Comparison of Accury Between The Exponential Model and The Logistic Model in

Indonesia’s Population Growth

Background
Population dynamics is the study of the growth of a population of organisms. While the population is
defined as a collection of individuals who inhabit a certain place that can mate with each other so that
they can develop into more. A population can grow fast or slow. The rate of population growth is
influenced by survival at a certain age and the number of deaths before reaching productive age.
Today the major issue of the world is the tremendous growth of the population especially in the
developing countries like Indonesia’s.

Year Indonesia Population’s

2010 238,5
2011 242
2012 245,4
2013 248,8
2014 252,2
2015 255,5
2016 258,7
2017 261,9
2018 265
2019 268,1
2020 270,2
2021 273,87
Source : BPS
Formulation of The Problem
1. Determine Indonesia's population growth in 5 consecutive years.
2. Compare accuracy using exponential model or logistic model.
Purpose
1. Can find out the growth of Indonesia's population in 5 consecutive years.
2. Can compare accuracy using exponential model or logistic model.
Theoretical Basis
The Exponential Growth Model
In 1798 Thomas R. Malthus proposed a mathematical model of population growth. He proposed by
the assumption that the population grows at a rate proportional to the size of the population. This is a
reasonable assumption for a population of a bacteria or animal under ideal conditions (unlimited
environment, adequate nutrition, absence of predators, and immunity from disease). Suppose we
know the population P0 at some given time t=t 0 , and we are interested in projecting the population
P, at some future time t=t 1 , In other words we want to find a population function P(t) for t 0 ≤ t ≤ t 1
dP
satisfying P(t 0) = P0 . Then considering the initial value problem =kP ( t ) t 0 ≤t ≤t 1 ; P ( t 0 )=P0 .
dt
Integrating by variable separable in (1)
Integrating by variable separable in (1)
 dP
∫ P
=k ∫ dt

InP=kt +c

P ( t ) =P 0 e or P ( t ) =P 0 exp {k ( t−t 0 ) }
kt
(2)

where k is a constant called the Malthus factor, is the multiple that determines the growth rate.
Equation (1) is the Exponential growth model with (2) as its solution. It is a differential equation
dP
because it contains an unknown function P and it derivative . Having formulated the model, we
dt
now look at its consequences. If we rule out a population of 0, then P(t) > 0 for all t .So if k > 0 then
dP
equation shows that > 0 for all t. This means that the population is always increasing. In fact, as
dt
dP
P(t) increases, equation (1) shows that becomes larger. In order words, the growth rate increases
dt
as the population increases. Equation (1) is appropriate for modeling population growth under ideal
conditions, thus we have to recognize that a more realistic must reflect the fact a given environment
has a limited resources.
The Logistic Growth Model
This model was proposed by the Belgianmathematical biologist Verhulst in the 1840s as model for
world population growth. His model incorporated the idea of carrying capacity. Thus the population
growth not only on how to depends on the population size but also on how far this size is from the its
upper limit i.e. (maximum supportable population. He modified Malthus’s Model to make a
population size proportional to both the previous population and a new term

a−bP(t)
a
(3)
where and are the vital coefficients of the population. This term depicts how far the population is from
a
its maximum limit. Now as the population value gets closer to , this new term will become very
b
small and tend to zero, providing the right feedback to limit the population growth. Thus the second
term models the competition for available resources, which tends to limit the population growth. So
the modified equation using this new term is:

d P aP( t)(a−bP ( t ) )
=
dt a
(4)
t 0 ≤ t ≤ t 1 ; P ¿)= P0

This equation is known as the Logistic Law of population growth. Solving (4) applying the initial
conditions, the (4) become
dP 2
=aP−b P
dt
(5)
By the application of separation of variables and integrating, we obtain ∫ (
1 1
+
b
a P a−bP )
dP=∫ dt

1
( InP−¿ ( a−b )) =t+ c
k
(6)

At t=0∧P=P0

1
c=
k
( ¿ P0−¿ ( a−b P 0 ) )
Substituting c into (6) and solving for P yields
a
b
P=

( )
a
b
1+ −1 e−at
P0
(7)

Now taking the limit as t → ∞ of (7)

a
Pmax =lim P= (k >0)
t→∞ b
(8)

Putting t = 1 and t = 2, the values of P are P1∧P2 respectively, then we obtain from (7) the following
−a
b( 1 e
1−e )= −
−a
k N1 N0
(9)
−2 a
b( 1 e
1−e )= −
−2 a
k N2 N 0
(10)
Dividing (0) by (9) we have

1 e−2 a
−
−a P 2 P0
1+e =
1 e−a
−
P 1 P0
(11)

Hence solving for e−a we have

P0 (P2 −P 1)
e−a =
P2 (P1−P 0)
(12)
 2
b P1−P0 P 2
=
a P1 (P 0 P1−2 P0 P1+ P 1 P2 )
(13)
Therefore the limiting value of P is giving by
a
Pmax =lim P=
t→∞ b
P 1( P 0 P1−2 P0 P1+ P 1 P2)
2
P 1−P0 P 2
(14)
Mean Absolute Percentage Error (MAPE)
It is an evaluation statistic which is used to assess the goodness of fit of different models in national
and sub national population projections. This statistic is expressed in percentage. The concept of mean
absolute percentage error (MAPE) seems to be the very simple but of great importance in the
selecting a parsimonious model than the other statistics. A model with smaller MAPE is preferred to
the others models.
The mathematical form of MAPE is given under

1 ˇ
y ∨¿
MAPE= ∑
n t=¿¿
n ¿ yt −
yt
×100 ¿

Where Y t , Y̌ and N are the actual, fitted and number of observation of the (dependent variable)
population respectively.
Lower MAPE values are better because they indicate that smaller percentages errors are produced by
the forecasting model. The following interpretation of MAPE values was suggested by Lewis (1982)
as follows: Less than 10% is highly accurate forecasting, 10% to 20% is good forecasting, 21% to
50%is reasonable forecasting and 51% and above is inaccurate forecasting
Results and Discussion
To estimate the future population of Indonesia, we need to determine growth rate of Indonesia using
the Exponential Growth model in (2). Using the actual population of Indonesia in million on table 1
below with t = 0 corresponding to the year 2010, we have P0=238,5 . We can solve for the growth
rate k , the fact that P1=242when t = 1
kt
P ( 1 )=P0 e
k
242=238,5 e

k =ln ( 238,5
242
)
k =0,0145
Hence the general solution
0,0145 t
P ( t ) =238,5 e
(15)
This suggests that the predicted rate of Indonesia population growth is 1% with the Exponential
growth model. With this we projected the population of Indonesia to 2025. Again, based on table 1,
let t = 0, 1, and 2 correspond to the years 2010, 2011 and 2012 respectively. Then P0 , P 1 ,∧P 2 also
correspond 238,5; 242; and 245,4.
a
Substituting the values of P0 , P 1 ,∧P 2 into (14) we get Pmax = =11.193,67313 . This is the
b
predicted carrying capacity of the population of Indonesia.

−a
P0 ( P2−P1 )
From equation (12), we obtain e = =0,94 hence
P2 ( P1−P0 )

a=−ln(0,94)
Therefore the value of a=¿0,0618. This also implies that the predicted rate of Ghana population
growth is approximately % with the Logistic growth model.
a
From =11.193,67313 and equation (15), we obtained b = 5,52098 ×10−6
b

−a a
Substituting the values of P0 , e and into equation (7) we obtain
b
11.193,67313
P (t)=
1+(45,93364) ×0,94 t
As the general solution and we use this to predict population of Indonesia to 2025. The predicted
populations of Indonesia with both models are presented on the table 1 below.

Year Actual Population (in Project Population (in millions)

millions)
Exponential Model Logistic Model
2010 238,5 238,5 238,5
2011 242 241,9 253,3
2012 245,4 245,5 269,1
2013 248,8 249,1 285,9
2014 252,2 252,7 303,6
2015 255,5 256,4 322,4
2016 258,7 260,1 342,4
2017 261,9 263,9 363,5
2018 265 267,8 385,9
2019 268,1 271,7 409,7
2020 270,2 275,7 434,8
2021 273,87 279,7 461,4
2022 283,8 489,6
2023 287,9 519,4
2024 292,1 550,9
2025 296,4 584,3
Mean Absolute (MAPE)
Percentage Error
Fig 1 depicts that from 2010 the population of Indonesia has increased throughout. This may be
attributed to the improvement in the education, agricultural productively, water and sanitation and
health services. The exponential model predicted Indonesia’s population to be 296,4 in 2050 whereas
the Logistic model projected it to be 584,3. This is presented on Fig 2. From equation (14) we
calculated the Mean Absolute Percentage Error (MAPE) of both models. The MAPE for Exponential
and the Logistic model are % and % respectively.

280

270

260
Actual Populations

250

240

230

220
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Year

Fig 1 : Graph of actual population from 2010 to 2021

700

600

500

400
Population

300
Exponential Model
200 Logistic Model

100

0
10 12 14 16 18 20 22 24
20 20 20 20 20 20 20 20

Year

Fig 2 : Graph of predicted population values

Figure 2 above shows the graph of the predicted population of Indonesia with both models. The
Logistic Model is in orange and it deviate far from the actual population. The blue line represents the
forecast of the exponential model which is quiet similar to the actual population graph.

Conclusion
In conclusion the Exponential Model predicted a growth rate of approximately % and predicted
Indonesia’s population to be 296,4 million in the year 2025 with a MAPE of %. The Logistic Model
on the other hand predicted a carrying capacity for the population of Indonesia to be 11.193,67313 .
Population growth of any country depends on the vital coefficients. Found here the vital coefficients a
and b are a=¿0,0618 and b = 5,52098 ×10−6 respectively. Thus the population growth rate of
Indonesia according to this model is approximately % per annum. It also predicted the population of
Indonesia to be 584,3 million in 2050 with a MAPE of %. Based on Lewis (1982) we can conclude
that the Exponential Model gave a good forecasting result as compared to the Logistic model.

References
Glen Ledder, (2005), Differential Equations: A modeling Approach. McGraw-Hill
Companies Inc. USA.
Lewis, C.D (1982). International and business forecasting method; A practical guide to
exponential smoothing and curve fitting. Butterworth Scientific, London.
Malthus T.R, (1987). An Essay on the Principle of Population (1st edition, plus excepts 1893
2nd edition), Introduction by Philip Appeman, and assorted commentary on Malthus edited by
Appleman, Norton Critical Edition, ISBN 0-393-09202-X.
Verhulst P. F., (1838). Noticesur la loique la population poursuitdans son Accroissement,
Correspondance, athematiqueet physique, 10.
World Development Indicators and Global Development Finance (WDIGDF): Google Public
Data Explorer. http://www.google.comgh/publicdata/explore
Ofori, T, dkk. 2013. Mathematical Model of Ghana’s Population Growth, International
Journal of Modern Management Sciences. 2(2). 57-66.