Math221: HW# 2 Solutions: Andy Royston October 12, 2005
Math221: HW# 2 Solutions: Andy Royston October 12, 2005
Math221: HW# 2 Solutions: Andy Royston October 12, 2005
Andy Royston
October 12, 2005
Z t Z t
d2 x 0 0
02
dt = ge−kt dt0 ⇒
0 dt
¯t 0
dx dx g ¯ g
(t) − (0) = − e−kt ¯¯ = (1 − e−kt ). (1)
dt dt k 0 k
dx dx
Since the object starts from rest, dt
(0) = v(0) = 0. Now integrate v(t) = dt
(t) to get the distance
traveled:
Z t t Z
g 0 0 0
v(t )dt = (1 − e−kt )dt0 ⇒
0 k
¯t 0
g 1 0 ¯ g g
d(t) ≡ x(t) − x(0) = (t0 + e−kt )¯¯ = t − 2 (1 − e−kt ). (2)
k k 0 k k
To look at the behavior for small t, Taylor expand and keep only the lowest order terms in t:
µ ¶
g g 1 22 3
d(t) = t + 2 1 − (1 − kt + k t − O(t ))
k k 2
g g 1 1
= t − 2 (kt − k 2 t2 + O(t3 )) = gt2 − O(t3 ), (3)
k k 2 2
or
1
d(t) ≈ gt2 , t small. (4)
2
So the motion is approximately free-fall for small t. For large t, we have e−kt → 0. Therefore, the
dominating term in the velocity is the constant:
g
v(t) → , for t large. (5)
k
1
8.1.7 We have F = dp/dt, where F is a constant. Integrate both sides from 0 to t:
Z t
dp 0 mv(t)
Ft = 0
dt = p(t) − p(0) = p − p(0). (6)
0 dt 1 − v(t)2 /c2
We could write out p(0) in terms of v(0), but it’s just a constant anyway. Now lets solve the above
for v(t).
v2 v2 2 2
(1 − )(F t + p0 )2 = m2 v 2 ⇒ (m c + (F tp0 )2 ) = (F t + p0 )2 ,
c2 c2
v(t) F t + p0
⇒ =p . (7)
c m2 c2 + (F t + p0 )2
(I am using the common notation q(0) = q0 for any quantity q). Let us look at this result in the
limit t → ∞. Rewrite this as
v(t) F t + p0 1
= q = p . (8)
c m2 c2
(F t + p0 ) 1 + (F t+p0 )2 1 + m 2 c2 /(F t + p )2
0
It is clear that this expression goes to 1 in the limit since the fraction in the denominator goes to
0. Therefore
v(t)
lim = 1, or lim v(t) = c. (9)
t→∞ c t→∞
Now we are asked to find the distance traveled, d(t), if the electron starts from rest. If the
electron starts from rest, then v0 = 0 ⇒ p0 = 0. Hence,
Z t Z t Z t
0 0 F t0 dt0 c 2F 2 t0 dt0
d(t) = v(t )dt = c √ = √
0 0 m2 c2 + F 2 t02 2F 0 m2 c2 + F 2 t02
¯t
c√ 2 2 ¯ c √
= m c + F 2 t02 ¯¯ = ( m2 c2 + F 2 t2 − mc). (10)
F 0 F
8.2.2
p √ xdx ydy
x 1 − y 2 dx + y 1 − x2 dy = 0 ⇒ √ = −p . (11)
1−x 2 1 − y2
Now integrate each side. Here, it is better to do indefinite integrals and add a constant to one side.
This is because we are not given initial data (such as y(x = 0) = y0 ); rather, we are given the value
at a pair of points. Then,
Z √ Z p
xdx ydy
√ = − 1 − x2 = − p +c= 1 − y2 + c
1 − x2 1 − y2
√ p
⇒ 1 − x2 + 1 − y 2 = c. (12)
2
1
1
0.5
0.5
-0.5
-0.5
-1
-1
Figure 1: Problem 8.2.2. The slope field and some solution curves. The values of c chosen for the
curves ranged between 1 and 2.
See Figure 1 for a plot of the slope field and some of the solution curves. Note that the result only
gives real curves for |x| ≤ 1 and |y| ≤ 1 This implies 0 ≤ c ≤ 2.
We are told that the solution passes through (x0 , y0 ) = (1/2, 1/2). Plugging these values into
the above equation:
p √
2 3/4 = 3 = c. (13)
Hence, the solution is
√ p √
1 − x2 + 1 − y2 = 3. (14)
One could solve this for y to get y as a function of x, but this is unnecessary.
8.2.4
Z Z
2 dx ydy
(1 + y )dx + xydy = 0 ⇒ =− +c
x 1 + y2
−1
⇒ ln x = ln (1 + y 2 ) + c. (15)
2
See Figure 2. We have (x0 , y0 ) = (5, 0), so
1
ln 5 = − ln 1 + c ⇒ c = ln 5. (16)
2
Hence,
3
3 3
2 2
1 1
-3 -2 -1 1 2 3 -3 -2 -1 1 2 3
-1 -1
-2 -2
-3 -3
Figure 2: Problem 8.2.4. Here the values of c were chosen between −3 and 3.
1 5
ln x = − ln (1 + y 2 ) + ln 5 ⇒ x= p . (17)
2 1 + y2
8.2.8
Z Z
dy 2 dy
+ 2xy = 0 ⇒ = − 2xdx + c
dx y2
1 1
⇒ − = −x2 + c, or y = 2 . (18)
y x −c
4
2
2
1
1
-2 -1 1 2
-2 -1 1 2
-1 -1
-2 -2
Figure 3: Problem 8.2.8. The c values were again taken from −3 to 3. For c = −3, . . . , 0, each
curve has two poles, where it blows up to ±∞. c = 1, 2, 3 correspond to the bounded curves.
and we are given the initial conditions N (t = 0) = 0. Let us first solve this by separation of
variables.
Z Z √
dN 1
√ = dt + c ⇒ 2 N = t + c ⇒ N (t) = (t + c)2 . (22)
N 4
The solution which satisfies N (0) = 0 is c = 0. Hence, we obtain N (t) = t2 /4. But this solution
is physically unreasonable. If the bacteria grows at a rate proportional to the square root of the
amount, and the amount is initially zero, it will not grow. The correct solution is N (t) = 0, ∀t.
Note that this is a solution to the differential
√ equation, though it is not in the family we obtained.
We essentially lost it when we divided by N .
8.2.21 We have
Z Z
dN dN
= kN ⇒ = kdt + c ⇒ ln N = kt + c
dt N
⇒ N (t) = Aekt . (23)
5
8.2.26
Z Z
dv dv
m = mg − kv ⇒ dv = (g − kv/m)dt ⇒ = dt + c
dt g − kv/m
−m −k
ln (g − kv/m) = t + c ⇒ ln (g − kv/m) = (t + c) ⇒ g − kv/m = Ae−kt/m
k m
mg m −kt/m
v(t) = − Ae . (25)
k k
Now, we are told that the object falls from rest, so
mg m
0 = v(0) = − A=0 ⇒ A = g. (26)
k k
Hence, the solution is
mg
v(t) = (1 − e−kt/m ). (27)
k
The limiting value of v as t → ∞ is the terminal speed. Since e−kt/m → 0, we have vterm = mg/k.
However, assuming a terminal speed exists, we can also find it directly from the differential equation.
If v → vterm , a constant, then we must have dv/dt → 0. Plugging this into the differential equation,
a Suppose a 100kg person parachutes from a plane. A reasonable value of vterm might be
1m/s. Then, in SI units,
(100)(9.8)
1= ⇒ k = 9, 800kg/s. (29)
k
b Note that we may write the solution for v(t) as v(t) = vterm (1 − e−kt/m ). We wish to solve
for the time such that v(t) = 0.99vterm :
m m
0.99vterm = vterm (1 − e−kt/m ) ⇒ e−kt/m = 0.01 ⇒ t=− ln 0.01 ≈ 4.605 . (30)
k k
8.2.31 We need to find the equation for the tangents of the orthogonal trajectories. Start by
taking the differential of the given equation, which represents circles centered at the origin,
dy x
2xdx + 2ydy = 0 ⇒ =− . (31)
dx y
This gives the tangents of the slope field. Note that the constant dropped out. Since the constant
tells us which curve (circle) we are looking at, this means that our expression above for the tangent
lines is correct for any curve, ie. for any value of (x, y). The tangents of the orthogonal trajectories
are the negative reciprocals:
6
1.5
0.5
-1
-1.5
Figure 4: Problem 8.2.31. Circles are original curves; straight line are orthogonal trajectories.
4
-3 -2 -1 1 2 3
-1
-2
Z Z
dyo.t. yo.t. dyo.t. dx
= ⇒ = +c ⇒ ln yo.t. = ln x + c
dx x yo.t. x
⇒ yo.t. = ec x = Ax. (32)
This makes sense. The orthogonal trajectories of circles centered about the origin are straight lines
passing through the origin. See Figure 4.
dy x
2(y − 1)dy = 2xdx ⇒ = . (33)
dx y−1
7
Again, the constant drops out. Hence,
Z Z
dyo.t 1 − yo.t. dyo.t. dx
= ⇒ = + c ⇒ − ln (1 − yc.t. ) = ln x + c
dx x 1 − yo.t. x
⇒ 1 − yo.t. = e− ln x e−c = A/x ⇒ yo.t. = 1 − A/x. (34)
See Figure 5.
d 1 1
ln (tan (x/2 + π/4)) = (sec (x/2 + π/4))2 ( )
dx tan (x/2 + π/4) 2
1 1 1
= = = . (37)
2 sin (x/2 + π/4) cos (x/2 + π/4) sin (x + π/2) cos x
Hence,
Z
dx
I= = ln (tan (x/2 + π/4)). (38)
cos x
R
For the general solution, we still need to compute the integral QeI :
Z Z
I
dxQe = dx cos x tan (x/2 + π/4). (39)
8
Thus,
Z Z Z
sin u
dx cos x tan (x/2 + π/4) = 4 du sin u cos u =4 du(sin u)2
cos u
Z
1
=2 du(1 − cos 2u) = 2(u −
sin (2u)) = 2u − sin (2u) = x + π/2 − sin (x + π/2)
2
= x − cos x + π/2. (41)
R
The general solution is given by y = e−I ( QeI + c). Hence,
µ ¶
y(x) = exp (− ln (tan (x/2 + π/4))) x − cos x + π/2 + c
µ ¶
0
= cot (x/2 + π/4) x − cos x + c . (42)
When I plug the original differential equation into Mathematica, I get the solution
cos (x/2 + π/4) cos (x/2) cos (π/4) − sin (x/2) sin (π/4)
cot (x/2 + π/4) = = . (44)
sin (x/2 + π/4) sin (x/2) cos (π/4) + cos (x/2) sin (π/4)
√
Using sin (π/4) = cos (π/4) (= 1/ 2), these factors cancel in the numerator and denominator.
Hence,
8.3.13 Note the hint at the bottom of page 403; we will solve for x as a function of y.
dx
dx + (x − ey )dy = 0 ⇒ = −(x − ey ) ⇒ x0 + x = e y . (46)
dy
R
So P (y) = 1 and therefore, I = dyP = y. Meanwhile
Z Z
1
dyQ(y)e = dye2y = e2y .
I
(47)
2
Therefore,
1 1
x(y) = e−y ( e2y + c) = ey + ce−y . (48)
2 2
This is exactly the form of the solution that Mathematica gives me.
9
8.3.20 See example 2 of the text, pages 402 − 403. Suppose polonium decays (into something
else) with decay constant λ3 . Polonium is being created at a rate equal to the rate at which radon is
decaying, but it is being destroyed at it’s decay rate. Therefore, the differential equation governing
the quantity of polonium is
dN3
= λ2 N2 − λ3 N3 ⇒ N30 + λ3 N3 = λ2 N2 (t), (49)
dt
where the right hand side of this equation has been solved for in the example and is given by
λ1 λ2 N0 −λ1 t
λ2 N2 (t) = (e − e−λ2 t ) ≡ A(e−λ1 t − e−λ2 t ). (50)
λ2 − λ1
This solution assumes λ1 6= λ2 . We will further assume λ3 6= λ1 and λ3 6= λ2 . Also, I’ve introduced
the constant A to avoid cluttered formulae in what follows. Returning to the diff. eq. for N3 (t),
we see that P = λ3 is a constant, so I = λ3 t. Then
Z Z
I
dtQ(t)e = A dt(e(λ3 −λ1 )t − e(λ3 −λ2 )t )
N0 λ 1 λ 2 N0 λ 1 λ 2 N0 λ1 λ2
N3 (t) = e−λ1 t − e−λ2 t + e−λ3 t . (54)
(λ2 − λ1 )(λ3 − λ1 ) (λ2 − λ1 )(λ3 − λ2 ) (λ3 − λ1 )(λ3 − λ2 )
If we define
λij = λi − λj , (55)
then this can be written in a more symmetric fashion:
−N0 λ1 λ2
N3 (t) = (λ23 e−λ1 t + λ31 e−λ2 t + λ12 e−λ3 t ). (56)
λ12 λ23 λ31
10
4
3
2
1
-4 -3 -2 -1 1 2 3 4
-1
-2
-3
-4
8.3.22 Take the differential to get an equation for the tangent lines of the slope field:
dy 1
dx = dy + cey dy = (1 + cey )dy ⇒= . (57)
dx 1 + cey
Note that the constant c does not drop out in this case and we need to eliminate it. The constant
c takes different values on different curves in the family, but we want a equation for the tangent
lines that is valid at any point (x, y). This can be done by solving the original equation for the
constant c in terms of x, y.
dyo.t. 0
= yo.t. − x ⇒ yo.t. − yo.t. = −x. (60)
dx
This is a first order linear equation. We have P = −1, so I = −x. Then
Z Z Z
I −x −x
dxQ(x)e = − dxxe = xe − dxe−x = xe−x + e−x
11
8.4.2 We have
1 √
y 0 + y = 2x2/3 y. (63)
x
√
This is Bernoulli equation with n = 1/2. Let z = y. Then,
dz dz dy 1 1
z0 = = = √ y0 = y0
dx dy dx 2 y 2z
⇒ y = 2zz 0 .
0
(64)
1 x3
z(x) = √ ( + c), (67)
x 3
and thus
1 x3
y(x) = z(x)2 = ( + c)2 . (68)
x 3
8.5.1 Using the notation D = d/dx, this equation has the form
12