MATH 218 Fall 2009

Assignment 1 Solutions

Part I: Problems from Problem Set 1 in the Course Notes

Question 1(iii) [5 marks]

Consider the family of functions given by

y = x + k e−x ,

where k is regarded as the constant of integration. Differentiating,

y 0 = 1 − k e−x .

Since k e−x = y − x, the required differential equation is

y 0 = 1 − y + x,

which is linear but not separable.

A sketch of the family is given in Figure 1. To help us make the sketch, we can make the
following observations.

• The line y = x is an exceptional solution (with k = 0).

• Since y 0 = 0 when y = x + 1, the line y = x + 1 is a horizontal isocline.

• Since y 0 < 0 when y > x + 1, solutions above the line y = x + 1 are decreasing. Similarly,
solutions below the line y = x + 1 are increasing.

• Since y 00 = k e−x , solutions are concave up when k > 0 and concave down when k < 0.

• All solutions asymptote to y = x as x → ∞.

• If k > 0 then lim = ∞. Similarly, if k < 0 then lim = −∞.

x→−∞ x→−∞

1
 y
y = x, k = 0

k>0 k increasing

k<0

y =x+1

Figure 1: Graph of the family of functions in Question 1(iii).

Question 2(viii) [5 marks]

d
Consider, where 0 = dx , the linear differential equation

y 0 = −2xy + 2x3 .

(a) An integrating factor is1 µZ ¶

2
µ(x) = exp 2x dx = ex .

So, write
³ 2 ´0 Z
x2 0 x2 3 x2 x 3 x2 x2 2
e y + 2x e y = 2x e =⇒ e y = 2x e =⇒ e y = 2x3 ex dx.

Then,
Z
−x2 2
y=e 2x3 ex dx. (1)

To evaluate the integral in (1), use the substitution u = x2 (which has du = 2xdx) to obtain
Z Z
3 x2
2x e dx = u eu du. (2)

Using integration by parts on the new integral in (2) with p = u and dq = eu du (which has
dp = du and q = eu ) gives
Z Z Z
3 x2 u u 2 ¡ ¢
2x e dx = u e du = u e − eu du = u eu − eu + C = ex x2 − 1 + C,

1
Note that exp(x) is another way of writing ex .

2
 y
y = x2 C>0 y = x2 − 1, C = 0

C<0 C increasing

Figure 2: Graph of the general solution in Question 2(viii).

where C ∈ R is arbitrary. It follows from (1) that the general solution is

2
y = x2 − 1 + C e−x .

(b) A sketch of the general solution is given in Figure 2. To help us make the sketch, we can make
the following observations.

• The solution y = x2 − 1 is an exceptional solution (with C = 0).

• Since y 0 = 0 when y = x2 or when x = 0, the curve y = x2 and the line x = 0 are horizontal
isoclines.

• Since y 0 > 0 when x < 0 and y > x2 , solutions are increasing in this region. Similarly,
solutions are increasing when x > 0 and y < x2 , solutions are decreasing when x < 0 and
y < x2 , and solutions are decreasing when x > 0 and y > x2 .

• All solutions asymptote to y = x2 − 1 as x → ±∞.

Question 2(ix) [5 marks]

Consider the separable differential equation

dy
= 1 − y2.
dx

3
(a) Observe that y = 1 and y = −1 are constant solutions. Assume that y 6= 1 and y 6= −1. Then,
we can write
Z Z
1
dy = 1 dx.
(1 − y) (1 + y)
Using partial fractions,
Z µ ¶ Z
1 1
+ dy = 2 dx.
1−y 1+y
Integrating and simplifying, we obtain
¯ ¯
¯1 + y ¯
¯
ln ¯ ¯ + C1 = 2x,
1 − y¯

where C1 ∈ R is arbitrary. Taking the exponential of both sides and removing the absolute
values, µ ¶
1+y
C = e2x ,
1−y
where C ∈ R\ {0} is arbitrary (note that C = ±eC1 ). Solving for y,

1 − C e−2x
y= .
1 + C e−2x

Since C = 0 accounts for the solution y = 1, the general solution is given by

1 − C e−2x
y= for C ∈ R and y = −1.
1 + C e−2x

To help us sketch the general solution, first recall that the hyperbolic tangent and cotangent
functions are given by

e−x − e−x 1 − e−2x e−x + e−x 1 + e−2x

tanh(x) = x −x
= and coth(x) = = .
e +e 1 + e−2x x
e −e −x 1 − e−2x

It follows that
¡ ¢ ¡ ¢
1 − e−2a e−2x 1 + e−2a e−2x
tanh(x + a) = and coth(x + a) = .
1 + (e−2a ) e−2x 1 − (e−2a ) e−2x

Comparing this with the general solution, we see that for |y| < 1 the solutions are horizontally-
translated hyperbolic tangents and for |y| > 1 the solutions are horizontally-translated hyper-
bolic cotangents. Note that there are other ways we could have sketched the general solution.

(b) A sketch of the general solution is given in Figure 3.

4
 y

y=1

y = −1

Figure 3: Graph of the general solution in Question 2(ix). The dashed lines are vertical asymptotes.

Question 4(iii) [4 marks]

Consider the initial value problem

2
y0 = y + x, y(1) = e,
x
d
where 0 = dx . Since the initial x-value is 1 and the differential equation is not defined at x = 0, we
will assume that x > 0. Now, an integrating factor is
µ Z ¶
2 1
µ(x) = exp − dx = 2 .
x x
Hence, write the differential equation as
· ¸0
y(x) 1
= .
x2 x
If we integrate with respect to x, we obtain
y(x)
= ln(x) + C.
x2
Setting x = 1 and y(x) = e tells us that C = e and thus the solution of the initial value problem is

y(x) = x2 [ln(x) + e] , x > 0.

5
Question 4(iv) [4 marks]

Consider the initial value problem

y 0 = ey−x , y(0) = ln(2) ,

d
where 0 = dx . The differential equation is separable and thus
Z Z
−y
e dy = e−x dx =⇒ −e−y + C = −e−x =⇒ e−y = e−x + C.

Set x = 0 and y = ln(2) to see that C = − 12 . Thus, the solution of the initial value problem is
¡ ¢
y = −ln e−x − 21 .

Since the argument of the natural logarithm must be strictly positive, the solution is valid for

x < ln(2) .

Question 10 [4 marks]

The velocity v of the projectile satisfies the initial value problem

dv gR2
v =− 2 , v(R) = vinit ,
dr r

where r is the distance from the centre of the planet, R is the radius of the planet, vinit > 0 is the
initial velocity, and g is the acceleration due to gravity.

(a) The differential equation is separable. Thus,

Z Z
2 1 gR2
v dv = −gR dr =⇒ 1
2 v2 = + C.
r2 r
1 2
Set r = R and v = vinit to get C = 2 vinit − gR. Thus, the velocity is given implicitly by

2gR2
v 2 = vinit
2
− 2gR + . (3)
r

(b) The escape velocity is the smallest initial velocity vesc such that v(r) > 0 for all r > 0. To find
this velocity, let vinit = vesc , r → ∞, and v(r) → 0 in (3) to obtain
p
vesc = 2gR.

6
Part II: Additional Problems

Question 1 [4 marks]

We want to solve the initial value problem

dy
+ ex y = 2 ex , y(0) = 3.
dx

An integrating factor is µZ ¶
µ(x) = exp e dx = exp(ex ) .
x

Now, write the differential equation as

d
(exp(ex ) y(x)) = 2 ex exp(ex ) .
dx

If we integrate with respect to x, we obtain

exp(ex ) y(x) = 2 exp(ex ) + C.

Set x = 0 and y = 3 (the initial condition) to obtain C = e. The solution is thus given by

y = exp(1 − ex ) + 2.

Question 2 [4 marks]

Consider the second-order linear differential equation

y 00 + y 0 = 1,

d
where 0 = dx . If we let u = y 0 then the differential equation becomes

u0 + u = 1 =⇒ ex u0 + ex u = ex (multiply by the integrating factor ex )

=⇒ (ex u)0 = ex (undo the Product Rule)

=⇒ ex u + A = ex (integrate with respect to x)

=⇒ u = y 0 = 1 − A e−x (solve for u)

=⇒ y = x + A e−x + B (integrate the directly-solvable DE).