Application of Differential Equations

Unlimited Population Growth

An elementary model of population growth is based on the assumption that
• The rate of growth of the population is proportional to the size of the population.
Note that the rate of change of a population depends on only the size of the population
and nothing else. In particular limitations of space or resources have no effect. This
assumption is reasonable for small populations in large environments — for example, the
first few spots of mold on a piece of bread or first European settlers in the United States.
Because the assumptio0n is so simple, we expect the model to be simple as well.
The quantities involved are
t= time (independent variable),
P = population (dependent variable),
k= proportionality constant (parameter) between the rate
of growth of the population and the size of the population.
The parameter k is often called the “growth-rate coefficient.”
Now let’s express our assumption using this notation. The rate of growth of the
population P is the derivative dP/dt. Being proportional to the population is expressed
as the product, kP , of the population P and the proportionality constant k. Hence our
assumption is expressed as the differential equation
dP
= kP .
dt
Solving the differential equation by separation of variables,

P (t) = P0 ekt , where P0 is the initial population.

Logistic Population Model

To adjust the exponential growth population model to account for a limited environment
and limited resources, we add the assumptions:
• If the population is small, the rate of growth of the population is proportional to
its size.
• If the population is too large to be supported by its environment and resources, the
population will decrease. That is, the rate of growth is negative.
For this model, we again use
t= time (independent variable),
P = population (dependent variable),
k= growth-rate coefficient for small populations (parameter).
However, our assumption about limited resources introduces another quantity, the
size of the population that corresponds to being “too large.” This quantity is a second
parameter, denoted by N , that we call the “carrying capacity” of the environment. In
terms of the carrying capacity, we are assuming that P (t) is increasing if P (t) < N .
However, if P (t) > N , we assume that P (t) is decreasing.

1
 Using this notation, we can restate our assumptions as:
dP
• ≈ kP if P is small (first assumption).
dt
dP
• If P > N , < 0 (second assumption).
dt
We also want the model to be “algebraically simple,” or at least as simple as possible,
so we try to modify the exponential model as little as possible. For instance, we might
look for an expression of the form
dP
= k · (something) · P .
dt
We want the “something” factor to be close to 1 if P is small, but if P > N we want
“something” to be negative. The simplest expression that has these properties is the
function  
P
(something) = 1 − .
N
Note that this expression equals 1 if P = 0, and it is negative if P > N . Thus our model
is  
dP P
=k 1− P.
dt N
This is called the logistic population model with growth rate k and carrying capacity
N . It is another first-order differential equation. This equation is said to be nonlinear
because its right-hand side is not a linear function of P as it was in the exponential
growth model.

Example
Consider the population model
 
dP P
= 0.4P 1− .
dt 230
where P (t) is the population at time t.
(a) For what values of P is the population in equilibrium?

(b) For what values of P is the population increasing?

(c) For what values of P is the population decreasing?

(d) Find an expression for P (t).

(a) If the population is in equilibrium, then the rate of change of the population is
zero. So,  
dP P
= 0.4P 1 − = 0 =⇒ P = 0 or P = 230.
dt 230
(b) If the population is increasing, then the rate of change of the population is positive.
So,  
P
0.4P 1 − > 0 =⇒ P (230 − P ) > 0 =⇒ 0 < P < 230.
230

2
 (c) If the population is decreasing, then the rate of change of the population is negative.
So,
 
P
0.4P 1 − < 0 =⇒ P (230 − P ) < 0 =⇒ P > 230 or P < 0 (rejected as P ≥ 0).
230

(d) Solving the differential equation by separation of variables,

 
dP P
= 0.4P 1 −
dt 230
dP 230 dP
=⇒ P

= = 0.4 dt
P 1 − 230 P (230 − P )
Z Z   Z
230 dP 1 1
=⇒ = + dP = 0.4 dt
P (230 − P ) P 230 − P
=⇒ ln |P | − ln |230 − P | = 0.4t + C
=⇒ − (ln |230 − P | − ln |P |) = −0.4t − C
230 − P
=⇒ ln = −0.4t − C
P
230 − P 230
=⇒ = − 1 = e−0.4t−C = e−C e−0.4t
P P
230
=⇒ − 1 = P0 e−0.4t , where P0 = e−C > 0
P
230
=⇒ = 1 + P0 e−0.4t
P
230
=⇒P (t) = .
1 + P0 e−0.4t

A Mixing Problem, Mixing in a Vat

Consider a large vat containing sugar-water that is to be made into soft drinks. Suppose:
• The vat contains 100 gallons of liquid. Moreover, the amount flowing in is the same
as the amount flowing out, so there are always 100 gallons in the vat.

• The vat is kept well mixed, so the concentration of sugar is uniform throughout the
vat.

• Sugar-water containing 5 tablespoons of sugar per gallon enters the vat through
pipe A at a rate of 2 gallons per minute.

• Sugar-water containing 10 tablespoons of sugar per gallon enters the vat through
pipe B at a rate of 1 gallons per minute.

• Sugar-water leaves the vat through pipe C at a rate of 3 gallons per minute.
To make the model, we let t be time measured in minutes (the independent variable).
For the dependent variable, we have two choices. We could choose either the total amount
of sugar, S (t), in the vat at time t measured in tablespoons, or C (t), the concentration
of sugar in the vat at time t measured in tablespoons per gallon. We will develop the
model for S, leaving the model for C as an exercise.

3
 Using the total sugar S (t) in the vat as the dependent variable, the rate of change
of S is the difference between the amount of sugar being added and the amount of sugar
being removed. The sugar entering the vat comes from pipes A and B and can be easily
computed by multiplying the number of gallons per minute of sugar mixture entering
the vat by the amount of sugar per gallon. The amount of sugar leaving the vat through
pipe C at any given moment depends on the concentration of sugar in the vat at that
moment. The concentration is given by S/100, so the sugar leaving the vat is the product
of the number of gallons leaving per minute (3 gallons per minute) and the concentration
(S/100). The model is

dS S
= 2·5 + |1 {z
· 10} − 3 · .
dt |{z}
| {z100}
sugar in sugar in
from pipe A from pipe B sugar out
from pipe C

That is,
dS 3S 2000 − 3S
= 20 − = .
dt 100 100
To solve this equation analytically, we separate and integrate. We find
dS dt
=
2000 − 3S 100
ln |2000 − 3S| t
= + c1
−3 100
3t
ln |2000 − 3S| = − − 3c1
100
ln |2000 − 3S| = − 0.03t + c2 ,

where c2 = −3c1 . Exponentiating we obtain

|2000 − 3S| = e(−0.03t+c2 ) = c3 e−0.03t ,

where c3 = ec2 . Note that this means that c3 is a positive constant. Now we must be
careful. Removing the absolute value signs yields

2000 − 3S = ±c3 e−0.03t ,

where we choose the plus sign if S (t) < 2000/3 and the minus sign if S (t) > 2000/3.
Therefore we may write this equation more simply as

2000 − 3S = c4 e−0.03t ,

where c4 is an arbitrary constant (positive, negative, or zero). Solving for S yields the
general solution
2000
S (t) = ce−0.03t + ,
3
where c = −c4 /3 is an arbitrary constant. We can determine the precise value of c if
we know the exact amount of sugar that is initially in the vat. Note that, if c = 0, the
solution is simply S (t) = 2000/3, an equilibrium solution.

4
Mixing Problems Revisited, A Polluted Pond
Consider a pond that has an initial volume of 10, 000 cubic meters. Suppose that
at time t = 0, the water in the pond is clean and that the pond has two streams flowing
into it, stream A and stream B, and one stream flowing out, stream C. Suppose 500
cubic meters per day of water flow into the pond from stream A, 750 cubic meters per
day flow into the pond from stream B, and 1250 cubic meters flow out of the pond via
stream C.
At time t = 0, the water flowing into the pond from stream A becomes contaminated
with road salt at a concentration of 5 kilograms per 1000 cubic meters. Suppose the water
in the pond is well mixed so the concentration of salt at any given time is constant. To
make matters worse, suppose also that at time t = 0 someone begins dumping trash into
the pond at a rate of 50 cubic meters a day. The trash settles to the bottom of the pond,
reducing the volume by 50 cubic meters per day. To adjust for incoming trash, the rate
that water flows out via stream C increases to 1300 cubic meters per day and the banks
of the pond do not overflow.
The description looks very much like the mixing problems we have already considered
(where “pond” replaces “vat” and “stream” replaces “pipe”). The new element here is
that the total volume is not constant. Because of the dumping of trash, the volume
decreases by 50 cubic meters per day.
If we let S (t) be the amount of salt (in kilograms) in the pond at time t, then dS/dt
is the difference between the rate that salt enters the pond and the rate that salt leaves
the pond. Salt enters the pond from stream A only, and the rate at which it enters is
the product of its concentration in the water and the rate at which the water flows in
through stream A. Since the concentration is 5 kilograms per 1000 cubic meters and
the rate that water flows into the pond from stream A is 500 cubic meters per day, the
rate at which salt leaves via stream C is the product of its concentration in the pond
and the rate at which water flows out of the pond. The rate at which water flows out
is 1300 cubic meters per day. To determine the concentration, we note that it is the
quotient of the amount S of salt in the pond by the volume V . Because the volume is
initially 10, 000 cubic meters and it decreases by 50 cubic meters per day, we know that
V (t) = 10, 000−50t. Hence, the concentration is S/(10, 000−50t), and the rate at which
salt flows out of the pond is
 
S 26S
1300 = .
10, 000 − 50t 200 − t
The differential equation that models the amount of salt in the pond is therefore
dS 5 26S
= − .
dt 2 200 − t
This model is valid only as long as there is water in the pond — that is, as long as
the volume V (t) = 10, 000 − 50t is positive. So the differential equation is valid for
0 ≤ t < 200. Because the water is clean at time t = 0, the initial condition is S (0) = 0.
Rewriting the differential equation as
dS 26S 5
+ =
dt 200 − t 2
indicates that the integrating factor is
R
= e−26 ln(200−t) = eln[(200−t) ] = (200 − t)−26 .
26
dt −26
µ (t) = e 200−t

5
 Multiplying both sides by µ (t) gives

dS 5
(200 − t)−26 + 26 (200 − t)−27 S = (200 − t)−26 .
dt 2
By the Product Rule, this equation is the same as the differential equation
d   5
(200 − t)−26 S = (200 − t)−26 .
dt 2
Integrating both sides yields
Z
5
(200 − t) −26
S= (200 − t)−26 dt
2
5 (200 − t)−25
= + c,
2 25
where c is an arbitrary constant. Solving for S, we obtain the general solution
200 − t
S= + c (200 − t)26 .
10
Using the initial condition S (0) = 0, we find that c = −20/20026 and the particular
solution for the initial-value problem is
 26
200 − t 200 − t
S= − 20 .
10 200