§1 Mathematical Modelling

Given a real world problem, we often try to simplify the problem to a more manageable
problem based on some sound assumptions and thus, propose a solution for the problem.
However, there will often be some limitations to the solution proposed. Thus, we may need to
review the process from which the solution is derived and refine the process.

The above mentioned process is known as Mathematical Modelling. By definition,

Mathematical modelling is the process of formulating a mathematical model, e.g. a
differential equation, to represent and solve a real world problem. The process involves
making assumptions, solving the model and interpreting the solution in the context of
the original problem.

Page 1 of 28
 The Modelling Process1
The generic process of modelling is illustrated by the flow chart below:

A real world problem

Step 1: Make assumptions Review assumptions

Step 2: Represent the problem in mathematical form

- State the quantities to be studied (including units and scales
used)
(a) Independent variable – e.g. time.
(b) Dependent variables – quantities that are
functions of the independent variable, e.g.
No
population size.
(c) Parameters – quantities that do not change
with the independent variable but can assume
different values under different conditions.
- Obtain an equation to relate the quantities based on
assumptions made

Step 3: Solve the mathematical problem

Is the
Step 4: Interpret the solution in the solution
context of the original problem satisfactory?

Yes

Use model (e.g. make prediction)

1
Modified from the UK Maths in Education and Industry (MEI) Structured Maths, administered by Oxford,
Cambridge and Royal Society of Arts Examinations (OCR).

Page 2 of 28
Newton’s Law of Cooling

Newton’s law of cooling states that the rate of change of the temperature  C of an object is
o

proportional to the difference in temperature between the object and the ambient temperature.

Example 1: 2010/P1/7 (Rate of Change of Temperature of a Body)

A bottle containing liquid is taken from a refrigerator and placed in a room where the
o
temperature is a constant 20 C . As the liquid warms up, the rate of increase of its

temperature  C after time t minutes is proportional to the temperature difference

o

 20   
o
C o
. Initially the temperature of the liquid is 10 C and the rate of increase of the
o
temperature is 1 C per minute. By setting up and solving a differential equation, show that

1
 t
  20  10e 10
. [7]

Find the time it takes the liquid to reach a temperature of 15 C , and state what happens to 
o

for large values of t. Sketch a graph of  against t. [4]

Solution:

Step 1: Make assumptions (Given in question)

Constant ambient temperature.
Step 2: Represent the problem in mathematical form

Rate of increase in temperature   20   

d
  20   
dt

d
  k  20    , k  0  liquid will increase in temperature 
dt

Page 3 of 28
Step 3: Solve the mathematical problem
d
 t  0 ,   10, 1
Given that initially dt
d
  10, 1
when dt
d
 k  20   
dt
1  k  20  10 
1
k
10
d 1
   20   
dt 10
d 1
  20   
dt 10
Alternatively, 1 1
1 1  20   d   10 dt  where temperature of liquid,   20 
 20   d   10 dt 1
 ln 20    t  c
t 10
 ln  20      c
10 1
ln 20     t  c
( 20    0) 10
1
 t c
20    e 10

1
 t
20    Ae 10
, where A  e  c
1
 t
  20  Ae 10

when t  0 ,   10
10  20  Ae0
A  10
1
 t
   20  10e 10
 shown 

when   15 ,
1
 t
15  20  10e 10

1
 t
10e 10
5

1 1
 t  ln
10 2
t  6.93 min

Page 4 of 28
 1
 t
As t   , e
10
 0 .   20 .

Step 4: Interpret the solution in the context of the original problem

 The temperature of the liquid will increase and approach room temperature,
20o C for large values of t.
Note for lecturer:
To highlight the use of
Window Settings using the
hints given in question.
20

Note:
For contextual question, the
values the variables can take
must be taken into account
O t when sketching the graph.

Page 5 of 28
Example 2: 2008 RJC Promo/8
Useful Result:
Overall rate of change = rate of in-flow/increase/birth – rate of out-flow/decrease/death
In an experiment, glucose is infused into the bloodstream at a constant rate of
B mg/min. Glucose is also converted and removed from the bloodstream at a rate proportional
to the amount of glucose, x mg, present in the bloodstream at time
t minutes after the start of the experiment. If x = 150, the amount of glucose in the
bloodstream remains constant.
dx B
  150  x 
(i) Show that dt 150 . [2]
(ii) Find x in terms of B and t, given that the bloodstream contains 70mg of glucose
when t = 0. [4]

Solution:
(i)
Step 1: Make assumptions
NA
Step 2: Represent the problem in mathematical form

Infuse Remove
Constant Rate Proportional  kx
B mg/min x mg

Overall Rate of Change = Rate of inflow – Rate of Outflow

dx
 B  kx ,k  0
dt

Page 6 of 28
Step 3: Solve the mathematical problem

Given that when x  150 ,

dx
 0
Amount of glucose in the bloodstream remains constant (no change in rate) dt

dx
 0
when x  150 , dt
dx
 B  kx
dt
0  B  k  150 
B
k
150
dx B B
  B x  150  x   shown 
dt 150 150
(ii)
dx B
  150  x 
dt 150
1 B
  150  x  dx   150 dt , x  150

B
 ln 150  x  tc
150
B
 t c
150  x  e 150

B
 t
150  x  Ae 150
, where A  e  c
B
 t
x  150  Ae 150

Given also when t  0 , x  70

70  150  A
A  80

B
 t
 x  150  80e 150

Page 7 of 28
Example 3: 2015/P2/1
As a tree grows, the rate of increase of its height, h m, with respect to time, t years after
planting, is modelled by the differential equation
dh 1 1
 16  h
dt 10 2

The tree is planted as a seedling of negligible height, so that h  0 when t  0 .

(i) State the maximum height of this tree, according to this model. [1]

(ii) Find an expression for t in terms of h, and hence find the time the tree takes to reach
half its maximum height. [5]

Solution:
(i)
dh 1 1
 16  h
dt 10 2
dh
0
For stationary point, dt

1 1
16  h  0
10 2
h  32

To verify that it is the maximum height: Note to Teachers:

For this question, we cannot
dh prove using differentiation that
For dt to be defined, it is the maximum point without
1 first solving the D.E.
16  h  0  h  32
2

 The maximum height is 32 m.

Page 8 of 28
(ii)
dh 1 1
 16  h
dt 10 2
1 1
 1
dh   dt
10
16  h
2
1

 h 2 1
  2  dh   10 dt
 16 
1
 h 2
 16  
1  2 1
 t c
1 2 12 10
1
 h 2 1
4  16    t  c
 2 10

when t = 0, h = 0,

1
4  16  0  2  0  c
c  16

1
 h 2 1
 4 16    t  16
 2  10
 h 1
4  16    t  16
 2  10
1  h
t  16  4 16  
10  2
 h
t  160  40 16  
 2

h  16  half of its maximum height of 32 

when ,

16
t  160  40 16 
2
 46.9

 The tree will take 46.9 years to reach half its maximum height.

Page 9 of 28
§2 Modelling for Population Dynamics

In modelling of population dynamics, we would like to investigate how a population size

changes over time with or without external influences. One important conclusion is the long-
term behaviour of the population under certain conditions.

We will study 3 types of population growth models:

1. Exponential model
2. Logistic Model (without harvesting)
3. Logistic Model (with harvesting)

Notation:
P = Population Size (dependent variable)
t = Time (independent variable)

Definition:
Per Capita Birth Rate is the number of births per unit of population per unit time.
Per Capita Death Rate is the number of deaths per unit of population per unit time.

Example:
Population, P = 1000
No. of births = 10 per year
10
 0.01
Per Capita Birth Rate = 1000 per person per year i.e. there are 0.01 births per person
per year.

Assumptions:
 Ignore random fluctuations.
 Births and Deaths are continuous in time.
 Immigration and Emigration are negligible.

Useful result:

Rate of change of population = Rate of Births – Rate of Deaths

Page 10 of 28
§2.1 Exponential Growth (Malthus Model)

The English demographer and political economist, Thomas Robert

Malthus, (1766 –1834), was best-known for his influential views on
population growth. He emphasized the potential for populations to rise
steeply.

Assumptions:

 The Per Capita Birth and Death Rate (  and  ) are constants.
 The birth rate and death rate are each proportional to the size of the population i.e.

birth rate =  P and death rate =  P

Since Rate of change of population = Rate of Births – Rate of Deaths, we get

dP
  P  P , where  and  are constants
dt
dP
 (   )P
dt

dP
 kP , where k    
dt

The parameter, k is known as the constant of proportionality, or growth rate

coefficient, or effective or nett growth rate.

Observations:

 If k  0 , there is growth in the population.

 If k  0 , there is no growth in the population.

 If k  0 , there is decay in the population.

Page 11 of 28
Solving the differential equation,

dP
 kP
dt
1
 P dP   kdt
ln P  kt  c

P  e kt  c

P  Ae kt , A  

If given the initial condition P  P0 when t  0 , we get

A  P0

 P  P0 ekt

Graphical representation:

>0

P0
=0
<0
0 t

When k  0 , the population will grow infinitely.

When k  0 , there is no growth and the population remains constant.

When k  0 , the population will shrink and tend to 0 as t increases i.e. the population will
become extinct over time. This can also be used to model the rate of natural decay for
radioactive substance.

Page 12 of 28
Think: When is it appropriate to use the Exponential Growth Model?
May be a reasonable model for a small population in a large environment with no space or
resource limitations, over a relatively short period of time. E.g. culture of bacteria under
controlled conditions.

In reality, population will eventually be limited by some factors e.g. limited food resources
and space. For this, we can consider the next model (Logistic Model).

The exponential model with a negative growth rate coefficient can be used to model the rate
of natural decay for radioactive substance (See Example 5).

Example 4:
In 2000, the world's population was approximately 6 billion. In 2010, it was 7 billion. If the

dP
 kP
population P at time t obeys the differential equation, dt , estimate the world's
population in the year 2050. Find the year in which the population reaches 8 billion.

Solution:
Let P be the population in billions and t the time in years measured from year 2000.
dP In 2010, t  10 and P  7 :
 kP
dt
1 7
1 7  6e k (10) i.e. k  ln
 P d P   k dt 10 6

ln P  kt  C ,P  0
In 2050, t  50,
P  Ae , where A  e
kt C
Note:
A  P0  1 7
 ln  (50)
 P  6e  10 6 
 13.0
Note: P = 0 is a trivial solution
when t  0, P  6 Hence, the world's population in the year
2050 is approximately 13.0 billions.
6  P0 e k (0) i.e. P0  6

 P  6e kt  1 7
 ln  ( t )
When P  8, 8  6e  10 6 
 t  18.9
The year is 2019.

Page 13 of 28
Example 5: Radioactive Decay (Exponential Growth Model in which growth rate
coefficient is negative)
A block of substance having an original mass of 100g is observed. After 40h, its mass is
observed to have reduced to 90g. The substance is decaying at a rate proportional to its mass.
Find (i) an expression for the mass of the substance at any time,
(ii) the time lapse before the block decays to half of its original mass.

Solution:
(i) (ii)
Let y be the mass of radioactive substance when y  50 ,
present at time t. t 9
ln  
y  100e 40  10 

Rate of decay of substance  its mass t 9

ln  
50  100e 40  10 
dy
 y t 9 1
dt ln    ln  
dy 40  10  2
  ky, k  0
dt t  263.15
 substance is decaying (decrease in mass)   263  3 s.f.
dy Note: This value of t is known as the
 ky
dt radioactive half-life of this substance.
1
 y dy   k dt
ln  y    kt  c  y  0
y  e  kt  c  e  kt e c

y  Ae  kt , where A  ec

when t  0, y  100
100  A

when t  40, y  90
90  100e  k  40
9
40k  ln  
 10 
1 9
k   ln  
40  10 
t 9
ln  
 y  100e 40  10 

Page 14 of 28
§2.2 Logistic Model (Verhulst Model)

In Malthus’ model, we do not take into consideration the idea of limited resources. In reality,
the population size would be limited by resources such as food and space where the
crowding effects become important. The logistic model, proposed by Verhulst, takes into
account such considerations.

Pierre-Francois Verhulst (1804 - 1849) was a Belgian mathematician

who generalized the Malthusian model by allowing for the fact that
populations encounter internal competition as they grow within a
closed environment, and this competition has a tendency to retard
the rate of growth. His idea is that while the population will continue
to grow as time goes on, the rate at which it does so will also get
slower.

To incorporate the crowding effect, instead of having a constant per capita death rate (as in
the Malthus model), we assume a linear dependence of the per capita death rate with
population size. That is, the larger the population size, the higher the per capita death rate.

Per Capita Death Rate     P

Using Rate of change of Population = Rate of births – Rate of deaths, we get

dP
  P      P P , where  ,  and  are constants
dt
dP
 (   )P   P2
dt
 P 
 (   ) P 1  
   

dP  P  
 kP  1   , where k     and N 
dt  N 

The parameter, k is known as the intrinsic growth rate coefficient, and

N is known as the carrying capacity of the environment.

Page 15 of 28
Observations:
P
0
 If the population is small compared to N (i.e N ), the rate of growth of the

dP
 kP
population is proportional to the size of the population i.e. dt .
 If the population is too large to be supported by the environment and resources (i.e

dP
0
P  N ), the population will decrease (i.e dt ).
dP
0
 At equilibrium (i.e. dt ) , we have P  0 or P  N .

The analytical solution for the differential equation is:

Suppose P  0, N ,

dP  P
 kP 1  
dt  N
N
 P  N  P  d P   k dt
1 1
 P N  P d P   k dt
+

ln P  ln N  P  kt  C
, where C is an arbitrary constant.
P
 e kt . eC
N P

 P 
  Ae
kt

where A   e
C
N P

Page 16 of 28
 P0
A
P  0   P0 N  P0
If given the initial condition , we get

 P  P0e kt
  
 N  P  N  P0

P  N  P0   NP0 e kt  PP0 e kt

NP0e kt
P
P0 e kt  N  P0
NP0
P
P0   N  P0  e  kt

NP0
 P
P0   N  P0  e  kt
,

where k  Intrinsic growth rate coefficient and N  carrying capacity

For P  0 ,

If k  0 , as t   , e  0 and thus P  N .
 kt


Case 1:

Case 2:

0 t

 If k  0 ,
P  P0 , that is, no growth in population size.

If k  0 , as t   , e   and thus P  0 i.e. lead to extinction of

 kt
 the
population.

Page 17 of 28
Example 6:
Given that the rate of change of population on an isolated island is given by
dP
 90 P  P 2
dt where P is the population at time t,
(i) Write down the carrying capacity of the population.

(ii) Given
P  P0 at time t  0 , solve for P in terms of t and P0 .

(iii) Sketch, on the same diagram, P against t for

P0  20, P0  60, P0  120.

Solution:
dP  P (iii)
 90 P  P 2  90 P  1  
(i) dt  90 
Carrying capacity is 90. P

dP
 90 P  P 2
(ii) dt
dP
 90 P  P 2  P  90  P 
dt
1 dP
1
P  90  P  dt
1
 P  90  P  
d P  1 dt
1 1 1
 
90 P 90  P 
d P  1 dt O t
1
 ln P  ln 90  P   t  c
90 
P
ln  90t  c '
90  P
P
 Ae90t , A   ec '
90  P

P0
A
When t = 0, 90  P0
90t
 P  P0 e
 
 90  P  90  P0

P  90  P0   90 P0 e90t  PP0e90t

90 P0 e90t
P
P0 e90t  90  P0

Page 18 of 28
Example 7:

The National Institute of Education has 4,000 trainee teachers. On the first day of the
semester, a group of 4 trainees thought they heard their mathematics lecturer say that
everyone would receive the A grade for the course. The next day, a carefully conducted
survey of the entire trainee population revealed that by now 80 students had heard this

dy
 ay  b  y 
rumour. If the rumour spreads according to the logistic equation, then dt , where
y is the number of trainees who have heard the rumour, t is the time in days measured from
the first day, and a and b are constants. At what time will 90% of the trainees have heard the
rumour?

dy  y
 ay  b  y   aby 1  
dt  b
dP  P
 kP 1  
Comparing with dt  N ,
b=4000 (the carrying capacity, which in this case is the total number of trainees).

dy  y 
 4000ay  1  
dt  4000 
4000
 y  4000  y  dy   4000a dt
1 1
 y 4000  y dy   4000a dt
+

ln y  ln 4000  y  4000at  C
, where C is an arbitrary constant.
y
 e 4000 at . eC
4000  y

 y 
  Ae
4000 at

 4000  y  where A   e
C

4 1
A 
Considering initial condition, when t = 0, y  4 , we get 4000  4 999

 y  e4000 at
  
 4000  y  999

Page 19 of 28
Given also when t =1 , y=80

4000 a
 80  e
  
 4000  80  999
1  999 
a ln       1
4000  49 

 y  e4000 at
 
 4000  y  999
999 y  4000e 4000 at  ye 4000 at

4000e 4000 at
y
e4000 at  999
4000
y
1  999e 4000 at
Sub (1)
4000
 y  999 
t ln  
1  999e  49 

For 90% of trainees to have heard the rumour, y  4000  0.9  3600
4000
3600   999 
t ln  
1  999e  49 

 999 
t ln   1
999e  49 

9
 999  1
t ln    ln
 49  8991
t  3.02 days

Page 20 of 28
§2.3 Logistic Model with Harvesting

Source: https://www.worldwildlife.org/threats/overfishing

Page 21 of 28
The removal of a constant number of populations during each time period is known as
harvesting or fishing. Ever since primitive humans began hunting and fishing thousands of
years ago, there has been a need to know how killing a certain number of animals will affect
the population at large. The fact that there are over 750 plants and animals on the endangered
species list indicates that humans are not always cognizant of how their actions will affect
plants and animals.

To include the effect of harvesting,

dP  P
 kP 1    H
dt  N
where H is a constant denoting the constant harvesting rate.

Similar to the above, we can solve the differential equation and get an analytical solution.
However, we are more interested in what will happen to the population size in the long run.

In this model, we consider the case where k  0 and N  0 .

dP
To do so, we analyse the graph of dt against P and by determining the equilibrium values.

dP
H=0
dt
H< kN/4

0 N/2 As H
increases
H= kN/4

H > kN/4

dP N N 2 NH
0 P  
By setting dt , we get the equilibrium values 2 4 k .

Page 22 of 28
 dP Equilibrium value(s) Limit Graph of P against t
Graph of dt against P
H 0 PN dP
0  P  N, 0 P
dP
For dt
dt
 P increases and approaches N.
P N
N dP
P  N, 0
For dt t
 P decreases and approaches N.

kN N N 2 NH dP
0 H  P   0 P
4 For 0  P   , dt
dP 2 4 k
dt
Let  P decreases to 0.
P dP 
N N 2 NH 0
    
2 4 k , For   P   , dt
N  P increases and approaches  . 
N N 2 NH
   dP
2 4 k 0 t
For P   , dt
 P decreases and approaches  .
kN N N dP
H P 0P , 0 P
4 2 For 2 dt
dP
dt N  P decreases to 0. N
2 P N dP 2
P , 0
For 2 dt t
N
.
 P decreases and approaches 2

Page 23 of 28
 kN
H
4 is known as the bifurcation value.

Page 24 of 28
Think about this:
kN
H
What happens when 4 ?
Though a little more fishing causes only a little more decline in the population size initially,

kN
this can change drastically when H reaches the bifurcation value of 4 . At this point, if we

kN
H
allow slightly more fishing i.e. when 4 , the fish population will decrease and become
extinct.
(Note: A bifurcation occurs when a small smooth change made to the parameter values of a
system causes a sudden qualitative change in its behaviour.)

Example 8:
dP P2
=2 P−
Consider the population model dt 50 for a species of fish in a lake. It is decided
that fishing will be allowed, but it is unclear how many fishing licenses should be issued. It is
known that the average catch of a fisherman with a license is 3 fish per year.
(a) What is the largest number of licenses that can be issued if the fish are to have a chance
to survive in the lake?
(b) Suppose the number of fishing licenses in part (a) is issued. Discuss the long term
behaviour of the fish population for different initial population.

Solution:
(a) Let n denote the number of licenses issued per year.
dP P2
 2P   3n
Then the differential equation becomes dt 50 .
dP P2
 2P   3n  0
Equilibrium points: dt 50

 P 2  100 P  150n  0
100  10000  600n
P 
 2 50  5 100  6n

For the fish to have a chance to survive in the lake (i.e. to have real roots), we need:

Page 25 of 28
 50
100  6n  0  n
3

As n   , the largest number of licenses that can be issued is 16.

Page 26 of 28
(b) When n = 16, equilibrium points are
P  50  2500  150(16)
 50  10  40 or 60

dP
(i) If initial population, P(0) > 60, dt < 0. The population of P will decrease to the
equilibrium value of 60.
dP
(ii) If 40 < P(0) < 60, dt > 0. The population of P will increase to the equilibrium
value of 60.
dP
(iii) If 0 < P (0) < 40, dt < 0. The population of P will decrease to 0 after a finite
period of time.
dP
(iv) If P(0) = 40 or 60, dt = 0. The population will remain constant.

Investigating slope fields:

https://www.geogebra.org/m/W7dAdgqc

Reference:
Parts of the materials are taken from the materials (Teaching H2 Further Mathematics:
Differential Equations and Mathematical Modelling) by Dr Ang Keng Cheng, NIE.

Page 27 of 28
Page 28 of 28