Part I Problems

Problem 1: Find the general solution by separation of variables:

dy
= 2 y,
dx

y (0) = 0

Problem 2: Find the general solution by separation of variables:

dy
( y 1)2
=
dx
( x + 1)2

Problem 3: The rate of change of a certain population is proportional to the square root
of its size. Model this situation with a differential equation.
Problem 4:
The rate of change of the velocity of an object is proportional to the square of the velocity.
Model this situation with a differential equation.
Problem 5:
In a population of xed size S, the rate of change of the number N of persons who have
heard a rumor is proportional to the number of those who have not yet heard it. Model
this situation with a differential equation.
Problem 6: The amount of a certain medicine in the bloodstream decays exponentially
with a half-life of 5 hours. In order to keep a patient safe during a one-hour procedure,
there needs to be at least 50 mg of medicine per kg of body weight. How much medicine
should be administered to a 60kg patient at the start of the procedure?
Problem 7: Early one morning it starts to snow. At 7AM a snowplow sets off to clear the
road. By 8AM, it has gone 2 miles. It takes an additional 2 hours for the plow to go another
2 miles. Let t = 0 when it begins to snow, let x denote the distance traveled by the plow at
time t. Assuming the snowplow clears snow at a constant rate in cubic meters/hour:
a) Find the DE modeling the value of x.
b) When did it start snowing?

Part I Problems

OCW 18.03SC

Problem 8: A tank holds 100 liters of water which contains 25 grams of salt initially. Pure
water then ows into the tank, and salt water ows out of the tank, both at 5 liters/minute.
The mixture is kept uniform at all times by stirring.
a) Write down the DE with IC for this situation.
b) How long will it take until only 1 gram of salt remains in the tank?

Part I Problems and Solutions

Problem 1: Find the general solution by separation of variables:
dy
= 2 y,
dx
Solution:

dy
= dx
2y

y (0) = 0

dy
=
2y

ln |2 y| = x + c |2 y| = Cex
y = 2 Ce x

dx

(with C = ec )

(with C any number)

IC: y(0) = 2 C = 0 C = 2 y = 2(1 e x )

Problem 2: Find the general solution by separation of variables:
dy
( y 1)2
=
dx
( x + 1)2
Solution:

dy
dx
=

( y 1)2
( x + 1)2

dy
=
( y 1)2

dx

( x + 1)2

1
1
=
+c
y1
x+1

Extra: solve for y as a function of x:

Answer: y = 1 +

x +1
1 c ( x +1)

Problem 3: The rate of change of a certain population is proportional to the square root of its size.
Model this situation with a differential equation.
Solution: Let P = P(t) be the size of the population as a function of time t. Then

dP
=k P
dt
where k > 0 is the constant of proportionality.

Part I Problems and Solutions

OCW 18.03SC

Problem 4:
The rate of change of the velocity of an object is proportional to the square of the velocity. Model
this situation with a differential equation.
Solution:

dv
= kv2
dt

with k > 0, constant.

Problem 5:
In a population of xed size S, the rate of change of the number N of persons who have heard a
rumor is proportional to the number of those who have not yet heard it. Model this situation with a
differential equation.
Solution:

dN
= k(S N )
dt

with k > 0, constant.

Problem 6: The amount of a certain medicine in the bloodstream decays exponentially with a
half-life of 5 hours. In order to keep a patient safe during a one-hour procedure, there needs to be
at least 50 mg of medicine per kg of body weight. How much medicine should be administered to a
60kg patient at the start of the procedure?
Solution: Let x (t) be the amount of the medicine in mg present in the bloodstream at time
t in hours.
1
Given information: x (t) = x0 ekt with k = ln52 hr
, since half-life is given as 5 hours.
Since x (t) = x0 ekt is decreasing, x (t) x (1) for t 1 hrs. So the patient will be (just)
safe if x (1) = x0 ek1 = 50 60 = 3000 mg = 3 g where k = ln52 . Thus, x0 = x (1) ek =
e(ln 2)/5 3000 3446 mg (or about 3.446 g).
Problem 7: Early one morning it starts to snow. At 7AM a snowplow sets off to clear the road.
By 8AM, it has gone 2 miles. It takes an additional 2 hours for the plow to go another 2 miles. Let
t = 0 when it begins to snow, let x denote the distance traveled by the plow at time t. Assuming
the snowplow clears snow at a constant rate in cubic meters/hour:
a) Find the DE modeling the value of x.
b) When did it start snowing?

Part I Problems and Solutions

OCW 18.03SC

Solution: a) One approach: let k1 be the rate (in height/hour) of snowfall and k2 be the
rate of snow clearance.
The height of snow is k1 t x k1 t k2 t
This is then

dx
dt

x
t

k2
k1 t

= kt , where k is a constant.

b) Solving by separation of variables, x = k ln t + C.

Let t = t1 at 7AM, so t = t1 + 1 at 8AM and t = t1 + 3 at 10AM.

2 miles between 7 and 8AM 2 = x (t1 + 1) x (t1 ) = k ln ((t1 + 1) /t1 )

4 miles between 7 and 10AM 4 = x (t1 + 3) x (t1 ) = k ln ((t1 + 3) /t1 )

2
Thus, ln t1t+1 3 = 2 ln t1t+1 1 t1t+1 3 = t1t+1 1
After a little algebra, t1 = 1, so the snow started at 6AM

Problem 8: A tank holds 100 liters of water which contains 25 grams of salt initially. Pure water
then ows into the tank, and salt water ows out of the tank, both at 5 liters/minute. The mixture
is kept uniform at all times by stirring.
a) Write down the DE with IC for this situation.
b) How long will it take until only 1 gram of salt remains in the tank?
Solution:
minutes.

Let x = x (t) be the amount of salt in the tank in grams, with t the time in

a) DE: net rate of change of salt.

x

0 5 100
.
DE:

dx
dt

= .05x,

dx
dt

= salt rate in - salt rate out =

IC x (0) = 25

b) Solution to DE: x (t) = Ce.05t . IC x (0) = 25 = c, so x (t) = 25e.05t .

x (t) = 1 when 25e.05t = 1 t =

ln 25
.05

64.38 min

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part II Problems
Problem 1: [Natural growth, separable equations] In recitation a population model was
studied in which the natural growth rate of the population of oryx was a constant k > 0,
so that for small time intervals t the population change x (t + t) x (t) is well approx
imated by kx (t)t. (You also studied the effect of hunting them, but in this problem we
will leave that aside.) Measure time in years and the population in kilo-oryx (ko).
A mysterious virus infects the oryxes of the Tana River area in Kenya, which causes the
growth rate to decrease as time goes on according to the formula k (t) = k0 /( a + t)2 for
t 0, where a and k0 are certain positive constants.
(a) What are the units of the constant a in a + t, and of the constant k0 ?
(b) Write down the differential equation modeling this situation.
(c) Write down the general solution to your differential equation. Dont restrict yourself to
the values of t and of x that are relevant to the oryx problem; take care of all values of these

dx
variables. Points to be careful about: use absolute values in
= ln | x | + c correctly, and
x
dont forget about any lost solutions.
(d) Now suppose that at t = 0 there is a positive population x0 of oryx. Does the progres
sive decline in growth rate cause the population stabilize for large time, or does it grow
without bound? If it does stabilize, what is the limiting population as t ?

Part II Problems and Solutions

Problem 1: [Natural growth, separable equations] In recitation a population model was studied
in which the natural growth rate of the population of oryx was a constant k > 0, so that for small
time intervals t the population change x (t + t) x (t) is well approximated by kx (t)t. (You
also studied the effect of hunting them, but in this problem we will leave that aside.) Measure time
in years and the population in kilo-oryx (ko).
A mysterious virus infects the oryxes of the Tana River area in Kenya, which causes the growth rate
to decrease as time goes on according to the formula k(t) = k0 /( a + t)2 for t 0, where a and k0
are certain positive constants.
(a) What are the units of the constant a in a + t, and of the constant k0 ?
(b) Write down the differential equation modeling this situation.
(c) Write down the general solution to your differential equation. Dont restrict yourself to the
values of t and of x that are relevant to the oryx problem; take care of all values of these variables.

dx
Points to be careful about: use absolute values in
= ln | x | + c correctly, and dont forget
x
about any lost solutions.
(d) Now suppose that at t = 0 there is a positive population x0 of oryx. Does the progressive decline
in growth rate cause the population stabilize for large time, or does it grow without bound? If it
does stabilize, what is the limiting population as t ?
Solution: (a) The growth rate k (t) has units years1 (so that k(t) x (t)t has the same units
as x (t)). The variable t has units years, so the a added to it must have the same units, and
k0 must have units years in order for the units of the fraction to work out.

(b) x (t + t) x (t) + k (t) x (t)t, so x = k0 x/( a + t)2 .

(c) Separate: dx/x = k0 ( a + t)2 dt. Integrate: ln | x | + c1 = k0 ( a + t)1 + c2 . Amal
gamate constants and exponentiate: | x | = ec ek0 /(a+t) . Eliminate the absolute value:
x = Cek0 /(a+t) , where C = ec . Reintroduce the solution we lost by dividing by x in
the rst step: allow C = 0. So the general solution is x = Cek0 /(a+t) . (Note that the
exponent k0 /( a + t) is dimensionless, as an exponent must be.)
(d) When t gets very large, the exponent gets very near to zero, so there is a nite limiting
population: x = C. Thus x (t) = x ek0 /(a+t) . Take t = 0 in the solution: x0 = x ek0 /a ,
or x = ek0 /a x0 .

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part I Problems
For each of the following ODEs, draw a direction eld by using about ve isoclines; the
picture should be square, using the intervals between -2 and 2 on both axes. Then sketch
in some integral curves, using the information provided by the direction eld. Finally, do
whatever else is asked.
y

Problem 1: y = x . Solve the equation exactly and compare your integral curves with
the correct ones.
Problem 2: y = 2x + y. Find a solution whose graph is also an isocline, and verify this
fact analytically (i.e., by calculation, and not from a picture).
Problem 3:
y = x+1 y .
Use the interval 3 to 3 on both axes; draw in the integral
curves that pass respectively through (0, 0), (1, 1), (0, 2). Will these curves cross the
line y = x 1? Explain by using the Intersection Principle.

Part I Problems and Solutions

For each of the following ODEs, draw a direction eld by using about ve isoclines; the picture
should be square, using the intervals between -2 and 2 on both axes. Then sketch in some integral
curves, using the information provided by the direction eld. Finally, do whatever else is asked.
Problem 1:
correct ones.

y = x .
y

Solve the equation exactly and compare your integral curves with the
y

Solution: y = x . Isoclines: x = m y = mx.

Solutions:

dy
y

c
= dx
x ln y = ln x + c y = x .

Problem 2: y = 2x + y. Find a solution whose graph is also an isocline, and verify this fact
analytically (i.e., by calculation, and not from a picture).
Solution: y = 2x + y has isoclines 2x + y = m y = 2x + m. Isocline y = 2x + m is
also a solution if y = 2 (from solution) and also y = 2x + y (from DE) y = 2x 2,
that is, the isocline with m = 2.

Problem 3: y = x+1 y . Use the interval 3 to 3 on both axes; draw in the integral curves that
pass respectively through (0, 0), (1, 1), (0, 2). Will these curves cross the line y = x 1?
Explain by using the Intersection Principle.

Part I Problems and Solutions

Solution: Isoclines x + y =

1
m

OCW 18.03SC

y = x + m1 .

y = x 1 is an integral curve (or solution) so other solutions cannot cross it.

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part II Problems
Problem 1: [Direction elds, isoclines] In this problem you will study solutions of the
differential equation
dy
= y2 x .
dx
Solutions of this equation do not admit expressions in terms of the standard functions of
calculus, but we can study them anyway using the direction eld.
(a) Draw a large pair of axes and mark off units from 4 to +4 on both. Sketch the direction
eld given by our equation. Do this by rst sketching the isoclines for slopes m = 1,
m = 0, m = 1, and m = 2. On this same graph, sketch, as best you can, a couple of
solutions, using just the information given by these four isoclines.
Having done this, you will continue to investigate this equation using one of the Mathlets. So invoke http://math.mit.edu/mathlets/mathlets in a web browser and select
Isoclines from the menu. (To run the applet from this window, click the little black box
with a white triangle inside.) Play around with this applet for a little while. The Mathlets have many features in common, and once you get used to one it will be quicker to
learn how to operate the next one. Clicking on Help pops up a window with a brief
description of the applets functionalities.
Select from the pull-down menu our differential equation y = y2 x. Move the m slider
to m = 2 and release it; the m = 2 isocline is drawn. Do the same for m = 0, m = 1,
and m = 2. Compare with your sketches. Then depress the mousekey over the graphing
window and drag it around; you see a variety of solutions. How do they compare with
what you drew earlier?
(b) A separatrix is a curve such that above it solutions behave (as x increases) in one way,
while below it solutions behave (as x increases) in quite a different way. There is a sepa
ratrix for this equation such that solutions above it grow without bound (as x increases)
while solutions below it eventually decrease (as x increases). Use the applet to nd its
graph, and submit a sketch of your result.
(c) Suppose y( x ) is a solution to this differential equation whose graph is tangent to the
m = 1 isocline: it touches the m = 1 isocline at a point ( a, b), and the two curves have
the same slope at that point. Find this point on the applet, and then calculuate the values
of a and b.
(d) Now suppose that y( x ) is a solution to the equation for which y( a) < b, where ( a, b) is
the point you found in (c). Whathappens to it as x ? I claim that its graph is asymp
totic to the graph of f ( x ) = x. Explain why this is so. For large x, is y( x ) > f ( x ),
y( x ) < f ( x ), or does the answer depend on the value of y( a)?

Part II Problems

OCW 18.03SC

The following observations will be useful in justifying your claims. Please explain as
clearly as you can why each is true.
(i) The graph of y( x ) cant cross the m = 1 isocline at a point ( x, y) with x > a.
(ii) If c > a and y(c) lies above the nullcline, then the graph of y( x ) continues to lie above
the nullcline for all x > c.
(iii) If c > a and y(c) lies below the nullcline, then the graph of y( x ) will cross the nullcline
for some x > c.
(e) Suppose a solution y( x ) has a critical point at (c, d)that is, y (c) = 0 and y(c) = d.
What can you say about the relationship between c and d? The applet can be very helpful
here, but verify your answer.
(f) It appears from the applet that all critical points are local maxima. Is that true?

Part II Problems and Solutions

Problem 1: [Direction elds, isoclines] In this problem you will study solutions of the differential
equation
dy
= y2 x .
dx
Solutions of this equation do not admit expressions in terms of the standard functions of calculus,
but we can study them anyway using the direction eld.
(a) Draw a large pair of axes and mark off units from 4 to +4 on both. Sketch the direction eld
given by our equation. Do this by rst sketching the isoclines for slopes m = 1, m = 0, m = 1,
and m = 2. On this same graph, sketch, as best you can, a couple of solutions, using just the
information given by these four isoclines.
Having done this, you will continue to investigate this equation using one of the Mathlets. So in
voke http://math.mit.edu/mathlets/mathlets in a web browser and select Isoclines from
the menu. (To run the applet from this window, click the little black box with a white triangle in
side.) Play around with this applet for a little while. The Mathlets have many features in common,
and once you get used to one it will be quicker to learn how to operate the next one. Clicking on
Help pops up a window with a brief description of the applets functionalities.
Select from the pull-down menu our differential equation y = y2 x. Move the m slider to
m = 2 and release it; the m = 2 isocline is drawn. Do the same for m = 0, m = 1, and m = 2.
Compare with your sketches. Then depress the mousekey over the graphing window and drag it
around; you see a variety of solutions. How do they compare with what you drew earlier?
(b) A separatrix is a curve such that above it solutions behave (as x increases) in one way, while
below it solutions behave (as x increases) in quite a different way. There is a separatrix for this
equation such that solutions above it grow without bound (as x increases) while solutions below it
eventually decrease (as x increases). Use the applet to nd its graph, and submit a sketch of your
result.
(c) Suppose y( x ) is a solution to this differential equation whose graph is tangent to the m = 1
isocline: it touches the m = 1 isocline at a point ( a, b), and the two curves have the same slope at
that point. Find this point on the applet, and then calculuate the values of a and b.
(d) Now suppose that y( x ) is a solution to the equation for which y( a) < b, where ( a, b) is the
point you found in
(c). What happens to it as x ? I claim that its graph is asymptotic to the
graph of f ( x ) = x. Explain why this is so. For large x, is y( x ) > f ( x ), y( x ) < f ( x ), or does

the answer depend on the value of y( a)?

The following observations will be useful in justifying your claims. Please explain as clearly as you

can why each is true.

(i) The graph of y( x ) cant cross the m = 1 isocline at a point ( x, y) with x > a.

Part II Problems and Solutions

OCW 18.03SC

(ii) If c > a and y(c) lies above the nullcline, then the graph of y( x ) continues to lie above the
nullcline for all x > c.
(iii) If c > a and y(c) lies below the nullcline, then the graph of y( x ) will cross the nullcline for
some x > c.
(e) Suppose a solution y( x ) has a critical point at (c, d)that is, y (c) = 0 and y(c) = d. What
can you say about the relationship between c and d? The applet can be very helpful here, but verify
your answer.
(f) It appears from the applet that all critical points are local maxima. Is that true?
Solution: (a) (The picture lacks the m = 1 isocline, sorry.)

(b) The separatrix is the elbow curve that other integral curves snuggle up to to the left
and separate from as things move to the right. [Here are some facts about solutions to this
differential equation, some of which are contained in later parts of this problem: Every
solution has a starting time (if I think of x as time) before which it does not existevery
integral curve is asymptotic to a vertical straight line to its left. Every solution above
the separatrix has an ending time after which it does not existit is asymptotic to a
different vertical straight line to its right. Every solution below the separatrix
survives

forever, and becomes asymptotic (from above) to the graph of y = x. These are also
exactly the solutions with a critical pointeach has one local maximum. The separatrix
itself is a solution, a very special one: it is
the only solution which lives forever
but has no
maximum and is not asymptotic to y = x. Instead it is asymptotic to y = x.]
(c) The point of tangency occurs where the tangent line to the m = 1 isocline has slope
1. The m = 1 isocline is the locus of x = y2 + 1. Differentiate implicitly (using the
chain rule) 1 = 2yy , so if y = 1 then y = 1/2. The equation then gives x = 5/4, so
( a, b) = (5/4, 1/2).
2

Part II Problems and Solutions

OCW 18.03SC

(d) (i) Follow the solution from the x = a line. Suppose that it does cross the m = 1
isocline, and suppose that the rst crossing is at a point ( x, y). Since it started below the
m = 1 isocline, it must be crossing it from below. This means that the slope of the solution
must be greater than the slope of the isocline at the crossing point. The slope of the isocline
at the crossing point is greater than 1, so the slope of the solution at the crossing point
must be greater than 1 too. But when a solution crosses the 1 isocline, it must cross it
with slope exactly 1. So it cant cross it at all.
(ii) Suppose that in fact the graph of y( x ) does cross the nullcline. The slope of the nullcline
is negative; the rst crossing must be from above; so at the crossing the slope of the solution
must be even more negative; but when it crosses the nullcline it must have slope zero.
(iii) Finally I claim that if ( a, y( a)) is below the nullcline, then the integral curve must
eventually cross the nullcline. Until it does, it is in the region where the direction eld has
positive slopes. So it is increasing. But the lower branch of the nullcline is is falling; so the
two curves must eventually intersect.
One can also argue that if the solution never crosses the nullcline then, after rising at the
start it must eventually start to fall, since the nullcline does. So it must reach a maximum
somewhere in between. But all maxima of solutions occur along the nullcline.
(e) Critical points of solutions of y = y2 x occur when y = 0, that is, along the null-cline:
so if there is one at (c, d) then c = d2 .
(f) Yes. Here are two arguments: All solutions are increasing for small values of x. From
what we can see, the solutions
with critical points are the ones which become asymototic

to the graph of y = x, which decreases. Also they seem to have just one critical point,
which must therefore be a maximum. More complete argument: Compute (as in (c)) y =
2yy 1. So at a critical point, where y = 0, we see that y = 1. This means maximum.

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part I Problems
Problem 1: Use the Euler method and the step size .1 on the IVP
y = x + y2 , y(0) = 1, to calculate an approximate value for the solution y( x ) when
x = .1, .2, .3. (Make a table.) Is your answer for y(.3) too high or low?

Part I Problems and Solutions

Problem 1: Use the Euler method and the step size .1 on the IVP

y = x + y2 , y(0) = 1, to calculate an approximate value for the solution y( x ) when x = .1, .2, .3.

(Make a table.) Is your answer for y(.3) too high or low?

Solution: Euler method formula: yn+1 = yn + h f ( xn , yn ).

xn
0
.1
.2
.3

yn
1
1.1
1.23
1.403

f ( xn , yn )
1
1.31
1.72

h f ( xn , yn )
.1
.131
.172

h = .1
f ( x, y) =
x + y2

Isoclines x + y2 = C (parabolas).
Solution curve through (0, 1) is convex (concave up), so Euler method gives too low a
result.

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part II Problems
Problem 1: [Eulers method] (a) Write y for the solution to y = 2x with y(0) = 0. What is
y(1)? What is the Euler approximation for y(1), using 2 equal steps? 3 equal steps? What
about n steps, where n can now be any natural number? (It will be useful to know that
1 + 2 + + (n 1) = n(n 1)/2.) As n , these approximations should converge to
y(1). Do they?
(b) In the text and in class it was claimed that for small h, Eulers method for stepsize h
has an error which is at most proportional to h. The n-step approximation for y(1) has
h = 1/n. What is the exact value of the difference between y(1) and the n-step Euler
approximation? Does this conform to the prediction?

Part II Problems and Solutions

Problem 1: [Eulers method] (a) Write y for the solution to y = 2x with y(0) = 0. What is
y(1)? What is the Euler approximation for y(1), using 2 equal steps? 3 equal steps? What about n
steps, where n can now be any natural number? (It will be useful to know that 1 + 2 + + (n
1) = n(n 1)/2.) As n , these approximations should converge to y(1). Do they?
(b) In the text and in class it was claimed that for small h, Eulers method for stepsize h has an error
which is at most proportional to h. The n-step approximation for y(1) has h = 1/n. What is the
exact value of the difference between y(1) and the n-step Euler approximation? Does this conform
to the prediction?
Solution: y = x2 , so y(1) = 1.

Eulers method with stepsize h for this equation: xk = kh, yk+1 = yk + 2xk h.

With n = 2, h = 1/2:

k
0
1
2

xk
0
1/2
1

yk
0
0
1/2

mk = yk
0
1

hmk
0
1/2

With n = 3, h = 1/3:

k
0
1
2
3

xk
0
1/3
2/3
1

yk
0
0
2/9
2/3

mk = yk
0
2/3
4/3

hmk
0
2/9
4/9

With n arbitrary, h = 1/n:

k
0
1
2
3
4
..
.

xk
0
h
2h
3h
4h
..
.

yk
0
0
2h2
2
2h + 4h2
2
2h + 4h2 + 6h2
..
.

mk = 2xk
0
2h
4h
6h
8h
..
.

mk h
0
2h2
4h2
6h2
8h2
..
.

So yn = 2(1 + 2 + + (n 1))h2 = n(n 1)h2 . With h = 1/n this gives our estimate for
y(1): n(n 1)/n2 = (n 1)/n. The limit of this as n is 1, which is good, and the
error is 1/n, which is exactly h.

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part I Problems
For each of the next three problems, solve the given linear DE. Give the general solution,
and also the specic solution satisfying the initial condition.
Problem 1:

Problem 2: xy y = x

dy
+y = 2
dx
and

Problem 3: y = 1 + x + y + xy,

y (0) = 0

x (1) = 7
y (0) = 0

Problem 4: Water ows into and out of a 100,000 liter () reservoir at a constant rate of
10 /min. The reservoir initially contains pure water, but then the water coming in has a
concentration of 10 grams/liter of a certain pollutant. The reservoir is well-stirred so that
the concentration of pollutant in it is uniform at all times.
a) Set up the DE for the concentration c = c(t) of salt in the reservoir at time t. Specify
units.
b) Solve for c(t) with the given initial condition, and graph the solution c vs. t.
g

c) How long will it take for the concentration of salt to be 5 ?

d) What happens in the long run?

Part I Problems and Solutions

For each of the next three problems, solve the given linear DE. Give the general solution,
and also the specic solution satisfying the initial condition.
Problem 1:

dy
+y = 2
dx

Solution: Integrating factor = e x = e

y (0) = 0

1dx

General solution y = 2 + ce x
Specic solution y = 2(1 e x )
Problem 2: xy y = x

x (1) = 7

and

Solution:
First need to bring to the form

P
(
x
)
dx
=e
.

dy
dx

+ P( x )y = Q( x ) in order to compute the IF

So y 1x y = 1 = e x dx = e ln x = 1x

y = x c + 1x 1dx y = x (c + ln x ) (general solution)

y(1) = 1(c + 0) = c = 7 y = x (7 + ln x ) (specic solution)
Problem 3: y = 1 + x + y + xy,

y (0) = 0

Solution: First rewrite as y (1 + x )y = 1 + x.

IF: = e

(1+ x )dx

= e(x+x

General solution: y = e x+ x

2 /2)

2 /2

2
c + (1 + x )e(x+ x /2) so
y = 1 + ce x+ x

2 /2

2
2
using (1 + x )e(x+ x /2) dx = e x+ x /2 .
Specic solution: y(0) = 1 + c 1 = 0 c = 1 so y = e x+ x

2 /2

Problem 4: Water ows into and out of a 100,000 liter () reservoir at a constant rate of 10 /min.
The reservoir initially contains pure water, but then the water coming in has a concentration of 10
grams/liter of a certain pollutant. The reservoir is well-stirred so that the concentration of pollutant
in it is uniform at all times.

Part I Problems and Solutions

OCW 18.03SC

a) Set up the DE for the concentration c = c(t) of salt in the reservoir at time t. Specify units.
b) Solve for c(t) with the given initial condition, and graph the solution c vs. t.
g

c) How long will it take for the concentration of salt to be 5 ?

d) What happens in the long run?

Solution: Let x = x (t) be the amount of salt in the reservoir at time t, with x in grams and
g
t in minutes. Then c(t) = x (t)/V or c(t)
=
105 x (t) in .
We will work with x (t), and

then get c(t) at the end.

a)

dx
dt

= salt rate in - salt rate out = net rate of change.

g

Rate in is 10 10
min
= 102 min

Rate out is 10 min

x (t)
V

Thus,

in

g
min .

= 10 min

x (t) g
105

g
= 104 x (t) min

dx
= 102 104 x
dt
The initial condition is x (0) = 0.

b) Can use linear or separable method.

Using separable: (Exercise: solve using linear method and compare results)

dx
= 104 dt
x
ln(106 x ) = 104 t + c
106

106 x = Ce10

4 t

x = 106 Ce10

4 t

x (0) = 0: x (0) = 106 c = 0 c = 106 .

x (t) = 106 (1 e10
Thus,
in

g
,

4 t

ingrams.

c(t) = 105 x (t) = 10(1 e10

with t in minutes

4 t

Part I Problems and Solutions

OCW 18.03SC

c
10

c) c(t) = 5 = 10(1 e10

min, or t 4.81 days.

4 t

1
2

= 1 e10

4 t

104 t = ln 2 t = 104 ln 2 6931.5

d) c(t) 10 = the input concentration as t .

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Part II Problems
Problem 1: [Linear models] Scrooge McDuck wants to set up a trust fund for his nephew
Don. He has fool-proof investments which make a constant interest rate of I, measured in
units of (years)1 (so I = 0.05 means 5% per year), and he proposes to dole out the money
to his proigate nephew at a constant rate q dollars per year.
(a) Model this process by a differential equation. (Use the symbols I and q, rather than
specic values for them.) Explain your steps.
(b) Then nd the general solution to this differential equation.
(c) Now take I = 0.05. If Uncle Scrooge wanted to fund the trust so as to provide his
nephew with $1000 per month in perpetuity, while maintaining a constant balance in the
fund, how much should he invest?
(d) But in fact Uncle Scrooge wants to teach his nephew the virtues of self-reliance, and
so plans on having the trust fund run entirely out of money in exactly twenty years. If he
wants to give his nephew $1000 per month, how much should he fund the trust with at
the outset? Give the answer to the nearest penny (as Scrooge would insist on).
Problem 2: [Solutions to linear equations] Almost all the radon in the world today was
created within the past week or so by a chain of radioactive decays beginning mainly from
uranium, which has been part of the earth since it was formed. This cascade of decay
ing elements is quite common, and in this problem we study a toy model in which the
numbers work out decently. This is about Tatooine, a small world endowed with unusual
elements.
A certain isotope of Startium, symbol St, decays with a half-life tS . Strangely enough, it
decays with equal probability into a certain isotope of either Midium, Mi, or into the little
known stable element Endium. Midium is also radioactive, and decays with half-life t M
into Endium. All the St was in the star-stuff that condensed into Tatooine, and all the Mi
and En arise from the decay route described. Also, t M = tS .
Use the notation x (t), y(t), and z(t), for the amount of St, Mi, and En on Tatooine, in units
so that x (0) = 1. Also, assume y(0) = 0 and z(0) = 0.
(a) Make rough sketches of graphs of x, y, z, as functions of t. What are the limiting values
as t ?
(b) Write down the differential equations controlling x, y, and z. Be sure to express the
constants that occur in these equations correctly in terms of the relevant decay constants.
Use the notation (Greek letter sigma) for the decay constant for St and (Greek letter
mu) for the decay constant for Mi. Your rst step is to relate to tS and to t M . A check
. . .
on your answers: the sum x + y + z is constant, and so we should have x + y + z = 0.

Part II Problems

OCW 18.03SC

(c) Solve these equations, successively, for x, y, and z.

(d) At what time does the quantity of Midium peak? (This will depend upon and .)
(e) Suppose that instead of x (0) = 1, we had x (0) = 2. What change will this make to x (t),
y(t), and z(t)?
(f) Unrelated question: Suppose that x (t) = et is a solution to the differential equation
.
tx + 2x = q(t). What is q(t)? What is the general solution?

Part II Problems and Solutions

Problem 1: [Linear models] Scrooge McDuck wants to set up a trust fund for his nephew
Don. He has fool-proof investments which make a constant interest rate of I, measured in
units of (years)1 (so I = 0.05 means 5% per year), and he proposes to dole out the money
to his proigate nephew at a constant rate q dollars per year.
(a) Model this process by a differential equation. (Use the symbols I and q, rather than
specic values for them.) Explain your steps.
(b) Then nd the general solution to this differential equation.
(c) Now take I = 0.05. If Uncle Scrooge wanted to fund the trust so as to provide his
nephew with $1000 per month in perpetuity, while maintaining a constant balance in the
fund, how much should he invest?
(d) But in fact Uncle Scrooge wants to teach his nephew the virtues of self-reliance, and
so plans on having the trust fund run entirely out of money in exactly twenty years. If he
wants to give his nephew $1000 per month, how much should he fund the trust with at
the outset? Give the answer to the nearest penny (as Scrooge would insist on).
Solution: (a) Pick a letter to denote the number of years after the fund is set upsay t.
Pick a letter to denote the function of t giving the value of the fund at time tsay x. In a
small time interval from t to t + t, the fund increases in value by Ix (t)t, but decreases in
.
value by qt: x (t + t) x (t) c Ix (t)t qt. Divide by t and take the limit: x = Ix q.
(b) Separate: dx/( Ix q) = dt. Integrate: I 1 ln | Ix q| + c1 = t + c2 . Amalgamate
constants and multiply by I: ln | Ix q| = It + c. Exponentiate: | Ix q| = ec e It . Eliminate
the absolute value and reintroduce the lost solution: Ix q = Ce It . Solve for x: x =
(q/I ) + Ce It (where this C is the earlier one divided by I).

(c) Constant trust value means x = 0, which says Ix = q or x = q/I. So with q = 12,000
dollars/year and I = 0.05, x = $240,000. (If Scrooge socks away more than this, then the
trust fund could pay out the $1000/month and still grow. But this wouldnt be Scrooge.)
(d) We want to nd the constant of integration which makes x ( T ) = 0, where T = 20:
0 = x ( T ) = (q/I ) + Ce IT , or C = (q/I )e IT . Thus x = (q/I )(1 e IT e It ). Now we can
set t = 0 to nd the required initial value of the trust: x (0) = (q/I )(1 e IT ). With T = 20
and I = 0.05, 1 e IT = 1 e1 c 0.63212056. Thus the initial funding is about 63% of
what it was in (c): x (0) c ($240,000)(.63212056) c $151,708.93.
Problem 2: [Solutions to linear equations] Almost all the radon in the world today was
created within the past week or so by a chain of radioactive decays beginning mainly from
uranium, which has been part of the earth since it was formed. This cascade of decay

Part II Problems and Solutions

OCW 18.03SC

ing elements is quite common, and in this problem we study a toy model in which the
numbers work out decently. This is about Tatooine, a small world endowed with unusual
elements.
A certain isotope of Startium, symbol St, decays with a half-life tS . Strangely enough, it
decays with equal probability into a certain isotope of either Midium, Mi, or into the little
known stable element Endium. Midium is also radioactive, and decays with half-life t M
into Endium. All the St was in the star-stuff that condensed into Tatooine, and all the Mi
and En arise from the decay route described. Also, t M = tS .
Use the notation x (t), y(t), and z(t), for the amount of St, Mi, and En on Tatooine, in units
so that x (0) = 1. Also, assume y(0) = 0 and z(0) = 0.
(a) Make rough sketches of graphs of x, y, z, as functions of t. What are the limiting values
as t ?
(b) Write down the differential equations controlling x, y, and z. Be sure to express the
constants that occur in these equations correctly in terms of the relevant decay constants.
Use the notation (Greek letter sigma) for the decay constant for St and (Greek letter
mu) for the decay constant for Mi. Your rst step is to relate to tS and to t M . A check
. . .
on your answers: the sum x + y + z is constant, and so we should have x + y + z = 0.
(c) Solve these equations, successively, for x, y, and z.
(d) At what time does the quantity of Midium peak? (This will depend upon and .)
(e) Suppose that instead of x (0) = 1, we had x (0) = 2. What change will this make to x (t),
y(t), and z(t)?
(f) Unrelated question: Suppose that x (t) = et is a solution to the differential equation
.
tx + 2x = q(t). What is q(t)? What is the general solution?

Solution: (b) Startium obeys the natural decay equation, x = x, with solution x =
x (0)et . To relate to its half-life, solve for it in x (0)/2 = x (0)etS to nd = (ln 2)/tS .
Similarly, = (ln 2)/t M .
Midium decays as well, but in each small time interval gets half the decayed Startium
.
added: so y(t + t) c y(t)t + 12 x (t)t. Thus y = y + 21 x. Endium receives half
.
the decayed Startium and all the decayed Midium: z = 12 x + y. Adding these three
. . .
equations gives x + y + z = 0.

(c) Using x (0) = 1, we know that x = et . Thus y + y = 12 et . An integrating

()t
factor is given by et : dtd (et y) = 21 e()t . Integrating, et y = 21
+ c or y =
e
1 t
2 e

+ cet . The initial condition is y(0) = 0, so c = 21 : y =

1
t
2 ( e

et ).

We could solve for z in the same way, but its easier to calculate z = 1 x y = 1 +

Part II Problems and Solutions

/2 t
e

OCW 18.03SC

/2 t
e

(d) From the differential equation for y, we know that a critical point occurs when y =
1
t . Substitute the value for y: 1 ( et et ) = 1 et . Some algebra leads to
2
2 e
2
et = et , so e()t = /, so tmax =

ln ln
.

(e) Everything gets doubled.

(f) If x = et then q(t) = tx + 2x = tet + 2et = (t + 2)et . The associated homogeneous

.
equation is tx + 2x = 0, which is separable: dx/x = 2dt/t, so ln | x | = 2 ln |t| + c =
ln(t2 ) + c and x = C/t2 . So the general solution of the original equation is et + C/t2 .

MIT OpenCourseWare
http://ocw.mit.edu

18.03SC Differential Equations

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.