chapter 8 mathematical modeling with di erential...
TRANSCRIPT
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 1 Page number 404 black
404
CHAPTER 8
Mathematical Modeling with DifferentialEquations
EXERCISE SET 8.1
1. y′ = 9x2ex3
= 3x2y and y(0) = 3 by inspection.
2. y′ = x3 − 2 sinx, y(0) = 3 by inspection.
3. (a) first order;dy
dx= c; (1 + x)
dy
dx= (1 + x)c = y
(b) second order; y′ = c1 cos t− c2 sin t, y′′ + y = −c1 sin t− c2 cos t+ (c1 sin t+ c2 cos t) = 0
4. (a) first order; 2dy
dx+ y = 2
(− c
2e−x/2 + 1
)+ ce−x/2 + x− 3 = x− 1
(b) second order; y′ = c1et − c2e−t, y′′ − y = c1e
t + c2e−t −
(c1e
t + c2e−t) = 0
5. False. It is a first-order equation, because it involves y and dy/dx, but not dny/dxn for n > 1.
6. True. y = −1/2 is a solution.
7. True. As mentioned in the marginal note after equation (2), the general solution of an n’th orderdifferential equation usually involves n arbitrary constants.
8. False. Every solution of the first order differential equation y′ = y has the form y = Aex = Aex+b
with b = 0.
9. (a) If y = e−2x then y′ = −2e−2x and y′′ = 4e−2x, soy′′ + y′ − 2y = 4e−2x + (−2e−2x)− 2e−2x = 0.
If y = ex then y′ = ex and y′′ = ex, so y′′ + y′ − 2y = ex + ex − 2ex = 0.(b) If y = c1e
−2x + c2ex then y′ = −2c1e−2x + c2e
x and y′′ = 4c1e−2x + c2ex, so
y′′ + y′ − 2y = (4c1e−2x + c2ex) + (−2c1e−2x + c2e
x)− 2(c1e−2x + c2ex) = 0.
10. (a) If y = e−2x then y′ = −2e−2x and y′′ = 4e−2x, soy′′ − y′ − 6y = 4e−2x − (−2e−2x)− 6e−2x = 0.
If y = e3x then y′ = 3e3x and y′′ = 9e3x, so y′′ − y′ − 6y = 9e3x − 3e3x − 6e3x = 0.(b) If y = c1e
−2x + c2e3x then y′ = −2c1e−2x + 3c2e3x and y′′ = 4c1e−2x + 9c2e3x, so
y′′ − y′ − 6y = (4c1e−2x + 9c2e3x)− (−2c1e−2x + 3c2e3x)− 6(c1e−2x + c2e3x) = 0.
11. (a) If y = e2x then y′ = 2e2x and y′′ = 4e2x, so y′′ − 4y′ + 4y = 4e2x − 4(2e2x) + 4e2x = 0.If y = xe2x then y′ = (2x+ 1)e2x and y′′ = (4x+ 4)e2x, soy′′ − 4y′ + 4y = (4x+ 4)e2x − 4(2x+ 1)e2x + 4xe2x = 0.
(b) If y = c1e2x + c2xe
2x then y′ = 2c1e2x + c2(2x + 1)e2x and y′′ = 4c1e2x + c2(4x + 4)e2x, soy′′ − 4y′ + 4y = (4c1e2x + c2(4x+ 4)e2x)− 4(2c1e2x + c2(2x+ 1)e2x) + 4(c1e2x + c2xe
2x) = 0.
12. (a) If y = e4x then y′ = 4e4x and y′′ = 16e4x, so y′′ − 8y′ + 16y = 16e4x − 8(4e4x) + 16e4x = 0.If y = xe4x then y′ = (4x+ 1)e4x and y′′ = (16x+ 8)e4x, soy′′ − 8y′ + 16y = (16x+ 8)e4x − 8(4x+ 1)e4x + 16xe4x = 0.
(b) If y = c1e4x + c2xe
4x then y′ = 4c1e4x + c2(4x+ 1)e4x and y′′ = 16c1e4x + c2(16x+ 8)e4x, soy′′−8y′+16y = (16c1e4x+c2(16x+8)e4x)−8(4c1e4x+c2(4x+1)e4x)+16(c1e4x+c2xe4x) = 0.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 2 Page number 405 black
Exercise Set 8.1 405
13. (a) If y = sin 2x then y′ = 2 cos 2x and y′′ = −4 sin 2x, so y′′ + 4y = −4 sin 2x+ 4 sin 2x = 0.
If y = cos 2x then y′ = −2 sin 2x and y′′ = −4 cos 2x, so y′′ + 4y = −4 cos 2x+ 4 cos 2x = 0.(b) If y = c1 sin 2x+ c2 cos 2x then y′ = 2c1 cos 2x− 2c2 sin 2x and y′′ = −4c1 sin 2x− 4c2 cos 2x,
so y′′ + 4y = (−4c1 sin 2x− 4c2 cos 2x) + 4(c1 sin 2x+ c2 cos 2x) = 0.
14. (a) If y = e−2x sin 3x then y′ = e−2x(−2 sin 3x+3 cos 3x) and y′′ = e−2x(−5 sin 3x−12 cos 3x), soy′′+ 4y′+ 13y = e−2x(−5 sin 3x−12 cos 3x) + 4e−2x(−2 sin 3x+ 3 cos 3x) + 13e−2x sin 3x = 0.
If y = e−2x cos 3x then y′ = e−2x(−3 sin 3x− 2 cos 3x) and y′′ = e−2x(12 sin 3x− 5 cos 3x), soy′′ + 4y′ + 13y = e−2x(12 sin 3x− 5 cos 3x) + 4e−2x(−3 sin 3x− 2 cos 3x) + 13e−2x cos 3x = 0.
(b) If y = e−2x(c1 sin 3x+ c2 cos 3x) then y′ = e−2x[−(2c1 + 3c2) sin 3x+ (3c1 − 2c2) cos 3x] andy′′ = e−2x[(−5c1 + 12c2) sin 3x− (12c1 + 5c2) cos 3x], soy′′ + 4y′ + 13y = e−2x[(−5c1 + 12c2) sin 3x− (12c1 + 5c2) cos 3x]
+ 4e−2x[−(2c1 + 3c2) sin 3x+ (3c1 − 2c2) cos 3x]+ 13e−2x(c1 sin 3x+ c2 cos 3x) = 0.
15. From Exercise 9, y = c1e−2x + c2e
x is a solution of the differential equation, withy′ = −2c1e−2x + c2e
x. Setting y(0) = −1 and y′(0) = −4 gives c1 + c2 = −1 and −2c1 + c2 = −4.So c1 = 1, c2 = −2, and y = e−2x − 2ex.
16. From Exercise 10, y = c1e−2x + c2e
3x is a solution of the differential equation, withy′ = −2c1e−2x + 3c2e3x. Setting y(0) = 1 and y′(0) = 8 gives c1 + c2 = 1 and −2c1 + 3c2 = 8.So c1 = −1, c2 = 2, and y = −e−2x + 2e3x.
17. From Exercise 11, y = c1e2x + c2xe
2x is a solution of the differential equation, withy′ = 2c1e2x + c2(2x+ 1)e2x. Setting y(0) = 2 and y′(0) = 2 gives c1 = 2 and 2c1 + c2 = 2,so c2 = −2 and y = 2e2x − 2xe2x.
18. From Exercise 12, y = c1e4x + c2xe
4x is a solution of the differential equation, withy′ = 4c1e4x + c2(4x+ 1)e4x. Setting y(0) = 1 and y′(0) = 1 gives c1 = 1 and 4c1 + c2 = 1,so c2 = −3 and y = e4x − 3xe4x.
19. From Exercise 13, y = c1 sin 2x+ c2 cos 2x is a solution of the differential equation, withy′ = 2c1 cos 2x− 2c2 sin 2x. Setting y(0) = 1 and y′(0) = 2 gives c2 = 1 and 2c1 = 2,so c1 = 1 and y = sin 2x+ cos 2x.
20. From Exercise 14, y = e−2x(c1 sin 3x+ c2 cos 3x) is a solution of the differential equation, withy′ = e−2x[−(2c1 + 3c2) sin 3x+ (3c1 − 2c2) cos 3x]. Setting y(0) = −1 and y′(0) = −1 givesc2 = −1 and 3c1 − 2c2 = −1, so c1 = −1 and y = −e−2x(sin 3x+ cos 3x).
21. y′ = 2−4x so y =∫
(2−4x) dx = −2x2+2x+C. Setting y(0) = 3 gives C = 3, so y = −2x2+2x+3.
22. (y′)′ = −6x so y′ =∫
(−6x) dx = −3x2 +C. Setting y′(0) = 2 gives C = 2, so y′ = −3x2 + 2 and
y =∫
(−3x2 + 2) dx = −x3 + 2x+D. Setting y(0) = 1 gives D = 1 so y = −x3 + 2x+ 1.
23. If the solution has an inverse function x(y) then, by equation (3) of Section 3.3,dx
dy=
1dy/dx
= y−2. So x =∫y−2 dy = −y−1+C. When x = 1, y = 2, so C =
32
and x =32−y−1.
Solving for y gives y =2
3− 2x. The solution is valid for x <
32
.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 3 Page number 406 black
406 Chapter 8
24. If the solution has an inverse function x(y) then, by equation (3) of Section 3.3,dx
dy=
1dy/dx
=1
1 + y2. So x =
∫dy
1 + y2= tan−1 y + C. When x = 0, y = 0,
so C = 0, x = tan−1 y, and y = tanx. The solution is valid for −π/2 < x < π/2.
25. By the product rule,d
dx(x2y) = x2y′ + 2xy = 0, so x2y = C and y = C/x2. Setting y(1) = 2
gives C = 2 so y = 2/x2. The solution is valid for x > 0.
26. By the product rule,d
dx(xy) = xy′ + y = ex, so xy =
∫ex dx = ex + C. Setting y(1) = 1 + e
gives C = 1, so xy = ex + 1 and y =ex + 1x
. The solution is valid for x > 0.
27. (a)dy
dt= ky2, y(0) = y0, k > 0 (b)
dy
dt= −ky2, y(0) = y0, k > 0
28. (a) Either y is always zero, or y is positive and increases at a rate proportional to the squareroot of y.
(b) Either y is always zero, or y is positive and decreases at a rate proportional to the cube of y,or y is negative and increases at a rate proportional to the cube of y.
29. (a)ds
dt=
12s (b)
d2s
dt2= 2
ds
dt
30. (a)dv
dt= −2v2 (b)
d2s
dt2= −2
(ds
dt
)2
31. (a) Since k > 0 and y > 0, equation (3) givesdy
dt= ky > 0, so y is increasing.
(b)d2y
dt2=
d
dt(ky) = k
dy
dt= k2y > 0, so y is concave upward.
32. (a) Both y = 0 and y = L satisfy equation (4).
(b) The rate of growth isdy
dt= k
(1− y
L
)y; we wish to find the value of y which maximizes
this. Sinced
dy
[k(
1− y
L
)y]
=k
L(L − 2y), which is positive for y < L/2 and negative for
y > L/2, the maximum growth rate occurs for y = L/2.
33. (a) Both y = 0 and y = L satisfy equation (6).
(b) The rate of growth isdy
dt= ky(L− y); we wish to find the value of y which maximizes this.
Sinced
dy[ky(L − y)] = k(L − 2y), which is positive for y < L/2 and negative for y > L/2,
the maximum growth rate occurs for y = L/2.
34. If T is constant thendT
dt= 0, so equation (7) gives T = Te. Hence T = Te is the unique constant
solution of (7).
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 4 Page number 407 black
Exercise Set 8.2 407
35. If x = c1 cos
(√k
mt
)+ c2 sin
(√k
mt
)then
dx
dt= c2
√k
mcos
(√k
mt
)− c1
√k
msin
(√k
mt
)
andd2x
dt2= −c1
k
mcos
(√k
mt
)− c2
k
msin
(√k
mt
)= − k
mx. So m
d2x
dt2= −kx; thus x satisfies
the differential equation for the vibrating string.
36. (a) From Exercise 35 we have x = c1 cos
(√k
mt
)+ c2 sin
(√k
mt
)and
x′ = c2
√k
mcos
(√k
mt
)− c1
√k
msin
(√k
mt
). Setting x(0) = x0 and x′(0) = 0 gives
c1 = x0 and c2
√k
m= 0, so c2 = 0 and x = x0 cos
(√k
mt
).
(b) From the discussion before Example 2 in Section 0.3, the amplitude is |x0|, the period is2π√k/m
= 2π√m
k, and the frequency is
√k/m
2π. The amplitude is the maximum displacement
of the mass from its rest position. The period is the length of time the mass takes to moveback and forth once. The frequency tells how often the mass moves back and forth in oneunit of time.
EXERCISE SET 8.2
1.1ydy =
1xdx, ln |y| = ln |x|+ C1, ln
∣∣∣yx
∣∣∣ = C1,y
x= ±eC1 = C, y = Cx
including C = 0 by inspection.
2.dy
1 + y2= 2x dx, tan−1 y = x2 + C, y = tan
(x2 + C
)
3.dy
1 + y= − x√
1 + x2dx, ln |1 + y| = −
√1 + x2 + C1, 1 + y = ±e−
√1+x2
eC1 = Ce−√
1+x2,
y = Ce−√
1+x2 − 1, C 6= 0
4. y dy =x3 dx
1 + x4,y2
2=
14
ln(1 + x4) + C1, 2y2 = ln(1 + x4) + C, y = ±√
[ln(1 + x4) + C]/2
5.2(1 + y2)
ydy = exdx, 2 ln |y|+ y2 = ex + C; by inspection, y = 0 is also a solution.
6.dy
y= −x dx, ln |y| = −x2/2 + C1, y = ±eC1e−x
2/2 = Ce−x2/2, including C = 0 by inspection.
7. eydy =sinx
cos2 xdx = secx tanx dx, ey = secx+ C, y = ln(secx+ C)
8.dy
1 + y2= (1 + x) dx, tan−1 y = x+
x2
2+ C, y = tan(x+ x2/2 + C)
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 5 Page number 408 black
408 Chapter 8
9.dy
y2 − y=
dx
sinx,
∫ [−1y
+1
y − 1
]dy =
∫cscx dx, ln
∣∣∣∣y − 1y
∣∣∣∣ = ln | cscx− cotx|+ C1,
y − 1y
= ±eC1(cscx− cotx) = C(cscx− cotx), y =1
1− C(cscx− cotx), C 6= 0;
by inspection, y = 0 is also a solution, as is y = 1.
10.1ydy = cosx dx, ln |y| = sinx+ C, y = C1e
sin x
11. (2y + cos y) dy = 3x2 dx, y2 + sin y = x3 + C, π2 + sinπ = C,C = π2, y2 + sin y = x3 + π2
12.dy
dx= (x+ 2)ey, e−ydy = (x+ 2)dx, −e−y =
12x2 + 2x+ C, −1 = C,
−e−y =12x2 + 2x− 1, e−y = −1
2x2 − 2x+ 1, y = − ln
(1− 2x− 1
2x2
)
13. 2(y − 1) dy = (2t+ 1) dt, y2 − 2y = t2 + t+ C, 1 + 2 = C,C = 3, y2 − 2y = t2 + t+ 3
14.dy
dxcosh2 x = y cosh 2x, y−1 dy =
cosh 2xcosh2 x
dx =2 cosh2 x− 1
cosh2 xdx = (2− sech2 x) dx,
ln y = 2x− tanhx+ C, y = eCe2x−tanh x, 3 = eCe2·0−tanh 0 = eC , y = 3e2x−tanh x
15. (a)dy
y=dx
2x, ln |y| = 1
2ln |x|+ C1,
|y| = C2|x|1/2, y2 = Cx;
by inspection y = 0 is also a solu-tion.
(b) 12 = C · 2, C = 1/2, y2 = x/2
–2 2
–2
2
x
yx = –0.5y2
x = –1.5y2x = y2 x = 2y2
x = 2.5y2
y = 0x = –3y2
16. (a) y dy = −x dx, y2
2= −x
2
2+ C1, y = ±
√C2 − x2
(b) y =√
25− x2
–3 3
–3
3
x
yy = √9 – x2
y = –√6.25 – x2
y = –√4 – x2
y = –√1 – x2
y = √2.25 – x2
y = √0.25 – x2
17.dy
y= − x dx
x2 + 4
ln |y| = −12
ln(x2 + 4) + C1
y =C√x2 + 4
1.5
–1
–2 2C = –1
C = 1C = 0C = 2
C = –2
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 6 Page number 409 black
Exercise Set 8.2 409
18. cos y dy = cosx dxsin y = sinx+ Cy = 2nπ + sin−1(sinx+ C) or y = (2n+ 1)π − sin−1(sinx+ C) for some integer n
For C = 0, the integral curves are lines of the form y =2nπ+x and y = (2n+1)π−x for integers n. These dividethe xy-plane into squares rotated 45◦ from the axes. ForC 6= 0, −2 < C < 2, each integral curve stays within eitherthe top half or the bottom half of one of these squares. Thefigure shows 5 such curves in the square x − 2π < y < x,π − x < y < 3π − x. From top to bottom, their equationsare y = π − sin−1(sinx+ 0.5), y = π − sin−1(sinx+ 1.5),y = sin−1(sinx + 1.7), y = sin−1(sinx + 1), and y =sin−1(sinx+ 0.3).
x
y
! 2!
!
19. (1− y2) dy = x2 dx,
y − y3
3=x3
3+ C1, x
3 + y3 − 3y = C
–2 2
–2
2
x
y
20.(
1y
+ y
)dy = dx, ln |y|+ y2
2= x+ C1,
yey2/2 = ±eC1ex = Cex including C = 0.
–5 5
–3
3
x
y
21. True. The equation can be rewritten as1
f(y)dy
dx= 1, which has the form (1).
22. False. The equation can be rewritten as1
g(y)dy
dx=
1h(x)
, which has the form (1).
23. True. After t minutes there will be 32 · (1/2)t grams left; when t = 5 there will be 32 · (1/2)5 = 1gram.
24. True. The population will quadruple in twice the doubling time.
25. Of the solutions y =1
2x2 − C, all pass through the point
(0,− 1
C
)and thus never through (0, 0).
A solution of the initial value problem with y(0) = 0 is (by inspection) y = 0. The method ofExample 1 fails in this case because it starts with a division by y2 = 0.
26. If y0 6= 0 then, proceeding as before, we get C = 2x2 − 1y, C = 2x2
0 −1y0
, and
y =1
2x2 − 2x20 + 1/y0
, which is defined for all x provided 2x2 is never equal to 2x20 − 1/y0; this
last condition will be satisfied if and only if 2x20−1/y0 < 0, or 0 < 2x2
0y0 < 1. If y0 = 0 then y = 0is, by inspection, also a solution for all real x.
27.dy
dx= xe−y, ey dy = x dx, ey =
x2
2+ C, x = 2 when y = 0 so 1 = 2 + C,C = −1, ey = x2/2− 1
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 7 Page number 410 black
410 Chapter 8
28.dy
dx=
3x2
2y, 2y dy = 3x2 dx, y2 = x3 + C, 1 = 1 + C,C = 0,
y2 = x3, y = x3/2 passes through (1, 1).
2
00 1.6
29. (a)dy
dt= 0.02y, y0 = 10,000 (b) y = 10,000et/50
(c) T =1
0.02ln 2 ≈ 34.657 h (d) 45,000 = 10,000et/50,
t = 50 ln45,00010,000
≈ 75.20 h
30. k =1T
ln 2 =120
ln 2
(a)dy
dt= ((ln 2)/20)y, y(0) = 1 (b) y(t) = et(ln 2)/20 = 2t/20
(c) y(120) = 26 = 64 (d) 1,000,000 = 2t/20,
t = 20ln 106
ln 2≈ 398.63 min
31. (a)dy
dt= −ky, y(0) = 5.0× 107; 3.83 = T =
1k
ln 2, so k =ln 23.83
≈ 0.1810
(b) y = 5.0× 107e−0.181t
(c) y(30) = 5.0× 107e−0.1810(30) ≈ 219,000
(d) y(t) = (0.1)y0 = y0e−kt, −kt = ln 0.1, t = − ln 0.1
0.1810= 12.72 days
32. (a) k =1T
ln 2 =1
140ln 2 ≈ 0.0050, so
dy
dt= −0.0050y, y0 = 10.
(b) y = 10e−0.0050t
(c) 10 weeks = 70 days so y = 10e−0.35 ≈ 7 mg.
(d) 0.3y0 = y0e−kt, t = − ln 0.3
0.0050≈ 240.8 days
33. 100e0.02t = 10,000, e0.02t = 100, t =1
0.02ln 100 ≈ 230 days
34. y = 10,000ekt, but y = 12,000 when t = 5 so 10,000e5k = 12,000, k =15
ln 1.2. y = 20,000 when
2 = ekt, t =ln 2k
= 5ln 2
ln 1.2≈ 19, in the year 2017.
35. y(t) = y0e−kt = 10.0e−kt, 3.5 = 10.0e−k(5), k = −1
5ln
3.510.0
≈ 0.2100, T =1k
ln 2 ≈ 3.30 days
36. y = y0e−kt, 0.7y0 = y0e
−5k, k = −15
ln 0.7 ≈ 0.07
(a) T =ln 2k≈ 9.90 yr
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 8 Page number 411 black
Exercise Set 8.2 411
(b) y(t) ≈ y0e−0.07t,y
y0≈ e−0.07t, so e−0.07t × 100 percent will remain.
38. (a) None; the half-life is independent of the initial amount.(b) kT = ln 2, so T is inversely proportional to k.
39. (a) T =ln 2k
; and ln 2 ≈ 0.6931. If k is measured in percent, k′ = 100k,
then T =ln 2k≈ 69.31
k′≈ 70k′
.
(b) 70 yr (c) 20 yr (d) 7%
40. Let y = y0ekt with y = y1 when t = t1 and y = 3y1 when t = t1 + T ; then y0e
kt1 = y1 (i) and
y0ek(t1+T ) = 3y1 (ii). Divide (ii) by (i) to get ekT = 3, T =
1k
ln 3.
41. From (19), y(t) = y0e−0.000121t. If 0.27 =
y(t)y0
= e−0.000121t then t = − ln 0.270.000121
≈ 10,820 yr, and
if 0.30 =y(t)y0
then t = − ln 0.300.000121
≈ 9950, or roughly between 9000 B.C. and 8000 B.C.
42. (a) 1
00 50000
(b) t = 1988 yields
y/y0 = e−0.000121(1988) ≈ 79%.
43. (a) Let T1 = 5730− 40 = 5690, k1 =ln 2T1≈ 0.00012182;T2 = 5730 + 40 = 5770, k2 ≈ 0.00012013.
With y/y0 = 0.92, 0.93, t1 = − 1k1
lny
y0= 684.5, 595.7; t2 = − 1
k2ln(y/y0) = 694.1, 604.1; in
1988 the shroud was at most 695 years old, which places its creation in or after the year 1293.
(b) Suppose T is the true half-life of carbon-14 and T1 = T (1 + r/100) is the false half-life. Then
with k =ln 2T, k1 =
ln 2T1
we have the formulae y(t) = y0e−kt, y1(t) = y0e
−k1t. At a certain
point in time a reading of the carbon-14 is taken resulting in a certain value y, which in thecase of the true formula is given by y = y(t) for some t, and in the case of the false formulais given by y = y1(t1) for some t1.
If the true formula is used then the time t since the beginning is given by t = −1k
lny
y0. If
the false formula is used we get a false value t1 = − 1k1
lny
y0; note that in both cases the
value y/y0 is the same. Thus t1/t = k/k1 = T1/T = 1 + r/100, so the percentage error inthe time to be measured is the same as the percentage error in the half-life.
44. If y = y0ekt and y = y1 = y0e
kt1 then y1/y0 = ekt1 , k =ln(y1/y0)
t1; if y = y0e
−kt and
y = y1 = y0e−kt1 then y1/y0 = e−kt1 , k = − ln(y1/y0)
t1.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 9 Page number 412 black
412 Chapter 8
45. (a) If y = y0ekt, then y1 = y0e
kt1 , y2 = y0ekt2 , divide: y2/y1 = ek(t2−t1), k =
1t2 − t1
ln(y2/y1),
T =ln 2k
=(t2 − t1) ln 2
ln(y2/y1). If y = y0e
−kt, then y1 = y0e−kt1 , y2 = y0e
−kt2 ,
y2/y1 = e−k(t2−t1), k = − 1t2 − t1
ln(y2/y1), T =ln 2k
= − (t2 − t1) ln 2ln(y2/y1)
.
In either case, T is positive, so T =∣∣∣∣ (t2 − t1) ln 2
ln(y2/y1)
∣∣∣∣.(b) In part (a) assume t2 = t1 + 1 and y2 = 1.25y1. Then T =
ln 2ln 1.25
≈ 3.1 h.
46. (a) In t years the interest will be compounded nt times at an interest rate of r/n each time. Thevalue at the end of 1 interval is P + (r/n)P = P (1 + r/n), at the end of 2 intervals it isP (1 + r/n) + (r/n)P (1 + r/n) = P (1 + r/n)2, and continuing in this fashion the value at theend of nt intervals is P (1 + r/n)nt.
(b) Let x = r/n, then n = r/x andlim
n→+∞P (1 + r/n)nt = lim
x→0+P (1 + x)rt/x = lim
x→0+P [(1 + x)1/x]rt = Pert.
(c) The rate of increase is dA/dt = rPert = rA.
47. (a) A = 1000e(0.08)(5) = 1000e0.4 ≈ $1, 491.82
(b) Pe(0.08)(10) = 10, 000, Pe0.8 = 10, 000, P = 10, 000e−0.8 ≈ $4, 493.29
(c) From (11), with k = r = 0.08, T = (ln 2)/0.08 ≈ 8.7 years.
48. Let r be the annual interest rate when compounded continuously and r1 the effective annualinterest rate. Then an amount P invested at the beginning of the year is worth Per = P (1 + r1)at the end of the year, and r1 = er − 1.
49. (a) Givendy
dt= k
(1− y
L
)y, separation of variables yields
(1y
+1
L− y
)dy = k dt so that
lny
L− y= ln y − ln(L − y) = kt + C. The initial condition gives C = ln
y0L− y0
so
lny
L− y= kt+ ln
y0L− y0
,y
L− y= ekt
y0L− y0
, and y(t) =y0L
y0 + (L− y0)e−kt.
(b) If y0 > 0 then y0 + (L − y0)e−kt = Le−kt + y0(1 − e−kt) > 0 for all t ≥ 0, so y(t) exists for
all such t. Since limt→+∞
e−kt = 0, limt→+∞
y(t) =y0L
y0 + (L− y0) · 0= L.
(Note that for y0 < 0 the solution “blows up” at t = −1k
ln−y0L− y0
, so limt→+∞
y(t) is undefined.)
50. The differential equation for the spread of disease can be rewritten asdy
dt= kL
(1− y
L
)y, which is
the logistic equation with k replaced by kL. Making this replacement in the solution from Exercise
49 gives y(t) =y0L
y0 + (L− y0)e−kLt.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 10 Page number 413 black
Exercise Set 8.2 413
51. (a) k = L = 1, y0 = 2
2
00 2
(b) k = L = y0 = 1
2
00 2
(c) k = y0 = 1, L = 2
2
00 5
(d) k = 1, y0 = 0.5, L = 10
00
10
10
52. y0 ≈ 400, L ≈ 1000; since the curve y =400,000
400 + 600e−ktpasses through the point (200, 600),
600 =400,000
400 + 600e−200k, 600e−200k =
8003, k =
1200
ln 2.25 ≈ 0.00405.
53. y0 ≈ 2, L ≈ 8; since the curve y =2 · 8
2 + 6e−ktpasses through the point (2, 4), 4 =
162 + 6e−2k
,
6e−2k = 2, k =12
ln 3 ≈ 0.5493.
54. This is the logistic equation (Equation (4) of Section 8.1) withk = 0.98, L = 5, and y0 = 1. From Exercise 49, the solution is
given by y(t) =5
1 + 4e−0.98t.
5
00 12
55. (a) y0 = 5 (b) L = 12 (c) k = 1
(d) L/2 = 6 =60
5 + 7e−t, 5 + 7e−t = 10, t = − ln(5/7) ≈ 0.3365
(e)dy
dt=
112y(12− y), y(0) = 5
56. (a) y0 = 1 (b) L = 1000 (c) k = 0.9
(d) 750 =1000
1 + 999e−0.9t, 3(1 + 999e−0.9t) = 4, t =
10.9
ln(3 · 999) ≈ 8.8949
(e)dy
dt=
0.91000
y(1000− y), y(0) = 1
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 11 Page number 414 black
414 Chapter 8
57. (a) Assume that y(t) students have had the flu t days after the break. If the disease spreadsas predicted by equation (6) of Section 8.1 and if nobody is immune, then Exercise 50 gives
y(t) =y0L
y0 + (L− y0)e−kLt, where y0 = 20 and L = 1000. So y(t) =
2000020 + 980e−1000kt
=
10001 + 49e−1000kt
. Using y(5) = 35 we find that k = − ln(193/343)5000
. Hence y =1000
1 + 49(193/343)t/5.
(b) ty(t)
0
20
1
22
2
25
3
28
4
31
5
35
6
39
7
44
8
49
9
54
10
61
11
67
12
75
13
83
14
93
(c)
3 6 9 12
25
50
75
100
t
y
58. If T0 < Ta thendT
dt= k(Ta − T ) where k > 0. If T0 > Ta then
dT
dt= −k(T − Ta) where k > 0;
both cases yield T (t) = Ta + (T0 − Ta)e−kt with k > 0.
59. (a) From Exercise 58 with T0 = 95 and Ta = 21, we have T = 21 + 74e−kt for some k > 0.
(b) 85 = T (1) = 21 + 74e−k, k = − ln6474
= − ln3237, T = 21 + 74et ln(32/37) = 21 + 74
(3237
)t,
T = 51 when3074
=(
3237
)t, t =
ln(30/74)ln(32/37)
≈ 6.22 min
60.dT
dt= k(70− T ), T (0) = 40; − ln(70− T ) = kt+ C, 70− T = e−kte−C , T = 40 when t = 0, so
30 = e−C , T = 70− 30e−kt; 52 = T (1) = 70− 30e−k, k = − ln70− 52
30= ln
53≈ 0.5,
T ≈ 70− 30e−0.5t
61. (a)dv
dt=
ck
m0 − kt− g, v = −c ln(m0 − kt)− gt+ C; v = 0 when t = 0 so 0 = −c lnm0 + C,
C = c lnm0, v = c lnm0 − c ln(m0 − kt)− gt = c lnm0
m0 − kt− gt.
(b) m0 − kt = 0.2m0 when t = 100 so
v = 2500 lnm0
0.2m0− 9.8(100) = 2500 ln 5− 980 ≈ 3044 m/s.
62. (a) By the chain rule,dv
dt=dv
dx
dx
dt=dv
dxv so m
dv
dt= mv
dv
dx.
(b)mv
kv2 +mgdv = −dx, m
2kln(kv2 +mg) = −x+ C; v = v0 when x = 0 so
C =m
2kln(kv2
0 +mg),m
2kln(kv2 +mg) = −x+
m
2kln(kv2
0 +mg), x =m
2klnkv2
0 +mg
kv2 +mg.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 12 Page number 415 black
Exercise Set 8.2 415
(c) x = xmax when v = 0 so
xmax=m
2klnkv2
0 +mg
mg=
3.56× 10−3
2(7.3× 10−6)ln
(7.3× 10−6)(988)2 + (3.56× 10−3)(9.8)(3.56× 10−3)(9.8)
≈1298 m
63. (a) A(h) = π(1)2 = π, πdh
dt= −0.025
√h,
π√hdh = −0.025dt, 2π
√h = −0.025t + C;h = 4 when
t = 0, so 4π = C, 2π√h = −0.025t+ 4π,
√h = 2− 0.025
2πt, h ≈ (2− 0.003979 t)2.
(b) h = 0 when t ≈ 2/0.003979 ≈ 502.6 s ≈ 8.4 min.
64. (a) A(h) = 6[2√
4− (h− 2)2]
= 12√
4h− h2,
12√
4h− h2dh
dt= −0.025
√h, 12
√4− h dh = −0.025dt,
−8(4− h)3/2 = −0.025t+ C; h = 4 when t = 0 so C = 0,
(4− h)3/2 = (0.025/8)t, 4− h = (0.025/8)2/3t2/3,
h ≈ 4− 0.021375t2/3 ft
h − 22
2√4 − (h − 2)2
h
(b) h = 0 when t =8
0.025(4− 0)3/2 = 2560 s ≈ 42.7 min
65.dv
dt= − 1
32v2,
1v2dv = − 1
32dt,−1
v= − 1
32t+ C; v = 128 when t = 0 so − 1
128= C,
−1v
= − 132t− 1
128, v =
1284t+ 1
cm/s. But v =dx
dtso
dx
dt=
1284t+ 1
, x = 32 ln(4t+ 1) + C1;
x = 0 when t = 0 so C1 = 0, x = 32 ln(4t+ 1) cm.
66.dv
dt= −0.02
√v,
1√vdv = −0.02dt, 2
√v = −0.02t+ C; v = 9 when t = 0 so 6 = C,
2√v = −0.02t+ 6, v = (3− 0.01t)2 cm/s. But v =
dx
dtso
dx
dt= (3− 0.01t)2,
x = −1003
(3− 0.01t)3 + C1; x = 0 when t = 0 so C1 = 900, x = 900− 1003
(3− 0.01t)3 cm.
67. Suppose that H(y) = G(x) + C. ThendH
dy
dy
dx= G′(x). But
dH
dy= h(y) and
dG
dx= g(x), hence
y(x) is a solution of (1).
68. Suppose that y(x) satisfies y(x0) = y0 and∫ y(x)
y0
h(r) dr =∫ x
x0
g(s) ds for all x in some interval
containing x0. Differentiate with respect to x; by the Fundamental Theorem of Calculus (Theorem
5.6.3) and the chain rule, we have h(y(x))dy
dx= g(x).
69. If h(y) = 0 then (1) implies that g(x) = 0, so h(y) dy = 0 = g(x) dx. Otherwise the slope of L
isdy
dx=g(x)h(y)
. Since (x1, y1) and (x2, y2) lie on L, we havey2 − y1x2 − x1
=g(x)h(y)
. So h(y)(y2 − y1) =
g(x)(x2 − x1); i.e. h(y) dy = g(x) dx.
70. It is true that the method may not give a formula for y as a function of x. But sometimes such aformula (in terms of familiar functions) does not exist. In such cases the method at least gives arelationship between x and y, from which we can find as good an approximation to y as we want.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 13 Page number 416 black
416 Chapter 8
71. Suppose that y = f(x) satisfies h(y)dy
dx= g(x). Integrating both sides of this with respect to x
gives∫h(y)
dy
dxdx =
∫g(x) dx, so
∫h(f(x))f ′(x) dx =
∫g(x) dx. By equation (2) of Section
5.3 with f replaced by h, g replaced by f , and F replaced by∫h(y) dy, the left side equals
F (f(x)) = F (y). Thus∫h(y) dy =
∫g(x) dx.
EXERCISE SET 8.3
1.
x
y
-2 -1 1 2
-2
-1
1
22.
x
y
1 2 3 4
1
2
3
43.
5
–1
2
x
y
y(0) = 1
y(0) = 2
y(0) = –1
4.dy
dx+ y = 1, µ = e
∫dx = ex,
d
dx[yex] = ex, yex = ex + C, y = 1 + Ce−x
(a) −1 = 1 + C,C = −2, y = 1− 2e−x
(b) 1 = 1 + C,C = 0, y = 1
(c) 2 = 1 + C,C = 1, y = 1 + e−x
5. limx→+∞
y = 1
6. (a) IV, since the slope is positive for x > 0 and negative for x < 0.
(b) VI, since the slope is positive for y > 0 and negative for y < 0.
(c) V, since the slope is always positive.
(d) II, since the slope changes sign when crossing the lines y = ±1.
(e) I, since the slope can be positive or negative in each quadrant but is not periodic.
(f) III, since the slope is periodic in both x and y.
7. y0 = 1, yn+1 = yn + 12y
1/3n
nxnyn
0
0
1
1
0.5
1.50
2
1
3 4
2
5
2.07
1.5
2.71 3.41
2.5
4.16
6
3
4.96
7
3.5
5.81
8
4
6.71
2 41 3
9
x
y
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 14 Page number 417 black
Exercise Set 8.3 417
8. y0 = 1, yn+1 = yn + (xn − y2n)/4
nxn
yn
0
0
1
1
0.25
0.75
2 3 4 5
0.50
0.67
0.75
0.68
1.00
0.75
1.25
0.86
6
1.50
0.99
7
1.75
1.12
8
2.00
1.24
0.5 1 1.5 2
0.5
1
1.5
2
x
y
9. y0 = 1, yn+1 = yn +12
cos yn
ntnyn
0
0
1
1
0.5
1.27
2
1
3 4
2
1.42
1.5
1.49 1.53
3
3
t
y
10. y0 = 0, yn+1 = yn + e−yn/10
ntnyn
0
0
0
1
0.1
0.10
2 3 4 5
0.2
0.19
0.3
0.27
0.4
0.35
0.5
0.42
6
0.6
0.49
7
0.7
0.55
8
0.8
0.60
9
0.9
0.66
10
1.0
0.71
1
1
t
y
11. h = 1/5, y0 = 1, yn+1 = yn +15
sin(πn/5)
ntnyn
0
0
0
1
0.2
0.12
0.4
0.31
2 3 4
0.6
0.50
0.8
0.62
1.0
0.62
5
12. False. This is only true if the slope at (x, y) does not depend on y.
13. True.dy
dx= exy > 0 for all x and y. So, for any integral curve, y is an increasing function of x.
14. True.d2y
dx2=
d
dx(ey) = ey
dy
dx= e2y > 0 for all y.
15. True. Every cubic polynomial has at least one real root. If p(y0) = 0 then y = y0 is an integralcurve that is a horizontal line.
16. (a) By inspection,dy
dx= e−x
2and y(0) = 0.
(b) yn+1 = yn + e−x2n/20 = yn + e−(n/20)2/20 and y20 = 0.7625. From a CAS, y(1) = 0.7468.
17. (b) y dy = −x dx, y2/2 = −x2/2 + C1, x2 + y2 = C; if y(0) = 1 then C = 1 so y(1/2) =
√3/2.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 15 Page number 418 black
418 Chapter 8
18. (a) Yes.(b) y = 0 is an integral curve of y′ = 2xy which is a horizontal line.
19. (b) The equation y′ = 1− y is separable:dy
1− y= dx, so
∫dy
1− y=∫
dx, − ln |1− y| = x+C.
Substituting x = 0 and y = −1 gives C = − ln 2, so x = ln 2−ln |1−y| = ln∣∣∣∣ 21− y
∣∣∣∣. Since the
integral curve stays below the line y = 1, we can drop the absolute value signs: x = ln2
1− yand y = 1− 2e−x. Solving y = 0 shows that the x-intercept is ln 2 ≈ 0.693.
20. (a) y0 = 1, yn+1 = yn + (√yn/2)∆x
∆x = 0.2 : yn+1 = yn +√yn/10; y5 ≈ 1.5489
∆x = 0.1 : yn+1 = yn +√yn/20; y10 ≈ 1.5556
∆x = 0.05 : yn+1 = yn +√yn/40; y20 ≈ 1.5590
(b)dy√y
=12dx, 2
√y = x/2 + C, 2 = C,
√y = x/4 + 1, y = (x/4 + 1)2,
y(1) = 25/16 = 1.5625
21. (a) The slope field does not vary with x, hence along a given parallel line all values are equalsince they only depend on the height y.
(b) As in part (a), the slope field does not vary with x; it is independent of x.
(c) From G(y)− x = C we obtaind
dx(G(y)− x) =
1f(y)
dy
dx− 1 =
d
dxC = 0, i.e.
dy
dx= f(y).
22. (a) Separate variables:dy√y
= dx, 2√y = x + C, y = (x/2 + C1)2 is a parabola that opens up,
and is therefore concave up.(b) A curve is concave up if its derivative is increasing, and y′ =
√y is increasing.
23. (a) By implicit differentiation, y3 + 3xy2 dy
dx− 2xy − x2 dy
dx= 0,
dy
dx=
2xy − y3
3xy2 − x2.
(b) If y(x) is an integral curve of the slope field in part (a), thend
dx{x[y(x)]3−x2y(x)} = [y(x)]3 + 3xy(x)2y′(x)− 2xy(x)−x2y′(x) = 0, so the integral curve
must be of the form x[y(x)]3 − x2y(x) = C.(c) x[y(x)]3 − x2y(x) = 2
24. (a) By implicit differentiation, ey + xeydy
dx+ ex
dy
dx+ yex = 0,
dy
dx= − e
y + yex
xey + ex
(b) If y(x) is an integral curve of the slope field in part (a), thend
dx{xey(x) + y(x)ex} = ey(x) + xy′(x)ey(x) + y′(x)ex + y(x)ex = 0 from part (a). Thus
xey(x) + y(x)ex = C.(c) Any integral curve y(x) of the slope field above satisfies xey(x) + y(x)ex = C; if it passes
through (1, 1) then e+ e = C, so xey(x) + y(x)ex = 2e defines the curve implicitly.
25. (a) For any n, yn is the value of the discrete approximation at the right endpoint, that, is anapproximation of y(1). By increasing the number of subdivisions of the interval [0, 1] onemight expect more accuracy, and hence in the limit y(1).
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 16 Page number 419 black
Exercise Set 8.4 419
(b) For a fixed value of n we have, for k = 1, 2, . . . , n, yk = yk−1 +yk−11n
=n+ 1n
yk−1. In partic-
ular yn =n+ 1n
yn−1 =(n+ 1n
)2
yn−2 = . . . =(n+ 1n
)ny0 =
(n+ 1n
)n. Consequently,
limn→+∞
yn = limn→+∞
(n+ 1n
)n= e, which is the (correct) value y = ex
∣∣∣∣x=1
.
26. Euler’s Method is repeated application of local linear approximation, each step dependent on theprevious step.
27. Visual inspection of the slope field may show where the integral curves are increasing, decreasing,concave up, or concave down. It may also help to identify asymptotes for the integral curves.
For example, in Exercise 3 we see that y = 1 is an integral curve that is an asymptote of all otherintegral curves. Those curves with y < 1 are increasing and concave down; those with y > 1 aredecreasing and concave up.
EXERCISE SET 8.4
1. µ = e∫4 dx = e4x, e4xy =
∫ex dx = ex + C, y = e−3x + Ce−4x
2. µ = e2∫x dx = ex
2,d
dx
[yex
2]
= xex2, yex
2=
12ex
2+ C, y =
12
+ Ce−x2
3. µ = e∫dx = ex, exy =
∫ex cos(ex)dx = sin(ex) + C, y = e−x sin(ex) + Ce−x
4.dy
dx+ 2y =
12
, µ = e∫
2dx = e2x, e2xy =∫
12e2xdx =
14e2x + C, y =
14
+ Ce−2x
5.dy
dx+
x
x2 + 1y = 0, µ = e
∫(x/(x2+1))dx = e
12 ln(x2+1) =
√x2 + 1,
d
dx
[y√x2 + 1
]= 0, y
√x2 + 1 = C, y =
C√x2 + 1
6.dy
dx+y = − 1
1− ex, µ = e
∫dx = ex, exy = −
∫ex
1− exdx = ln(1−ex)+C, y = e−x ln(1−ex)+Ce−x
7.dy
dx+
1xy = 1, µ = e
∫(1/x)dx = eln x = x,
d
dx[xy] = x, xy =
12x2 + C, y =
x
2+C
x,
2 = y(1) =12
+ C,C =32, y =
x
2+
32x
8. Divide by x to put the differential equation in the form (3):dy
dx−x−1y = x. We have p(x) = − 1
xand
q(x) = x, so∫p(x) dx = − ln |x|. So we may take e− ln |x| = |x−1| as an integrating factor. Since
integrating factors are only determined up to a constant factor, we may drop the absolute value
signs and simply take µ = x−1. We haved
dx(x−1y) = x−1 dy
dx− x−2y = x−1
(dy
dx− x−1y
)= 1, so
x−1y = x+ C and y = x2 + Cx. Since y(1) = −1, C = −2 and y = x2 − 2x.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 17 Page number 420 black
420 Chapter 8
9. µ = e−2∫x dx = e−x
2, e−x
2y =
∫2xe−x
2dx = −e−x
2+ C,
y = −1 + Cex2, 3 = −1 + C, C = 4, y = −1 + 4ex
2
10. µ = e∫dt = et, ety =
∫2et dt = 2et + C, y = 2 + Ce−t, 1 = 2 + C, C = −1, y = 2− e−t
11. False. If y1 and y2 both satisfydy
dx+p(x)y = q(x) then
d
dx(y1 +y2)+p(x)(y1 +y2) = 2q(x). Unless
q(x) = 0 for all x, y1 + y2 is not a solution of the original differential equation.
12. True. If y = C is a solution, then dy/dx = 0, so p(x)C = q(x).
13. True. The concentration in the tank will approach the concentration in the solution flowing intothe tank.
14. False. Although equation (18) implies that vτ =mg
c, where mg is the weight of the object, the
model does not specify how c depends on the object.
15.
-2 2
–10
10
x
y
y(0) = –1
y(–1) = 0y(1) = 1
16.dy
dx− 2y = −x, µ = e−2
∫dx = e−2x,
d
dx
[ye−2x
]= −xe−2x,
ye−2x =14
(2x+ 1)e−2x + C, y =14
(2x+ 1) + Ce2x
(a) 1 = 3/4 + Ce2, C = 1/(4e2), y =14
(2x+ 1) +14e2x−2
(b) −1 = 1/4 + C,C = −5/4, y =14
(2x+ 1)− 54e2x
(c) 0 = −1/4 + Ce−2, C = e2/4, y =14
(2x+ 1) +14e2x+2
17. It appears that limx→+∞
y =
{+∞, if y0 ≥ 1/4;
−∞, if y0 < 1/4.
To confirm this, we solve the equation using the method of integrating factors:dy
dx− 2y = −x, µ = e−2
∫dx = e−2x,
d
dx
[ye−2x
]= −xe−2x, ye−2x =
14
(2x + 1)e−2x + C,
y =14
(2x+ 1) + Ce2x. Setting y(0) = y0 gives C = y0 −14
, so y =14
(2x+ 1) +(y0 −
14
)e2x.
If y0 = 1/4, then y =14
(2x + 1) → +∞ as x → +∞. Otherwise, we rewrite the solution as
y = e2x(y0 −
14
+2x+ 14e2x
); since lim
x→+∞
2x+ 14e2x
= 0, we obtain the conjectured limit.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 18 Page number 421 black
Exercise Set 8.4 421
18. (a) y0 = 1, yn+1 = yn + (2yn − xn)(0.1) = (12yn − xn)/10
n 0 1 2 3 4 5xn 0 0.1 0.2 0.3 0.4 0.5yn 1 1.2 1.43 1.696 2.0052 2.36624
(b) Less. The integral curve appears to be concave up, so each yn is an underestimate of theactual value of y(xn).
(c) From Exercise 16, we have y =14
(2x + 1) + Ce2x for some constant C. Since y(0) = 1 we
find C =34
, so y =3e2x + 2x+ 1
4and y
(12
)=
3e+ 24≈ 2.53871.
19. (a) y0 = 1,yn+1 = yn + (xn + yn)(0.2) = (xn + 6yn)/5
nxn
yn
0
0
1
1
0.2
1.20
2 3 4 5
0.4
1.48
0.6
1.86
0.8
2.35
1.0
2.98
(b) y′ − y = x, µ = e−x,d
dx
[ye−x
]= xe−x,
ye−x = −(x+ 1)e−x + C, 1 = −1 + C,
C = 2, y = −(x+ 1) + 2ex
xn
y(xn)
abs. error
perc. error
0
1
0
0 3
0.2
1.24
0.04
6 9 11 13
0.4
1.58
0.10
0.6
2.04
0.19
0.8
2.65
0.30
1.0
3.44
0.46
(c)
0.2 0.4 0.6 0.8 1
3
x
y
20. h = 0.1, yn+1 = (xn + 11yn)/10
nxn
yn
0
0
1.00
1
0.1
1.10
2 3 4 5
0.2
1.22
0.3
1.36
0.4
1.53
0.5
1.72
6
0.6
1.94
7
0.7
2.20
8
0.8
2.49
9
0.9
2.82
10
1.0
3.19
With ∆x = 0.2, Euler’s method gives y(1) ≈ 2.98; with ∆x = 0.1, it gives y(1) ≈ 3.19. The truevalue is y(1) = 2e− 2 ≈ 3.44; so the absolute errors are approximately 0.46 and 0.25 respectively.
21.dy
dt= rate in − rate out, where y is the amount of salt at time t,
dy
dt= (4)(2)−
( y50
)(2) = 8− 1
25y, so
dy
dt+
125y = 8 and y(0) = 25.
µ = e∫(1/25)dt = et/25, et/25y =
∫8et/25dt = 200et/25 + C,
y = 200 + Ce−t/25, 25 = 200 + C, C = −175,
(a) y = 200− 175e−t/25 oz (b) when t = 25, y = 200− 175e−1 ≈ 136 oz
22.dy
dt= (5)(20)− y
200(20) = 100− 1
10y, so
dy
dt+
110y = 100 and y(0) = 0.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 19 Page number 422 black
422 Chapter 8
µ = e∫(1/10)dt = et/10, et/10y =
∫100et/10dt = 1000et/10 + C,
y = 1000 + Ce−t/10, 0 = 1000 + C, C = −1000;
(a) y = 1000− 1000e−t/10 lb (b) when t = 30, y = 1000− 1000e−3 ≈ 950 lb
23. The volume V of the (polluted) water is V (t) = 500 + (20− 10)t = 500 + 10t; if y(t) is the number
of pounds of particulate matter in the water, then y(0) = 50 anddy
dt= 0−10
y
V= − y
50 + t. Using
the method of integrating factors, we havedy
dt+
150 + t
y = 0; µ = e∫
dt50+t = 50+t;
d
dt[(50+t)y] = 0,
(50 + t)y = C, 2500 = 50y(0) = C, y(t) = 2500/(50 + t). (The differential equation may also besolved by separation of variables.)
The tank reaches the point of overflowing when V = 500 + 10t = 1000, t = 50 min, soy = 2500/(50 + 50) = 25 lb.
24. The volume of the lake (in gallons) is V = 264πr2h = 264π(15)23 = 178,200π gals. Let y(t)
denote the number of pounds of mercury salts at time t; thendy
dt= 0 − 103 y
V= − y
178.2πlb/h
and y0 = 10−5V = 1.782π lb;dy
y= − dt
178.2π, ln y = − t
178.2π + C1, y = Ce−t/(178.2π), and
C = y(0) = y0 = 1.782π, y = 1.782πe−t/(178.2π) lb of mercury salts.
t 1 2 3 4 5 6 7 8 9 10 11 12y(t) 5.588 5.578 5.568 5.558 5.549 5.539 5.529 5.519 5.509 5.499 5.489 5.480
We assumed that the mercury is always distributed uniformly throughout the lake, and doesn’tsettle to the bottom.
25. (a)dv
dt+
c
mv = −g, µ = e(c/m)
∫dt = ect/m,
d
dt
[vect/m
]= −gect/m, vect/m = −gm
cect/m + C,
v = −gmc
+Ce−ct/m, but v0 = v(0) = −gmc
+C,C = v0+gm
c, v = −gm
c+(v0 +
gm
c
)e−ct/m
(b) Replacemg
cwith vτ and −ct/m with −gt/vτ in (16).
(c) From part (b), s(t) = C − vτ t− (v0 + vτ )vτge−gt/vτ ; s0 = s(0) = C − (v0 + vτ )
vτg
,
C = s0 + (v0 + vτ )vτg
, s(t) = s0 − vτ t+vτg
(v0 + vτ )(
1− e−gt/vτ)
26. (a) Let t denote time elapsed in seconds after the moment of the drop. From Exercise 25(b),while the parachute is closedv(t) = e−gt/vτ (v0 + vτ ) − vτ = e−32t/120 (0 + 120) − 120 = 120
(e−4t/15 − 1
)and thus
v(25) = 120(e−20/3 − 1
)≈ −119.85, so the skydiver is falling at a speed of 119.85 ft/s
when the parachute opens. From Exercise 25(c), s(t) = s0 − 120t +12032
120(
1− e−4t/15),
s(25) = 10000− 120 · 25 + 450(
1− e−20/3)≈ 7449.43 ft.
(b) If t denotes time elapsed after the parachute opens, then, by Exercise 25(c),
s(t) ≈ 7449.43−24t+2432
(−119.85+24)(1−e−32t/24). A calculating utility finds that s(t) = 0
for t ≈ 307.4 s, so the skydiver is in the air for about 25 + 307.4 = 332.4 s.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 20 Page number 423 black
Review Exercises, Chapter 8 423
27.dI
dt+R
LI =
V (t)L
, µ = e(R/L)∫dt = eRt/L,
d
dt(eRt/LI) =
V (t)L
eRt/L,
IeRt/L = I(0) +1L
∫ t
0
V (u)eRu/Ldu, I(t) = I(0)e−Rt/L +1Le−Rt/L
∫ t
0
V (u)eRu/Ldu.
(a) I(t) =15e−2t
∫ t
0
20e2udu = 2e−2te2u]t0
= 2(1− e−2t
)A.
(b) limt→+∞
I(t) = 2 A
28. From Exercise 27 and Endpaper Table #42,
I(t) = 15e−2t +13e−2t
∫ t
0
3e2u sinu du = 15e−2t + e−2t e2u
5(2 sinu− cosu)
]t0
= 15e−2t +15
(2 sin t− cos t) +15e−2t.
29. (a) Let y =1µ
[H(x) + C] where µ = eP (x),dP
dx= p(x),
d
dxH(x) = µq, and C is an arbitrary
constant. Thendy
dx+ p(x)y =
1µH ′(x)− µ′
µ2[H(x) + C] + p(x)y = q − p
µ[H(x) + C] + p(x)y = q
(b) Given the initial value problem, let C = µ(x0)y0−H(x0). Then y =1µ
[H(x)+C] is a solution
of the initial value problem with y(x0) = y0. This shows that the initial value problem hasa solution.
To show uniqueness, suppose u(x) also satisfies (3) together with u(x0) = y0. Following the
arguments in the text we arrive at u(x) =1µ
[H(x) + C] for some constant C. The initial
condition requires C = µ(x0)y0 −H(x0), and thus u(x) is identical with y(x).
30. (a) y = x and y = −x are both solutions of the given initial value problem.
(b)∫y dy = −
∫x dx, y2 = −x2 + C; but y(0) = 0, so C = 0. Thus y2 = −x2, which is
impossible.
REVIEW EXERCISES, CHAPTER 8
2. (a) yes (b) yes (c) no (d) yes
3.dy
1 + y2= x2 dx, tan−1 y =
13x3 + C, y = tan
(13x3 + C
)
4.1
tan ydy =
3secx
dx,cos ysin y
dy = 3 cosx dx, ln | sin y| = 3 sinx+ C1,
sin y = ±e3 sin x+C1 = ±eC1e3 sin x = Ce3 sin x, C 6= 0,
y = sin−1(Ce3 sin x
), as is y = 0 by inspection
5.(
1y
+ y
)dy = exdx, ln |y|+ y2/2 = ex + C; by inspection, y = 0 is also a solution
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 21 Page number 424 black
424 Chapter 8
6.dy
y2 + 1= dx, tan−1 y = x+ C, π/4 = C; y = tan(x+ π/4)
7.(
1y5
+1y
)dy =
dx
x, −1
4y−4 + ln |y| = ln |x|+ C; −1
4= C, y−4 + 4 ln(x/y) = 1
8.dy
y2= 4 sec2 2x dx, −1
y= 2 tan 2x+ C, −1 = 2 tan
(2π
8
)+ C = 2 tan
π
4+ C = 2 + C, C = −3,
y =1
3− 2 tan 2x
9.dy
y2= −2x dx, −1
y= −x2 + C, −1 = C, y = 1/(x2 + 1)
–1 1
1
x
y
10. 2y dy = dx, y2 = x + C; if y(0) = 1 then C = 1, y2 = x + 1, y =√x+ 1; if y(0) = −1 then
C = 1, y2 = x+ 1, y = −√x+ 1.
–1 1
–1
1
x
y
–1 1
–1
1
x
y
11.
x
y
1 2 3 4
1
2
3
4
12.dy
y=
18x dx, ln |y| = 1
16x2 + C, y = C1e
x2/16
13. y0 = 1, yn+1 = yn +√yn/2
n 0 1 2 3 4 5 6 7 8xn 0 0.5 1 1.5 2 2.5 3 3.5 4yn 1 1.50 2.11 2.84 3.68 4.64 5.72 6.91 8.23
2 41 3
9
x
y
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 22 Page number 425 black
Review Exercises, Chapter 8 425
14. y0 = 1, yn+1 = yn +12
sin yn
n 0 1 2 3 4tn 0 0.5 1 1.5 2yn 1 1.42 1.92 2.39 2.73
3
3
t
y
15. h = 1/5, y0 = 1, yn+1 = yn +15
cos(2πn/5)
n 0 1 2 3 4 5tn 0 0.2 0.4 0.6 0.8 1.0yn 1.00 1.20 1.26 1.10 0.94 1.00
16. yn+1 = yn + 0.1(1 + 5tn − yn), y0 = 5
n 0 1 2 3 4 5 6 7 8 9 10tn 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2yn 5.00 5.10 5.24 5.42 5.62 5.86 6.13 6.41 6.72 7.05 7.39
17. From formula (19) of Section 8.2, y(t) = y0e−0.000121t, so 0.785y0 = y0e
−0.000121t, t = − ln 0.785/0.000121 ≈2000.6 yr
18. (a)d
dty(t) = 0.01y, y(0) = 5000
(b) y(t) = 5000e0.01t
(c) 2 = e0.01t, t = 100 ln 2 ≈ 69.31 h(d) 30,000 = 5000e0.01t, t = 100 ln 6 ≈ 179.18 h
19. µ = e∫3 dx = e3x, e3xy =
∫ex dx = ex + C, y = e−2x + Ce−3x
20.dy
dx+ y =
11 + ex
, µ = e∫dx = ex, exy =
∫ex
1 + exdx = ln(1 + ex) + C, y = e−x ln(1 + ex) + Ce−x
21. µ = e−∫x dx = e−x
2/2, e−x2/2y =
∫xe−x
2/2dx = −e−x2/2 + C,
y = −1 + Cex2/2, 3 = −1 + C, C = 4, y = −1 + 4ex
2/2
22.dy
dx+
2xy = 4x, µ = e
∫(2/x)dx = x2,
d
dx
[yx2]
= 4x3, yx2 = x4 + C, y = x2 + Cx−2,
2 = y(1) = 1 + C, C = 1, y = x2 + 1/x2
23. By inspection, the left side of the equation isd
dx(y coshx), so
d
dx(y coshx) = cosh2 x =
12
(1 + cosh 2x) and y coshx =12x+
14
sinh 2x+ C =12
(x+ sinhx coshx) + C.
When x = 0, y = 2 so 2 = C, and y = 2 sech x+12
(x sech x+ sinhx).
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 23 Page number 426 black
426 Chapter 8
24. (a) µ = e−∫dx = e−x,
d
dx
[ye−x
]= xe−x sin 3x,
ye−x =∫xe−x sin 3x dx =
(− 3
10x− 3
50
)e−x cos 3x+
(− 1
10x+
225
)e−x sin 3x+ C;
1 = y(0) = − 350
+ C, C =5350, y =
(− 3
10x− 3
50
)cos 3x+
(− 1
10x+
225
)sin 3x+
5350ex
(c)
–10 –2
–2
4
x
y
25. (a) linear (b) both (c) separable (d) neither
26.dy
dx− 4xy = x
(a) IF: e∫(−4x)dx = e−2x2
,d
dx[ye−2x2
] = xe−2x2, ye−2x2
=∫xe−2x2
dx = −14e−2x2
+ C,
y = −14
+ Ce2x2
(b)dy
dx= 4xy + x,
dy
4y + 1= x dx,
14
ln(4y + 1) =12x2 + C, ln(4y + 1) = 2x2 + C1,
4y + 1 = C2e2x2, y = 1
4 (C2e2x2 − 1) = C3e
2x2 − 14 , same as in part (a)
27. Assume the tank contains y(t) oz of salt at time t. Then y0 = 0 and for 0 < t < 15,dy
dt= 5 · 10− y
100010 = (50− y/100) oz/min, with solution y = 5000 +Ce−t/100. But y(0) = 0 so
C = −5000, y = 5000(1− e−t/100) for 0 ≤ t ≤ 15, and y(15) = 5000(1− e−0.15). For 15 < t < 30,dy
dt= 0− y
10005, y = C1e
−t/200, C1e−0.075 = y(15) = 5000(1−e−0.15), C1 = 5000(e0.075−e−0.075),
y = 5000(e0.075 − e−0.075)e−t/200, y(30) = 5000(e0.075 − e−0.075)e−0.15 ≈ 646.14 oz.
28. (a) Assume the air contains y(t) ft3 of carbon monoxide at time t. Then y0 = 0 and for
t > 0,dy
dt= 0.04(0.1)− y
1200(0.1) = 1/250− y/12000,
d
dt
[yet/12000
]=
1250
et/12000,
yet/12000 = 48et/12000 + C, y(0) = 0, C = −48; y = 48(1− e−t/12000). Thus the percentage
of carbon monoxide is P =y
1200100 = 4(1− e−t/12000) percent.
(b) 0.012 = 4(1− e−t/12000), t = 36.05 min
MAKING CONNECTIONS, CHAPTER 8
1. (a) u(x) = q − p y(x) sodu
dx= −pdy
dx= −p(q − py(x)) = (−p)u(x).
If p < 0 then −p > 0 so u(x) grows exponentially.If p > 0 then −p < 0 so u(x) decays exponentially.
November 10, 2008 19:17 ”ISM ET chapter 8” Sheet number 24 Page number 427 black
Making Connections, Chapter 8 427
(b) From (a), u(x) = 4 − 2y(x) satisfiesdu
dx= −2u(x), so equation (14) of Section 8.2 gives
u(x) = u0e−2x for some constant u0. Since u(0) = 4 − 2y(0) = 6, we have u(x) = 6e−2x;
hence y(x) = 2− 3e−2x.
2. (a)du
dx=
d
dx(ax+ b y(x)+ c) = a+ b
dy
dx= a+ b f(ax+ by+ c) = a+ b f(u), so
1a+ b f(u)
du
dx= 1.
(b) From (a) with a = b = 1, c = 0, f(t) = 1/t, we have1
1 + 1/udu
dx= 1, where u = x + y. So
u
u+ 1du = dx,
∫u
u+ 1du =
∫dx, u− ln |u+ 1| = x+C, x+ y− ln |x+ y+ 1| = x+C, and
y − ln |x+ y + 1| = C.
3. (a)du
dx=
d
dx
(yx
)=xdy
dx− y
x2=x f(yx
)− y
x2. Since y = ux,
du
dx=x f(u)− ux
x2=f(u)− u
x
and1
f(u)− udu
dx=
1x
.
(b)dy
dx=x− yx+ y
=1− y/x1 + y/x
has the form given in (a), with f(t) =1− t1 + t
. So1
1− u1 + u
− u
du
dx=
1x
,
1 + u
1− 2u− u2du =
dx
x,∫
1 + u
1− 2u− u2du =
∫dx
x, −1
2ln |1 − 2u − u2| = ln |x| + C1, and
|1 − 2u − u2| = e−2C1x−2. Hence 1 − 2u − u2 = Cx−2 where C is either e−2C1 or −e−2C1 .
Substituting u =y
xgives 1− 2y
x− y2
x2= Cx−2, and x2 − 2xy − y2 = C.
4. (a)du
dx= (1 − n)y−n
dy
dx= (1 − n)y−n[q(x)yn − p(x)y] = (1 − n)q(x) − (1 − n)p(x)y1−n =
(1− n)q(x)− (1− n)p(x)u. Hencedu
dx+ (1− n)p(x)u = (1− n)q(x).
(b)dy
dx− 1xy = −2y2, so this has the form given in (a) with p(x) = −1/x, q(x) = −2, and n = 2.
So u = y−1 satisfiesdu
dx+
1xu = 2. An integrating factor is given by µ = e
∫dx/x = eln x = x.
Sod
dx(xu) = x
du
dx+u = 2x, xu = x2+C, u = x+Cx−1, and y = u−1 =
1x+ Cx−1
=x
x2 + C.
Since y(1) =12
, C = 1 and y =x
x2 + 1.