See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326177816

A Comparative Study on Numerical Solution of Initial Value Problem by Using

Euler's Method, Modified Euler's Method and Runge-Kutta Method

Article in Journal of Computer and Mathematical Sciences · May 2018

DOI: 10.29055/jcms/784

CITATIONS READS

25 5,631

2 authors:

Md Kamruzzaman C.N. Mithun

Presidency University, Dhaka Presidency University, Dhaka
2 PUBLICATIONS 29 CITATIONS 5 PUBLICATIONS 72 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Md Kamruzzaman on 04 July 2018.

The user has requested enhancement of the downloaded file.

Journal of Computer and Mathematical Sciences, Vol.9(5), 493-500 May 2018 ISSN 0976-5727 (Print)
(An International Research Journal), www.compmath-journal.org ISSN 2319-8133 (Online)

A Comparative Study on Numerical Solution of Initial

Value Problem by Using Euler’s Method, Modified
Euler’s Method and Runge-Kutta Method
Md. Kamruzzaman1 and Mithun Chandra Nath2
1
Lecturer, Department of Civil Engineering
Presidency University, Dhaka, BANGLADESH.
2
Lecturer, Primeasia University, Dhaka, BANGLADESH.
email: mdkamruzzaman94@gmail.com, mithundevftb@gmail.com.

(Received on: May 16, 2018)

ABSTRACT

In this article, we discussed the numerical solutions of ordinary differential

equations with initial value problems by Euler’s Method, Modified Euler’s Method
and Runge- Kutta Methods. Here the solutions of some numerical examples have been
obtained with the help of MATLAB program as well as we determined the exact
analytic solutions. In order to validate the accuracy, we compared the exact solutions
with the numerical approximate solutions. We observed that when the step size is very
small the accuracy of the solution become more correct. And the relative error is the
difference of the approximate solution and analytic solution. Among these three
proposed methods we observed that only the relative error is nominal for Runge-Kutta
fourth order method.

Keywords: Initial value problem, Euler method, Modified Euler’s method, Runge-
Kutta method, Error analysis.

INTRODUCTION

In mathematical modeling differential equations are generally used in the field of

science and engineering. In the field of mathematical physics different problems are arise as
the form of differential equations. These types of differential equations may the formation
either ordinary differential equations or partial differential equations. Practically, most of the
models of the problem which are formulated by means of these equations are so complicated
to determine the exact solution and one of two approaches is taken to approximate the solution.

493
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

The first technique is to reduce the differential equation in to one that can be solved exactly
and then use the result of the reduced equation to approximate the solution to the original
problem. Another technique, which we will verify in this article, uses methods of
approximations the solution of original problem. This is the technique that is generally taken
as the approximation methods give more perfect results and relative error information.
Numerical methods are usually used for solving mathematical problems that are articulated in
the field of science and engineering in which case the determination of the exact solution is so
hard or impossible. Only a few numbers of differential equations can be solved analytically.
Consequently, to obtain the analytical solution for differential equation there exist different
methods. A huge number of differential equations are unable to determine the solution in
closed form using familiar analytical methods, in which case we apply numerical technique
for solving a differential equation under certain initial restriction or restrictions. There exist
different kinds of practical numerical methods for finding the solution of initial value problem
of ordinary differential equations. In this research article we present three outstanding
numerical methods such as Euler’s method, Modified Euler’s method and Runge-Kutta
method for finding the solution of initial value problems for ordinary differential equations.
From the literature analysis, we may understand that numerous works in numerical
solutions of initial value problems applying Euler’s method, Modified Euler’s method and
Runge-Kutta method have been performed. Various authors have endeavored to solve initial
value problems to determine high correctness promptly by using several methods, such as
Euler’s method, Modified Euler’s method and Runge-Kutta method and so on. In1 the author
deliberated the accurate solution of initial value problems for ordinary differential equation by
using Runge-Kutta fourth order method. In2-3 discussed on accurate solution of initial value
problems for ordinary differential equations.3 considered on some numerical methods for
solving initial value problems in ordinary differential equations. Also 4-16 studied numerical
solutions of initial value problems for ordinary differential equations using a variety of
numerical methods. In this work Euler method, Euler modified method and Runge Kutta
method are performed without any discretization, alteration or limiting assumption for solving
initial value problems of ordinary differential equations. The Euler method is conventionally
the first numerical method. It is extraordinarily easy to realize and geometrically simple to
expressive however not very handy; the technique has restricted accuracy for further intricate
functions. A more vigorous and complicated numerical method is the Runge Kutta method.
This method is the most commonly used one since it provides consistent initial values and is
especially suitable when the calculation of higher derivatives are intricate. The numerical
outcomes are very hopeful. As a final point, a pair of example including different kinds of
ordinary differential equations is given to validate the anticipated formula. The solutions of
each numerical illustration specify that the convergence and error analysis which are discussed
demonstrate the effectiveness of the methods. In case of Euler method, Euler modified method
to determine the solution of the differential equation numerically is less resourceful since it
requires h to be tiny for finding logical accuracy. However in Runge-Kutta method, the
derivatives of superior order are not prerequisite and they are intended to provide better
494
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

accuracy with the benefit of requiring only the functional values at some preferred points on
the sub-interval. Runge- Kutta method is a more common and spontaneous method as
compared to that of the Euler method, Euler modified method. We detect that in the Euler
method, Euler modified method extremely small step size converges to analytical solution. So,
more number of approximation is needed. On the contrary, Runge Kutta method gives
improved results and it converges closer to analytical solution and has less iteration to get
exactitude solution. This article is rearranged as follows: segment 2: problem formulations;
segment 3: error analysis; segment 4: numerical examples; segment 5: discussion of results;
and the final segment: the conclusion of the article.
SEGMENT:2 Problem Formulation
In this segment we consider three numerical methods for obtaining the numerical
solutions of the initial value problem (IVP) of the first-order ordinary differential equation is
of the form
= , ( ) , ( , ), ( ) = (1)
where = and , ( ) is the given function and ( ) is the solution of the equation
(1). In this article, we find the solution of this equation on a finite interval ( , ), starting
with the initial point . A continuous approximation to the solution ( ) will not be obtained;
instead, approximations to y will be generated at various values, called mesh points, in the
interval ( , ). Numerical methods employ the Equation (1) to obtain approximations to the
values of the solution corresponding to various selected values of the solution corresponding
to different selected values of = = + h, = 0,1,2, … … . The parameter h is called
the step size. The numerical solution of (1) is given by a set of points {( , ): =
0,1,2,3, … … } and each point ( , ) is an approximation to the corresponding point
{ , ( )} on the solution curve.
2.1 Euler Method
Euler’s method is the simplest one-step method. It is basic explicit method for
numerical integration of ordinary differential equations. Euler proposed his method for initial
value problems (IVP) in 1768. It is first numerical method for solving IVP and serves to
illustrate the concepts involved in the advanced methods. It is important to study because the
error analysis is easier to understand. The general formula for Euler approximation is
( ) = ( ) + ℎ ( , ); = 0,1,2, … …

2.2 Modified Euler Method

In this method the curve in the interval ( , ) where = + ℎ is approximated

by the line through ( , ) with the slope + , ( , ) , which is the slope as
the middle point whose abscissa is average of and . A generalized form of Euler’s
Modified formula is

495
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

= +ℎ ( + , + ( , )

2.3 Runge-Kutta Method

This method was devised by two German mathematicians, Runge about 1894 and
extended by Kutta a few years later. The Runge-Kutta method is most familiar because it is
pretty accurate, steady and simple to program. This method is notable by their order in the
logic that they concur with Taylor’s series solution up to terms of ℎ where r is the order of
the method. It do not require previous computational of higher derivatives of ( ) as in
Taylor’s series method. The fourth order Runge-Kutta method (RK4) is broadly used for
solving initial value problems (IVP) for ordinary differential equation (ODE). The general
formula for Runge-Kutta method is
( +2 +2 + )
= + 6 = 0,1,2, … …
=ℎ ( , )
=ℎ + , +
ℎ
=ℎ + , +
2 2
=ℎ ( + ℎ, + )
SEGMENT: 3 Error Analysis
There exist two kinds of errors in numerical solution of ordinary differential equations.
Round-off errors and Truncation errors happen when ordinary differential equations are solved
numerically. Rounding errors initiate from the verity that computers can only characterize
numbers using a fixed and restricted number of important figures. Thus, such numbers cannot
be represented accurately in computer memory. The inconsistency introduced by this
restriction is call Round-off error. Truncation errors in numerical study occur when
approximations are used to determine a number of quantities. The exactness of the solution
will rely on how miniature we make the step size, h. A numerical method is said to be
convergent if
→
| ( ) − | = 0.
Where y(xn) denotes the approximate solution and yn denotes the exact solution. In this work
we consider two initial value problems to examine accuracy of the proposed methods. The
Approximated solution is determined by using MATLAB software for three proposed
numerical methods at different step size. The maximum error is defined by
= → | ( )− |

SEGMENT: 4 Numerical Examples

In this section we consider two numerical examples to verify which numerical

methods converge faster to analytical solution. Numerical solutions and errors are computed.

496
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

Example 1: We consider the initial value problem ( ) = + , (0) = 1 on the

interval 0 ≤ ≤ 1. Then the exact solution of the given problem is as ( ) =
/ /
erf +e − x . The approximate solutions and maximum errors are obtained
√
and shown in tables 1(a)-(d).
Table 1 (a) Numerical approximations and maximum errors for step; size h=0.1; (b)
Numerical approximations and maximum errors for step size h=0.05; (c) Numerical
approximations and maximum errors for step size h=0.025; (d) Numerical approximations and
maximum errors for step size h=0.0125.

1(a): Numerical approximations and maximum errors for step; size h=0.1;
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 1.0000000000000000 5.34652E−03 1.005250000000000 9.65218E-05 1.0053464802083334 4.16045E−08 1.0053465218128410
0.2 1.0110000000000000 1.18895E−02 1.022661643750000 0.000227819 1.0228893798037348 8.26716E−08 1.0228894624752929
0.3 1.0352199999999998 1.99720E−02 1.054783850254688 0.000408114 1.0551918407370900 1.23029E−07 1.0551919637660336
0.4 1.0752765999999998 3.00424E−02 1.104662546534985 0.000656406 1.1053187896458685 1.63325E−07 1.1053189529706604
0.5 1.1342876640000000 4.26873E−02 1.175976557420941 0.000998415 1.1769747667144460 2.05805E−07 1.1769749725189769
0.6 1.2160020472000000 5.86769E−02 1.273209735845547 0.001469256 1.2746787363539485 2.55624E−07 1.2746789919776722
0.7 1.3249621700320000 7.90261E−02 1.401871127660406 0.002117191 1.4039879953710888 3.23030E−07 1.4039883184007750
0.8 1.4667095219342400 1.05078E−01 1.568778873945045 0.003008896 1.5717873427344033 4.26941E−07 1.5717877696756601
0.9 1.6480462836889793 1.38620E−01 1.782428926401787 0.004236927 1.7866652528501639 6.00769E−07 1.7866658536190383
1.0 1.8773704492209877 1.82037E−01 2.053477058070325 0.005930347 2.0594065035273252 9.01815E−07 2.059407405342576

1(b) : Numerical approximations and maximum errors for step size h=0.05;
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 1.002625000000000 2.72152E−03 1.0052500 00000000 9.65218E-05 1.00534648 0208333 4.1605E-08 1.00534652 18128410
0.2 1.0168241609375002 6.06530E−03 1.022661643750000 0.000227819 1.0228893798037348 8.2672E-08 1.0228894624752929
0.3 1.0449798075787111 1.02122E−02 1.054783850254688 0.000408114 1.0551918407370900 1.2303E-07 1.0551919637660336
0.4 1.0899197085245087 1.53992E−02 1.104662546534985 0.000656406 1.1053187896458685 1.6332E-07 1.1053189529706604
0.5 1.1550367600056364 2.19382E−02 1.175976557420941 0.000998415 1.1769747667144460 2.058E-07 1.1769749725189769
0.6 1.244439027678436 3.02400E−02 1.273209735845547 0.001469256 1.2746787363539485 2.5562E-07 1.2746789919776722
0.7 1.3631397949603248 4.08485E−02 1.401871127660406 0.002117191 1.4039879953710888 3.2303E-07 1.4039883184007750
0.8 1.517300301075834 5.44875E−02 1.568778873945045 0.003008896 1.5717873427344033 4.2694E-07 1.5717877696756601
0.9 1.7145419864264193 7.21239E−02 1.782428926401787 0.004236927 1.7866652528501639 6.0077E-07 1.7866658536190383
1.0 1.9643507036668488 9.50567E−02 2.053477058070325 0.005930347 2.0594065035273252 9.0182E-07 2.059407405342576

1(c): Numerical approximations and maximum errors for step size h=0.025;
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 1.0039732143920899 1.37331E-03 1.0053404 3378088 6.08803E-06 1.0053465216505684 1.6228E-10 1.005346521812841
0.2 1.0198254164252103 3.06405E−03 1.022875020804634 1.44417E-05 1.0228894621535374 3.2176E-10 1.0228894624752929
0.3 1.050026859341876 5.16510E−03 1.055165966239071 2.59975E-05 1.0551919632892464 4.7679E-10 1.0551919637660336
0.4 1.097520387412045 7.79857E−03 1.105276938021313 4.20149E-05 1.105318952342068 6.286E-10 1.1053189529706604
0.5 1.1658497569818174 1.11252E−02 1.176910764603475 6.42079E-05 1.1769749717346203 7.8435E-10 1.1769749725189769
0.6 1.259321437932899 1.53576E−02 1.274584063574121 9.49284E-05 1.2746789910151588 9.6252E-10 1.2746789919776722
0.7 1.3832106899613061 2.07776E−02 1.403850894637291 0.000137424 1.4039883171991199 1.2017E-09 1.403988318400775
0.8 1.5440262079869167 2.77616E−02 1.571591569133329 0.000196201 1.5717877681003285 1.5753E-09 1.5717877696756601
0.9 1.7498524246222582 3.68134E−02 1.786388311742869 0.000277542 1.786665851402671 2.2164E-09 1.7866658536190383
1.0 2.0107951384702343 4.86123E−02 2.059017158076531 0.000390247 2.0594074019860655 3.3565E-09 2.059407405342576

497
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

1 (d): Numerical approximations and maximum errors for step size h=0.0125.
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 1.0046566697375803 6.89852E−04 1.0053404 3378088 6.08803E-06 1.0053465218027011 1.01399E−11 1.005346521812841
0.2 1.0213494287582197 1.54003E−03 1.022875020804634 1.44417E-05 1.0228894624551952 2.00999E−11 1.0228894624752929
0.3 1.0525943319732851 2.59763E−03 1.055165966239071 2.59975E-05 1.0551919637362754 2.97600E−11 1.0551919637660336
0.4 1.1013943547977403 3.92460E−03 1.105276938021313 4.20149E-05 1.1053189529314784 3.91900E−11 1.1053189529706604
0.5 1.1713723532944522 5.60262E−03 1.176910764603475 6.42079E-05 1.1769749724701777 4.88001E−11 1.1769749725189769
0.6 1.2669392048911032 7.73979E−03 1.274584063574121 9.49284E-05 1.274678991917932 5.97400E−11 1.2746789919776722
0.7 1.3935085610750229 1.04798E−02 1.403850894637291 0.000137424 1.403988318326378 7.44000E−11 1.403988318400775
0.8 1.5577734062756774 1.40144E−02 1.571591569133329 0.000196201 1.571787769578305 9.73599E−11 1.5717877696756601
0.9 1.7680647857193632 1.86011E−02 1.786388311742869 0.000277542 1.786665853482107 1.36930E−10 1.7866658536190383
1.0 2.034820184163635 2.45872E−02 2.059017158076531 0.000390247 2.0594074051349014 2.07670E−10 2.059407405342576

Example 2: We consider the initial value problem = − , (0) = 1 on the interval

0≤ ≤ 1. The exact solution of the given problem is ( ) = . The approximate

√
√
solutions and maximum errors are obtained and shown in tables 2(a)-(d).
2(a): Numerical approximations and maximum errors for step; size h=0.1
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 0.9000000000000000 1.35091E−02 0.9145000 0000000 0.000990872 0.9135089320204528 1.95878E−07 0.913509127898782
0.2 0.8280000000000001 2.12185E−02 0.850700872600742 0.001482354 0.8492181710604544 3.47642E−07 0.849218518702443
0.3 0.7760016000000001 2.58218E−02 0.803539722296333 0.001716324 0.8018229448618213 4.53096E−07 0.8018233979576023
0.4 0.7390637996797441 2.87198E−02 0.769598236661347 0.001814651 0.7677830621260258 5.24033E−07 0.7677835861595071
0.5 0.7140048216672278 3.06849E−02 0.746530852829335 0.001841152 0.7446891282464022 5.72232E−07 0.7446897004786337
0.6 0.6987247742141842 3.21636E−02 0.732718569364669 0.001830167 0.730887796150559 6.06628E−07 0.7308884027785085
0.7 0.691826629656969 3.34247E−02 0.727051749410847 0.00180045 0.725250665872579 6.33400E−07 0.7252512992720983
0.8 0.6923920851827047 3.46350E−02 0.728789350891342 0.001762265 0.7270264295821635 6.56635E−07 0.7270270862176577
0.9 0.6998427720349557 3.59008E−02 0.737464721311611 0.001721133 0.7357429095800243 6.78965E−07 0.7357435885449581
1.0 0.7138506309611446 3.72897E−02 0.752820255668837 0.001679904 0.751139649932897 7.02025E−07 0.7511403519579868
2(b): Numerical approximations and maximum errors for step size h=0.05;
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 0.9072500000000000 6.25913E−03 0.9137386 3169801 0.000229504 0.9135091213176563 6.58113E−09 0.913509127898782
0.2 0.8392609277701965 9.95759E−03 0.849562976531398 0.000344458 0.8492185048488676 1.38536E−08 0.849218518702443
0.3 0.7895884571598809 1.22349E−02 0.802223388565394 0.000399991 0.8018233782217195 1.97359E−08 0.8018233979576023
0.4 0.7540743267065178 1.37093E−02 0.768207590571715 0.000424004 0.7677835620326319 2.41269E−08 0.767783586159507
0.5 0.7299570754427195 1.47326E−02 0.745120887406124 0.000431187 0.7446896730961782 2.73825E−08 0.744689700478633
0.6 0.7153744093633583 1.55140E−02 0.731317887808574 0.000429485 0.7308883728944706 2.98840E−08 0.730888402778508
0.7 0.7090695033855021 1.61818E−02 0.725674565817607 0.000423267 0.7252512673367656 3.19353E−08 0.725251299272098
0.8 0.7102098229998883 1.68173E−02 0.727442024142068 0.000414938 0.7270270524626191 3.37550E−08 0.727027086217657

2(c) Numerical approximations and maximum errors for step size h=0.025;
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 0.9104875290532907 3.02160E−03 0.9137386 3169801 0.000229504 0.9135091276397532 2.59029E−10 0.913509127898782
0.2 0.8443847744779372 4.83374E−03 0.849562976531398 0.000344458 0.8492185180509749 6.51469E−10 0.849218518702443
0.3 0.7958587285640485 5.96467E−03 0.802223388565394 0.000399991 0.8018233969598182 9.97784E−10 0.8018233979576023
0.4 0.7610777171130718 6.70587E−03 0.768207590571715 0.000424004 0.7677835848899629 1.26955E−09 0.7677835861595071
0.5 0.7374639944962212 7.22571E−03 0.745120887406124 0.000431187 0.7446896989999765 1.47866E−09 0.7446897004786337
0.6 0.7232631491061028 7.62525E−03 0.731317887808574 0.000429485 0.7308884011344937 1.64401E−09 0.7308884027785085
0.7 0.7172840690076906 7.96723E−03 0.725674565817607 0.000423267 0.7252512974898684 1.78223E−09 0.7252512992720983
0.8 0.7187354764318908 8.29161E−03 0.727442024142068 0.000414938 0.7270270843117828 1.90588E−09 0.7270270862176577
0.9 0.7271193149751998 8.62427E−03 0.736149394556421 0.000405806 0.7357435865210894 2.02387E−09 0.7357435885449581
1.0 0.7421585135282368 8.98184E−03 0.751536906604233 0.000396555 0.7511403498157103 2.14228E−09 0.7511403519579868

498
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

2(d): Numerical approximations and maximum errors for step size h=0.0125.
Euler’s Method Euler’s Modified Runge-Kutta Method Exact Solution
Method
xn ER ER ER
0.1 0.9120236443372686 1.48548E−03 0.9135226 4673271 1.35188E-05 0.9135091278869912 1.17910E−11 0.913509127898782
0.2 0.846835976864134 2.38254E−03 0.849238872455001 2.03538E-05 0.8492185186679586 3.44851E−11 0.849218518702443
0.3 0.7988774812685904 2.94592E−03 0.801847096047129 2.36981E-05 0.8018233979021255 5.54771E−11 0.8018233979576023
0.4 0.7644662941812183 3.31729E−03 0.767808764293526 2.51781E-05 0.7677835860871474 7.23600E−11 0.7677835861595071
0.5 0.7411106710960523 3.57903E−03 0.744715355453644 2.5655E-05 0.7446897003930565 8.55770E−11 0.7446897004786337
0.6 0.7271075635837666 3.78084E−03 0.730913999935655 2.55972E-05 0.7308884026823441 9.61640E−11 0.7308884027785085
0.7 0.721297592738653 3.95371E−03 0.725276562956256 2.52637E-05 0.7252512991670091 1.05089E−10 0.7252512992720983
0.8 0.722909634547887 4.11745E−03 0.727051884359835 2.47981E-05 0.7270270861045542 1.13103E−10 0.7270270862176577
0.9 0.7314586485260627 4.28494E−03 0.735767867647858 2.42791E-05 0.7357435884242071 1.20751E−10 0.7357435885449581
1.0 0.7466759122017882 4.46444E−03 0.751164100109758 2.37482E-05 0.751140351829581 1.28405E−10 0.7511403519579868

SEGMENT: 5 DISCUSSION OF RESULTS

The results which are obtained shown in Tables 1(a)-(d) and Tables 2(a)-(d). The
numerical approximate solution is calculated with step sizes 0.1, 0.05, 0.025 and 0.0125 and
utmost errors also are evaluated at particular step size. From the tables for each technique we
state that a numerical solution converges to the accurate solution if the step size leads to
decreased errors such that in the limit when the step size to zero the errors also tends to zero.
We saw that the Euler method, Euler modified method iterations using the step size 0.1 and
0.05 didn’t converge to exact solution but for step size 0.025 and 0.0125 converge gradually
to exact solution. Also we observed that the Runge-Kutta approximations for same step size
converge firstly to exact solution. This shows that the small step size gives the improved
estimation. The Runge-Kutta method of order four requires four evaluations per step, so it
should give more accurate results in compare with Euler method, Euler modified method with
one-fourth the step size if it is to be superior. Lastly we observed that the fourth order Runge-
Kutta method is converging quicker than the Euler method, Euler modified method and it is
the best efficient method for solving initial value problems for ordinary differential equations.

SEGMENT: 6 CONCLUSION

In this paper, Euler improved, Euler modified method and Runge-Kutta method are
used for solving ordinary differential equation (ODE) in initial value problems (IVP). Finding
more accurate results needs the step size smaller for all methods. From the figures we can see
the accuracy of the methods for decreasing the step size h. The numerical solutions obtained
by the three proposed methods are in good agreement with exact solutions. Comparing the
results of the three methods under investigation, we observed that the rate of convergence of
Euler improved, Euler modified method is O(h) and the rate of convergence of fourth-order
Runge Kutta method is O(h ) . The Euler improved, Euler modified method was found to be
less accurate due to the inaccurate numerical results that were obtained from the approximate
solution in comparison to the exact solution. From the study the Runge-Kutta method was
found to be generally more accurate and also the approximate solution converged faster to the
exact solution when compared to the Euler improved, Euler modified method. It may be
499
 Md. Kamruzzaman, et al., Comp. & Math. Sci. Vol.9 (5), 493-500 (2018)

concluded that the Runge Kutta method is powerful and more efficient in finding numerical
solutions of initial value problems (IVP).

REFERENCES

1. Islam, Md. A. Accurate solution of initial problems for ordinary differential equation with
fourth order Runge- Kutta method. Journal of Mathematical Research (2015).
2. Ogunrinde, R.B., Fadugba, S.E. and Okunlola, J.T. On Some Numerical Methods for
Solving Initial Value Problems in Ordinary Differential Equations. IOSR Journal of
Mathematics, 1, 25-31 (2012).
3. Islam, Md.A. Accuracy Analysis of Numerical solutions of Initial Value Problems (IVP)
for Ordinary Differential Equations (ODE). IOSR Journal of Mathematics, 11, 18-23
(2015).
4. Shampine, L.F. and Watts, H.A. Comparing Error Estimators for Runge-Kutta Methods.
Mathematics of Computation, 25, 445-455. http://dx.doi.org/10.1090/S0025-5718-1971-
0297138-9 (1971).
5. Eaqub Ali, S.M. A Text Book of Numerical Methods with Computer Programming.
Beauty Publication, Khulna (2006).
6. Akanbi, M.A. Propagation of Errors in Euler Method, Scholars Research Library. Archives
of Applied Science Research, 2, 457-469 (2010).
7. Kockler, N. (1994) Numerical Method for Ordinary Systems of Initial value Problems.
John Wiley and Sons, New York.
8. Lambert, J.D. Computational Methods in Ordinary Differential Equations. Wiley, New
York (1973).
9. Gear, C.W. Numerical Initial Value Problems in Ordinary Differential Equations.
Prentice-Hall, Upper Saddle River (1971).
10. Hall, G. and Watt, J.M. Modern Numerical Methods for Ordinary Differential Equations.
Oxford University Press, Oxford (1976).
11. Hossain, Md.S., Bhattacharjee, P.K. and Hossain, Md.E. Numerical Analysis. Titas
Publications, Dhaka (2013).
12. Balagurusamy, E. Numerical Methods. Tata McGraw-Hill, New Delhi (2006).
13. Sastry, S.S. Introductory Methods of Numerical Analysis. Prentice-Hall, India (2000).
14. Burden, R.L. and Faires, J.D. Numerical Analysis. Bangalore, India (2002).
15. Gerald, C.F. and Wheatley, P.O. Applied Numerical Analysis. Pearson Education, India
(2002).
16. Ogunrinde, R.B., Fadugba, S.E. and Okunlola, J.T. On Some Numerical Methods for
Solving Initial Value Problems in Ordinary Differential Equations. IOSR Journal of
Mathematics (2012).

500

View publication stats