − 2k)!
P2n+1 (x) = (−1)k x2n+1−2k .
22n+1 k!(2n + 1 − k)!(2n + 1 − 2k)!
k=0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.2. LEGENDRE POLYNOMIALS 413
so
∞
(4n − 2k)!
P2n (x) = (−1)k x2n−2k .
22n k!(2n − k)!(2n − 2k)!
k=0
(−1)n (2n)!
P2n (0) = .
22n (n!)2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
414 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.5
0
-1 -0.5 0 0.5 1
x
-0.5
-1
c0 = c2 = c4 = 0, c1 = 1.215854203, c3 = −0.2248913308.
Figure 15.8 shows a graph of sin(πx/2) and this partial sum. These graphs
are nearly identical in the scale of the graphics.
10. Write
∞
e−x = cn Pn (x)
n=0
Figure
4 15.9 compares a graph of e−x with a graph of this partial sum
n=0 cn Pn (x).
Within the scale of the graph, e−x and this partial sum are nearly indis-
tinguishable.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.2. LEGENDRE POLYNOMIALS 415
2.5
1.5
0.5
-1 -0.5 0 0.5 1
x
11. Compute
c0 = 0.2726756433, c1 = 0, c2 = 0.4961198722,
c3 = 0, c4 = −0.06335726400.
Figure 15.10 shows a graph of this partial sum and the function.
13. Compute
c0 = 0, c1 = 1.500000000, c2 = 0,
c3 = −0.8750000000, c4 = 0.
Figure 15.12 shows a graph of this partial sum and the function. In this
case, many more terms of the eigenfunction expansion are needed to ap-
proximate the function with any accuracy. Figure 15.13 shows the function
and the fortieth partial sum of this expansion. This is a much better fit.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
416 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-1 -0.5 0 0.5 1
x
1.2
0.8
0.4
0
-1 -0.5 0 0.5 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.2. LEGENDRE POLYNOMIALS 417
0.5
0
-1 -0.5 0 0.5 1
x
-0.5
-1
0.5
x
-1 -0.5 0 0.5 1
0
-0.5
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
418 CHAPTER 15. EIGENFUNCTION EXPANSIONS
1.2
0.8
0.6
0.4
0.2
0
-1 -0.5 0 0.5 1
x
14. Compute
c0 = 0.8411909850, c1 = 0.7174008810, c2 = −0.3101752600,
c3 = −0.1820611100, c4 = −0.00909900000.
Figure 15.14 shows a graph of this partial sum and the function.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 419
c2 xa−2 [(bxc )2 Jν (bxc ) + bxc Jν (bxc ) + ((bxc )2 − ν 2 )Jν (bxc )] = 0.
2. We need
1 2 2 7
2a = , b c = 1, 2c − 2 = 0, and a2 − ν 2 c2 = .
3 144
Then
1 1
a=
, c = b = 1, ν =
3 4
so the general solution of the differential equation is
3. We need
4
1 − 2a = 1, b2 c2 = 4, 2c − 2 = 2, a2 − ν 2 c2 = − ,
9
so
1
a = 0, c = 2, b = 1, ν = .
3
The general solution is
For Problems 4 - 7, we will just give the values of a, b, c and ν and the general
solution.
4. a = 3, c = 4, b = 2, ν = 1/2, so
6. a = 2, c = 3, b = 1, ν = 2/3, so
7. a = 4, c = 3, b = 2, ν = 3/4, so
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
420 CHAPTER 15. EIGENFUNCTION EXPANSIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 421
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
422 CHAPTER 15. EIGENFUNCTION EXPANSIONS
Then
y(x) = c1 x2 J1/3 (x) + c2 x2 Y1/3 (x).
Then
y(x) = c1 x2/3 J1/2 (x) + c2 x2/3 Y1/2 (x).
√
16. We have u = y x, so y = x−1/2 u. The differential equation transforms to
2 −1/2 −3/2 3 −5/2
4x x u −x u + x u
4
1
+ 8x x1/2 u − x−1/2 x−1/2 u + (4x2 − 35)u = 0.
2
√
Collect terms and multiply by x/4 to obtain
Then
y(x) = c1 x−1/2 J3 (x) + c1 x−1/2 Y3 (x).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 423
Then
xv + v + β 2 xv = 0.
(b) Multiply the first equation by v and the second by u and then subtract
the resulting first equation from the second to obtain
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
424 CHAPTER 15. EIGENFUNCTION EXPANSIONS
By equation (15.21),
20. These are immediate from Problem 19. First, we know from the first
conclusion of Problem 19 that
sn Jn−1 (s) ds = sn Jn (s).
Let s = αx to obtain
αn xn Jn−1 (αx)α dx = αn xn Jn (αx).
21. Define 1
In,k = (1 − x2 )k xn+1 Jn (αx) dx.
0
Replace n with n + 1:
sn+1 Jn (s) ds = sn+1 Jn+1 (s).
Then α !
sn+1 Jn (s) ds = sn+1 Jn+1 (s) 0
α
= αn+1 Jn+1 (α).
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 425
Then
1
1
xn+1 Jn (αx) dx = Jn+1 (α).
0 α
But
1
In,0 = xn+1 Jn (αx) dx.
0
This completes part (b). Now, for (c), integrate the expression of (b) by
parts:
1
1 n+1 d
In,k = (1 − x2 )k
x Jn+1 (αx) dx
0 α dx
n+1
1
x
= (1 − x2 )k Jn+1 (αx)
α 0
1
1 n+1
− x Jn+1 (αx)k(1 − x2 )k−1 (−2x) dx
0 α
2k 1
= (1 − x2 )k−1 xn+2 Jn+1 (αx) dx
α 0
2k
= In+1,k−1 .
α
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
426 CHAPTER 15. EIGENFUNCTION EXPANSIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 427
Then
π/2
xm
Jm (x) = (cos2 (t))m−1/2 cos(x sin(θ)) cos(t) dt
2m−1 Γ(m + 1/2) 0
π/2
xm
= cos2m (θ) cos(x sin(θ)) dθ.
2m−1 Γ(m + 1/2) 0
cn J1 (jn x),
n=1
24. With f (x) = e−x , the fifth partial sum (Figure 15.15) of the Fourier-Bessel
function bears little resemblance to the function. Figure 15.16 shows the
thirty-fifth partial sum of this expansion.
25. With f (x) = x, then nth Fourier-Bessel coefficient for expanding in a
series of eigenfunctions J1 (jn x) is
1
2
cn = xJ1 (jn x) dx.
J2 (jn )2 0
Figure 15.17 shows the fifth partial sum, compared to the function in
this expansion. Clearly this fifth partial sum does not approximate the
function at all well. Figure 15.18 shows the function and the thirty-fifth
partial sum, suggesting convergence of the Fourier-Bessel expansion to the
function as more terms are included in the expansion.
26. Figures 15.19 and 15.20 show the fifth and thirty-fifth partial sums of this
Fourier-Bessel expansion.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
428 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 429
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
430 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0 0.2 0.4 0.6 0.8 1
x
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 431
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
x
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
432 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.2
x
0 0.2 0.4 0.6 0.8 1
0
-0.2
-0.4
-0.6
-0.8
-1
0.2
x
0 0.2 0.4 0.6 0.8 1
0
-0.2
-0.4
-0.6
-0.8
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 433
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
27. Figures 15.21 and 15.22 show the fifth and thirty-fifth partial sums of this
Fourier-Bessel expansion of f (x) = xe−x .
28. Figures 15.23 and 15.24 show the fifth and thirty-fifth partial sums of this
Fourier-Bessel expansion of f (x) = x cos(πx).
29. Figure 15.25 shows the fifth partial sum of the Fourier-Bessel expansion
of f (x) = sin(πx). This partial sum appears to be a good approximation
to the function.
cn J2 (jn x),
n=1
30. The fifth and twenty-fifth partial sums are given in Figures 15.26 and
15.27, respectively.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
434 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 435
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
31. Figures 15.28 and 15.29 show the fifth and twenty-fifth partial sums of
this expansion.
32. Figures 15.30 and 15.31 show the fifth and twenty-fifth partial sums.
33. The fifth and twenty-fifth partial sums are shown in Figures 15.32 and
15.33.
34. The fifth and twenty-fifth partial sums are given in Figures 15.34 and
15.35.
35. The fifth and twenty-fifth partial sums are shown in Figures 15.36 and
15.37.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
436 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 437
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
x
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
438 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
x
0.2
x
0 0.2 0.4 0.6 0.8 1
0
-0.2
-0.4
-0.6
-0.8
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 439
0.2
x
0 0.2 0.4 0.6 0.8 1
0
-0.2
-0.4
-0.6
-0.8
-1
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
440 CHAPTER 15. EIGENFUNCTION EXPANSIONS
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
15.3. BESSEL FUNCTIONS 441
t → 1, and
1
dt = du.
(1 + u)2
We obtain
1
B(x, y) = tx−1 (1 − t)y−1 dt
0
∞ x−1 y−1
u u 1
= 1− du
0 1+u 1+u (1 + u)2
∞
ux−1
= du,
0 (1 + u)x+y
Γ(x)Γ(y)
B(x, y) = .
Γ(x + y)
This can be done by a clever use of double integrals, but here is a proof
using the convolution operation of the Laplace transform. First, it is
routine to check that
Γ(x)
L[tx−1 ](s) = x .
s
This is consistent with the result obtained in Chapter Three for the case
that x = n, an integer greater than 1, since in that case Γ(x) = (x − 1)!.
From this, we have
−1 1 tx−1
L = .
sx Γ(x)
Using this, we will derive the result by computing an inverse Laplace
transform, first directly, then using the convolution theorem:
1 tx+y−1
L−1 (t) =
sx+y Γ(x + y)
1 1
= L−1 x ∗ L−1 y
s s
t x−1
u 1
= (t − u)y−1 du
0 Γ(x) Γ(y)
t
1
= ux−1 (t − u)y−1 du
Γ(x)Γ(y) 0
t
1 u y−1 y−1
= ux−1 1 − t du.
Γ(x)Γ(y) 0 t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
442 CHAPTER 15. EIGENFUNCTION EXPANSIONS
Now make the change of variables w = u/t in the last integral to continue
the last equation:
1
1
(wt)x−1 (1 − w)y−1 ty−1 dw
Γ(x)Γ(y) 0
1
1
= tx+y−1 wx−1 (1 − w)y−1 dw
Γ(x)Γ(y) 0
tx+y−1
= B(x, y).
Γ(x)Γ(y)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 16
∂2z
2 2 2
= −(n + m )c sin(nx) cos(my) cos( n2 + m2 ct),
∂t2
∂2z
2
= −n sin(nx) cos(my) cos( n2 + m2 ct),
∂x2
∂2z
2
= −m sin(nx) cos(my) cos( n2 + m2 ct).
∂y 2
Therefore
∂2y ∂2y ∂2y
= c2 + .
∂t2 ∂x2 ∂y 2
∂2y 1
= (c2 f (x + ct) + c2 f (x − ct))
∂t2 2
and
∂2y 1
= (f (x + ct) + f (x − ct)).
∂x2 2
443
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
444 CHAPTER 16. THE WAVE EQUATION
Then
∂2y ∂2y
2
= c2 2 .
∂t ∂x
1
y(0, t) = y(2π, t) = sin(ct)
c
∂y
(x, 0) = −c sin(x) sin(ct) + cos(x) cos(ct)|t=0 = cos(x).
∂t
for 0 < x < a, 0 < y < b and t > 0. Finally, the initial conditions are
∂z
z(x, y, 0) = f (x, y), (x, y, 0) = g(x, y).
∂t
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 445
where L
2
an = f (ξ) sin(nπξ/L) dξ
L 0
for n = 1, 2, · · · and
L
2
bn = g(ξ) sin(nπξ/L) dξ
L 0
for n = 1, 2, · · · . f is the initial position function and g is the initial velocity
function.
To write the solution in specific instances, we need only identify c and L and
compute the coefficients for the particular initial position and velocity functions.
Figure 16.3 shows the solution waves increasing in amplitude over times
t = 0.1, 0.3 and 0.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
446 CHAPTER 16. THE WAVE EQUATION
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2
x
1.5
0.5
0
0 1 2 3 4
-0.5 x
-1
-1.5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 447
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
x
4. The solution is
∞
4 1
y(x, t) = sin(x) cos(3t) + sin((2n − 1)x) sin(3(2n − 1)t).
3π n−1 (2n − 1)2
Figure 16.4 shows the wave profile at t = 0.4 (highest wave), t = 0.7
(lowest), and t = 0.9 (middle wave).
5. The solution is
∞
24 (−1)n+1 √
y(x, t) = 2
sin((2n − 1)x/2) cos((2n − 1) 2t).
π n=1 (2n − 1)
Figure 16.5 shows the waves moving downward through times t = 0.3, 0.5,
0.9 and 1.4.
6. The solution is
∞
5 1
y(x, t) = sin(nπx/5) [5 sin(4nπ/5) + nπ cos(4nπ/5)] sin(2nπt/5)
π 3 n=1 n3
Figure 16.6 shows the waves moving to the left through times t = 0.3, 0.7
and 1.2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
448 CHAPTER 16. THE WAVE EQUATION
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
-0.2 x
-0.4
-0.6
0
0 1 2 3 4 5 6
x
-2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 449
0.12
0.1
0.08
0.06
0.04
0.02
0
0 1 2 3 4 5
x
7. The solution is
∞
32 1
y(x, t) = − sin((2n − 1)πx/2) cos(3(2n − 1)πt/2)
π 3 n=1 (2n − 1)3
∞
4 1
+ [cos(nπ/4) − cos(nπ/2)] sin(nπx/2) sin(3nπt/2)
π 2 n=1 n2
Waves for this solution are shown in Figure 16.7, increasing in amplitude
for times t = 0.3, 0.4 and 0.5.
8. The solution is
∞
2 1
y(x, t) = sin(2x) cos(10t) + sin(nx) sin(5nt)
5 n=1 n2
In Figure 16.8, the lowest wave (near the origin) is at t = 0.3, then the
higher one at t = 0.5, and the middle wave at 0.7.
9. The differential equation is not separable, due to the 2x forcing term. Let
y(x, t) = Y (x, t) − h(x) and choose h to obtain a problem for Y that we
have solved. Substitute y into the partial differential equation to obtain
∂2Y ∂2Y
=3 − h + 2x.
∂t2 ∂x2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
450 CHAPTER 16. THE WAVE EQUATION
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2
x
0.5
x
0 0.5 1 1.5 2 2.5 3
0
-0.5
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 451
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2
x
This is the standard wave equation for Y if 3h (x) = 2x. For homogeneous
boundary conditions at 0 and 2, we need h(0) = h(2) = 0. Solve for h(x)
by two integrations to obtain
1 3
h(x) = (x − 4x).
9
Then Y (x, t) satisfies the standard problem
∂2Y ∂2Y
2
=3 2,
∂t ∂x
Y (0, t) = Y (2, t)0,
1 ∂Y
Y (x, 0) = h(x) = (x3 − 4x), (x, 0) = 0.
9 ∂t
Write the solution Y (x, t) and then
The waves increase in amplitude in Figure 16.9 through times t = 0.3, 0.5,
0.7 and 1.4.
10. Follow the ideas of Example 16.7 and the solution to Problem 9. Let
y(x, t) = Y (x, t) − h(x) and choose h to obtain a standard problem that
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
452 CHAPTER 16. THE WAVE EQUATION
x
0 1 2 3 4
0
-0.5
-1
-1.5
we have solved for Y . Substituting y(x, t) into the wave equation and the
boundary and initial conditions, we obtain
1
h(x) = (64x − x4 )
108
Y must satisfy the wave equation with c = 3 and the conditions
∂Y
Y (0, t) = Y (4, t) = 0, Y (x, 0) = h(x), (x, 0) = 0.
∂t
Solve this standard problem for Y and obtain
1
y(x, t) = (x4 − 64x)
108
∞
512 2(1 − (−1)n ) + n2 π 2 (−1)n
− 5 sin(nπx/4) cos(3nπt/4).
9π n=1 n5
Waves move downward in Figure 16.10 through times t = 0.3, 0.5, 0.7 and
1.6.
11. Let y(x, t) = Y (x, t) − h(x). Substitute y(x, t) into the wave equation and
use the boundary conditions to obtain a simpler problem for Y (that is,
one we have already solved). This occurs if
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 453
0.3
0.2
0.1
0
0 1 2 3 4 5 6
x
-0.1
-0.2
Then h(x) = cos(x) − 1 and Y (x, t) satisfies the wave equation with c = 1
and the conditions
∂Y
Y (0, t) = Y (2π, t) = 0, Y (x, 0) = cos(x) − 1, (x, 0) = 0.
∂t
Solve this familiar problem for Y to obtain
y(x, t) = 1 − cos(x)
∞
16 1
+ sin((2n − 1)x/2) cos((2n − 1)t/2).
π n=1 (2n − 1)((2n − 1)2 − 4)
a2 X (4) − λX = 0, T + a2 λT = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
454 CHAPTER 16. THE WAVE EQUATION
−A+C =0
− A cos(απ) − B sin(απ) + C cosh(απ) + D sinh(απ) = 0
−B+D =0
A sin(απ) + B cos(απ) + C sinh(απ) + D cosh(απ) = 0.
From the first and third equations, A = C and B = D. The second and
fourth equations become
cos(απ) cosh(απ) = 1.
where
sin(αn π) − sinh(αn π)
rn = .
cosh(αn π) − cos(αn π)
For each αn , we obtain
Case 3: λ < 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 455
B−D =0
A−B−C −D =0
− Aeαπ sin(απ) + Beαπ cos(απ) + Ce−απ sin(απ) − De−απ cos(απ) = 0
− Aeαπ (cos(απ) + sin(απ)) + Beαπ (cos(απ) − sin(απ))
− Ce−απ (cos(απ) − sin(απ)) − De−απ (cos(απ) + sin(απ)) = 0.
Coordinate these conditions with the conclusions of the above analysis for
part (b). There are three cases.
λ=0
As noted above, X(x) = A + Bx + Cx2 + Dx3 in this case. The boundary
conditions give us
and
X (π) = −B sin(απ) + D sinh(απ) = 0.
This 2 × 2 system has determinant sinh2 (απ), and this is zero exactly
when α = n = 1, 2, · · · . This gives eigenvalues λn = n4 and eigenfunctions
Xn (x) = sin(nx). We also obtain Tn (t) = An cos(an2 t) + Bn sin(an2 t).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
456 CHAPTER 16. THE WAVE EQUATION
λ<0
Using the previous analysis in this case, we have
X + λX = 0, X(0) = X(L) = 0,
and
T + AT + (B + c2 λ)T = 0, T (0) = 0.
Eigenvalues and eigenfunctions for X are
n2 π 2
λn = , Xn (t) = sin(nπx/L).
L2
With these eigenvalues, the characteristic equation of the differential equa-
tion for T is
r2 + Ar + (B + c2 n2 π 2 /L2 ) = 0,
with roots
−A 1 c2 n2 π 2
r= ± A2 − 4 B + .
2 2 L2
The given condition A2 L2 < 4(BL2 + c2 π 2 ) ensures that these roots are
complex. Let
rn2 = 4(BL2 + c2 n2 π 2 ) − A2 L2 .
The roots are then
A rn
r=− ± i.
2 2L
Then
Tn (t) = e−At/2 [an cos(rn t/2L) + bn sin(rn t/2L)] .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 457
14. Let y(x, t) = Y (x, t) + ψ(x). Substitute this into the wave equation to
obtain 2
∂2Y ∂ Y
= 9 + ψ (x) + 5x3 .
∂t2 ∂t2
This is simplified if ψ(x) is chosen so that
5
ψ (x) + x3 = 0.
9
Further,
∂2Y ∂2Y
2
=9 2,
∂t ∂x
Y (0, t) = Y (4, t) = 0,
∂Y
Y (x, 0) = cos(πx) − ψ(x), (x, 0) = 0.
∂t
Then
∞
1
y(x, t) = Y (x, t) + ψ(x) = an sin(nπx/4) cos(3nπt/4) + x(256 − x4 ),
n=1
36
where
4
2 1
an = cos(πx) − x(256 − x4 ) sin(nπx/4) dx
4 0 36
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
458 CHAPTER 16. THE WAVE EQUATION
2n(1−(−1)n ) 10240(−1)n (n2 π 2 −6)
π(n2 −16) + 9n5 π 5 for n = 4,
= 10(16π 2 −6)
9π 5 for n = 4.
Therefore the solution of the forced problem is
y(x, t) =
∞
2n(1 − (−1)n ) 10240(−1)n (n2 π 2 − 6)
+ sin(nπx/4) cos(3nπt/4)
π(n2 − 16) 9n5 π 5
n=1,n=4
10(16π 2 − 6) 1
+ 5
sin(πx) cos(3πt) + x(256 − x4 ).
9π 36
Without the forcing term, the problem has a solution of the form
∞
y(x, t) = αn sin(nπx/4) cos(3nπt/4)
n=1
where 4
2
αn = cos(πx) sin(nπx/4) dx
4 0
2n(1−(−1)n )
π(n2 −16) for n = 4,
=
0 for n = 4.
Figure 16.12 compares the forced wave (upper graph) with the unforced
solution at t = 0.4. Figure 16.13 does this for t = 0.8, and Figure 16.14
for t = 1.4.
15. Set y(x, t) = Y (x, t) + ψ(x). To simplify the problem for Y (x, t), choose
ψ(x) to satisfy
1
ψ (x) = − cos(πx), ψ(0) = ψ(4) = 0.
9
By integrating we find that
1
ψ(x) = (cos(πx) − 1).
9π 2
The solution Y (x, t) has the form
∞
Y (x, t) = an sin(nπx/4) cos(3nπt/4),
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 459
10
0
0 1 2 3 4
x
Figure 16.12: Forced and unforced motion in Problem 14, Section 16.2, at t =
0.4.
20
15
10
0
0 1 2 3 4
x
Figure 16.13: Forced and unforced motion in Problem 14, Section 16.2, at t =
0.8.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
460 CHAPTER 16. THE WAVE EQUATION
25
20
15
10
0
0 1 2 3 4
x
Figure 16.14: Forced and unforced motion in Problem 14, Section 16.2, at t =
1.4.
where
4
2 1
an = x(4 − x) − 2 (cos(πx) − 1) sin(nπx/4) dx
4 0 9π
−32(1−(−1)n )(288−17n2 )
9n3 π 3 (n2 −16) for n = 4,
=
0 for n = 4.
where 4
1 64(1 − (−1)n )
αn = x(4 − x) sin(nπx/4) dx = .
2 0 n3 π 3
Thus the solution for the unforced motion is
∞
64(1 − (−1)n )
y(x, t) = sin(nπx/4) cos(3nπt/4).
n=1
n3 π 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.2. WAVE MOTION ON AN INTERVAL 461
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4
x
Figure 16.15: Forced and unforced motion in Problem 15, Section 16.2, at t =
0.6.
Forced and unforced solutions at t = 0.6, 1 and 1.4 are shown, respectively,
in Figures 16.15, 16.16 and 16.17. In this example the forced motion is
very similar to the unforced motion.
16. Let y(x, t) = Y (x, t) + ψ(x) and obtain a simpler problem for Y (x, t) by
choosing ψ(x) to satisfy
1 −x
ψ (x) = e , ψ(0) = ψ(4) = 0.
9
Then
1
ψ(x) = (4e−x + (1 − e−4 )x − 4).
36
The solution for Y has the form
∞
Y (x, t) = an sin(nπx/4) cos(3nπt/4),
n=1
where, for n = 4,
4
1
an = (sin(πx) − ψ(x)) dx
2 0
32(1 − (−1)n e−4 )(n2 − 16)
=
9nπ(16n2 − 16n2 π 2 + n4 π 2 − 256)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
462 CHAPTER 16. THE WAVE EQUATION
x
0 1 2 3 4
0
-0.5
-1
-1.5
-2
-2.5
-3
Figure 16.16: Forced and unforced motion in Problem 15, Section 16.2, at t = 1.
x
0 1 2 3 4
0
-1
-2
-3
-4
Figure 16.17: Forced and unforced motion in Problem 15, Section 16.2, at t =
1.4.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.3. WAVE MOTION IN AN INFINITE MEDIUM 463
and, for n = 4,
1 − e−4 + 18π + 18π 3
a4 = .
18π(1 + π 2 )
The solution for the forced motion is
∞
y(x, t) = an sin(nπx/4) cos(3nπt/4)
n=1,n=4
1
+ a4 sin(πx) cos(3πt) + (4e−x + (1 − e−4 )x − 4).
36
Without the forcing term, the solution has the form
∞
y(x, t) = αn sin(nπx/4) cos(3nπt/4),
n=1
where
4
1 0 for n = 4,
αn = sin(πx) sin(nπx/4) dx =
2 0 1 for n = 1.
where
1 ∞
aω = f (ξ) cos(ωξ) dξ,
π −∞
1 ∞
bω = f (ξ) sin(ωξ) dξ,
π −∞
∞
1
αω = g(ξ) cos(ωξ) dξ,
πωc −∞
∞
1
βω = f (ξ) sin(ωξ) dξ.
πωc −∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
464 CHAPTER 16. THE WAVE EQUATION
0.8
0.4
x
0 1 2 3 4
0
-0.4
-0.8
Figure 16.18: Forced and unforced motion in Problem 16, Section 16.2, at t =
0.6.
0.5
x
0 1 2 3 4
0
-0.5
-1
Figure 16.19: Forced and unforced motion in Problem 16, Section 16.2, at t = 1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.3. WAVE MOTION IN AN INFINITE MEDIUM 465
0.8
0.4
x
0 1 2 3 4
0
-0.4
-0.8
Figure 16.20: Forced and unforced motion in Problem 16, Section 16.2, at t =
1.4.
1. Compute ∞
1 10
aω = e−5|ξ| cos(ωξ) dξ = .
π −∞ (25 + ω 2 )π
Immediately bω = 0 because e−5|ω| sin(ωω) is an odd function. With the
zero initial velocity condition, these are all the coefficients and the solution
is
10 ∞ 1
y(x, t) = cos(ωx) cos(12ωt) dω.
π 0 25 + ω 2
If we use the Fourier transform in x, take the transform of the wave
equation to obtain
ŷ + 144ω 2 ŷ = 0;
∞
10
ŷ(ω, 0) = e−5|ξ| e−iωξ dξ = ,
−∞ 25 + ω 2
ŷ (ω, 0) = 0.
The solution of this problem is
10
ŷ(ω, t) = cos(12ωt).
25 + ω 2
Invert this to obtain the solution
∞
1 10 iωx
y(x, t) = Re cos(12ωt)e dω .
2π −∞ 25 + ω 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
466 CHAPTER 16. THE WAVE EQUATION
and 8
1 8ω − sin(8ω)
(8 − ξ) sin(ωξ) dξ = .
π 0 πω 2
The solution is
∞
1 − cos(8ω) 8ω − sin(8ω)
y(x, t) = cos(ωx) + sin(ωx) cos(8ωt) dω.
0 πω 2 πω 2
To solve this problem using the Fourier transform, apply the transform to
the initial-boundary value problem to obtain:
ŷ + 64ω 2 ŷ = 0;
8
1 − 8ωi − e−8ωi
ŷ(ω, 0) = (8 − ξ)e−iωξ dξ = ;
0 ω2
ŷ (ω, 0) = 0.
1 − 8ωi − e−8ωi
ŷ(ω, t) = cos(8ωt).
ω2
Invert and take the real part to obtain the solution
∞
1 1 − 8ωi − e−8ωi iωx
y(x, t) = Re cos(8ωt)e dω .
2π −∞ ω2
and π
1 sin(πω)
sin(ξ) sin(ωξ) dξ = − .
4πω −π 2πω(ω 2 − 1)
The solution is
∞
sin(πω)
y(x, t) = − sin(ωx) sin(4ωt) dω.
0 2πω(ω 2 − 1)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.3. WAVE MOTION IN AN INFINITE MEDIUM 467
ŷ + 16ω 2 ŷ = 0;
ŷ(ω, 0) = 0;
π
2i sin(πω)
ŷ(ω, 0) = sin(ξ)e−iωξ dξ = .
−π ω2 − 1
and 2
1
(2 − |ξ|) sin(ωξ) = 0.
π −2
The solution is
∞
2
y(x, t) = (2 − cos(2ω)) cos(ωx) cos(ωt) dω.
0 πω 2
ŷ + ω 2 ŷ = 0;
2
2
ŷ(ω, 0) = (2 − |ξ|)e−iωξ dξ = 2 (1 − cos(2ω));
−2 ω
ŷ (ω, 0) = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
468 CHAPTER 16. THE WAVE EQUATION
To obtain the solution using the Fourier transform, first transform the
problem to obtain
ŷ + 9ω 2 ŷ = 0;
ŷ(ω, 0) = 0;
(2 − iω)e−(2+iω)
ŷ (ω, 0) = F(e−2x H(x − 1)) = .
4 + ω2
This problem has solution
(2 − iω)e−(2+iω)
ŷ(ω, t) = sin(3ωt).
3ω(4 + ω 2 )
Invert this to obtain the solution
∞
1 (2 − iω)e−(2+iω) iωx
y(x, t) = Re sin(3ωt)e dω .
2π −∞ 3ω(4 + ω 2 )
and 2
1 1 − cos(2ω)
g(ξ) sin(ωξ) dξ = .
2πω −2 πω 2
The solution is
∞
1 − cos(2ω)
y(x, t) = sin(ωx) sin(2ωt) dω.
0 πω 2
To solve the problem using the Fourier transform, first obtain the trans-
formed problem
ŷ + 4ω 2 ŷ = 0;
ŷ(ω, 0) = 0;
2
2(1 − cos(2ω))
ŷ (ω, 0) = g(ξ)e−iωξ dξ = .
−2 ω
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.4. WAVE MOTION IN A SEMI-INFINITE MEDIUM 469
where ∞
2
aω = f (ξ) sin(ωξ) dξ
π 0
and ∞
2
bω = g(ξ) sin(ωξ) dξ.
πcω 0
To solve the problem using the Fourier sine transform, first take the trans-
form of the initial-boundary value problem to obtain
ŷS + 9ω 2 ŷS = 0;
1
2(1 − cos(ω)) − ω sin(ω)
ŷS (ω, 0) = ξ(1 − ξ) sin(ωξ) dξ = ;
0 ω3
ŷS (ω, 0) = 0.
The solution of this transformed problem is
2(1 − cos(ω)) − ω sin(ω)
ŷS (ω, t) = cos(3ωt).
ω3
Invert this to obtain the solution
2 ∞ 2(1 − cos(ω)) − ω sin(ω)
y(x, t) = sin(ωx) cos(3ωt) dω.
π 0 ω3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
470 CHAPTER 16. THE WAVE EQUATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.4. WAVE MOTION IN A SEMI-INFINITE MEDIUM 471
4. Compute bω = 0 and
∞
2 4ω
aω = −2e−ξ sin(ωξ) dξ = − .
π 0 π(1 + ω 2 )
This yields the solution
∞
4 ω
y(x, t) = − sin(ωx) cos(6ωt) dω.
π 0 1 + ω2
For the solution by Fourier sine transform, first transform the problem to
obtain
To use the Fourier sine transform, first transform the problem to obtain
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
472 CHAPTER 16. THE WAVE EQUATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.5. LAPLACE TRANSFORM TECHNIQUES 473
Then
Y − 6sY + 9s2 Y = 0.
This has characteristic equation
Now
L[y(0, t)](s) = Y (0, s) = c1
so
Y (x, s) = c2 xe3xs .
Next,
L[y(2, t)](s) = F (s) = 2c2 e6s .
Then
1 −6s
c2 = e F (s),
2
and
1 −6s 1
Y (x, s) = e F (s)xe3xs = xF (s)e(3x−6)s .
2 2
The solution is
1
y(x, t) = xf (t − (6 − 3x))H(t − (6 − 3x)).
2
lim Y (x, s) = 0.
s→∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
474 CHAPTER 16. THE WAVE EQUATION
A
Y (x, s) = c2 e−sx/c − .
s4
Next, y(0, t) = 0, so
A
Y (0, s) = c2 − ,
s4
so
A
c2 = .
s4
Finally we have
A −sx/c A
Y (x, s) = e − 4.
s4 s
Then
A
x 3
x A 3
y(x, t) = t− H t− − t .
6 c c 6
Then
s2
y − Y = 0,
c2
with general solution
lim Y (x, s) = 0
x→∞
Y (x, s) = c2 e−sx/c .
Y (0, s) = c2 = F (s)
so
Y (x, s) = e−sx/c F (s).
The solution is
x
x
y(x, t) = f t − H t− .
c c
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.6. D’ALEMBERT’S SOLUTION 475
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
476 CHAPTER 16. THE WAVE EQUATION
5.
1 x−14t
u(x, t) = e + ex+14t + xt = ex cosh(14t) + xt
2
6.
u(x, t) = x2 + 144t2 − 5x + 3t
7.
1 1 x+4t −ξ
u(x, t) = (f (x − 4t) + f (x + 4t)) + e dξ
2 8 x−4t
1 t x+4t−4η
+ (ξ + η) dξ dη
8 0 x−4t+4η
1 1 1
= x + e−x sinh(4t) + xt2 + t3
4 2 6
8.
1 1 x+2t
u(x, t) = (f (x − 2t) + f (x + 2t)) + 2ξ dξ
2 4 x−2t
1 t x+2t−2η
+ 2ξη dξ dη
4 0 x−2t+2η
1 1
= (sin(x − 2t) + sin(x + 2t)) + 2xt + xt3
2 3
9.
x+8t
1 1
u(x, t) = (f (x − 8t) + f (x + 8t)) + cos(2ξ) dξ
2 16 x−8t
t x+8t−8η
1
+ η cos(ξ) dξ dη
16 0 x−8t+8η
1 1
= x2 + 64t2 − x + (sin(2(x + 8t)) − sin(2(x − 8t))) + xt4
32 12
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.6. D’ALEMBERT’S SOLUTION 477
10.
1 1 x+4t −ξ
u(x, t) = (f (x − 4t) + f (x + 4t)) + ξe dξ
2 8 x−4t
1 t x+4t−4η
+ ξ sin(η) dξ dη
8 0 x−4t+4η
1 1
= x2 + 16t2 + xe−x sinh(4t) + e−x sinh(4t)
2 4
− te−x cosh(4t) − x sin(t) + xt
11.
x+3t
1 1
u(x, t) = (f (x − 3t) + f (x + 3t)) + dξ
2 6 x−3t
t x+3t−3η
+ 3ξη 3 dξ dη
0 x−3t+3η
1 9
= (cosh(x − 3t) + cosh(x + 3t)) + t + xt5
2 10
12.
x+7t
1 1
u(x, t) = (f (x − 7t) + f (x + 7t)) + sin(ξ) dξ
2 14 x−7t
t x+7t−7η
1
+ (ξ − cos(η)) dξ dη
14 0 x−7t+7η
1 1
= 1 + x − (cos(x − 7t) − cos(x + 7t)) + xt2 + cos(t)
14 2
For each of Problems 13 - 16, we give graphs of the wave position at selected
times.
13. In Figures 16.21 through 16.25, the wave is shown at times t = 1/2, 1, 2,
3 and 4.
14. Figures 16.26 through 16.30 show the wave profile at times t = 1/2, 2/3,
7/8, 1.2 and 3.
15. Figures 16.31 through 16.34 show graphs of the solution at times t = 1/2,
t = 0.9, t = 1.3 and t = 1.8.
16. Figures 16.35 through 16.38 show wave positions at times t = 1/2, t = 0.7,
t = 0.9 and t = 1.3.
17. Figures 16.39 through 16.41 show wave positions at times t = 1, 1.4, 1.7.
18. Figures 16.42 through 16.45 show the wave at times t = 0.7, 1.4, 1.7, 2.2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
478 CHAPTER 16. THE WAVE EQUATION
0.4
0.2
x
-6 -4 -2 0 2 4 6
0
-0.2
-0.4
0.4
0.2
0
-6 -4 -2 0 2 4 6
x
-0.2
-0.4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.6. D’ALEMBERT’S SOLUTION 479
0.6
0.4
0.2
0
-6 -4 -2 0 2 4 6
x
-0.2
-0.4
-0.6
0.4
0.2
0
-6 -4 -2 0 2 4 6
x
-0.2
-0.4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
480 CHAPTER 16. THE WAVE EQUATION
0.4
0.2
x
-6 -4 -2 0 2 4 6
0
-0.2
-0.4
0.6
0.5
0.4
0.3
0.2
0.1
0
-4 -2 0 2 4
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.6. D’ALEMBERT’S SOLUTION 481
0.5
0.4
0.3
0.2
0.1
0
-4 -2 0 2 4
x
0.5
0.4
0.3
0.2
0.1
0
-3 -2 -1 0 1 2 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
482 CHAPTER 16. THE WAVE EQUATION
0.4
0.3
0.2
0.1
0
-4 -2 0 2 4
x
0.5
0.4
0.3
0.2
0.1
0
-6 -4 -2 0 2 4 6
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.6. D’ALEMBERT’S SOLUTION 483
0.8
0.6
0.4
0.2
0
-4 -2 0 2 4
x
0.6
0.5
0.4
0.3
0.2
0.1
0
-4 -2 0 2 4
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
484 CHAPTER 16. THE WAVE EQUATION
0.5
0.4
0.3
0.2
0.1
0
-4 -2 0 2 4
x
0.5
0.4
0.3
0.2
0.1
0
-4 -2 0 2 4
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.6. D’ALEMBERT’S SOLUTION 485
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-3 -2 -1 0 1 2 3
x
0.5
0.4
0.3
0.2
0.1
0
-3 -2 -1 0 1 2 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
486 CHAPTER 16. THE WAVE EQUATION
0.5
0.4
0.3
0.2
0.1
0
-3 -2 -1 0 1 2 3
x
0.5
0.4
0.3
0.2
0.1
0
-3 -2 -1 0 1 2 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.7. VIBRATIONS IN A CIRCULAR MEMBRANE I 487
x
-4 -2 0 2 4
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
where jn is the nth zero of J0 (x). For a given initial displacement f (r), the
coefficients are 1
2
z(r, t) = sf (s)J0 (jn s)
J1 (jn )2 0
for n = 1, 2, · · · .
Figure 16.46 shows the displacement at times t = 0.05, 0.25, 0.5, 0.75 and
1.25.
2. With f (r) = 1 − r2 the coefficients are approximately
Figure 16.47 shows the displacement at times t = 0.05, 0.25, 0.5, 0.75 ad
1.25.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
488 CHAPTER 16. THE WAVE EQUATION
x
-4 -2 0 2 4
0
-0.2
-0.4
-0.6
-0.8
-1
x
-4 -2 0 2 4
0
-0.2
-0.4
-0.6
-0.8
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.7. VIBRATIONS IN A CIRCULAR MEMBRANE I 489
0
-4 -2 0 2 4
x
2.5
1.5
0.5
0
-4 -2 0 2 4
x
-0.5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
490 CHAPTER 16. THE WAVE EQUATION
2.5
1.5
0.5
0
-4 -2 0 2 4
x
-0.5
2.5
1.5
0.5
0
-4 -2 0 2 4
x
-0.5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.7. VIBRATIONS IN A CIRCULAR MEMBRANE I 491
0.6
0.4
0.2
x
0 0.2 0.4 0.6 0.8 1
0
-0.2
-0.4
-0.6
0.5
x
0 0.2 0.4 0.6 0.8 1
0
-0.5
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
492 CHAPTER 16. THE WAVE EQUATION
0.5
x
0 0.2 0.4 0.6 0.8 1
0
-0.5
-1
-1.5
Figure 16.48 shows the displacement at times t = 0.05, 0.25, 0.5, 0.75 ad
1.25.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.8. VIBRATIONS IN A CIRCULAR MEMBRANE II 493
and
∞
4 − r2 j2k
− = α2 (r) = a2k J2 r .
2 2
k=1
Further, αn (r) = 0 for n = 2 and βn (r) = 0 for n ≥ 0, from which it
follows that
ank = 0 for n = 0, n = 2, k ≥ 1
and
bnk = 0 for n ≥ 0, k ≥ 1.
Finally, using the orthogonality of the Bessel functions J0 (j0k r/2) for k =
1, 2, · · · , and J2 (j0k r/2), we can calculate the coefficients as
1
2
a0k = ξ(1 − ξ 2 )J0 (j0k ξ) dξ for k ≥ 1
[J1 (j0k )]2 0
and 1
4
a2k = ξ(ξ 2 − 1)J2 (j2k ξ) dξ for k ≥ 1.
[J3 (j2k )]2 0
Using MAPLE, we can carry out numerical approximations of coefficients
in the solution. Some of the terms are
z(r, θ, t) ≈ 1.108022J0 (1.202413r) cos(2.404826t) − 0.139778J0 (2.760039r) cos(5.520078t)
+ 0.045476J0 (4.326864r) cos(8.653728t) + · · ·
− 2.976777J2 (2.567811r) cos(5.135622t) cos(2θ)
− 1.434294J2 (4.208622r) cos(8.417244t) cos(2θ)
− 1.140494J2 (5.809921r) cos(11.619841t) cos(2θ) + · · · .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
494 CHAPTER 16. THE WAVE EQUATION
We need
∞
∞
z(x, y, 0) = cnm sin(nx/2) sin(my/2) = x2 sin(y).
n=1 m=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
16.9. VIBRATIONS IN A RECTANGULAR MEMBRANE 495
The solution is
z(x, y, t)
∞
−8 1 n 2 2 n
= (2(1 − (−1) ) + n π (−1) ) sin(nx/2) sin(y) cos( 4 + n2 t/2).
n=1
π n3
z(x, y, t) =
∞ ∞
sin(nx) sin(my)(anm cos(3 n2 + m2 t)
n=1 m=1
+ bnm sin(3 n2 + m2 t)).
∂z ∞ ∞
(x, y, 0) = 3bnm n2 + m2 sin(nx) sin(my) = xy.
∂t n=1 m=1
Then
π
1 2 π 2
bnm = √ x sin(nx) dx y sin(my) dy
3 n2 + m 2 π 0 π 0
4 π(−1)n+1 π(−1)m+1
= √
3π 2 n2 + m2 n m
4(−1)n+m
= √ .
3nm n2 + m2
The solution is
z(x, y, t) =
∞ ∞
4 (−1)n+m
√ sin(nx) sin(my) sin(3 n2 + m2 t).
3 n=1 m=1 nm n2 + m2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
496 CHAPTER 16. THE WAVE EQUATION
z(x, y, t) =
∞ ∞
(anm cos( n2 + m2 t) + bnm sin( n2 + m2 t)) sin(nx/2) sin(my/2).
n=1 m=1
∂z ∞ ∞
(x, y, 0) = bnm n2 + m2 sin(nx/2) sin(my/2) = 1.
∂t n=1 m=1
Then
1 1 2π 1 2π
bnm = √ sin(nx/2) dx sin(my/2) dy
n2 + m2 π 0 π 0
1 2(1 − (−1)n ) 2(1 − (−1)m )
= √ .
π n2 + m2 n m
z(x, y, t) =
∞ ∞
16 √
+ 2 cnm sin((2n − 1)x/2) sin((2m − 1)y/2) sin( αnm t),
π n=1 m=1
where
1
cnm = √
(2n − 1)(2m − 1) αnm
and
αnm = (2n − 1)2 + (2m − 1)2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 17
∂u ∂2u
= k 2 for t > 0, 0 < x < L,
∂t ∂x
u(0, t) = α(t), u(L, t) = β(t) for t > 0,
u(x, 0) = f (x) for 0 < x < L.
∂u ∂2u
= k 2 for t > 0, 0 < x < L,
∂t ∂x
with
∂u
(0, t) = 0, u(L, t) = β(t) for t > 0,
∂x
u(x, 0) = f (x) for 0 < x < L.
497
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
498 CHAPTER 17. THE HEAT EQUATION
hence ∞
2
cn = f (ξ) sin(nπξ/L) dξ.
L 0
Figure 17.2 shows this temperature function at times t = 0.2, 0.4 and 1.3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 499
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
0
0 0.5 1 1.5 2 2.5 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
500 CHAPTER 17. THE HEAT EQUATION
0
0 0.5 1 1.5 2 2.5 3
x
c4 = c6 = · · · = ceven = 0.
u(x, t) =
∞
16L 1 2 2 2
− sin((2n − 1)πx/L)e−3(2n−1) π t/L .
π n=1 (2n − 1)((2n − 1)2 − 4)
Figure 17.3 shows the temperature function at times t = 0.2, 0.5 and 1.1.
n2 π 2
λn = , Xn (x) = cos(nπx/L).
L2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 501
where L
2
cn = f (ξ) cos(nπξ/L) dξ for n = 0, 1, 2, · · · .
L 0
4. Compute
2 L
c0 = f (ξ) cos(nπξ/L) dξ
L 0
π
2 4
= sin(ξ) dξ = , and
π 0 π
2 π
cn = sin(ξ) cos(nξ) dξ
π 0
=
0 for n = 1, 3, 5, · · · ,
4 1
− π n2 −1 for n = 2, 4, · · · .
The solution is
2 4 1 2
u(x, t) = − 2
cos(2nx)e−4n t .
π π 4n − 1
Figure 17.4 shows the temperature function at times t = 0.2, 0.4 and 0.7.
5. Compute
2π
1 4π 2
c0 = ξ(2π − ξ) dξ = ,
π 0 3
and
2π
1 4
cn = ξ(2π − ξ) cos(nx) dξ = − for n = 1, 2, · · · .
π 0 n2
The solution is
∞
2π 2 1 2
u(x, t) = −4 2
cos(nx)e−4n t .
3 n=1
n
Figure 17.5 shows the solution at times t = 0.2, 0.4 and 0.7.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
502 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
x
10
0
0 1 2 3 4 5 6
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 503
0
0 0.5 1 1.5 2 2.5 3
x
6. Compute
3
2
c0 = ξ 2 dξ = 6
3 0
and, for n = 1, 2, · · · ,
3
2 36(−1)n
cn = ξ 2 cos(nπξ/3) dξ = .
3 0 n2 π 2
The solution is
∞
36 (−1)n 2 2
u(x, t) = 3 + cos(nπx/3)e−4n π t/9 .
π 2 n=1 n2
The solution is shown in Figure 17.6 at times t = 0.2, 0.4 and 0.8.
7. Compute
6
2 1
c0 = e−ξ dξ = 1 − e−6 ,
6 0 3
and, for n = 1, 2, · · · ,
6
1
cn = e−ξ cos(nπξ/6) dξ
3 0
12
= 1 − (−1)n e−6 .
36 + n2 π 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
504 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 1 2 3 4 5 6
x
The solution is
∞
1
12
2 2
u(x, t) = 1 − e−6 + 2 2
1 − (−1)n e−6 cos(nπx/6)e−n π t/18 .
6 n=1
36 + n π
Figure 17.7 shows the temperature function at t = 0.2, 0.4 and 0.8.
8. The initial-boundary value problem is
∂u ∂2u
= k 2,
∂t ∂x
∂u ∂u
(0, t) = (L, t) = 0,
∂x ∂x
u(x, 0) = B.
The coefficients in the series solution are
2 L
c0 = B dξ = 2B
L 0
and
2 L
cn = B cos(nπξ/L) dξ = 0
L 0
for n = 1, 2, · · · . The solution is
u(x, t) = B.
This is consistent with intuition.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 505
∂u ∂2u
= k 2,
∂t ∂x
∂u
u(0, t) = (L, t) = 0,
∂x
u(x, 0) = B.
X + λX = 0, X(0) = 0, X (L) = 0
and
T + λkT = 0.
The problem for X(x) is routine and we obtain the eigenvalues and eigen-
functions
(2n − 1)2 π 2
λn = and Xn (x) = sin((2n − 1)πx/2L).
4L2
Then
2
π 2 kt/4L2
Tn (t) = e−k(2n−1) .
By superposition, the solution has the form
∞
2
π 2 t/4L2
u(x, t) = cn sin((2n − 1)πx/2L)e−k(2n−1) .
n=1
The solution is
∞
4B 1 2 2 2
u(x, t) = sin((2n − 1)πx/2L)e−k(2n−1) π t/4L .
π n=1 2n − 1
∂u ∂2u
= 9 2,
∂t ∂x
∂u
u(0, t) = (L, t) = 0,
∂x
u(x, 0) = x2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
506 CHAPTER 17. THE HEAT EQUATION
where
2
cn = ξ 2 sin((2n − 1)πξ/4) dξ
0
64 2 + (−1)n (2n − 1)π
=− 3
π (2n − 1)3
for n = 1, 2, · · · .
For given x in [0, 2], limt→∞ u(x, t) = 0.
11. Let u(x, t) = eαx+βt v(x, t) to transform the given problem. Substitute
this into the heat equation and divide out the common exponential factor
to obtain
∂v ∂v ∂2v ∂v
βv + = k α2 v + 2α + + Aαv + A + Bv .
∂t ∂x ∂x2 ∂x
The idea is to choose α and β to obtain a standard heat equation for v.
To do this, we must eliminate terms containing v or ∂v/∂x. Thus choose
2α + A = 0,
2
k(α + Aα + B) − β = 0.
Then
A A2
α=− and β = k B − .
2 4
With these choices, v satisfies
∂v ∂2v
=k 2
∂t ∂x
v(0, t) = v(L, t) = 0
v(x, 0) = e−αx u(x, 0).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 507
where
π
2
cn = ξ(π − ξ)e2ξ sin(nξ) dξ
π 0
4n
= ((−1)n e2π (12 + 8π − n2 (1 + 2π)) − (12 + 8π − n2 (1 − 2π))).
π(n2 + 1)
The solution for the original problem is
∞
2
u(x, t) = e−2x−2t cn sin(nx)e−n t .
n=1
∂v ∂2v
= ,
∂t ∂x2
v(0, t) = v(4, t) = 0,
v(x, 0) = e3x u(x, 0) = e3x .
where
4
1
cn = e3ξ sin(nπξ/4) dξ
2 0
2nπ
= (1 − e12 (−1)n ).
144 + n2 π 2
The solution for the original problem for u is
∞
2
π 2 t/16
u(x, t) = e−3x−9t cn sin(nπx/4)e−n .
n=1
Figure 17.8 shows the solution at times t = 0.2, 0.4, 0.7 and 1.1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
508 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 1 2 3 4
x
∂v ∂2v
= ,
∂t ∂x2
v(0, t) = v(π, t) = 0,
v(x, 0) = e−3x u(x, 0) = x(π − x)e−3x .
where
π
2
cn = e−3ξ ξ(π − ξ) sin(nξ) dξ
π 0
4n
= (1 − (−1)n e−3π )(3π(n2 + 9) + n2 − 27).
π(n2 + 9)3
Then
u(x, t) = e3x−9t v(x, t).
Figure 17.9 shows the solution at times t = 0.2, 0.4 and 0.6.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 509
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
x
∂v ∂2v
= 16 2 + 16L (x),
∂t ∂x
v(0, t) + L(0) = 2, v(1, t) + L(1) = 5,
v(x, 0) + L(x) = x2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
510 CHAPTER 17. THE HEAT EQUATION
where
1
cn = 2 (ξ 2 − 3ξ − 2) sin(nπξ) dξ
0
4
= (−1)n (1 + 2n2 π 2 ) − (1 + n2 π 2 ) .
n3 π 3
The original problem has the solution
∞
2
π2 t
u(x, t) = 3x + 2 + cn sin(nπx)e−16n .
n=1
∂v ∂2v
=k 2
∂t ∂x
v(0, t) = 0, v(L, t) = 0,
T 1
v(x, 0) = x(L − x) − T + x = (Lx − T )(L − x).
L L
This has the solution
∞
2
π 2 t/L2
v(x, t) = cn sin(nπx/L)e−kn ,
n=1
where
L
2 1
cn = (Lξ − T )(L − ξ) sin(nπξ/L) dξ
L 0 L
2
= 3 3 (2L2 (1 − (−1)n ) − n2 π 2 T ).
n π
With this solution for v(x, t), then
T
u(x, t) = v(x, t) + T − x.
L
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 511
16
12
0
0 2 4 6 8
x
17. Let u(x, t) = e−αt w(x, t) and substitute into the heat equation, choosing
α to eliminate the −Aw term. This requires that
∂w −αt ∂2w
−αwe−αt + e = 4 2 e−αt − Awe−αt .
∂t ∂x
Thus choose α = A. Then w satisfies
∂w ∂2w
=4 2,
∂t ∂x
w(0, t) = w(9, t) = 0,
w(x, 0) = 3x.
The diagrams show the solution at various times for A = 1/4, 1/2, 1 and
3. Figure 17.10 is for t = 0.2, Figure 17.11 for t = 0.7, and Figure 17.12
for t = 1.4.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
512 CHAPTER 17. THE HEAT EQUATION
10
0
0 2 4 6 8
x
0
0 2 4 6 8
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 513
18. Let u(x, t) = v(x, t) + h(x). Substitute this into the problem for u and
choose h to obtain homogeneous boundary conditions. This gives us
x
h(x) = T 1 − .
L
The problem for v is
∂v ∂2v
= 9 2,
∂t ∂x
v(0, t) = v(L, t) = 0,
x
v(x, 0) = −T 1 − .
L
By separation of variables we obtain the solution
∞
2
π 2 t/L2
v(x, t) = an sin(nπx/L)e−9n ,
n=1
where
L
2 ξ 2T
an = −T 1− sin(nπξ/L) dξ = − .
L 0 L nπ
Then
x 2T 1
∞
2 2 2
u(x, t) = T 1 − − sin(nπx/L)e−9n π t/L .
L π n=1 n
where L
2
bn = f (ξ) sin(nπξ/L) dξ
L 0
n2 π 2
Tn (t) + k Tn (t) = Bn (t); Tn (0) = bn ,
L2
with L
2
Bn (t) = F (ξ, t) sin(nπξ/L) dξ
L 0
for n = 1, 2, · · · .
Note that the second term in this solution for u(x, t) is the solution to the
problem without the forcing term.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
514 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
x
Figure 17.13: Problem 19, Section 17.2, at t = 0.2, with and without the source
term.
4
bn = (1 − (−1)n ),
πn3
and
1 2
Tn (t) = (1 − (−1)n )(−1 + 4n2 t + e−4n t ).
8πn5
The solution is
∞
1 2
u(x, t) = 5
(1 − (−1)n )(−1 + 4n2 t + e−4n t ) sin(nx)
n=1
8πn
∞
4 2
+ 3
(1 − (−1)n ) sin(nx)e−4n t .
n=1
πn
Figure 17.13 shows the solution with and without the source term, at time
t = 0.2. Figure 17.14 is at t = 0.5, and Figure 17.15 at t = 1.1.
20. Compute
4
1 8(−1)n+1
Bn (t) = ξ sin(t) sin(nπξ/4) dξ = sin(t),
2 0 nπ
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 515
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3
x
Figure 17.14: Problem 19, Section 17.2, at t = 0.5, with and without source
term.
0.25
0.2
0.15
0.1
0.05
0
0 0.5 1 1.5 2 2.5 3
x
Figure 17.15: Problem 19, Section 17.2, at t = 1.1, with and without source
term.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
516 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 1 2 3 4
x
4
1 2
bn = sin(nπξ/4) dξ = (1 − (−1)n ),
2 0 nπ
and
128(−1)n 2 2
Tn (t) = (16 cos(t) − n2 π 2 sin(t) − 16e−n π t/16 ).
nπ(n4 π 4 + 256)
The solution is
128(−1)n 2 2
u(x, t) = (16 cos(t) − n2 π 2 sin(t) − 16e−n π t/16 ) sin(nπx/4)
nπ(n4 π 4 + 256)
∞
2 2 2
+ (1 − (−1)n ) sin(nπx/4)e−n π t/16 .
n=1
nπ
Figures 17.16 through 17.20 show the solution with and without the source
term, at times t = 0.3, 0.7, 1.8, 3.9 and 4.6, respectively.
21. Compute
5
2
Bn (t) = t cos(ξ) sin(nπξ/5) dξ
5 0
2t
= ((−1)n+1 (5 + nπ) + nπ),
n2 π 2 − 25
5
2 500
bn = ξ 2 (5 − ξ) sin(nπξ/5) dξ = ((−1)n+1 − 1),
5 0 n3 π 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 517
1.2
0.8
0.6
0.4
0.2
0
0 1 2 3 4
x
2.5
1.5
0.5
0
0 1 2 3 4
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
518 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 1 2 3 4
x
x
0 1 2 3 4
0
-0.2
-0.4
-0.6
-0.8
-1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 519
0.2
x
0 1 2 3 4 5
0
-0.2
-0.4
-0.6
Figure 17.21: Problem 21, Section 17.2, with and without source term, at t =
1.5.
and
50(1 − cos(5)(−1)n ) 2 2 2 2
Tn (t) = (n π t − 25 + 25e−n π t/25 ).
n3 π 3 (n2 π 2 − 25)
The solution is
∞
50(1 − cos(5)(−1)n ) 2 2 2 2
u(x, t) = 3 3 2 2
(n π t − 25 + 25e−n π t/25 ) sin(nπx/5)
n=1
n π (n π − 25)
∞
500 2 2
+ 3 3
((−1)n+1 − 1) sin(nπx/5)e−n π t/25 .
n=1
n π
Figures 17.21, 17.22 and 17.23 show the solution, with and without source
term, at times t = 1.5, 2.5 and 2.9, respectively.
22. Compute
1
2K
Bn (t) = K sin(nπξ/2) dξ = (1 − cos(nπ/2)),
0 nπ
b1 = 1 and bn = 0 for n = 1,
2K 2 2
Tn (t) = 3 3
(1 − cos(nπ/2))(1 − e−n π t ),
n π
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
520 CHAPTER 17. THE HEAT EQUATION
x
0 1 2 3 4 5
0
-0.5
-1
-1.5
x
0 1 2 3 4 5
0
-0.5
-1
-1.5
-2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.2. THE HEAT EQUATION ON [0, L] 521
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2
x
Figure 17.24: Problem 22, Section 17.2, with and without source term, at t =
0.05.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
522 CHAPTER 17. THE HEAT EQUATION
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2
x
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0 0.5 1 1.5 2
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.3. SOLUTIONS IN AN INFINITE MEDIUM 523
0.25
0.2
0.15
0.1
0.05
0
0 0.5 1 1.5 2
x
Figures 17.28, 17.29 and 17.30 show the solution, with and without source
term, at times t = 0.1, 0.2 and 0.3, respectively. K = 4 is used in these
graphs.
where
∞ ∞
1 1
aω = f (ξ) cos(ωξ) dξ and bω = f (ξ) sin(ωξ) dξ.
π −∞ π −∞
To write the solution of this problem using the Fourier transform, first trans-
form the problem to obtain
dû
+ kω 2 û = 0; û(ω, 0) = fˆ(ω).
dt
The solution of this transformed problem is
2
û(ω, t) = fˆ(ω)e−kω t .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
524 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
x
Figure 17.28: Problem 23, Section 17.2, with and without source term, at t =
0.1.
0.25
0.2
0.15
0.1
0.05
0
0 0.5 1 1.5 2 2.5 3
x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.3. SOLUTIONS IN AN INFINITE MEDIUM 525
0.12
0.1
0.08
0.06
0.04
0.02
0
0 0.5 1 1.5 2 2.5 3
x
Recover the solution u(x, t) of the original problem by taking the Fourier trans-
form of this solution of the transformed problem. Use the result that
2
1 2
F −1 e−kω t = √ e−x /4kt ,
2 πkt
1. Compute
∞
1 8 1
aω = e−4|ξ| cos(ωξ) dξ = and bω = 0.
π −∞ π 16 + ω 2
The solution is
∞
8 1 2
u(x, t) = cos(ωx)e−ω kt dω.
π 0 16 + ω 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
526 CHAPTER 17. THE HEAT EQUATION
2. Compute π
1
aω = sin(ξ) cos(ωξ) dξ = 0,
π −π
and π
1 2 sin(ωπ)
bω = sin(ξ) sin(ωξ) dξ = .
π −π π(ω 2 − 1)
The solution is
∞
2 sin(ωπ) 2
u(x, t) = 2
sin(ωx)e−ω kt dω.
π 0 ω −1
to obtain
∞
1 2
u(x, t) = √ sin(ξ)(H(ξ + π) − H(ξ − π))e−(x−ξ) /4kt dξ
2 πkt −∞
π
1 2
= √ sin(ξ)e−(x−ξ) /4kt dξ.
2 πkt −π
3. Compute
4
1 1 4ω sin(4ω) + cos(4ω) − 1
aω = ξ cos(ωξ) dξ =
π 0 π ω2
and 4
1 1 sin(4ω) − 4ω cos(ω)
bω = ξ sin(ωξ) dξ = .
π 0 π ω2
The solution is
∞
2
u(x, t) = (aω cos(ωx) + bω sin(ωx))e−ω kt
dω.
0
to obtain
∞
1 2
u(x, t) = √ ξ(H(ξ) − H(ξ − 4))e−(x−ξ) /4kt dξ
2 πkt −∞
4
1 2
= √ ξe−(x−ξ) /4kt dξ.
2 πkt 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.3. SOLUTIONS IN AN INFINITE MEDIUM 527
4. Compute
1
1 2 cos(ω) sinh(1) + ω sin(ω) cosh(1)
aω = e−ξ cos(ωξ) dξ =
π −1 π ω2 + 1
and
1
1 2 ω cos(ω) sinh(1) − sin(ω) cosh(1)
bω = e−ξ sin(ωξ) dξ = .
π −1 π ω2 + 1
The solution is
∞
2 2
u(x, t) = (aω cos(ωx) + bω sin(ωx))e−ω kt
dω.
π 0
To use the Fourier transform, write f (x) = e−x (H(x + 1) − H(x − 1)) to
obtain
∞
1 2
u(x, t) = √ e−ξ (H(ξ + 1) − H(ξ − 1))e−(x−ξ) /4kt dξ
2 πkt −∞
1
1 2
= √ e−ξ e−(x−ξ) /4kt dξ.
2 πkt −1
In Problems 5 through 8 the problems are stated on the half line and are
solved by separation of variables.
5. Compute ∞
2 2 ω
bω = e−αξ sin(ωξ) dξ = .
π 0 π ω 2 + α2
The solution is
∞
2 ω 2
u(x, t) = sin(ωx)e−kω t dω.
π 0 ω 2 + α2
6. Compute ∞
2 2 2αω
bω = ξe−αξ sin(ωξ) dξ = .
π 0 π (α + ω 2 )2
2
The solution is
∞
2 2αω 2
u(x, t) = sin(ωx)e−kω t dω.
π 0 (α2 + ω 2 )2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
528 CHAPTER 17. THE HEAT EQUATION
8. Compute
2
2 2 sin(2ω) − 2ω cos(2ω)
bω = ξ sin(ωξ) dξ = .
π 0 π ω2
The solution is
∞
2 sin(2ω) − 2ω cos(2ω) 2
u(x, t) = sin(ωx)e−kω t dω.
π 0 ω2
9. Apply the Fourier sine transform with respect to x to the given problem
to obtain:
11. Let ∞
2
F (x) = e−ζ cos(xζ) dζ.
0
Think of this as a function of x. Differentiate under the integral sign to
obtain ∞
2
F (x) = −ζe−ζ sin(xζ) dζ.
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.4. LAPLACE TRANSFORM TECHNIQUES 529
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
530 CHAPTER 17. THE HEAT EQUATION
Now u(0, t) = 0, so
U (0, s) = c1 + c2 = 0
so c2 = −c1 . Then
√ √
s
U (x, s) = c1 e s/kx − e− s/kx = c sinh x .
k
Next, u(L, t) = T0 , so U (L, s) = T0 /s, so
s T0
c sinh L =
k s
so
T0 sinh( s/kx)
U (x, s) = .
s sinh( s/kL)
The solution u(x,
t) is the inverse transform of U (x, t). To compute this
inverse, let α = s/k and write
sinh( s/kx) eαx − e−αx
=
s sinh( s/kL) s(eαL − e−αL )
eα(x−L) − e−α(x+L)
= .
s(1 − e−2αL )
Now essentially duplicate the calculation done in the section (with cosh
in place of sinh) to obtain the solution
∞
(2n + 1)L − x (2n + 1)L + x
u(x, t) = T0 erfc √ − erfc √ .
n=0 2 kt 2 kt
2. Take the Laplace transform (in t) of the heat equation and use the initial
condition to obtain
s
U − U = 0,
k
with general solution
√ √
U (x, s) = c1 e s/kx + c2 e− s/kx .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.4. LAPLACE TRANSFORM TECHNIQUES 531
so
2 −√s/kx
U (x, s) = e .
s3
Write this as
2 1 −√s/kx
U (x, s) = 2 e .
s s
By writing U (x, s) in this way, we are able to take the inverse transform
of both factors. Now use the convolution theorem to write the solution
x
u(x, t) = 2t ∗ erfc √ .
2 kt
Then
s 1
U −
U = − e−x .
k k
This is a linear, second-order, nonhomogeneous differential equation. The
general solution of the associated homogeneous equation is
√ √
Uh (x, s) = c1 e s/kx + c2 e− s/kx .
s 1
A− A=− ,
k k
so
1
A=
s−k
and we obtain
√ √ 1 −x
s/kx
U (x, s) = Uh (x, s) + Up (x, s) = c1 e + c2 e− s/kx + e .
s−k
1
U (0, s) = c + .
s−k
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
532 CHAPTER 17. THE HEAT EQUATION
By carrying out manipulations like those done in the text, and using the
geometric series, we can write
1 √
∞
1
U (x, s) = 2 (−1)n e s/k(x−nL) + .
n=1
s s
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.5. HEAT CONDUCTION IN AN INFINITE CYLINDER 533
where 1
2
an = ξf (Rξ)J0 (jn ξ) dξ,
J1 (jn )2 0
Figure 17.31 shows the sum of these five terms for times t = 0.001, 0.025,
0.1, 0.3 and 0.5.
2. The coefficients are
1
2
an = ξe3ξ J0 (jn ξ) dξ.
J1 (jn )2 0
Figure 17.32 shows the sum of these five terms for times t = 0.0025, 0.001,
0.005, 0.01 and 0.2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
534 CHAPTER 17. THE HEAT EQUATION
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
r
12
10
0
0 0.5 1 1.5 2 2.5 3
r
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.6. HEAT CONDUCTION IN A RECTANGULAR PLATE 535
0
0 0.5 1 1.5 2 2.5 3
r
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
536 CHAPTER 17. THE HEAT EQUATION
and
m2 π 2
μm = , Ym (y) = sin(mπy/K).
K2
The solution has the form of a double superposition
∞
∞
u(x, y, t) = cnm sin(nπx/L) sin(mπy/K)e−kαnm t ,
n=1 m=1
in which
n2 π 2 m2 π 2
+ αnm =
.
L2 K2
The coefficients must be chosen so that
∞
∞
u(x, y, 0) = f (x, y) = cnm sin(nπx/L) sin(mπy/K).
n=1 m=1
Thus choose
L K
4
cnm = f (ξ, η) sin(nπξ/L) sin(mπη/K) dξ dη.
LK 0 0
2. With the given constants and initial position function, the coefficients are
3
2 2 2
cnm = ξ (2 − ξ) sin(nπξ/2) dξ sin(η)(3 − η) sin(mπη/3) dη
3 0 0
2 −32 n 54mπ m
= (1 + 2(−1) ) (1 − (−1) mπ cos(3)) .
3 n3 π 3 (m2 π 2 − 9)2
The solution is
∞
∞
2
u(x, y, t) = cnm sin(nπx/2) sin(mπy/3)e−4αnm π t ,
n=1 m=1
in which
n2 m2
αnm = + .
4 9
3. The coefficients in the series solution are
π π
4
cnm = 2 sin(ξ) sin(nξ) dξ cos(η/2) sin(mη) dη.
π 0 0
Now
π
0 for n = 1,
sin(η) sin(nη) dη =
0 π/2 for n = 1,
then the only nonzero coefficients are
2 4m
c1m = .
π 4m2 − 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
17.6. HEAT CONDUCTION IN A RECTANGULAR PLATE 537
The solution is
∞
8 m 2
u(x, y, t) = sin(x) 2−1
sin(my)e−(1+m )t .
π m=1
4m
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
538 CHAPTER 17. THE HEAT EQUATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 18
so αf + βg is harmonic on D.
2. (a)
(x3 − 3xy 2 )xx + (x3 − 3xy 2 )yy = 6x − 6x = 0
(b)
(3x2 y − y 3 )xx + (3x2 y − y 3 )yy = 6y − 6y = 0
(c)
(d)
(4x3 y − 4xy 3 )xx + (4x3 y − 4xy 3 )yy = 24xy − 24xy = 0
(e)
(f)
539
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
540 CHAPTER 18. THE POTENTIAL EQUATION
(g)
2x 2y 2 − 2x2
fx = , fxx = 2 ,
x2 +y 2 (x + y 2 )2
and
2y 2x2 − 2y 2
fy = , fyy =
x2 + y 2 (x2 + y 2 )2
so fxx + fyy = 0 on the plane with the origin removed.
X + λX = 0; X(0) = X(1) = 0
and
Y − λY = 0; Y (π) = 0.
The regular Sturm-Liouville problem for X has been solved in connection
with the heat and wave equations. The eigenvalues and eigenfunctions are
λn = n2 π 2 , Xn (x) = sin(nπx).
The problem for Y has solutions that are constant multiples of hyperbolic
sines of the form sinh(nπ(π − y)). For each n, we have functions
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.2. DIRICHLET PROBLEM FOR A RECTANGLE 541
Y + λY = 0; Y (0) = Y (2) = 0.
n2 π 2
λn = , Yn (y) = sin(nπy/2).
4
The problem for X is
n2 π 2
X − X = 0; X(3) = 0,
4
with solutions that are of the form Xn (x) = bn sinh(nπ(3−x)/2). Attempt
a solution of the form
∞
u(x, y) = bn sinh(nπ(3 − x)/2) sin(nπy/2).
n=1
The solution is
∞
16 1 − (−1)n sinh(nπ(3 − x)/2)
u(x, y) = sin(nπy/2).
π 3 n=1 n3 sinh(3nπ/2)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
542 CHAPTER 18. THE POTENTIAL EQUATION
where
and
We need
∞
v(0, y) = sin(y) = an sin(ny).
n=1
sinh(π − x)
v(x, y) = sin(y) .
sinh(π)
We need
∞
w(x, 0) = x(π − x) = bn sin(nx).
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.2. DIRICHLET PROBLEM FOR A RECTANGLE 543
Thus choose
π
2
bn = ξ(π − ξ) sin(nξ) dξ
π 0
4
= (1 − (−1)n ).
n3 π
Then
∞
4 (1 − (−1)n ) sinh(n(π − y))
w(x, y) = sin(x) .
π n=1 n3 sinh(nπ)
and
These are defined on 0 < x < 2, 0 < y < π. The solution for w has the
form
∞
sinh(nx)
w(x, y) = bn sin(ny) .
n=1
sinh(2n)
We need
∞
w(2, y) = bn sin(ny) = sin(y)
n=1
We need
∞
v(x, π) = x sin(πx) = an sin(nπx/2).
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
544 CHAPTER 18. THE POTENTIAL EQUATION
Then
sinh(πy)
v(x, y) = sin(πx)
sinh(π 2 )
∞
16 n sinh(nπy/2)
+ ((−1)n − 1) sin(nπx/2) .
π2 (n2 − 4)2 sinh(nπ 2 /2)
n=1,n=2
We need
∞
u(a, y) = g(y) = an sin((2n − 1)πy/2b).
n=1
Then
b
2
an = g(η) sin((2n − 1)πη/2b) dη.
b 0
We need
∞
u(x, b) = f (x) = an sin((2n − 1)πx/2a),
n1
so choose a
2
cn = f (ξ) sin((2n − 1)πξ/2a) dξ.
a 0
Then
∞
u(x, b) = x(x − a)2 = an sin(nπx/a).
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.2. DIRICHLET PROBLEM FOR A RECTANGLE 545
Then
a
2
an = ξ(ξ − a)2 sin(nπξ/a) dξ
a 0
4
= 3 3 (1 + 2nπ(−1)n )
n π
for n = 1, 2, · · · . The solution is
∞
4 1 + 2nπ(−1)n sinh(nπy/a)
u(x, y) = sin(nπx/a) .
π 3 n=1 n3 sinh(nπb/a)
9. Decompose the problem into two problems, in each of which the boundary
data is homogeneous on three sides. Let u(x, y) = v(x, y) + w(x, y), where
and
These problems are defined on 0 < x < 4, 0 < y < 1. The solution for v
has the form
∞
sinh(nπ(4 − x))
v(x, y) = sin(nπy) .
n=1
sinh(4nπ)
Then
∞
v(0, y) = sin(πy) = an sin(nπy),
n=1
so a1 = 1 and, for n = 2, 3, · · · , an = 0. Then
sinh(π(4 − x))
v(x, y) = sin(πy) .
sinh(4π)
The solution for w has the form
∞
w(x, y) = bn sin(nπy) sinh(nπx).
n=1
Then
∞
w(4, y) = y(1 − y) = bn sinh(4nπ) sin(nπy),
n=1
so
1
2
bn = ξ(1 − ξ) sin(nπξ) dξ
sinh(4nπ) 0
4(1 − (−1)n )
= 3 3 .
n π sinh(4nπ)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
546 CHAPTER 18. THE POTENTIAL EQUATION
where π
1
an = f (ξ) cos(nξ) dξ
πRn −π
for n = 0, 1, 2, · · · and
π
1
bn = f (ξ) sin(nξ) dξ
πRn −π
for n = 1, 2, · · · .
4. Compute
π
1
an = ξ cos(ξ) sin(nξ) dξ = 0 for n = 0, 1, 2, · · · ,
5n π −π
π
1 2(−1)n n
bn = n ξ cos(ξ) sin(nξ) dξ = 2 for n = 2, 3, · · · ,
5 π −π (n − 1)5n
∞
1 1
b1 = ξ cos(ξ) sin(ξ) dξ = − .
5π −∞ 10
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.3. DIRICHLET PROBLEM FOR A DISK 547
The solution is
(−1)n n r n
∞
−r
u(r, θ) = sin(θ) + 2 sin(nθ).
10 n=2
n2 − 1 5
5. Compute
π
1 2 sinh(π)
a0 = e−ξ dξ = ,
π −π π
π
1 −ξ 2 sinh(π) (−1)n
an = e cos(nξ) dξ = for n ≥ 1
4n π −π π 4n (n2 + 1)
π
1 2 sinh(π) n(−1)n
bn = n e−ξ sin(nξ) dξ = for n ≥ 1.
4 π −π π 4n (n2 + 1)
The solution is
2 (−1)n r n
∞
sinh(π)
u(r, θ) = + sinh(π)(cos(nθ) + n sin(nθ)).
π π n=1 n2 + 1 4
6. Write sin2 (θ) = (1 − cos(2θ))/2 and we can identify the only nonzero
coefficients as a0 = −1/2 and a2 = −1. The solution is
1 r
u(r, θ) = − cos(2θ).
2 2
7. Compute
π
1 2
a0 = (1 − ξ 2 ) dξ = 2 − π 2 ,
−π π 3
π
1 4(−1)n+1
an = n (1 − ξ 2 ) cos(nξ) dξ = , n ≥ 1,
8 π −π 8n n2
π
1
bn = n (1 − ξ 2 ) sin(nξ) dξ = 0, n ≥ 1.
8 π −π
The solution is
4(−1)n+1 r n
∞
1
u(r, θ) = 1 − π 2 + cos(nθ).
3 n=1
n2 8
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
548 CHAPTER 18. THE POTENTIAL EQUATION
9. Letting U (r, θ) = u(r cos(θ, r sin(θ)), this Dirichlet problem in polar coor-
dinates is
∇2 U (r, θ) = 0, U (4, θ) = 16 cos2 (θ),
for −π ≤ θ ≤ π and 0 ≤ r < 3. Write 16 cos2 (θ) = 8(1 + cos(2θ)) to
recognize that
1
a0 = 8, a2 (42 ) = 8,
2
and all other an = 0. The solution is
r 2
U (r, θ) = 8 + 8 cos(2θ).
4
Convert this solution back to rectangular coordinates using x = r cos(θ)
and y = r sin(θ) and the identity cos(2θ) = 2 cos2 (θ) − 1 to obtain
1
u(x, y) = 8 + (x2 − y 2 ).
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.4. POISSON’S INTEGRAL FORMULA 549
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
550 CHAPTER 18. THE POTENTIAL EQUATION
u(7, 0) ≈ −0.628310(10−11 ).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.5. DIRICHLET PROBLEM FOR UNBOUNDED REGIONS 551
2. From the integral formula for the upper half plane, the solution is
y ∞ e−|ξ|
u(x, y) = dξ
π −∞ y + (ξ − x)2
2
X − ω 2 Y = 0, Y + ω 2 Y = 0.
Use the condition u(x, 0) = X(x)Y (0) = 0 and the condition that X(x)
remains bounded as x → ∞ to obtain
Now ∞
u(x, 0) = g(y) = Bω sin(ωy) dω.
0
This is the Fourier sine expansion of g(y), hence choose
2 ∞
Bω = g(η) sin(ωη) dη.
π 0
Given g, this yields an integral formula for the solution u(x, y).
We can also approach this problem using the Fourier sine transform in
y. Let the Fourier sine transform of u(x, y) be ûS (x, ω). Transform the
differential equation to obtain and boundary condition to obtain
We also require that ûS (x, ω) remains bounded as x increases. This prob-
lem for the transformed function has solution
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
552 CHAPTER 18. THE POTENTIAL EQUATION
2 ∞
u(x, y) = g(ξ) sin(ξω) dξ sin(ωy)e−ωx dω.
π 0 0
5. To solve this problem, split it into two problems, in each of which there
is a single nonhomogeneous boundary condition on one edge. One of the
problems thus formed is exactly Problem 4. For the other, exchange x and
y and the names of f and g. Finally, add these two solutions to obtain
∞
2 ∞
u(x, y) = f (ξ) sin(ωξ) dξ sin(ωx)e−ωy dω
π 0 0
∞
2 ∞
+ g(ξ) sin(ωξ) dξ sin(ωy)e−ωx dω.
π 0 0
6. If u(x, y) is harmonic on the upper half plane and u(x, 0) = f (x) along the
real axis, then for y < 0 define v(x, y) = u(x, −y). It is routine to check
that
∇2 v = 0; v(x, 0) = u(x, 0) = f (x).
This defines a problem for v on the upper half plane. This problem for v
has solution
y ∞ f (ξ)
v(x, y) = − dξ.
π −∞ y 2 + (ξ − x)2
Since u(x, y) = v(x, −y), this solution for v on the upper half plane yields
a solution for u on the lower half plane.
1
ûS − ω 2 ûS = 0; ûS (0, ω) = .
1 + ω2
This problem is easily solved to obtain
e−ωx
ûS (x, ω) = .
1 + ω2
Invert this to obtain the solution
2 ∞ e−ωx
u(x, y) = sin(ωy) dω.
π 0 1 + ω2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.5. DIRICHLET PROBLEM FOR UNBOUNDED REGIONS 553
Here −∞ < x < ∞, and we require that û(ω, y) be bounded. Thus write
û(ω, y) = Cω e−|ω|y
and choose
1 a − iω
Cω = = 2 .
a + iω a + ω2
Now invert the transform to obtain
∞ ∞ −|ω|y
1 e
u(x, y) = eiωx dω.
2π −∞ −∞ a2 + ω 2
To simplify this expression, break the integral into integrals over (−∞, 0]
and [0, ∞). Make the change of variable ω = −η in the first integral,
rename ω = η in the second, and recombine the integrals to obtain the
solution
1 ∞ e−ηy
u(x, y) = (a cos(ηx) + η sin(ηx)) dη.
π 0 a2 + η 2
X − λX = 0, Y + λY = 0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
554 CHAPTER 18. THE POTENTIAL EQUATION
Then
Xn (x) = sinh((2n − 1)πx/4).
This suggests a solution of the form
∞
u(x, y) = cn sinh((2n − 1)πx/4) cos((2n − 1)πy/4).
n=1
11. The solution for the right half plane can be obtained from the integral
formula for the upper half plane by interchanging x and y. We obtain
x 1 1
u(x, y) = dη
π −1 x2 + (η − y)2
1 1−y 1+y
= arctan + arctan
π x x
for x > 0 and −∞ < y < ∞.
12. With the zero function values given on the left edge, we use a Fourier sine
transform in x. Let ûS (ω, y) be the transformed of u(x, y). Transform the
problem to obtain
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.6. A DIRICHLET PROBLEM FOR A CUBE 555
Use a superposition
∞
∞
u(x, y, z) = cnm sin(nπx) sin(mπy) sinh(π n2 + m2 z).
n=1 m=1
We need
∞
∞
u(x, y, 1) = xy = cnm sin(nπx) sin(mπy) sinh(π n2 + m2 ).
n=1 m=1
As we have done with wave and heat equations in two space variables,
choose
1 1
4
cnm = √ ξ sin(nπξ) dξ η sin(mπη) dη
sinh(π n2 + m2 ) 0 0
4(−1)n+m
= √ .
nmπ sinh( n2 + m2 π)
2
The solution is
u(x, y, z) =
∞ ∞
4 (−1)n+m
√ sin(nπx) sin(mπy) sinh( n2 + m2 πz).
π2 n=1 m=1
nm sinh( n2 + m2 π)
where
1
1 1 2π
cnm = √ 2 sin(nη/2) dη 2 sin(mπξ) dξ
sinh( n2 + 4m2 π 2 π) π 0 0
4
= √ (1 − (−1)n )(1 − (−1)m ).
π sinh( n2 + 4m2 π 2 π)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
556 CHAPTER 18. THE POTENTIAL EQUATION
in which
1 2π
1 1
anm = √ 2 sin(nπξ) dξ sin(mη/2) dη
2 2 2
sinh( 4n π + m π/2) 0 0 π
4 1 − (−1)n 1 − (−1)m
= √ .
sinh( 4n2 π 2 + m2 π/2) nπ mπ
Next, we obtain
∞
∞
v(x, y, z) = bnm sin(nπx) sin(mz) sinh( n2 π 2 + m2 y),
n=1 m=1
where
1
1 2 π
bnm = √ (2) sin(nπξ) dξ sin(mη) dη
sinh(2 n2 π 2 + m2 π) 0 π 0
8 1 − (−1)n 1 − (−1)m
= √ .
sinh(2 n2 π 2 + m2 π) nπ mπ
∇2 v = 0,
v(0, y, z) = 0, v(1, y, z) = sin(πy) sin(z),
v(x, 0, z) = v(x, 2, z) = 0,
v(x, y, 0) = v(x, y, π) = 0,
and
∇2 w = 0,
w(0, y, z) = w(1, y, z) = 0,
w(x, 0, z) = w(x, 2, z) = 0,
w(x, y, 0) = x2 (1 − x)y(2 − y), w(x, y, π) = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.7. STEADY-STATE HEAT EQUATION FOR A SPHERE 557
and
∞
∞
w(x, y, z) = bnm sin(nπx) sin(mπy/2) sinh( (nπ)2 + (mπ/2)2 (π−z)).
n=1 m=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
558 CHAPTER 18. THE POTENTIAL EQUATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.7. STEADY-STATE HEAT EQUATION FOR A SPHERE 559
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
560 CHAPTER 18. THE POTENTIAL EQUATION
Then dn = 0 if n ≥ 2, and
4
d1 = − .
π sinh(π)
A solution is given by
4
u(x, y) = c0 − cosh(π(1 − y)) cos(πx).
π sinh(π)
Since c0 is arbitrary, this solution is not unique.
2. First check that
π
y
y− dy = 0,
0 2
so it is worthwhile to look for a solution. From the zero boundary condi-
tions on the opposite edges y = 0 and y = π we expect to see a solution
of the form
∞
u(x, y) = c0 + [cn cosh(nx) + dn cosh(n(1 − x))] cos(ny).
n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.8. THE NEUMANN PROBLEM 561
and
∞
∂u
(1, y) = cos(y) = ncn sinh(n) cos(ny).
∂x n=1
and
−1 2 π π 2(1 − (−1)n )
dn = y− dy =
n sinh(n) π 0 2 πn3 sinh(n)
for n ≥ 1. We have the solution
cosh(x)
u(x, y) = c0 +
sinh(1)
∞
2((−1)n − 1)
+ cosh(n(1 − x)) cos(ny).
n=1
πn2 sinh(n)
π
3. A solution may exist because 0 cos(3x) dx = 0. From the zero boundary
conditions on edges x = 0 and x = π, separation of variables will yield a
solution of the form
∞
u(x, y) = c0 + [cn cosh(ny) + dn cosh(n(π − y))] cos(nx).
n=1
Now
∞
∂u
(x, 0) = cos(3x) = −ndn sinh(nπ) cos(nx),
∂y n=1
so
1
c3 = − and dn = 0 for n = 3.
3 sinh(3π)
The boundary condition at y = π gives us
∞
∂u
(x, π) = 6x − 3π = ncn sinh(nπ) cos(nx),
∂y n=1
so
1 2 π π
cn = (6x − 3π) cos(nx) dx
n sinh(nπ) π 0 0
1 12
= ((−1)n − 1)
n sinh(nπ) n2 π
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
562 CHAPTER 18. THE POTENTIAL EQUATION
4. Write u(x, y) = X(x)Y (y) and separate the variables in Laplace’s equa-
tion, then use the boundary conditions to obtain
X − λX = 0, X (0) = X (π) = 0,
Y + λY = 0.
The solutions for X are X0 (x) = 1 and Xn (x) = cos(nx) for n = 1, 2, · · · .
We find that
Yn (y) = cn cosh(ny) + dn cosh(n(π − y))
for n = 1, 2, · · · . Thus we seek a solution of the form
∞
u(x, y) = c0 + [cn cosh(ny) + dn cosh(n(π − y))] cos(nx).
n=1
Thus, for a solution, the coefficients must be chosen to satisfy the equations
c0 = 0,
cn cosh(nπ) + dn = 0,
2 π
cn + dn cosh(nπ) = f (x) cos(nx) dx
π 0
for n = 1, 2, · · · . Now, the determinant of the matrix of coefficients of this
system (for n ≥ 1) is
cosh(nπ) 1
= cosh2 (nπ) − 1 = sinh2 (nπ) = 0.
1 cosh(nπ)
Thus there is a unique solution of these algebraic equations for the cn s
and dn s, for each positive integer n, so the problem for u has a unique
solution.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.8. THE NEUMANN PROBLEM 563
and
∞
∂u
(0, y) = 3y 2 − 2y = −nπ sinh(nπ) sin(nπy).
∂x n=1
Then each cn = 0 and
1
−2
dn = (3y 2 − 2y) sin(nπy) dy
nπ sinh(nπ) 0
2
= 4 4 [n2 π 2 (−1)n + 6(1 − (−1)n )]
n π sinh(nπ)
for n = 1, 2, · · · . This yields the solution
∞
2
u(x, y) = [n2 π 2 (−1)n +6(1−(−1)n )] cosh(nπ(1−x)) sin(nπy).
n=1
n4 π 4 sinh(nπ)
π
6. Since −π sin(3θ) dθ = 0, a solution is possible. Any such solution will
have the form
∞
1
u(r, θ) = a0 + [an rn cos(nθ) + bn rn sin(nθ)].
2 n=1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
564 CHAPTER 18. THE POTENTIAL EQUATION
π
7. Check that −π cos(2θ) dθ = 0, so there may be a solution. Any such
solution must have the form
∞
1
r(r, θ) = a0 + [an rn cos(nθ) + bn rn sin(nθ)].
2 =1
10. A solution of the Dirichlet problem for the lower half-plane was obtained in
Problem 6 of Section 18.5. Using that result and the technique discussed
for solving a Neumann problem in the upper half-plane, we can write the
solution for the lower half-plane as
∞
1
u(x, y) = − ln(y 2 + (ξ − x)2 )f (ξ) dξ + c.
2π −∞
Notice that we can also observe that the sign of the outer normal derivative
changes along the real axis as we move from the upper half-plane to the
lower half-plane, accounting for the negative sign in the integral.
11. We will apply a Fourier cosine transform (with respect to x) to this prob-
lem. Let
UC (ω, y) = FC [u(x, y)](ω, y).
Using the operational formula for the cosine transform, we obtain
∂u
−ω 2 UC − (0, y) + UC = 0,
∂x
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
18.8. THE NEUMANN PROBLEM 565
UC (ω, y) = bω e−ωy .
US (ω, y) = bω e−ωy .
Now calculate
US (ω, 0) = −ωbω = fˆS (ω)
to obtain ∞
2
bω = − f (ξ) sin(ωξ) dξ.
πω 0
The solution is ∞
u(x, y) = bω e−ωy sin(ωx) dω.
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
566 CHAPTER 18. THE POTENTIAL EQUATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 19
2. √ √
i(6 − 2i) + |1 − i| = 6i + 2 + 1+1=2+ 2 + 6i
3.
2+i (2 + i)(4 + 7i) 1
= = (1 + 18i)
4 − 7i (4 − 7i)(4 + 7i) 65
4.
(2 + i) − (3 − 4i) (−1 + 5i)(16 − 2i) −3 + 41i
= =
(5 − i)(3 + i) (16 + 2i)(16 − 2i) 130
5.
(17 − 6i)−3 − 12i = (17 − 6i)(−3 + 12i) = 4 + 228i
6.
3i |3i| 3 3
= =√ = √
−4 + 8i | − 4 + 8i| 80 4 5
7.
i3 − 4i2 + 2 = −i + 4 + 2 = 6 − i
8.
567
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
568 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
9.
2 2
−6 + 2i (−6 + 2i)(1 + 8i)
=
1 − 8i (1 − 8i)(1 + 8i)
(−22 − 46i)2 −1632 + 2024i
= =
652 4225
10.
(−1 − 8i)(2i)(4 − i) = (−3 − 8i)(2 + 8i) = 58 − 40i
11.
π
|3i| = 3, arg(3i) = + 2nπ
2
12.
√ √ 3π
| − 2 + 2i| = 8 = 2 2 and arg(−2 + 2i) = + 2nπ
4
13.
√
| − 3 + 2i| = 13 and arg(−3 + 2i) = − arctan(2/3) + (2n + 1)π
14. √
|8 + i| = 65 and arg(8 + i) = arctan(1/8) + 2nπ
15.
| − 4| = 4 and arg(−4) = (2n + 1)π
16. √
|3 − 4i| =25 = 5 and arg(3 − 4i) = − arctan(4/3) + 2nπ
√
17. Since | − 2 + 2i| = 2 2 and 3π/4 is an argument of 2 + 2i, the polar form
is √
z = 2 2e3πi/4 .
Here there is no point in adding 2nπ to the argument to obtain all argu-
ments, since e2nπi = 1 for any integer n.
18. | − 7i| = 7 and an argument of −7i is 3π/2 (or, just as good, −π/2). The
polar form is
−7i = 7e3πi/2 .
We could also write
−7i = 7e−πi/2 .
√
19. |5 − 2i| = 29 and an argument of 5 − 2i is − arctan(2/5), so the polar
form of 5 − 2i is √
5 − 2i = 29e− arctan(2/5)i .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.1. GEOMETRY AND ARITHMETIC OF COMPLEX NUMBERS 569
√
20. | − 4 − i| = 17 and an argument of −4 − i is π + arctan(1/4), so the polar
form is √ (π+arctan(1/4))i
−4 − i = 17e .
√
21. |8 + i| = 65 and an argument of 8 + i is arctan(1/8), so the polar form is
√
8 + i = 65earctan(1/8)i .
√
22. | − 12 + 3i| = 153 and an argument of −12 + 3i is − arctan(3/12), so the
polar form is √
−12 + 3i = 153e− arctan(1/4)i .
25. Suppose first that z, w, u form the vertices of a triangle, labeled in the
clockwise order around the triangle. The sides of this triangle are vectors
represented by the complex numbers w − z, u − w and z − u. This triangle
is equilateral if and only if
|w − z| = |u − w| = |z − u|
and each of the vector sides can be rotated by θ = 2π/3 radians clockwise
to align with the next side. This occurs exactly when
w2 − 2wu + u2 = zw + uz − uv − z 2 .
z 2 + w2 + u2 = zw + zu + wu.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
570 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
Now, 2xy = 0 can occur in only two ways. Either x = 0, in which case z
is pure imaginary, or y = 0, in which case z is real.
27. Suppose first that |z| = 1. Then
|z| = |zz| = 1.
28. Compute
|z + w|2 + |z − w|2
= (z + w)(z + w) + (z − w)(z − w)
= zz + zw + wz + w(w) + z(z) + zw − wz − zw
= 2zz + 2ww
= 2 |z|2 + |w|2 .
29. M consists of all x + iy with y < 7. This is the half-plane lying below the
horizontal line y = 7. M is open and the boundary points are all complex
numbers x + 7i on the ”edge” of M . None of the boundary points of M
belong to M .
30. S consists of all points outside the circle of radius 2 about the origin.
This set is open, and the boundary points are all the points on the circle
|z| = 2. No boundary points of S are in S.
31. U consists of all points x + iy with 1 < x ≤ 3. This is the vertical strip
between the lines x = 1 and x = 3, including points on the line x = 3, but
not those on x = 1. The boundary points are the points on these vertical
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.2. COMPLEX FUNCTIONS 571
lines. these are all points 1 + iy and 3 + iy. Only the boundary points
3 + iy are in U . The boundary points 1 + iy are not in U . U is not open,
since it contains some boundary points. U is not closed, because U does
not contain all of its boundary points.
32. V consists of all points x + iy with 2 < x ≤ 3 and −1 < y < 1. These are
points inside the rectangle having vertices (2, 1), (3, 1), (2, −1) and (3, −1).
The boundary points are the points on the edges of this rectangle. Of these
points, only the points 3 + iy with −1 < y < 1 are in V . V is not open
because V contains some of its boundary points. V is not closed because
V does not contain all of its boundary points.
33. W consists of all points x+iy with x > y 2 . These are points ”enclosed by”
the parabola x = y 2 . Boundary points are the points on this parabola,
which are complex numbers y 2 +iy. Since W contains no boundary points,
W is open. Since W does not contain all of its boundary points, W is not
closed.
34. The boundary points of R are all points of the form 0+(1/m)i and (1/n)+
0i), that is, all pure imaginary points z = i/m and all real points z = 1/n.
Since none of these boundary points are in R, R is open. Since there are
boundary points of R not in R, R is not closed.
∂u ∂v
=1=
∂x ∂y
and
∂u ∂v
=− = 0.
∂y ∂x
Since u, v and their first partial derivatives are continuous for all x + iy,
f is differentiable for all z.
2.
Then
u(x, y) = x2 − y 2 + y, v(x, y) = 2xy − x.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
572 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
Then
∂u ∂v
= 2x =
∂x ∂y
and
∂u ∂v
= −2y + 1 = − .
∂y ∂x
Since u, v and their first partial derivatives are continuous for all z, f is
differentiable for all z.
3. f (z) = |x + iy| = x2 + y 2 , so
u(x, y) = x2 + y 2 , v(x, y) = 0.
If x and y are not both zero, then the partial derivatives are
∂u x ∂u y
= , = ,
∂x x2 + y 2 ∂y x2 + y 2
∂v ∂v
= = 0.
∂x ∂y
The Cauchy-Riemann equations are not satisfied at any point with both
x = 0 and y = 0. The only point left to check is z = 0, where x = y = 0.
Now the above expressions for the partial derivatives of v are still valid,
but those for the partial derivatives of u are not, and we must fall back
on the definition of the partial derivatives:
∂u u(h, 0) − u(0, 0)
(0, 0) = lim
∂x h→0 h
√
h 2
= lim
h→0 h
|h|
= lim .
h→0 h
|h| 1 if h > 0,
=
h −1 if h < 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.2. COMPLEX FUNCTIONS 573
u(x, y) = x + y, v(x, y) = y.
Then
∂u ∂u ∂v ∂v
= = = 1, = 0.
∂x ∂y ∂y ∂x
The Cauchy-Riemann equations do not hold at any z, so f is not differ-
entiable anywhere.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
574 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
7.
x + iy y
f (z) = = 1 + i,
x x
so
y
u(x, y) = 1, v(x, y) =
x
for x = 0. Then
∂u ∂u ∂v ∂v
= = 0, = −y/x2 , = 1/x.
∂x ∂y ∂x ∂y
The Cauchy-Riemann equations do not hold at any point at which the
function is defined, so f is not differentiable anywhere.
8. Write f (z) = u(x, y) + iv(x, y). Then
f (z) = (x + iy)3 − 8(x + iy) + 2
= z 3 − 8x + 2 = u(x, y) + iv(x, y),
so
u(x, y) = x3 − 3xy 2 − 8x + 2, v(x, y) = 3x2 y − y 3 − 8y.
Then, at all z = x + iy,
∂u ∂v
= 3x2 − 3y 2 − 8 =
∂x ∂y
and
∂u ∂v
= −6xy = − .
∂y ∂x
The Cauchy-Riemann equations hold for all z. Since u, v and its first
partial derivatives are continuous everywhere, f is differentiable for all z.
9.
f (z) = (z)2 = (x − iy)2 = x2 − y 2 − 2xyi.
Then
u(x, y) = x2 − y 2 , v(x, y) = −2xy.
Then
∂u ∂u
= 2x, = −2y,
∂x ∂y
∂v ∂v
= −2y, = −2x.
∂x ∂y
The Cauchy-Riemann equations hold only at z = 0. But
f (h) − f (0) (h)2
lim = lim
h→0 h h→0 h
h
= lim h
h→0 h
=0
because h/h has magnitude 1, and h → 0 as h → 0. Therefore f (0) = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.2. COMPLEX FUNCTIONS 575
10.
f (z) = iz + |z| = ix − y + x2 + y 2 .
Then
u(x, y) = x2 + y 2 − y, v(x, y) = x.
Now
∂u x
= ,
∂x x2 + y 2
∂u y
= −1 + ,
∂y x + y2
2
∂v ∂v
= 1, = 0.
∂x ∂y
The Cauchy-Riemann equations are not satisfied at any z, so f is not
differentiable anywhere.
11.
1 1 x − iy
f (z) = −4z + = −4x − 4iy + = −4x − 4yi + 2
z x + iy x + y2
for z = 0. Then
x y
u(x, y) = −4x + , v(x, y) = −4y − 2 .
x2 +y 2 x + y2
Compute
∂u y 2 − x2 ∂u −2xy
= −4 + 2 , = ,
∂x (x + y 2 )2 ∂y (x2 + y 2 )2
∂v 2xy ∂v y 2 − x2
= 2 2 2
, = −4 + 2 .
∂x (x + y ) ∂y (x + y 2 )2
The Cauchy-Riemann equations hold for all z = 0, so f is differentiable
at all z at which it is defined (z = 0).
12.
z−i x + i(y − 1)
f (z) = =
z+i x + i(y + 1)
x2 + y 2 − 1 − 2xi
= .
x2 + (y + 1)2
Then
x2 + y 2 − 1 −2x
u(x, y) = and v(x, y) 2 .
x2 + (y − 1)2 x + (y + 1)2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
576 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
Compute
∂u 4x(y + 1)
= ,
∂x (x2 + (y + 1)2 )2
∂u 2(y + 1)2 − 2x2
= 2 ,
∂y (x + (y + 1)2 )2
∂v 2x2 − 2(y + 1)2
= 2 ,
∂x (x + (y + 1)2 )2
∂v 4x(y + 1)
= 2 .
∂y (x + (y + 1)2 )2
13. Let zn = xn + iyn and z0 = x0 + iy0 . Write f (z) = u(x, y) + iv(x, y). Since
u and v are continuous at (x0 , y0 ), then
2. There are several ways we can proceed. Perhaps the easiest is to use the
fact that
Then
Here we have also used the fact that cosh(−4) = cosh(4) and sinh(−4) =
− sinh(4).
Another approach is to begin with the definition of sin(z) and use Euler’s
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.3. THE EXPONENTIAL AND TRIGONOMETRIC FUNCTIONS 577
formula:
1 i(1−4i)
sin(1 − 4i) = e − e−i(1−4i)
2i
1 4+i
= e − e−4−i
2i
1 4 i
= e e − e−4 e−i
2i
1 4
= e (cos(1) + i sin(1)) − e−4 (cos(1) − i sin(1))
2i
1 1
= (e4 − e−4 ) cos(1) + (e4 + e−4 ) sin(1)
2i 2
1
= cosh(4) sin(1) + sinh(4) cos(1)
i
= sin(1) cosh(4) − i cos(1) sinh(4).
to write
cos(3 + 2i) = cos(3) cosh(2) − i sin(3) sinh(2).
Then
sin(3i) −i sinh(−3)
tan(3i) = =
cos(3i) cosh(−3)
i sinh(3)
= = i tanh(3).
cosh(3)
5.
e5+2i = e5 (cos(2) + i sin(2))
6.
cos(1 − πi/4)
cot(1 − πi/4) =
sin(1 − πi/4)
cos(1) cosh(π/4) + i sin(1) sinh(π/4)
= .
sin(1) cosh(π/4) − i cos(1) sinh(π/4)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
578 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
cot(1 − πi/4)
(cos(1) cosh(π/4) + i sin(1) sinh(π/4))(sin(1) cosh(π/4) + i cos(1) sinh(π/4))
=
sin2 (1) cosh2 (π/4) + cos2 (1) sinh2 (π/4)
sin(1) cos(1)(cosh2 (π/4) − sinh2 (π/4)) + i sinh(π/4) cosh(π/4)(sin2 (1) + cos2 (1))
=
sin2 (1) cosh2 (π/4) + cos2 (1) sinh2 (π/4)
sin(1) cos(1) + i sinh(π/4) cosh(π/4)
=
sin2 (1) cosh2 (π/4) + (1 − sin2 (1)) sinh2 (π/4)
sin(1) cos(1) + i sinh(π/4) cosh(π/4)
=
sin2 (1) cosh2 (π/4) + (1 − sin2 (1)) sinh2 (π/4)
sin(1) cos(1) + i sinh(π/4) cosh(π/4)
= .
sin2 (1) sinh2 (π/4)
7.
1
sin2 (1 + i) = [1 − cos(2(1 + i))]
2
1
= [1 − cos(2) cosh(2) + i sin(2) sinh(2)]
2
1 i
= [1 − cos(2) cosh(2)] + [sin(2) sinh(2)].
2 2
8.
cos(2 − i) − sin(2 − i)
= cos(2) cosh(1) + i sin(2) sinh(1) − sin(2) cosh(1) − i cos(2) sinh(1)
= cosh(1)[cos(2) − sin(2)] − i sinh(1)[cos(2) − sin(2)]
9.
eπi/2 = cos(π/2) + i sin(π/2) = i
10.
11. First,
2 2 2
−y 2 +2ixy
ez = e(x+iy) = ex
2
−y 2
= ex [cos(2xy) + i sin(2xy)].
Then
2
−y 2 2
−y 2
u(x, y) = ex cos(2xy) and v(x, y) = ex sin(2xy).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.3. THE EXPONENTIAL AND TRIGONOMETRIC FUNCTIONS 579
Compute
∂u 2
−y 2
= ex [2x cos(2xy) − 2y sin(2xy)],
∂x
∂u 2
−y 2
= ex [−2y cos(2xy) − 2x sin(2xy)],
∂y
∂v 2
−y 2
= ex [2x sin(2xy) + 2y cos(2xy)],
∂x
∂v 2
−y 2
= ex [−2y sin(2xy) + 2x cos(2xy)].
∂y
The Cauchy-Riemann equations are satisfied for all x + iy.
12. Write
1 1 x − iy
= = 2 ,
z x + iy x + y2
so
2 2 y y
e1/z = ex/(x +y ) cos − i sin .
x2 + y 2 x2 + y 2
Then
2
+y 2 ) y 2
+y 2 ) y
u(x, y) = ex/(x cos , v(x, y) = −ex/(x sin .
x2 + y 2 x2 + y 2
Then
2 2
∂u ex/(x +y ) 2 2 y y
= 2 (y − x ) cos + 2xy sin ,
∂x (x + y 2 )2 x2 + y 2 x2 + y 2
2 2
∂u ex/(x +y ) y 2 2 y
= 2 −2xy cos − (x − y ) sin ,
∂y (x + y 2 )2 x2 + y 2 x2 + y 2
2 2
∂v ex/(x +y ) 2 2 y y
= 2 (x − y ) sin + 2xy cos ,
∂x (x + y 2 )2 x2 + y 2 x2 + y 2
2 2
∂v ex/(x +y ) y 2 2 y
= 2 2xy sin − (x − y ) cos .
∂y (x + y 2 )2 x2 + y 2 x2 + y 2
The Cauchy-Riemann equations hold for all z = 0.
13.
so
u(x, y) = xex cos(y) − yex sin(y), v(x, y) = yex cos(y) + xex sin(y).
Then
∂u ∂v
= ex (cos(y) + x cos(y) − y sin(y)) =
∂x ∂y
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
580 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
and
∂u ∂v
= ex (−x sin(y) − sin(y) − y cos(y)) = − .
∂y ∂x
The Cauchy-Riemann equations hold for all z.
14. Write
1
f (z) = cos2 (z) = (1 − cos(2z))
2
1 1
= − (cos(2x) cosh(2y) − i sin(2x) sinh(2y)).
2 2
Then
1 1 1
u(x, y) = − cos(2x) cosh(2y), v(x, y) = sin(2x) sinh(2y).
2 2 2
Then
∂u
= sin(2x) cosh(2y),
∂x
∂u
= − cos(2x) sinh(2y),
∂y
∂v
= cos(2x) sinh(2y),
∂x
∂v
= sin(2x) cosh(2y).
∂y
The Cauchy-Riemann equations are satisfied at every z.
15. If ez = ex+iy = 2i, then
ex cos(y) = 0, ex sin(y) = 2.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.3. THE EXPONENTIAL AND TRIGONOMETRIC FUNCTIONS 581
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
582 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.4. THE COMPLEX LOGARITHM 583
1. In polar form,
z = −4i = 4e3nπi/2 .
Then
3π
log(−4i) = ln(4) + + 2nπ i.
2
2. Write √
2 − 2i = 2 2e7πi/4
so
√ 7π
log(2 − 2i) = ln(2 2) + + 2nπ i.
4
4. Write √
1 + 5i = 26ei arctan(5)
to obtain
1
log(1 + 5i) = ln(26) + (arctan(5) + 2nπ)i.
2
5. Write √
−9 + 2i = 85e(arctan(−2/9)+π)i
to obtain
1
log(−9 + 2i) = ln(85) + (− arctan(2/9) + (2n + 1)π)i.
2
Notice that there are infinitely many complex logarithms of 5, even though
5 is a positive real number with a natural logarithm.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
584 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
and
zw = |zw|e(θ1 +θ2 +2kπ)i .
Thus
log(zw) = ln(|zw|) + (θ1 + θ2 + 2kπ)i,
while
19.5 Powers
In these problems, n always denotes an arbitrary integer.
1.
2.
√
(1 + i)2i = e2i log(1+i) = e2i[ln( 2)+i(π/4+2nπ)]
3.
ii = ei log(i) = ei(i(π/2+2nπ)) = e−(π/2+2nπ)
This is consistent with Problem 1, since i1+i = i(ii ).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.5. POWERS 585
4.
5.
6.
√ 1/3
(1 − i)1/3 = 2e−i(π/4+2nπ)
= 21/6 e−i(π/12+2nπ/3)
7.
1/4
i1/4 = ei(π/2+2nπ)
= ei(π/8+nπ/2)
8.
1/4
161/4 = 16e2nπi = 2enπ/2 ,
9.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
586 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
10.
11. 1/4
(−16)1/4 = 16ei(π+2nπ) = 2ei(π/4+nπ/2) ,
for n = 0, 1, 2, 3. These values are
√ √ √ √
2(1 + i), 2(−1 + i), 2(−1 − i), 2(1 − i).
14.
15. Let ω be any nth root of unity different from 1. The numbers ω j , for
j = 0, 1, · · · , n − 1, are distinct, hence are all the nth roots of unity. Thus
it is enough to show that
n−1
ω j = 0.
j=0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
19.5. POWERS 587
But,
n−1
1 − ωn
ωj = = 0.
j=0
1−ω
Apply this with a = e2πi/n and use the fact that an = e2πi = 1 to write
n−1
1 − (−1)n 0 if n is even,
j 2πij/n
(−1) e = 2πi/n
= 2πi/n
j=0
1 + e 2/(1 + e ) if n is odd.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
588 CHAPTER 19. COMPLEX NUMBERS AND FUNCTIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 20
Complex Integration
589
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
590 CHAPTER 20. COMPLEX INTEGRATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
20.1. THE INTEGRAL OF A COMPLEX FUNCTION 591
Therefore
1
|z|2 dz = [16(t − 1)2 + t2 ](4 + i) dt
γ 0
4+i 1
= 16(t − 1)3 + t3 0
3
17
= (4 + i).
3
12.
−4−i
eiz dz = −ieiz −2
γ
= −i(e1−4i − e−2i )
= −e sin(4) + sin(2) + [cos(2) − e cos(4)]i
14. Since Im(z) has no antiderivative, parametrize the curve by γ(t) = 4eit ,
0 ≤ t ≤ 2π. Then
2π
Im(z) dz = 4 sin(t)4ieit dt
γ 0
2π
= 16(− sin2 (t) + i cos(t) sin(t)) dt = −16π
0
15. Since f has no antiderivative, write the curve as γ(t) = (1 + i)t − i for
0 ≤ t ≤ 1. Then
1
2
|z|2 dz = [t2 + (t − 1)2 ](1 + i) dt = (1 + i)
γ 0 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
592 CHAPTER 20. COMPLEX INTEGRATION
| cos(z 2 )| ≤ M for z on γ.
There
is no claim that this huge upper bound is close to the actual value
of | γ cos(z 2 ) dz|. We made very crude but quick estimates to derive a
number M , but much smaller numbers might also work. Often the point
to obtaining a bound is to take a limit in which the integral appears, and
then it is often enough to know that the integral is bounded.
√
17. The length of γ is 5. Now we need a number M so that
1
≤ M for z on γ.
1+z
Now, the point on γ closest to z = −1 is 2 + i, so for z on γ,
√
|z + 1| = |z − (−1)| ≥ |2 + i + 1| = 10.
Then 1
1 1
= ≤√
z+1 |z + 1| 10
√
and we can choose M = 1/ 10. Then
√
1 5 1
, dz ≤ √ = √ .
γ 1+z 10 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
20.2. CAUCHY’S THEOREM 593
by Cauchy’s theorem.
2. The circle encloses i, at which f (z) is not defined. Thus Cauchy’s theorem
does not apply. Evaluate the integral by parametrizing the curve z =
i + 3eit for 0 ≤ t ≤ 2π. Then
2π
2z 2i + 6eit
dz = 3ieit dt
γ z−i 0 3eit
2π
= (−2 + 6ieit ) dt = −4π.
0
3. γ encloses 2i, at which the function is not defined, hence not differentiable.
Parametrize γ by
Then
2π
1 1
dt = 2ieit dt
γ (z − 2i)3 0 (2eit )2
2π
i
= e−2it dt = 0.
4 0
This integral turns out to be 0, but we could not have concluded this from
Cauchy’s theorem, which does not apply to this integral.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
594 CHAPTER 20. COMPLEX INTEGRATION
by Cauchy’s theorem.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
20.3. CONSEQUENCES OF CAUCHY’S THEOREM 595
We need only evaluate C Im(z) dz. Cauchy’s theorem does not apply to
this integral because Im(z) is not a differentiable function. Parametrize
each side of the square. Let S1 bes the left side (0 to −2i), S2 the lower
side (−2i to 2 − 2i), S3 the right side (2 − 2i to 2), and S4 the top side (2
to 0), oriented counterclockwise. We can parametrize
S1 : z = −2it,
S2 : z = 2t − 2i,
S3 : z = 2 − 2i(1 − t),
S4 : z = 2(1 − t),
Therefore
f (z) dz = −4.
C
1. Since 2i is the center of the circle γ, we can apply the Cauchy integral
formula with f (z) = z 4 to write
z4
dz = 2πif (2i) = 2πi(2i)4 = 32πi.
γ z − 2i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
596 CHAPTER 20. COMPLEX INTEGRATION
4. Apply the Cauchy integral formula for derivatives, with n = 1 and f (z) =
2z 3 , to write
2z 3
2
dz = 2πif (2) = 48πi.
γ (z − 2)
5. We can use the Cauchy integral formula for the derivative of a function
(n = 1). With f (z) = iez , we have
iez
2
dz = 2πif (2 − i)
γ (z − 2 + i)
8. γ is not a closed curve and the Cauchy integral formulas do not apply.
Parametrize γ by γ(t) = 1 − t − it for 0 ≤ t ≤ 1. On the curve,
f (z) = 2iz|z| = 2i[(1 − t) + it] 1 − 2t + 2t2
so
1
2iz|z| dz = 2i[1 − t + it] 1 − 2t + 2t2 (−1 − i) dt
γ 0
1
1
=2 1 − 2t + 2t2 dt + 2i (2t − 1) 1 − 2t + 2t2 dt
0 0
√
√ 2+1
= 1 + 24 ln √ .
2−1
These integrations can be done using MAPLE, or the indefinite integrals
−1 + 4t
1 − 2t + 2t2 dt = 1 − 2t + 2t2
8
√
7 2 4 1
+ arcsinh √ t−
32 7 4
and
1
(2t − 1) 1 − 2t + 2t2 dt = (1 − 2t + 2t2 )3/2 .
3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
20.3. CONSEQUENCES OF CAUCHY’S THEOREM 597
9.
−(2 + i) sin(z 4 ) d 4
dz = −2πi(2 + i) (sin(z )
γ (z + 4)2 dz z=−4
3 4
= 2πi(1 − 2i) 4z cos(z ) z=−4
= −512π(1 − 2i) cos(256)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
598 CHAPTER 20. COMPLEX INTEGRATION
Equate the real part of the left side of this equation to the real part of the
right side to conclude that
2π
ecos(t) cos(sin(t)) dt = 2π.
0
However, we did not need this calculation to evaluate this integral, because
this integral, from 0 to π, is the negative of the integral from π to 2π, hence
the integral is zero.
14. First observe that
z − 4i z − 4i
f (z) = 3
= .
z + 4z z(z − 2i)(z + 2i)
Now let γ1 , γ2 and γ3 be nonintersecting circles enclosed by γ, which also
do not intersect γ, and having centers, respectively, 0, 2i and −2i. By the
extended deformation theorem,
3
f (z) dz = f (z) dz.
γ j=1 γj
Consider each term in the sum on the right. On and in the interior of γ1 ,
write
(z − 4i)/(z − 2i)(z + 2i)
f (z) =
z
is differentiable, and we can apply the Cauchy integral formula to write
z − 4i
f (z) dz = 2πi
γ (z − 2i)(z + 2i) z=0
1
−4i
= 2πi = 2πi(−i) = 2π.
(−2i)(2i)
Similarly, for the integral over γ2 , write
(z − 4i)/z(z + 2i)
f (z) =
z − 2i
and apply the Cauchy integral formula to write
−2i π
f (z) dz = 2πi =− .
γ2 (2i)(4i) 2
And, for the integral over γ3 , write
(z − 4i)/z(z − 2i)
f (z) =
z + 2i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
20.3. CONSEQUENCES OF CAUCHY’S THEOREM 599
to write
−6i 3π
f (z) dz = 2πi = .
γ3 (−2i)(−4i) 2
Then
π 3π
f (z) dz = 2π − − = 0.
γ 2 2
15. We will use the notation of the theorem, Figure 20.14 of the text, and the
hint outlined in the problem. In the text it was shown that it is sufficient
to show that
f (z)
dz
C ∗ z − z0
can be made arbitrarily small by choosing b larger. Recall that, for some
positive integer n and some positive number M |z n f (z)| ≤ M for z suffi-
ciently large. By choosing z far enough away from z0 , we can make
z n f (z)
≤ |z n f (z)| ≤ M.
z − z0
Then, for z on C ∗ ,
f (z) M M
≤ n | = n+1 .
z − z0 |z (z − z0 ) |z ||1 − z0 /z|
But
z0 z 1
0
1 − ≥ 1 − >
z z 2
if |z0 /z| < 1/2, that is, if |z0 | < 2|z|, so
1
<2
|1 − z0 /z|
and then f (z)
2M 2M
≤ n+1 < n+1 .
z − z0 z b
Now, the length of C ∗ is
Then,
f (z)
2M
dz ≤ n+1 (4b − 2σ) ,
C∗ z − z0 b
or, equivalently,
2M
f (z) 2σ
dz ≤ n 4− ,
C∗ z − z0 b b
and this approaches 0 as b → ∞.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
600 CHAPTER 20. COMPLEX INTEGRATION
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 21
Series Representations of
Functions
601
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
602 CHAPTER 21. SERIES REPRESENTATIONS OF FUNCTIONS
3.
(n + 1)n+1 /(n + 2)n+1
(z − 1 + 3i)
nn /(n + 1)n
(n + 1)2n+1
|z − 1 + 3i|
nn (n + 2)n+1
n n+1
n+1 n+1
= |z − 1 + 2i|
n n+2
n n+1
1 1 + 1/n
= 1+ |z − 1 + 3i|
n 1 + 2/n
→ e|z − 1 + 3i| < 1
if
1
.
|z − 1 + 3i| <
e
The radius of convergence is 1/e and the open disk of convergence is
|z − 1 + 3i| < 1/e.
4.
(2i/(5 + i))n+1
(z + 3 − 4i)
(2i/(5 + i))n
2i
= (z + 3 − 4i)
5+i
2
= √ |z + 3 − 4i| < 1
26
if √
26
|z + 3 − 4i| < .
2
√
The radius of √
convergence is 26/2 and the open
√ disk of convergence is
|z + 3 − 4i| < 26/2, the open disk of radius 26/2 about −3 + 4i.
6.
(1 − i)n+1 /(n + 3) n + 2 √
(z − 3) = 2|z − 3|
(1 − i)n /(n + 2) n+3
√
→ 2|z − 3| < 1
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
21.1. POWER SERIES 603
if
1
|z − 3| < √ .
2
√
The radius √of convergence is 1/ 2 and the open disk of convergence is
|z − 3| < 1/ 2.
7. No. The the power series has center 2i. If the series converged at 0, it
would converge also at the point i that is closer to 2i than 0 is.
8. The center of this power series is 4 − 2i. Now 1 + i is closer to 4 − 2i than
i is, so if the series converged at i, it would also have to converge at 1 + i.
9. Since we know the series for cos(z) about 0, replace z with 2z to obtain
∞
(−1)n
cos(2z) = (2z)2n ,
n=0
(2n)!
or
∞
(−1)n 22n 2n
cos(2z) = (z) .
n=0
(2n)!
This series has infinite radius of convergence, since the series for cos(z) as
infinite radius of convergence. This means that both series are valid for
all complex z.
10. Using the series for ez , we obtain the series for e−z , and then
z 2 − 3z + i = −3 + (1 − 2i)(z − 2 + i) + (z − 2 + i)2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
604 CHAPTER 21. SERIES REPRESENTATIONS OF FUNCTIONS
Since the sine series contains only odd powers of z while the exponential
series contains all powers, it is awkward to attempt to combine these two
series under one summation. Both series converge for all z.
13. Like Problem 11, this is an algebraic rearrangement of a second degree
polynomial in powers of z − 1 − i. If we use the Taylor coefficients, with
f (z) = (z − 9)2 , then
1
c0 = f (1 + i), c1 = f (1 + i) and c2 = f (1 + i).
2
We get
f (0) = 2f (0) + 1 = 3,
f (3) (0) = 2f (0) = 2i,
f (4) (0) = 2f (0) = 6,
f (5) (0) = 2f (3) (0) = 4i.
For the first six terms of the Maclaurin expansion, these numbers enable
us to compute the Taylor coefficients f k (0)/k!. We obtain
3 2i 6 4i
f (z) = 1 + iz + z 2 + z 3 + z 4 + z 5 + · · · .
2 3! 4! 5!
In this problem we can write the entire Maclaurin expansion, since it is
not difficult to show by induction that
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
21.1. POWER SERIES 605
(a) From these derivatives, compute the first seven terms of the Maclaurin
expansion, obtaining
1 2
sin2 (z) = z 2 − z 4 + z 6 + · · · .
3 45
(b) Multiply
2 1 3 1 5 1 3 1 5
sin (z) = z − z + z + ··· z− z + z + ···
6 120 6 120
1 1 1 1 1
= z2 − + z4 + + + z6 + · · ·
6 6 120 36 120
1 2
= z2 − z4 + z6 + · · · .
3 45
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
606 CHAPTER 21. SERIES REPRESENTATIONS OF FUNCTIONS
17. Begin with z a given complex number, and consider the integral
1 zn
ezw dw
2πi γ n!wn+1
with γ the unit circle about the origin (oriented counterclockwise). First
expand ezw in its Maclaurin series and then parametrize γ by γ(t) = eiθ
to write
∞
zn zw 1 z n (zw)k
n+1
e dw = dw
γ n!w 2πi γ n!wn+1 k!
k=0
∞
1 z n+k wk−n−1
= dw
2πi γ n!k!
k=0
2π ∞
1 z n+k ei(k−n−1)θ iθ
= ie dθ
2πi 0 n!k!
k=0
∞ 2π n+k
1 z
= ei(k−n)θ dθ.
2π 0 n!k!
k=0
Now,
2π
i(k−n)θ 0 if k = n,
e dθ =
0 2π if k − n.
Therefore we have
1 zn zw (z n )2
e dw = .
2πi γ n!wn+1 (n!)2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
21.1. POWER SERIES 607
18. f (z) has a zero of order 3 at 0, because z 3 has a zero of order 3 at 0, and
cos(0) = 0.
19. f (z) has a zero of order 4 at 0, because z 2 has a zero of order 2, and sin(z)
has a zero of order 1 at 0, so sin2 (z) has a zero of order 2 there.
20. f (z) = (z − π/2)2 cos(z) has a zero of order 3 at π/2 because cos(z) has a
simple zero there, and (z − π/2)2 has a zero of order 2 at π/2.
21. f (z) has a zero of order 3 at 3π/2 because cos(z) has a simple zero there.
22. f (z) has a zero of order 4 at 0 because cos(z) has a simple zero at −π/2,
so cos4 (z) has a fourth order zero there.
23. f (z) is not defined at zero. However, as we will see in the next section,
we can write, for z = 0,
1 1
f (z) = sin(z 4 )/z 2 = 2 z 4 − z 12 + · · ·
z 6
2 1 10
= z − z + ··· .
6
Since the right side of this equation is a power series about 0, it defines
a differentiable function g(z) which is equal to f (z) if z = 0. We can
therefore extend f (z) to a differentiable function by setting f (z) = g(z) if
z = 0, and f (0) = g(0) = 0. If we do this, then the extended function g(z)
has a zero of order 2 at 0. This is the idea behind a removable singularity,
which is discussed in the next chapter.
24. We can treat this function like that of Problem 23, since
1 1
sin(z − π) = (z − π) − (z − π)3 + (z − π)5 + · · · ,
3! 5!
so if (z − π)5 is divided by sin2 (z − π) we obtain a Maclaurin expansion
whose first term is (z − π)3 . We can extend f (z) to this function, defining
f (0) = 0, and this extended function has a zero of order 3 at π.
25. If we compute the kth derivative of f (z) at z0 by differentiating each series
term by term, we obtain
∞
(k) n−k
f (z0 ) = an (n)(n − 1) · · · (n − k + 1)(z − z0 )
n=0 z=z0
= ak k!
∞
n−k
= bn (n)(n − 1) · · · (n − k + 1)(z − z0 )
n=0 z=z0
= bk k!.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
608 CHAPTER 21. SERIES REPRESENTATIONS OF FUNCTIONS
Then
f (k) (z0 )
ak = = bk .
k!
The coefficients of the power series expansion of f (z) about z0 must be
the Taylor coefficients at z0 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
21.2. THE LAURENT EXPANSION 609
This expansion represents of f (z) in the annulus 0 < |z| < ∞, which is
the complex plane with the origin removed.
3. If z = 0, then
∞
1 − cos(2z) 1 (−1)n 2n
= 2 1− (2z)
z2 z n=0
(2n)!
∞
(−1)n+1 4n 2n−2
= z .
n=1
(2n)!
for 0 < |z| < ∞. Notice that in this case the Laurent expansion about 0
is actually a Taylor series about 0 and converges at 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
610 CHAPTER 21. SERIES REPRESENTATIONS OF FUNCTIONS
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
21.2. THE LAURENT EXPANSION 611
Substitute these into the integrals in the sum representing f (z) and inter-
change the integrals and the summations to obtain
∞
1 f (w)
f (z) = dw (z − z0 )n
2πi γ2 n=0 (w − z0 )n+1
∞
n+1
1 n 1
+ f (w)(w − z0 ) dw
2πi γ1 n=0 (z − z0 )
∞
1 f (w)
= n+1
dw (z − z0 )n
n=0
2πi γ2 (w − z0 )
∞
1 1
+ f (w)(w − z0 )n dw n+1
.
n=0
2πi γ1
(z − z 0)
Finally, use the deformation theorem to replace these integrals over γ1 and
γ2 with integrals over Γ, which is any path in the annulus and enclosing
z0 . This gives us
∞
f (z) = cn (z − z0 )n ,
n=−∞
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
612 CHAPTER 21. SERIES REPRESENTATIONS OF FUNCTIONS
where
1 f (w)
cn = dw,
2πi Γ (w − z0 )n+1
completing the proof.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 22
22.1 Singularities
1. cos(z)/z 2 has just one singularity, z = 0, and this is a pole of order 2
because cos(0) = 0.
2. f (z) has a double pole at −i and a simple pole at i, since the numerator
is differentiable and nonzero at these points.
3. e1/z (z + 2i) has a essential singularity at z = 0. The factor z + 2i makes
no contribution to singularities of this function.
4. The function is differentiable at all z = π, so we need only concern our-
selves with what is happening at π. Now
so
sin(z) sin(z − π)
=− .
z−π z−π
Then
sin(z) sin(z − π)
lim = − lim
z→π z−π z→π z−π
sin(z)
= − lim = −1.
z→0 z
The fact that this limit is nonzero means that sin(z)/(z − π) has a remov-
able singularity at π.
5. The only singularities are 1, i, −i, and 1 is a double pole, while i and −i
are simple poles.
613
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
614 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
7. Write
z−i z−i 1
2
= = .
z +1 (z + i)(z − i) z+i
Then f (z) has a removable singularity at i and a simple pole at −i.
1 z
sinh(z) = e − e−z = 0.
2
Then ez − e−z = 0, so
e2z = 1.
We know where the complex exponential function equals 1. e2z = 1 exactly
when 2z = 2nπi, or z = nπi, with n any integer. Since sin(nπi) = 0 unless
n = 0, then the numbers nπi n = 0 are singularities of sin(z)/ sinh(z).
Further, cosh(nπi) = 0, so these are simple poles of sin(z)/ sinh(z). If
n = 0, then we have the origin 0. But sin(z) and sinh(z) have simple zeros
at 0, so 0 is a removable singularity of this function.
9. Write
z z
= .
z4 − 1 (z − 1)(z + 1)(z − i)(z + i)
This function has simple poles at ±1, ±i.
10. tan(z) = sin(z)/ cos(z) has simple poles at the simple zeros (2n + 1)π/2
of cos(z), in which n is any integer.
11. sec(z) = 1/ cos(z) has simple poles at the zeros of cos(z). These are
(2n + 1)π/2, with n any integer.
12. e1/z(z+1) has essential singularities at 0 and −1. One way to see this is to
write
1 1 1
= − ,
z(z + 1) z z+1
so
e1/z(z+1) = e1/z e−1/(z+1) .
Now look at z = 0, to be specific. We know that e1/z has an essential
singularity at 0, and e−1/(z+1) is differentiable at 0, so the product will
have an essential singularity there. Similar reasoning applies at −1, since
e1/z is differentiable at −1.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.2. THE RESIDUE THEOREM 615
Compute
d 1 + z2
Res(f, 1) = lim
z→1 dz z + 2i
(z + 2i)(2z) − (1 + z 2 )
= lim
z→1 (z + 2i)2
4i
=
−3 + 4i
and
1 + z2 −3
Res(f, −2i) = lim 2
= .
z→−2i (z − 1) −3 + 4i
Then
1 + z2 4i 3
dz = 2πi − = 2πi.
γ (z − 1)2 (z + 2i) −3 + 4i −3 + 4i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
616 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
2. γ encloses i, which is a double pole and the only singularity of f (z). Then
2z d
2
dz = 2πiRes(f, i) = lim (2z) = 4πi.
γ (z − i) z→i dz
4. The only singularities of f (z) are ±2i, which are simple poles enclosed by
γ. Then
cos(z)
2
dz = 2πiRes(f, 2i) + 2πiRes(f, −2i)
γ 4+z
cos(2i) cos(−2i)
= 2πi + 2πi
4i −4i
= 0.
√ √
5. The function being integrated has simple poles at 6i and − 6i, both
enclosed by γ. Then
√ √
z+i
2
dz = 2πiRes(f, 6i) + 2πiRes(f, − 6i)
γ z +6
√ √
6+1 6−1
= 2πi √ + 2πi √ = 2πi.
2 6 2 6
7. z/ sinh2 (z) has a simple pole at z = 0, and double poles at the nonzero
zeros of sinh(z), which occur at z = nπi for integer values of n. The only
pole of z/ sinh2 (z) enclosed by γ is z = 0. Therefore
z
2 dz = 2πiRes(f, 0)
γ sinh (z)
Compute this residue as
z2
Res(f, 0) = lim (zf (z)) = lim
z→0 z→0 sinh2 (z)
2
z
= lim
z→0z 2 + 16 z 4 + · · ·
1
= lim = 1.
z→0 1 + 1 z 2 + · · ·
6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.2. THE RESIDUE THEOREM 617
Therefore
z
dz = 2πi.
γ sinh2 (z)
9. The integrand has simple poles at i, 3i and −3i. Only the pole at −3i is
enclosed by γ, so
iz
2
dz = 2πiRes(f, −3i)
γ (z + 9)(z − i)
iz 1 πi
= 2πi lim = 2πi − =− .
z→−3i (z − 3i)(z − i) 8 4
2
10. e2/z has an essential singularity at z = 0. Using the exponential series,
∞ n
2 1 2
e2/z = 2
.
n=0
n! z
11. The integrand has a simple pole at z = −4i, which is outside γ. Since
f is differentiable on and in the region bounded by γ, then by Cauchy’s
theorem,
8z − 4i + 1
dz = 0.
γ z + 4i
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
618 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
14. The integrand has simple poles at the cube roots of 8, which are 2, 2e2πi/3 ,
2e4πi/3 . Only 2 is enclosed by γ, so
(1 − z)2
3
dz = 2πiRes(f, 2)
γ z −8
(1 − z)2
= 2πi lim 2
z→2
3z
1 πi
= 2πi = .
12 6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.2. THE RESIDUE THEOREM 619
Now,
h (z) = 2(z − z0 )ϕ(z) + (z − z0 )2 ϕ (z),
h (z) = 2ϕ(z) + 4(z − z0 )ϕ (z) + (z − z0 )2 ϕ (z).
and
h(3) (z) = 6ϕ (z) + 6(z − z0 )ϕ (z) + (z − z0 )2 ϕ(3) (z).
Therefore
1 1
h (z0 ) and ϕ (z0 ) = h(3) (z0 ).
ϕ(z0 ) =
2 6
Substituting these into the above expression for the residue, we obtain
Res(f /f, z0 ) = k.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
620 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
It is important
in this calculation
of an integral to be clear on the difference
between γ g(z) dz, and γ (f (z)/f (z)) dz. In this example f (z)/f (z) =
2g(z).
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.2. THE RESIDUE THEOREM 621
20. Let g(z) = tan(z). g has simple poles at ±π/2, enclosed by γ. By the
residue theorem,
tan(z) dz = 2πi [Res(g, π/2) + Res(g, −π/2)]
γ
sin(π/2) sin(−π/2)
= 2πi +
− sin(π/2) − sin(−π/2)
= 2πi [−1 − 1] = −4πi.
To use the argument principle, notice that g(z) = −f (z)/f (z), where
f (z) = cos(z). Now f has no poles, and simple zeros at ±π/2 enclosed by
γ. Then
sin(z)
tan(z) dz = dz
γ γ cos(z)
f (z)
=− dz
γ f (z)
= −2πi(Z − P ) = −4πi.
√
21. First, g(z) = (z + 1)/(z 2 + 2z + 4) has simple poles at −1 ± 3i, enclosed
by γ. Then
√ √
z+1
2
dz = 2πi Res(g, −1 − 3i) + Res(g, −1 + 3i)
γ z + 2z + 4
√ √
1 − 1 − 3i −1 + 3i + 1
= 2πi √ + √
2(−1 − 3i) + 2 2(−1 + 3i) + 2
1 1
= 2πi + = 2πi.
2 2
z+1 1 f (z)
= ,
z 2 + 2z + 4 2 f (z)
22. Because p has exactly n simple zeros enclosed by γ, then p (z)/p(z) has
simple poles at z1 , · · · , cn , and
p (zj )
Res(p /p, zj ) = = 1.
p (zj )
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
622 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
If we use the argument principle, then p(z) has exactly n simple zeros
enclosed by γ, so Z = n, and a polynomial has no poles, so P = 0, hence
p (z)
dz = 2πi(Z − P ) = 2nπi.
γ p(z)
Then
2π
1 −1 2π
dθ = 2i(2πi) √ =√ .
0 2 − cos(θ) 2 3 3
Note that the integral must be real and positive, since the integrand is
positive, and this checks out.
2. f (z) = 1/(z 4 + 1) has two simple poles in the upper half-plane, at fourth
roots of −1 having positive imaginary parts. these are
1 1
z1 = √ (1 + i) and z2 = √ (−1 + i).
2 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.3. EVALUATION OF REAL INTEGRALS 623
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
624 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
to obtain
∞ √ √ √
x sin(2x) e−2 2 e2 2i − e−2 2i
dx = Im 2πi
−∞ x4 + 16 8 2i
√
πe−2 2 √
= sin(2 2).
4
2
6. f (z) = 1/(z√ − 2z + 6) has one simple pole in the upper half-plane and it
is z1 = 1 + 5i. Compute
1 1
Res(f, z1 ) = = √ .
2z1 − 1 2 5i
Then ∞
1 π
dx = √ .
−∞ x2 − 2x + 6 5
7. First use the identity
1
cos2 (x) = (1 + cos(2x))
2
to write the integral as
∞
cos2 (x) 1 ∞ 1 + cos(2x)
2 2
dx = dx.
−∞ (x + 4) 2 −∞ (x2 + 4)2
Let
1 + e2iz
f (z) = .
(z 2 + 4)2
Then f has a pole of order 2 in the upper half-plane at 2i, and
d 1 + e2iz 1 + 5e−4
Re(f, 2i) = lim 2
= .
z→2i dz (z + 2i) 32i
Then
∞
cos2 (x) 1 1 + 5e−4
dx = Re 2πi
−∞ (x2 + 4)2 2 32i
π −4
= (1 + 5e ).
3
8. Complex methods will work for this integral, but it is easier to observe
that, with the change of variables θ = 2π − ϕ,
2π π
sin(θ) sin(ϕ)
dθ = − dϕ.
π 2 + sin(θ) 0 2 + sin(ϕ)
Then 2π
sin(θ)
dθ = 0.
0 2 + sin(θ)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.3. EVALUATION OF REAL INTEGRALS 625
9. Let f (z) = z 2 /(z 2 + 4)2 . The only singularity of f in the upper half-plane
is 2i, which is a double pole. Compute
d z2 i
Res(f, 2i) = lim =− .
z→2i dz (z + 2i)2 8
Then
∞
x2 i π
dx = 2πi − = .
−∞ (x2 + 4)2 8 4
10. Let
eiβx
f (z) = .
(z 2 + α2 )2
Then f has only one singularity in the upper half-plane, a double pole at
αi. Compute
d eiβz
Res(f, αi) = lim
z→αi dz (z + αi)2
(αβ + 1)e−αβ
=− i.
4α3
Then
∞
cos(βx) (αβ + 1)e−αβ
dx = 2πi − i
∞ (x2 + α2 )2 4α3
(αβ + 1)e−αβ π
= .
2α3
11. Let
eiαz
f (z) = .
z2 + 1
The only singularity f has in the upper half-plane is a simple pole at i.
Compute
e−α
Res(f, i) = .
2i
Then
∞
cos(αx) e−α
dx = 2πi = πe−α .
−∞ x2 + 1 2i
12. Let
z 2 eiαz
f (z) = .
(z 2 + β 2 )2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
626 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
Then f (z) has a double pole in the upper half-plane at βi. Compute
d z 2 eiαz
Res(f, βi) = lim
z→βi dz (z + βi)2
= lim 2zeiαz (z + βi)−2 + iz 2 αz(z + βi)−2 − 2z 2 eiαz (z + βi)−3
z→βi
Then
∞ −αβ
x2 cos(αx) e
dx = 2πi i(αβ − 1)
−∞ (x2 + β 2 )2 4β
π −αβ
= e (1 − αβ).
2β
Solving for the zeros of the denominator of the integrand, we find that the
singularities satisfy
β−α β+α
z2 = or z 2 = .
β+α β−α
The simple poles enclosed by the unit circle are the square roots z1 and z2
of (β − α)/(β + α). The residue of the integrand at each of these poles are
obtained by a straightforward computation using Corollary 22.1. After
some computation, we obtain
1
Res(f, zj ) =
8αβ
for j = 1, 2. Then
2π
1 4 2 2π
dθ = (2πi) = .
0 α2 cos2 (θ) + β 2 sin2 (θ) i 8αβ αβ
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.3. EVALUATION OF REAL INTEGRALS 627
14. Let
1
g(θ) = .
α + sin2 (θ)
Write
2π π/2 π
g(θ) dθ = g(θ) dθ + g(θ) dθ
0 0 π/2
3π/2 2π
+ g(θ) dθ + g(θ) dθ.
π 3π/2
The integrand of the last integral has simple poles at z1 and z2 , where
zj = (1 + 2α) − 2 α2 + α.
Then
π/2
1 −2
dθ = i(2πi) √
0 α + sin2 (θ) 8 α2 + α
π
= √ .
2 α2 + α
15. Let Γ denote the suggested rectangular path. The four sides are
The intervals for the parameters on the sides are chosen to maintain coun-
2
terclockwise orientation around Γ. Now observe that e−z is differentiable
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
628 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
on and in the region bounded by Γ, and use Cauchy’s theorem and the
parametrization on each side of the rectangle to write
4
2 2
e−z dz = 0 = e−z dz.
Γ j=1 Γj
Next,
β β
−z 2 −(R2 +2Rti−t2 ) −R2 2
e dz = e i dt = ie et [cos(2Rt)−i sin(2Rt)] dt.
Γ2 0 0
−x2
Now e sin(2βx) is an odd function on the real line, so
∞
2
e−x sin(2βx) dx = 0.
−∞
Therefore,
∞ ∞
2 2 2
eβ e−x cos(2βx) dx = e−x dx.
−∞ 0
Finally, use the known result that
∞
2 √
e−x dx = π
−∞
to obtain ∞
2 √ 2
e−x cos(2βx) dx = πe−β .
−∞
Finally, because the integrand is an even function, then
∞ √
2 π −β 2
e−x cos(2βx) dx = e .
0 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.3. EVALUATION OF REAL INTEGRALS 629
16. Let Γ be the path indicated in Figure 22.3 of the text. By Cauchy’s
theorem,
2
eiz dz = 0.
Γ
Now examine the integral on the left over the three pieces of Γ consisting
of the segment on the x− axis (Γ1 ), the circular arc (Γ2 ), then the segment
from the end of this arc back to the origin (Γ3 ).
On Γ1 , z = x and
R R
2 2
eiz dz = eix dx = [cos(x2 ) + i sin(x2 )] dx.
Γ1 0 0
On Γ2 , z = Reiθ and
π/4
iz 2 2
e dz = eiR e2iθ dθ.
Γ2 0
iπ/4
On Γ3 , z = re , so
0
2 2
eiz dz = e−r eiπ/4 dr.
Γ3 R
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
630 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
Now
zeiαz
f (z) =
z4 + β4
has simple poles in the upper half-plane at z1 = βeiπ/4 and z2 = βe3πi/4 .
Compute the residues of f at these poles. In general,
iαz
ze eiαzk
Res(f, zk ) = = .
4z 3
z=zk 4zk2
Then,
iπ/4 3πi/4
eiαβe eiαβe
Res(f, z1 ) = and Res(f, z2 ) = .
4β 2 i −4β 2 i
Then
∞
x sin(αx) 1 2πi iαβ(1+i)/√2 √
iαβ(−1+i)/ 2 1
dx = Im e − e
0 x4 + β 4 2 4β 2 i
√
−αβ/ 2
πe αβ
= sin √ .
2β 2 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.4. RESIDUES AND THE INVERSE LAPLACE TRANSFORM 631
In the last integral on the right, put θ = 2π − u to show that the two
integrals on the right are equal, hence
π
1 1 2π 1
dθ = dθ
0 (α + β cos(θ))2 2 0 (α + β cos(θ))2
1 1 1
= 2
dz
2 γ (α + β(z + 1/z)/2) iz
2 z
= dz.
i γ (βz 2 + 2αz + β)2
Now
z
f (z) =
(βz 2 + 2αz + β)2
has double poles at the zeros of βz 2 + 2αz + β, which are
−α + α2 − β 2 −α − α2 − β 2
z1 = and z2 = .
β β
Then
π
1 2 α πα
dθ = (2πi) = 2 .
0 (α + β cos(θ))2 i 4(α2 − β 2 )3/2 (α − β 2 )3/2
1 3i 1
Res(etz F (z), 3i) = e and Res(etz F (z), −3i) = e−3i .
2 2
Then
1 3i
L−1 [F (s)](t) = (e + e−3i ) = cos(3t).
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
632 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
d tz
Res(etz F (z), −3) = lim (e ) = te−3t .
z→−3 dz
Then
L−1 [F (s)](t) = te−3t .
3. Let
1
F (z) = .
(z − 2)2 (z + 4)
F has a double pole at 2 and simple pole at −4. Compute
tz
d e
Res(etz F (z), 2) = lim
z→2 dz z+4
tz
te etz
= lim −
z→2 z + 4 (z + 4)2
1 1
= te2t − e2t .
6 36
Next,
e−4t
Res(etz F (z), −4) = .
36
Then
−1 1 1 1 −4t
L [F (s)](t) = t− e2t + e .
6 36 36
4. Let
1
F (z) = .
(z 2 + 9)(z − 2)2
F has simple poles at ±3i and a double pole at 2. Compute
2 5
Res(etz F (z), 3i) = − i e3it ,
169 1014
tz 1
Res(e F (z), −3i) = e−3it ,
72 + 30i
1 2t 4 2t
Res(etz F (z), 2) = te − e .
13 169
A routine but lengthy (by hand) calculation of the sum of these residues
yields
−1 −4 1
L [F (s)](t) = + t e2t
169 13
4 5
+ cos(3t) − sin(3t).
169 507
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
22.4. RESIDUES AND THE INVERSE LAPLACE TRANSFORM 633
7. Let F (z) = 1/(1 + z 4 . Then F has simple poles at the fourth roots of −1,
which are
1 1
z1 = √ (1 + i), z2 = √ (−1 + i),
2 2
1 1
z3 = √ (1 − i), z4 = √ (−1 − i).
2 2
The residues are
1 √
Res(etz F (z), z1 ) = √ e(1+i)t/ 2 ,
2 2(−1 + i)
1 √
Res(etz F (z), z2 ) = √ e(−1+i)t/ 2 ,
2 2(1 + i)
1 √
Res(etz F (z), z3 ) = √ e(1−i)t/ 2 ,
2 2(−1 − i)
1 √
Res(etz F (z), z4 ) = √ e(−1−i)t/ 2 .
2 2(1 − i)
Upon rearranging the sum of these residues, we obtain
1 t t 1 t t
L−1 [F (s)](t) = − √ sinh √ cos √ + √ cosh √ sin √ .
2 2 2 2 2 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
634 CHAPTER 22. SINGULARITIES AND THE RESIDUE THEOREM
9. Let F (z) = z 2 /(z − 2)3 . Then F has a pole of order 3 at 2. The residue is
1 d2 2 tz
Res(etz F (z), 2) = lim (z e )
z→2 2 dz 2
= lim (2e + 4tze2t + t2 z 2 e2t )
2t
z→2
= (1 + 4t + 2t2 )e2t .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
Chapter 23
so
u = cos(x) cosh(y), v = − sin(x) sinh(y).
We will examine the image of a vertical or horizontal line under this map-
ping. First consider the vertical line x = a. An image point has the form
(cos(a) cosh(y) − sin(a) sinh(y)). If a is not a zero of cos(x) or sin(x), then
u2 v2
− = 1.
cos (a) sin2 (a)
2
This is the equation of a hyperbola in the w− plane, but the image is only
one branch of this hyperbola, because cosh(y) > 0 for all real y.
If a = nπ for an integer n, then sin(a) = 0 and the image point of a point
on the line is (cos(nπ) cosh(y), 0), or ((−1)n cosh(y), 0). Now, cosh(y) ≥ 1
for all real y. Therefore, depending on whether n is even or odd, the image
635
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
636 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
22
20
18
16
14
12
10
0
-20 -15 -10 -5 0 5 10 15 20
2.5
1.5
0.5
0
0 0.4 0.8 1.2 1.6 2 2.4
-0.5
-1
-1.5
-2
-2.5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 637
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
0
-6 -4 -2 0 2 4 6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
638 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
0
0 1 2 3 4 5 6 7
-2
-4
-6
of the line x = nπ) is either the interval [1, ∞) or the interval (−∞, −1)
on the real axis in the w− plane.
If a = (2n + 1)π/2, for n an integer, then cos(a) = 0, so the image of a
point on the line is (0, − sin((2n + 1)π/2) sinh(y)). The image of the line is
the imaginary axis in the w0 plane, since sinh(y) varies over all real values
as y varies from −∞ to ∞.
For a horizontal line y = b, if b = 0, the image of the line y = b is given
by points
w = cos(x) cosh(b) − i sin(x) sinh(b).
If b = 0, this is the ellipse
u2 v2
+ = 1.
cosh (b) sinh2 (b)
2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 639
-0.4
-0.8
-1.2
-1.6
-2
-2.4
-2.8
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
640 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
-11
3.5
2.5
1.5
0.5
0
-3 -2 -1 0 1 2 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 641
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.7
-0.8
-0.9
-1
-1.1
0
0 1 2 3 4 5 6 7 8 9 10
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
642 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
20
18
16
14
12
10
0
0 2 4 6 8 10 12 14
2.2
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 643
0
-10 -8 -6 -4 -2 0 2 4 6 8 10
-2
-4
-6
-8
0
5 6 7 8 9 10 11 12 13 14 15
-2
-4
-6
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
644 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
π/2 to 5π/2, an interval of length 2π. Therefore the image of the sector
π/4 ≤ θ ≤ 5π/4 is the entire w− plane.
Assuming that cos(k) and sin(k) are not zero, then a little algebraic ma-
nipulation gives us
u2 v2
− =1
cos (k) sin2 (k)
2
which is the equation of a hyperbola. The foci are (±c, 0), where
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 645
10
-10 -5 0 5 10
0
-5
-10
v2
u = 2a2 −
8a2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
646 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
10
-10 -5 0 5 10
0
-5
-10
having intercepts at (2a2 , 0) and opening to the left. The horizontal line
y = 0 maps to u = 2y 2 ≥ 0, v = 0, the positive u− axis. Other horizontal
lines y = b map onto the parabolas
v2
u= − 2b2
8b2
having intercepts (−2b2 , 0) and opening right.
Figure 23.18 shows the image of the rectangle defined by 0 ≤ x ≤ 3/2, −3/2 ≤
y ≤ 3/2.
10. Let w = ez = ex+iy for all real x and for 0 ≤ y ≤ 2π. If we write
then the fact that y varies over an entire period of the sine and cosine
functions means that every point of the w− plane, except 0, is the image
of a point in the z− plane (let z = log(w). Thus ez maps the z− plane to
the entire w− plane with the origin removed.
11. If Re(z) = −4, then (z + z)2 = −4, so z + z = −8. Now, if w = 2i/z, then
z = 2i/w, so
2i 2i
z+z = − = −8.
w w
Multiply this by ww and rearrange terms to obtain
8ww − 2i(w − w) = 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 647
0
-4 -3 -2 -1 0 1 2 3 4
-2
-4
-6
-8
2(u2 + v 2 ) + v = 0.
This is the equation of a circle of radius 1/2 centered at (0, −1/4) in the
w− plane, and is the image of the vertical line x = −4 under the given
mapping.
12. Solve w = 2iz − 4 to write
w+4
w= .
2i
Now Re(z) = (z + z)/2 = 5 becomes
w+4 w+4
− = 10.
2i 2i
Set w = u + iv to obtain
w−w
= Im(w) = v = 10,
2i
a horizontal line in the w− plane.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
648 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
4(u2 + v 2 ) + 7v + u = 3.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 649
Next,
1 (2 − u)u + (5 − v)v
(z − z) = Im(z) = .
2i (u − 2)2 + v 2
Substitute these into the equation of the given line and clear fractions to
obtain
−2
z= .
w − 3i − 1
Solve for w:
(1 + 4i)z − (3 + 8i)
w= .
(2 + 3i)z − (4 + 7i)
18. Substitute the given values in equation (23.1) and solve for w to obtain
(1 + i)z − (2 + 2i)
w= .
(3 − i)z − 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
650 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
19. Since w3 = ∞, substitute the given values into equation (23.1), but leave
out the terms involving w3 to obtain
Solve for w:
(33 + i)z − (48 + 16i)
w= .
5(z − 4)
21. Substitute these values into equation (23.1) and solve for w to obtain
(3 + 22i)z + 4 − 75i
w= .
(2 + 3i)z − (21 − 4i)
az + b
T (z) = .
cz + d
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.1. CONFORMAL MAPPINGS 651
az + b
T (z) = z = .
cz + d
But then
cz 2 + (d − a)z − b = 0.
This is a quadratic equation if c = 0. T has two fixed points (if this
quadratic equation has distinct roots), or one fixed point (if the quadratic
equation has repeated roots).
This leaves the case that c = 0. In this case
a b
T (z) = z+ .
d d
If b = 0, then this is a translation, contrary to our assumption. Therefore
in this case b = 0 and T (z) = kz, where k = a/d. If a = d, this is the
identity mapping, and we assumed that it is not. Therefore a = d and T
is a magnification/rotation, which has exactly one fixed point, z = 0.
Therefore every bilinear transformation that is neither a translation nor
the identity mapping has one or two fixed points. A translation has the
form T (z) = az + b with b = 0, and leaves no point unmoved. Thus a
translation has no fixed point. The identity mapping T (z) = z leaves
every point unmoved, so every point is a fixed point.
25. Let
az + b
T (z) = .
cz + d
If T is not a translation or the identity mapping, then by the argument
used for Problem 24, T can have at most two fixed points. Therefore,
if T has three fixed points, then either T is a translation or the identity
mapping. But a translation has no fixed point, hence T is the identity
mapping.
az + b
T (z) =
cz + d
is a bilinear mapping, then ad − bc = 0, which guarantees that T has an
inverse mapping. It is routine to solve T (z) = w for z in terms of w to
obtain the inverse transformation
−wd + b
T −1 (w) = ,
wc − a
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
652 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
S = S ◦ I = S ◦ (S −1 ◦ T )
= (S ◦ S −1 ) ◦ T = I ◦ T = T.
z2 → 1, z3 → 0, z4 → ∞.
Then
[z1 , z2 , z3 , z4 ] = P (z1 ).
Now let T be any bilinear transformation. Then
Then R ◦ T = P . Then
28. Let
z3 − z4 z − z2
w = T (z) = 1 − .
z3 − z2 z − z4
A routine calculation yields
Since three points and their images uniquely determine a bilinear transfor-
mation, as noted in the solution to Problem 26. T is this unique bilinear
transformation, so
[z1 , z2 , z3 , z4 ] = T (z1 ).
29. In the definition of cross ratio, w2 , w3 , w4 all lie on an (extended) line, the
real axis. Since circles/lines map to circles/lines under bilinear transfor-
mations, then [z1 , z2 , z3 , z4 ] is real if and only if z1 , z2 , z3 , z4 all lie on the
same line or circle.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.2. CONSTRUCTION OF CONFORMAL MAPPINGS 653
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
654 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
5. We can map the line Re(z) = 0 onto the circle |w| = 4 by a bilinear
transformation. The domain Re(z) < 0 consists of all numbers to the left
of the imaginary axis, which is the boundary. Choose three points on this
axis, ordered upward so the region Re(z) < 0 is on the left as we walk up
the line. Choose three points on the image circle |w| = 4, counterclockwise
so the interior of this circle is on our left as we walk around it in this order.
Convenient choices are
z1 = −i, z2 = 0, z3 = i, w1 = −4i, w2 = 4, w3 = 4i.
The bilinear transformation mapping zj → wj
1+z
w = T (z) = 4 .
1−z
As a check, z = −1, which has negative real part, maps to 0, interior to
the circle |w| < 4. Thus w maps Re(z) < 0 to |w| < 4.
Of course, other choices for the zj s and wj s may result in different
mappings between the two given domains.
6. The domain Im(z) > −4 consists of all x + iy lying above the horizontal
line y = −4. This has as boundary the line y = −4. We want to make
this domain to |w − i| > 2, the exterior of the circle of radius 2 centered at
i. This domain has boundary |w − i| = 2. Choose three points on the line
y = −4 in the z− plane, ordered from left to right so that the domain is
to the left. Choose three points on the circle in the w− plane, clockwise
so as we walk around the boundary the region (exterior to the circle) is
on the left. Convenient choices are
z1 = −1 − 4i, z2 = −4i, z3 = 1 − 4i, w1 = 3i, w2 = 2 + i, w3 = −i.
Find the bilinear transformation mapping zj → wj by solving for w in the
equation
(3i − w)(−2 − 2i)(−1)(1 − 4i − z) = (−1 − 4i − z)(−2 + 2i)(−i − w)
to obtain
(−2 + i)z − (3 + 10i)
w= .
z + 3i
7. Because the boundary of the wedge in the w− plane is not a line or circle,
we cannot construct a bilinear mapping to solve this problem. However,
wedges suggest using polar representations. Let z = reiθ for 0 < θ < π.
These are points in the upper half-plane. Let
w = z 1/3 = r1/3 eiθ/3 = ρeiϕ ,
where ρ > 0 and 0 < ϕ < π/3. This mapping is conformal because
dw 1
= z −2/3 = 0
dz 3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.2. CONSTRUCTION OF CONFORMAL MAPPINGS 655
for z in the upper half-plane, and the mapping takes the open upper half-
plane onto the open wedge 0 < θ < π/2.
d 1
(log(z)) = = 0.
dz z
9. The solution to this problem requires some familiarity with the gamma
and beta functions, which are discussed in Section 15.3.
To show that f maps the upper half-plane onto the given rectangle, we
will evaluate the function at −1, 0, 1 and ∞ and then show that these are
the vertices of that rectangle.
First, it is obvious that f (0) = 0. Next,
1
f (1) = 2i (ξ 2 − 1)−1/2 ξ −1/2 dξ
0
1 1
(1 − ξ 2 )−1/2 −1/2
2i ξ dξ = 2 (1 − ξ 2 )−1/2 ξ −1/2 dξ.
0 i 0
Let ξ = u1/2 to obtain (in terms of the beta and gamma functions)
1
f (1) = (1 − u)−1/2 u−3/4 du
0
Γ(1/4)Γ(1/2)
= B(1/4, 1/2) = = c.
Γ(3/4)
Next calculate
−1
f (−1) = 2i (ξ 2 − 1)−1/2 ξ −1/2 dξ.
0
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
656 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
Let ξ = −u to obtain
1
f (−1) = 2i (1 − u2 )−1/2 u−1/2 du
0
iΓ(1/4)Γ(1/2)
= iB(1/4, 1/2) = = ic.
Γ(3/4)
Finally, calculate
∞
f (∞) = 2i (ξ + 1)−1/2 (ξ − 1)−1/2 ξ −1/2 dξ
0
1
= 2i (ξ + 1)−1/2 (ξ − 1)−1/2 ξ −1/2 dξ
0∞
+ 2i (ξ + 1)−1/2 (ξ − 1)−1/2 ξ −1/2 dξ.
1
The first integral on the last line is B(1/4, 1/2). In the second integral,
put ξ = 1/u to obtain
0 −1/2 −1/2
1+u 1−u 1/2 1
f (∞) = c + 2i u du
1 u u u2
1
= c + 2i (1 − u2 )−1/2 u−1/2 du = (1 + i)c.
0
1. Begin by mapping the upper half-plane Im(z) > 0 to the unit disk |w| < 1.
One such mapping is
i−z
w = T (z) = .
i+z
The solution of this dirichlet problem for the upper half-plane is
where C is the boundary of the upper half-plane (the real line) and
1 T (ξ) + T (z) T (ξ)
f (z) = g(ξ) dξ.
2πi C T (ξ) − T (z) T (ξ)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.3. CONFORMAL MAPPING SOLUTIONS OF DIRICHLET PROBLEMS657
−2(1 + tz)
.
(z − t)(1 + t2 )
(1 + tx)(x − t) − ty 2 − iy(1 + t2 ) −2
.
(x − t)2 + y 2 1 + t2
Substitute this into the integral and extract the real part, recalling that
g(t) is real-valued, to obtain
y ∞ g(t)
u(x, y) = dt.
π −∞ (x − t)2 + y 2
This agrees with the solution of the Dirichlet problem for the upper half-
plane obtained in Chapter 18.
2. The mapping
i − z2
w = T (z) =
i + z2
takes the first quadrant onto the unit disk. (Note that this is not a bilinear
mapping). Compute
T (z) 2iz
= .
T (z) 1 + z4
The boundary of the right quarter-plane (first quadrant) consists of L1 ,
the nonnegative real axis, and L2 , the nonnegative imaginary axis. On
L2 , ξ = it for t varying from ∞ to 0 (down this axis to maintain positive
orientation on the boundary of the first quadrant). Put ξ = it on L2 to
compute
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
658 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
1 t2 z 2 + 1 2it
+ g(t) 2 dt .
2πi 0 i(t − z 2 ) 1 + t4
To determine this real part of these integrals, recall that g(it) = g(0, t) and
g(t) = g(t, 0) are both real-valued. Further, both integrals have a factor
of i2 in the denominator (including the 2πi factor). Thus each integral is
left with a factor i, and we must find:
2
t (x + iy)2 − 1
Im
t2 + (x + iy)2
2 2
t (x − y 2 ) − 1 + 2xyt2 i
= Im
t2 + x2 − y 2 + 2xyi
2xy(1 + t4 )
= 2
(t + x2 + y 2 )2 + 4x2 y 2
and, omitting some computational details,
2 2
t z +1 2xy(1 + t4 )
Im = .
t2 − z 2 t2 − x2 + y 2 + 4xy
We therefore obtain the solution
u(x, y) =
2xy ∞ tg(0, t)
dt
π 0 (t2 + x2 − y 2 )2 + 4x2 y 2
∞
2xy tg(t, 0)
+ dt.
π 0 t − x + y 2 + 4x2 y 2
2 2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.3. CONFORMAL MAPPING SOLUTIONS OF DIRICHLET PROBLEMS659
4. From Example 23.19, the integral solution for the right half-plane is
1 ∞ xg(it)
u(x, y) = dt.
π −∞ x + (t − y)2
2
6. By Poisson’s formula,
π/4
1 1 − r2
u(r cos(θ), r sin(θ)) = dϕ.
2π 0 1+ r2 − 2r cos(ϕ − θ)
7. First construct a conformal mapping of the strip S onto the unit circle.
Begin with w1 = πiz/2, which rotates the strip π/2 radians counterclock-
wise and expands it to the strip −π/2 ≤ Re(w1 ) ≤ π/2. The reason
for doing this is to exploit the mapping of Example 23.3. From this, put
w2 = sin(w1 ). This maps the w1 − strip onto the upper half-plane. Finally,
find the bilinear mapping that maps
−1 → −i, 0 → 1, 1 → i
to obtain
i − w2
w= ,
i + w2
mapping the upper half-plane of the w2 − plane to the unit disk in the w−
plane. The end result of this sequence of mappings is
i − sin(πiz/2)
w = T (z) = .
i + sin(πiz/2)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
660 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
Since g(ξ) = 0 along the upper and lower edges of S, the solution simplifies
to
and
i + sin(πt/2)
T (ξ) = T (it) = .
i − sin(πt/2)
We need to compute
T (it) π cos(πt/2)
d(it) = dt,
T (it) 1 + sin2 (πt/2)
and
T (ξ) + T (z) |T (ξ)|2 + 2iIm(T (z)T (ξ)) + |T (z)|2
= .
T (ξ) − T (z) |T (ξ)|2 − 2Re(T (z)T (ξ)) + |T (z)|2
Now |T (ξ)|2 = 1, since T maps the boundary of S onto the unit circle
|w| = 1. Finally, we can write the solution
−1
(1 − |t|) cos(πt/2) Im(T (z)T (it))
u(x, y) = dt.
1 1 + sin2 (πt/2) 1 − 2Re(T (z)T (it)) + |T (z)|2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 661
y = cot(θ)x + b
y = − tan(θ)x + c,
2. Write
f (z) = z 3 = (x + iy)3 = (x3 − 3xy 2 ) + i(3x2 y − y 3 ).
Then
ϕ(x, y) = x3 − 3xy 2 and ψ(x, y) = 3x2 y − y 3 .
Equipotential curves are graphs of curves
x3 − 3xy 2 = c
3x2 y − y 3 = k.
√
If c = 0, then x = 0 (the y− axis) or y = ±(1/ 3)x. These lines divide the
plane into six wedge-shaped regions meeting at the origin. Equipotential
curves occur in these regions and are asymptotic to its boundary lines.
Figure 23.19 shows some of these equipotential curves, and Figure 23.20
some streamlines.
Note that the streamlines can be obtained by rotating equipotential curves
π/2 radians clockwise.
The velocity of this flow is
V (x, y) = f (z)
= 3z 2 = 3(x2 − y 2 ) − 6xyi = u(x, y) + iv(x, y)
with
u(x, y) = 3(x2 − y 2 ) and v(x, y) = −6xy.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
662 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
y 0
-3 -2 -1 0 1 2 3
x
-1
-2
-3
y 0
-3 -2 -1 0 1 2 3
x
-1
-2
-3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 663
y 0
-4 -2 0 2 4
x
-1
-2
so the velocity is increasing with distance from the origin. We can envision
the potential f (z) = z 3 as describing
√ fluid motion along the streamlines,
with the straight lines y = ±(1/ 3)x and y = 0 acting as barriers of the
flow (such as sides of a container). As fluid particles near the origin they
slow down, and speed up again as they move away from the origin.
3. Begin with
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
664 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
y 0
-6 -4 -2 0 2 4 6
x
-2
-4
The velocity is
Then
u(x, y) = − sin(x) cosh(y), v(x, y) = cos(x) sinh(y).
This has divergence zero. Further, using Green’s theorem, it is routine to
check that the flux of the flow across any closed path is zero, so the flow
is solenoidal.
The circulation is also zero about any closed path, so there is no source
or sink for this flow.
4. First write
so
ϕ(x, y) = x − 2xy and ψ(x, y) = y + x2 − y 2 .
Equipotential lines (Figure 23.23) are graphs of curves
x − 2xy = c
y + x2 − y 2 = k.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 665
y 0
-3 -2 -1 0 1 2 3
x
-1
-2
-3
y 0
-3 -2 -1 0 1 2 3
x
-1
-2
-3
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
666 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
K K
= − 2 (r sin(t))(−r sin(t)) + 2 (r cos(t))(r cos(t)) dt
0 r r
= 2πK.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 667
1 2
(a1 − b1 )x + (a2 − b2 )y = [(a + a22 ) − (b21 + b22 )].
2 1
If c = 1, we get an equation
2 2
a1 − cb1 a2 − cb2
x− + y− = r2 ,
1−c 1−c
where
c
r2 = [(a1 − b1 )2 + (a2 − b2 )2 ].
(1 − c)2
These are circles if c > 0. Notice that the centers of these circles all lie on
the line
(a2 − b2 )x − (a1 − b1 )y + a1 b2 − a2 b1 = 0.
This line connects a = a1 + a2 i and b = b1 + b2 i in the complex plane.
This line containing the centers of the equipotential curves (for c = 1) is
orthogonal to the equipotential curve obtained when c = 1, and these two
lines intersect at ((a1 + b1 )/2, (a2 + b2 )/2), which is the midpoint of the
segment connecting a and b in the complex plane.
For the streamlines, write
z−a (z − a)(z − b)
=
z−b |z − b|2
|z|2 − (az + bz) + ab
=
|z − b|2
2 2
x + y − [(a1 + b1 )x + (a2 + b2 )y] + a1 b1 + a2 b2
=
(x − b1 )2 + (y − b2 )2
a2 b1 − a1 b2 − x(a2 − b2 ) + y(a1 − b1 )
+i .
(x − b1 )2 + (y − b2 )2
a2 b1 − a1 b2 − (a2 − b2 )x + y(a1 − b1 )
= k.
x2 + y 2 − [(a1 + b1 )x + (a2 + b2 )y] + a1 b1 + a2 b2
(a2 − b2 )x − (a1 − b1 )y + a1 b2 − a2 b1 = 0,
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
668 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
2
y
1
0
-2 -1 0 1 2 3 4 5
x
-1
-2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 669
2.5
y 1.5
0.5
0.5 1 1.5 2 2.5
x
Kx(x2 + y 2 + 1)
ϕ(x, y) = = c1
x2 + y 2
Ky(x2 + y 2 − 1)
ψ(x, y) = = c2 .
x2 + y 2
x(x2 − bx + 1)
y2 = − .
x−b
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
670 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
0.5
y 0
-1 -0.5 0 0.5 1
x
-0.5
-1
0.4
0.2
y 0
-0.6 -0.4 -0.2 0 0.2 0.4 0.6
x
-0.2
-0.4
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 671
Compute
m − ik a−b
f (z) = .
2π (z − a)(z − b)
Since the velocity is
then
f (z) = u(x, y) − iv(x, y)
so
The integral on the left is easily evaluated using the residue theorem. First
write
m − ik a−b
f (z) dz = 2πi Res ,
γ 2π (z − a)(z − b)
where the sum is over residues at the poles enclosed by γ. The integrand
has simple poles at a and b, and
a−b a−b
Res , a = 1, Res , b = −1.
(z − a)(z − b) (z − a)(z − b)
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
672 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
9. Let z = x + iy to obtain
1 ib
f (z) = K x + iy + + Log(x + iy)
x + iy 2π
Kx(x2 + y 2 + 1) b
= − arg(x + iy)
x2 + y 2 2π
Ky(x2 + y 2 − 1) b 2 2
+i + ln(x + y ) .
x2 + y 2 4π
Kx(x2 + y 2 + 1) b
ϕ(x, y) = − arg(x + iy) = c1 .
x2 + y 2 2π
Ky(x2 + y 2 − 1) b
ψ(x, y) = + ln(x2 + y 2 ) = c2 .
x2 + y 2 4π
Some equipotential lines are shown in Figure 23.29 for K = 1 and b = 2π.
Streamlines are shown in Figure 23.30 for K = 1 and b = 4π.
Compute
1 ib 1 2 ib
f (z) = K 1 − 2 + = 2 kz + z−k .
z 2πz z 2π
These points lie on the unit circle symmetrically across the imaginary axis.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 673
0.2
0.1
x
-0.4 -0.2 0 0.2 0.4
0
-0.1
y
-0.2
-0.3
-0.4
-0.5
0.6
0.4
y 0.2
0
-0.4 -0.2 0 0.2 0.4
x
-0.2
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
674 CHAPTER 23. CONFORMAL MAPPINGS AND APPLICATIONS
10. Because for this Log function, the argument is restricted to −π ≤ arg(z) <
π, we can write
√ √
√ ia 3 ia 3
f (z) = iKa 3 Log z − − Log z + .
2 2
Compute
√
√ ia 3 −3Ka2 ((z)2 + 3a2 /4)
f (z) = iKa 3 √ √ = 2 .
(z − ia 3/2)(z + ia 3/2) (z + 3a2 /4)((z)2 + 3a2 /4)
because the integral is positive. Since the circulation of the flow about
(0, a) is not zero, the flow is not irrotational.
Next compute the flux
−v dx + u dy
γ
2π
6K cos(θ) a 6K sin(θ) a
= { − sin(θ) + cos(θ) } dθ
0 2 + sin(θ) 2 2 + sin(θ) 2
= 0.
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
23.4. MODELS OF PLANE FLUID FLOW 675
9a4 K 2
(f (z))2 = √ √ .
(z − ia 3/2)2 (z + ia 3/2)2
By Blasius’s theorem, the thrust of the fluid outside the barrier 4x2 +
4(y − a)2 = a2 is the vector Ai + Bj, where
1
A − Bi = iρ (f (z))2 dz
2 γ
iρ 9a4 K 2
= √ √ dz
2 γ (z − ia 3)/22 (z + ia 3/2)2
√
= πρRes (f (z))2 , ia 3/2
⎡ ⎤
√ −2
d ia 3
= −πρ(9a4 K 2 ) ⎣ z + ⎦
dz 2 √
z=ia 3/2
√
= −9πa4 K 2 ρ(−2(ia 3)−3 )
18πa4 K 2 ρ
=− √ i.
3 3a3
The vertical component of the thrust is
√
B = 2 3πaρK 2 .
© 2012 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
www.librosysolucionarios.net
www.librosysolucionarios.net