Aplication Adams-Bashforth-Moulton Methods in Logistic

Equation in Prediction Population Growth in Province Daerah

Istimewa Yogyakarta
Az-zakira Fitri Firdaussy1, Muhammad Yasin2, Nia Karlina3, Yunifar Setyo R4
1
Departemen Of Nuclear and Physics Engineering, Universitas Gadjah Mada
2
Mrs. Fitrotun Aliya as Lecture of Numerical Method in Departemen of Nuclear and Physics
Engineering, Universitas Gadjah Mada
Undergraduate Student in Departemen of Nuclear and Physics Engineering, Universitas
Gadjah mada

Abstract – Population growth of a region is an indicator of the population density of the

area. Predicted population growth can be an effort to prevent problems related to population
density. In this paper a aplication of numerical methods in logistic equation in prediction
population growth in province daerah istimewa Yogyakarta, to solve this case, population
growth can be described by logistic model which is a first order non-linear differential
equation. For example p(t) is the total population in a city. Birth and death rates proportional
to the current population. The numerical Method used to solve Runge-Kutta 4th Order
Method for Ordinary Differential Equations to derive four initial solutions, followed by the
Adams-Bashforth-Moulton method It’s consists of two step, predictor and corrector. Adam-
Bashforth Moulton method becomes predictor and Adams-Moulton method as corrector. The
data used is the data of the population provice Daerah Istimewa Yogyakarta in 2005 to 2016
taken from the Central Statistics Agency of Province Daerah Istimewa Yogyakarta.
Keyword: Ordinary Differential Equation, Adam-Bashforth Moulton method, Runge
Kutta Method.
I. INTRODUCTION
The population is all persons domiciled in period of time. The value of the population
an area for 6 months or more and or those growth can be an indicator of an area will
who live less than 6 months but aims to the population density of that area because
settle. As for population growth is the of the amount of population change while
change in the population resident at any the area of the fixed area. Predictions of
time, and can be calculated as the change in population growth in the future can help the
the number of individuals in a population prevention of problems related to
per unit of time. Population growth can be population problem such as the problem of
interpreted as a number that indicates the the narrowing of the land of residence,
percentage of the population in a certain employment, poverty, preservation of
nature as well as other problems related to higher order ordinary differential
population growth and population density equations or coupled (simultaneous)
of an area. In this paper the question area is differential equations.
the Province of Daerah
IstimewaYogyakarta.
The initial phase of numerical settlement is
II. METHODOLOGY by specifying the points in the distance the
same in the interval [a,b], is by applying
In this paper we used two equation, the first
we use Runge-Kutta Method to solve 𝑥𝑟 = 𝑥0 + 𝑟ℎ
numerical differential equations or Where r = 1, 2,3,...n
approach to solve more significat. And the
second we used Adam-Basforth-Moulton to Where ℎ represents the distance between
correct the first step(Runge-Kutta that point formulated by
method),because Runge-Kutta method has 𝑏−𝑎
large error. ℎ=
ℎ

RUNGE-KUTTA METHOD Runge-Kutta Methods is a method to

numerically solve differential equation,
Runge Kutta method is a method used to
solve numerical differential equations or its given by
approach to solve more significant. The 1
Runge Kutta method is a composite of a 𝑃𝑡+1 = 𝑃𝑡 + (𝑘1 + 2𝑘2 + 2𝑘3 + 𝑘4 )
6
large class method of one-step approach
(Euler, Heun, and midpoint). This method Where,
achieves the accuracy of a Taylor approach 𝑘1 = 𝑓(𝑡𝑡 , 𝑃𝑡 )
without requiring high-level derivatives.
Population growth can be decribed by 1 1
𝑘2 = 𝑓 (𝑡𝑡 + ℎ , 𝑃𝑡 + 𝑘1 ℎ)
logictic model which is first order non linier 2 2
differential. The numerical method use to 1 1
solve the differential equation is Runge- 𝑘3 = 𝑓 (𝑡𝑡 + ℎ , 𝑃𝑖 + 𝑘2 ℎ)
2 2
kutta 4th order method to derive first
solution 𝑘4 = 𝑓(𝑡𝑡 + ℎ , 𝑃𝑡 + 𝑘3 ℎ)

Runge-Kutta 4th order method is a ADAM-BASFORTH-MOULTON

numerical technique used to solve ordinary METHOD
differential
equation of the form In this paper, population growth can be
modeled as logistic mode, where it was
𝑑𝑥 included in the first order of non-linear
= 𝑓(𝑥, 𝑦), 𝑎 ≤ 𝑥 ≤ 𝑏
𝑑𝑡 Differential Equation i.e :
𝑦(𝑥)𝑜= 𝑦0 dP P
P′ = = m (1 − ) P = f(t, P)
So only first order ordinary differential dt K
equations can be solved by using the where P(t0) = P0 at the interval [a,b] with
Runge-Kutta 4th order method. In other the size oh ‘h’ remains the same and tr+1 =
sections, we have discussed how Euler and tr + h.
Runge-Kutta methods are used to solve
We use the fourth order of Runge Kuta BPS Data for Population Size of DIY
Methode to calculate the first four Year Population
solutions. Proceed with Adam-Bashforth- 2005 3,281,800
Moulton Methode to calculate the numericl 2006 3,325,850
solutions, we repeat it until the ‘r’ iteration 2007 3,359,404
fullfil : 2008 3,393,003
(1) (0)
|Pr+1 − Pr+1 | 2009 3,426,637
(1)
<ε 2010 3,457,491
|Pr+1 |
2011 3,509,997
This is done because if in predicting the 2012 3,552,462
population growth is done only with Runge
2013 3,594,854
Kuta Methode, the results obtained have
2014 3,637,116
large errors, so we need a methode to
correct this, namely a multistep methode, 2015 3,679,176
Adam-Bashforth-Moulton Methode. Adam- 2016 3,710,443
Bashforth-Moulton Methode itself is a Data Population growth from BPS(2005-
method of multistep that is, where it 2016)
consists of two step, namely prediction and Population growth modeled as logistic function
correction. Method Adam-Bashforth-
𝑑𝑃
Moulton requires three initial values
because the approach is done with the 𝑑𝑡
𝑃
Lagrange interpolation of degree three. The = 𝑚 (1
where, − )𝑃
𝐾
results of interpolation into an predictor 1 𝑃(𝑡)
equation. 𝑚 = ln ( )
𝑡 𝑃(0)
In order to predict yr+1 we use the predictor
equation (first approximation). Predictor
equation : P is population size, K is capacity constant
∗
h and m is population growth constant.
yr+1 = yr + (−9fr−3 + 37fr−2 − 59fr−1
24
+ 55fr )
When we want to get the value of the The value of population growth constant
correction of yr+1 we use the corrector Year t Population m
equation.. Corrector equation : 2005 0 3,281,800 -
h 𝑑𝑃
2006 1= 0.01143,325,850 𝑃
0.0133
yr+1 = yr + (fr−2 − 5fr−1 (1 − )𝑃
24
∗
2007 𝑑𝑡 2 3,359,404140000000
0.0117
+ 19fr + 9yr+1 ) 2008 3 3,393,003 0.0111
2009 4 3,426,637 0.0108
2010 5 3,457,491 0.0104
2011 6 3,509,997 0.0112
III. RESULT AND DISCUSSION 2012 7 3,552,462 0.0113
2013 8 3,594,854 0.0114
The data population growth in 2014 9 3,637,116 0.0114
Yogyakarta ,we get from BPS(Badan 2015 10 3,679,176 0.0114
Pusat Statistik). We get data population 2016 11 3,710,443 0.0112
growth from 2005 to 2016, and we analyze (m) is the average value from data 0.0114.
it. To predict population growth use the Value of capacity constant (K) assumed to
manual calculate of Runge-Kutta,Adam-
Basforth-Moulton and we using excel to be 140000000. Value of m and K
calculate the data. substituted into logistic function and
became the mathematical model of ℎ
𝑃𝑡+1 ∗ = 𝑃𝑡 + (−9𝑓𝑡−3 + 37𝑓𝑡−2
population growth. 24
𝑃 − 59𝑓𝑡−1 + 55𝑓𝑡 )
𝑓(𝑡, 𝑃) = 0.0114 (1 − )𝑃
140000000 After that, the predicted solution substituted
into Adam-Moulton equation to get the
𝑃 corrected solution.
𝑓(𝑡, 𝑃) = 0.0114 (1 − )𝑃
140000000
ℎ
𝑃𝑡+1 = 𝑃𝑡 + (𝑓 − 5𝑓𝑡−1 + 19𝑓𝑡
24 𝑡−2
+ 9𝑓𝑟+1 ∗ )
Where,
Four first solution of the mathematical
model solved using 4th order Runge-Kutta 𝑃𝑡
𝑓𝑡 = 0.0114 (1 − )𝑃
method. 140000000 𝑡

With f(0,P)= 3,281,800 and step size h=1 The relative error (𝜀) calculated and
four first solution counted. iteration could be continued if error less
than 5 x 10-9
1
𝑃𝑡+1 = 𝑃𝑡 + (𝑘1 + 2𝑘2 + 2𝑘3 𝑃𝑡+1 −𝑃𝑡+1 ∗
6 𝜀=| |
With + 𝑘4 ) 𝑃𝑡+1
𝑘1 = 𝑓(𝑡𝑡 , 𝑃𝑡 )
1
𝑘2 = 𝑓 (𝑡𝑡 + ℎ , 𝑃𝑡
2 Numerical solution show in table bellow
1
+ 𝑘1 ℎ)
2 Year t Pt
1 1
𝑘3 = 𝑓 (𝑡𝑡 + ℎ , 𝑃𝑖 + 𝑘2 ℎ) 2005 0 3,281,800
2 2
𝑘4 = 𝑓(𝑡𝑡 + ℎ , 𝑃𝑡 + 𝑘3 ℎ) 2006 1 3,318,499
2007 2 3,355,599
Growth constant (m) = 2008 3 3,393,103
Capacity constant (K) = 2009 4 3,431,015
2010 5 3,469,341
Year t P(t) 2011 6 3,508,083
2012 7 3,547,248
2005 0 3,281,800
2013 8 3,586,837
2006 1 3,318,499
2014 9 3,626,857
2007 2 3,355,599
2015 10 3,667,311
2008 3 3,393,103
2016 11 3,708,205
2017 12 3,749,542
First four numerical solution shown at table 2018 13 3,791,326
above 2019 14 3,833,564
2020 15 3,876,258
The four first solution substituted into
2021 16 3,919,414
Adam-Bashford equation to get the
2022 17 3,963,037
predicted solution
2023 18 4,007,131
2024 19 4,051,701
 2025 20 4,096,752
2026 21 4,142,288 IV. CONCLUSION
2027 22 4,188,315
2028 23 4,234,838 With Adam Bashforth-Moulton Method in
2029 24 4,281,861 logistic equation we can predicting the
2030 25 4,329,389 population’s growth.
Adam Bashforth-Moulton Method in
logistic equation in prediction population
The predicted data compared into real data
size in provinsi DIY show an increase of
and the error calculated
population size every years.
Year t Real data Prediction Error Population Growh in Yogyakarta always
2005 0 3,281,800 3,281,800 - increasing. The paper aim to predict
2006 1 3,325,850 3,318,499 0.222% population growth in next years. From the
2007 2 3,359,404 3,355,599 0.133% result and disccussion, we get population
2008 3 3,393,003 3,393,103 0.003% growth from 2017 until 2030. Population
2009 4 3,426,637 3,431,015 0.128% growh in 2017 is 3.749.542 , in 2018 is
2010 5 3,457,491 3,469,341 0.342% 3.791.326,and etc. In 2018 population
2011 6 3,509,997 3,508,083 0.055% growth increase 41.784 people.
2012 7 3,552,462 3,547,248 0.147%
2013 8 3,594,854 3,586,837 0.224% V. REFERENCE
2014 9 3,637,116 3,626,857 0.283%
2015 10 3,679,176 3,667,311 0.324% Bayu Prihandono, Apriadi, dan Evi
3,710,443 0.060% Noviani, “Metode Adams-
2016 11 3,708,205
BashforthMoulton dalam Penyelesaian
Persamaan Diferensial Non Linear”,
Graph Population growth we get from excel
Buletin Ilmiah Mat. Stat dan terapanya, no.
2(2014).
4,000,000 Population size
http://www.jurnal-msa.com
3,500,000

3,000,000
2000 2005 2010 2015 2020