• No results found

The Effect of Richardson Extrapolation for Solving IVP on Accuracy Using Matlab Software

N/A
N/A
Protected

Academic year: 2022

Share "The Effect of Richardson Extrapolation for Solving IVP on Accuracy Using Matlab Software"

Copied!
11
0
0

Loading.... (view fulltext now)

Full text

(1)

ISSN 2319-8133 (Online)

(An International Research Journal), www.compmath-journal.org

The Effect of Richardson Extrapolation for Solving IVP on Accuracy Using Matlab Software

Yilekal Yewlu Taddesse

Ambo University Woliso Campus, ETHIOPIA.

email:[email protected]

(Received on: September 27, 2018) ABSTRACT

Initial value problems are frequently occurring in mathematical models that arise in many branches of science, engineering and economics. The study focuses on solving initial value problems with a combination of Richardson extrapolation method with Euler’s, Runge kutta 2 nd , 3 rd and 4 th , order methods using MATLAB program. The amount of error obtained with each numerical methods and their combinational result with Richardson extrapolation was analyzed. The numerical results of each method were expressed using tables and graphs. In order to achieve higher accuracy in the solution of initial value problems it is better to use Runge Kutta 4 th order method together with Richardson extrapolation method.

Keywords: Initial Value Problem (IVP), Euler Method, Higher Order of Runge Kutta Method, Truncation error(TE), MATLAB Software, Richardson Extrapolation.

1. INTRODUCTION

Differential equations are mathematical descriptions of how the variables and their derivatives (rates of change) with respect to one or more independent variable affect each other in a dynamical way. Their solutions show us how the dependent variable(s) will change with the independent variable(s).

Ordinary differential equations are frequently occurring in mathematical models that arise in many branches of science, engineering and economics. Unfortunately it is seldom that these equations have solutions which can be expressed in closed form, so it is common to seek approximate solutions by means of numerical methods. Nowadays this can usually be achieved very inexpensively to high accuracy and with a reliable bound on the error between the analytical solution and its numerical approximation 1 .

Richardson extrapolation is an accelerating method to improve the rate of

convergence of a sequence of solutions. It is named after Lewis Fry Richardson, who

(2)

introduced the technique in the early 20th century 2 , which can successfully be used in the efforts to improve the accuracy of the approximate solutions of systems of ordinary differential equations (ODEs) obtained by different numerical methods 3 .

Richardson extrapolation can be used numerical methods such as Runge kutta 2 nd , 3 rd and 4 th , Euler’s forward and backward, method, Adam’s-moulton etc 4 . Richardson extrapolation does not depend too much on the particular numerical method to solve initial value problems 3 .

Due to this the study focuses on solving initial value problems with a combination of Richardson extrapolation method with Euler’s, Runge kutta 2 nd , 3 rd and 4 th , order methods using MATLAB program.

1.1 Euler’s Method

Euler’s method is the simplest numerical method for solving initial value problem in ordinary differential equation. By taking different value of 𝜃 in equation𝑦 𝑛+1 = 𝑦 𝑛 + ℎ[(1 − 𝜃)𝑓(𝑡 𝑛 , 𝑦 𝑛 ] + 𝜃𝑓(𝑡 𝑛+1 , 𝑦 𝑛+1 ), 𝜃 ∈ [0, 1] this method divided in to three, for 0 , 0.5 and 1 are the properties of Euler’s forward method, trapezoidal rule and Euler’s backward method respectively.

In this work we shall only consider forward Euler’s method, it is one of the single step method expressed as y n 1 y n hf t y ( , n n ) , Where h is the step size. According to this equation the slope estimate f is used to extrapolate from an old value y n to a new value y n 1 over a distance h 5 .

1.2 Runge Kutta Methods

The Runge-Kutta’s method is popular in Engineering, especially in the numerical analysis due to its degree of approximation, in which it allows numerical calculus with more accuracy due to having second, third and fourth order formulation, and if we compared Euler’s, Rung kutta 2 nd , Rung kutta 3 rd and Runge kutta 4 th then Runge kutta 4 th is the most accurate and the methods were executed using MATLAB 6 .

a. Runge kutta second order method.

𝑦 𝑛+1 = 𝑦 𝑛 +

2 (𝑘 1 + 𝑘 2 ), Where 𝑘 1 = 𝑓(𝑥 𝑖 , 𝑦 𝑖 ) and 𝑘 2 = 𝑓(𝑥 𝑖 + ℎ, 𝑦 𝑖 + 𝑘 1 ℎ) b. Runge kutta third order method

𝑦 𝑛+1 = 𝑦 𝑛 + ℎ

6 (𝑘 1 + 4𝑘 2 + 𝑘 3 )

Where 𝑘 1 = 𝑓(𝑥 𝑖 , 𝑦 𝑖 ) , 𝑘 2 = 𝑓 (𝑥 𝑖 + 2 , 𝑦 𝑖 + 2 𝑘 1 ) , 𝑘 3 = 𝑓(𝑥 𝑖 + ℎ, 𝑦 𝑖 + (2𝑘 2 − 𝑘 1 )ℎ) c. Runge kutta 4 th order method

𝑦 𝑛+1 = 𝑦 𝑛 + 1

6 (𝑘 1 + 2𝑘 2 + 2𝑘 3 + 𝑘 4 ) Where 𝑘 1 = 𝑓(𝑥 𝑖 , 𝑦 𝑖 ) ,

𝑘 2 = 𝑓 (𝑥 𝑖 + ℎ

2 , 𝑦 𝑖 + ℎ

2 𝑘 1 )

(3)

𝑘 3 = 𝑓 (𝑥 𝑖 + ℎ

2 , 𝑦 𝑖 + ℎ 2 𝑘 2 ) 𝑘 4 = 𝑓(𝑥 𝑖 + ℎ, 𝑦 𝑖 + ℎ𝑘 3 ) 1.3 Richardson extrapolation Method

Due to wide application of Richardson extrapolation method different attempts have been made to solve some practical problems in science and engineering. Let us consider the classical initial value problem for systems of s s ( 1) ordinary differential equations.

0

( , ), [ , ], , ( )

dy f t y t a b a b f a y

dt (1) Where, the unknown function y a b :[ , ] s is continuously differentiable, while right- hand-side function f (t, y) is continuous. However, it is often necessary to introduce much more restrictive assumptions when numerical methods of order p such as 2 nd , 3 rd and 4 th Runge kutta methods, Euler’s forward and backward, method, Adam’s-moulton are used in the treatment of (1).

In such cases it is necessary to assume that the function y is continuously differentiable up to order p .

Richardson Extrapolation can be introduced in the following way. Assume that [ , ]

t n a b is a time-point and given that y t ( ) n is the value of the exact solution of (1) at t t n . Assume also that two approximations of y t ( ) n have been obtained by applying the above numerical methods of order p and by using two time-step sizes h and 0.5h . More precisely, starting from a time-point (grid point) t t n 1 where t n 1 t n h , the two approximations are calculated by using first one large time-step and, after that, two small time-steps. Denoting these two approximations with z n and w n respectively, we can write:

( 1)

p p

n n

y t z h k o h

(2)

( ) o 1

2

p

p

n n

y t w h k h

(3) Where k is some quantity depending on the numerical method applied in the calculation of

z n and w n but not on the step size h. solving this gives

2 1

( ) ( )

2 1

p

n n p

n p

w z

y t o h

Define y = 2

2 1

p

n n

n p

w z

(4 )

(4)

It is clear that the approximation y n being of order h p 1 , will in general be more accurate than both z n and w n (at least the step size h is sufficiently small) 7 . The value of 𝑝 in (4) is the order of the numerical method.

2. ERROR ANALYSIS

There are two things to consider when applying numerical methods to solve a problem.

The first and most obvious is to obtain the approximation. The equally important second objective is to determine a safety factor for the approximation: some assurance, or at least a sense, of the accuracy of the approximation 8 . There are five types of errors can occur in the numerical solution of differential equations 9 . These are error in the initial data, algebraic errors, Truncation errors, Round of errors and inherited errors, however mostly the first two are nonexistence. Truncation error is the error occurred in a single step caused by truncating the Taylor series approximations for the exact derivatives. Truncation error depends on the step size h and it decreases when the step size decreases. Inherited error is the sum of all accumulated errors from all steps, the presence of inherited error means that the initial condition for the next step is incorrect 9 . The error that is produced when a calculator or computer is used to perform real-number calculations is called round-off error. It occurs because the arithmetic performed in a machine involves numbers with only a finite number of digits; with the result that calculations are performed with only approximate representations of the actual numbers 8 . For the rest of the paper assume that error in the initial data and algebraic errors are nonexistence. The truncation errors of all the above methods are as follows: TE y t ( ) n y n

Where, 𝑦(𝑡 𝑛 ) is the exact and 𝑦 𝑛 is approximation values.

3. METHODOLOGY 3.1 Materials

For solving initial problems the researcher uses MATLAB software. MATLAB is software that uses an extremely effective and interactive language, which is widely used in several applications in engineering, especially aimed at numerical calculation.

3.2 Methods

The scientific methodology used in the study was based on applied research, in which

the acquired knowledge is used for practical application in the software aiming at solving the

proposed Ordinary Differential Equations. The MATLAB cod of numerical methods like

Euler’s, Runge kutta 2 nd , Runge kutta 3 rd and Runge kutta 4 th was developed and the solution

of some selected initial value problems was solved. Then the researcher developed the

MATLAB code of Richardson extrapolation method together with other methods like Euler’s

method, Runge kutta 2 nd , Runge kutta 3 rd and Runge kutta 4 th methods. The result of the

(5)

MATLAB code was described using tables and graphs. Lastly the absolute error of all the above methods was analyzed.

4. NUMERICAL RESULTS

In this section we consider numerical examples to show the effect of Richardson Extrapolation method combined with other the numerical methods (Euler’s method, Runge kutta 2 nd , Runge kutta 3 rd and Runge kutta 4 th methods). [6] Solved the following examples by using Euler’s method, Runge kutta 2 nd , Runge kutta 3 rd and Runge kutta 4 th methods with MATLAB program . However this study focuses on the effect of Richardson extrapolation Using MATLAB program.

Example 1: Consider 𝑦 = 𝑦 − 𝑥 2 + 1 , 0 ≤ 𝑥 ≤ 3.2 , 𝑦(0) = 0.5

With analytical solution 𝑦 = (𝑥 + 1) 2 − 0.5𝑒 𝑥

The graphical representation of the solution of the given IVP using Euler’s, Runge kutta 2 nd , Runge kutta 3 rd and Runge kutta 4 th methods were as follows.

Figure 1: Solution of the given IVP using Euler’s, Runge kutta 2, Runge Kutta3 and Runge kutta order 4

Runge kutta order 4 is more accurate than Eulers, rung kutta order 2 and Rung kutta order 3 methods.

The solution of the given IVP using the combination of Richardson extrapolation together with Euler’s, Runge kutta 2 nd , Runge kutta 3 rd and Runge kutta 4 th order methods were as follows.

0 0.5 1 1.5 2 2.5 3 3.5

0 1 2 3 4 5 6 7

x-value dy/dx=y-x2+1

Euler Rung 2 Rung 3 Rung 4 Exact value

(6)

Table 1: The combination of Richardson extrapolation with Euler’s, runge2,3 and 4 methods

X- values Richardson Extrapolation with

Exact Euler Runge kutta 2 Runge kutta 3 Runge kutta 4

0 0.5000000 0.5 0.5 0.5 0.5

0.2 0.82800000 0.829246667 0.82929803 0.829298606 0.829298621 0.4 1.21108000 1.213975128 1.214086429 1.214087619 1.214087651 0.6 1.643726800 1.648757678 1.648938716 1.648940549 1.6489406 0.8 2.119213428 2.126965068 2.12722698 2.127229464 2.127229536 1.0 2.629333048 2.64050052 2.640855872 2.640858993 2.640859086 1.2 3.164074748 3.179474857 3.179937723 3.179941424 3.179941539 1.4 3.711228557 3.731809727 3.732395712 3.732399879 3.732400017 1.6 4.255904288 4.282752981 4.283479185 4.283483629 4.283483788 1.8 4.77994547 4.814286784 4.815171669 4.815176089 4.815176268 2.0 5.261215556 5.30440469 5.305467805 5.305471758 5.305471951 2.2 5.672728668 5.726228697 5.727490205 5.72749305 5.72749325 2.4 5.981591103 6.04693084 6.04841078 6.048411613 6.04841181 2.6 6.147712531 6.226416065 6.228133232 6.228130807 6.228130982 2.8 6.122236919 6.215713539 6.21768387 6.217676484 6.217676614 3.0 5.84563238 5.95501185 5.957246131 5.957231487 5.957231538 3.2 5.245366028 5.371259274 5.373759949 5.373734975 5.373734901

The combination of Richardson extrapolation with Euler’s, runge2,3 and 4 is as follows graphically

Figure 2: Combination of Richardson extrapolation together with Euler’s, Runge 2, Runge 3 and Runge 4

As we see table 1 and figure 2 the solution of the given IVP becomes more accurate if Richardson extrapolation together with Euler’s, Runge 2, Runge 3 and Runge 4 is combined.

In Runge kuta 4 th order method is more accurate if it is compared with Euler’s, Runge 2, and Runge 3 methods. However combination of Richardson extrapolation method with Runge kutta 4 th order method is the most accurate than the other methods and the truncation error of each methods are as follows.

0 0.5 1 1.5 2 2.5 3 3.5

0 1 2 3 4 5 6 7

x-value

dy/dx=y-x2+1

Euler & rich R2 and Rich R3 and Rich R4 and Rich Exact Value

(7)

Table 2: The truncation error of the given IVP.

X-values Euler’s Rung 2 Rung 3 Rung 4 Euler & Rich Rung 2 & Rich Rung 3 &rich Rung 4& Rich

0 0 0 0 0 0 0 0 0

0.2 0.029298621 0.003298621 9.86209E-05 5.28759E-06 0.001298621 5.19543E-05 5.90761E-07 1.54116E-08 0.4 0.062087651 0.007167651 0.000211385 1.14405E-05 0.003007651 0.000112523 1.2222E-06 3.24739E-08 0.6 0.0985406 0.0116982 0.000339719 1.85828E-05 0.0052138 0.000182922 1.88331E-06 5.11625E-08 0.8 0.138749536 0.016993808 0.000484994 2.68508E-05 0.008016108 0.000264467 2.55619E-06 7.13609E-08 1.0 0.182683086 0.023171498 0.000648419 3.6393E-05 0.011526038 0.000358566 3.21354E-06 9.28217E-08 1.2 0.230130339 0.030362681 0.00083091 4.73684E-05 0.015866791 0.000466681 3.81541E-06 1.15113E-07 1.4 0.280626577 0.03871381 0.001032903 5.99437E-05 0.02117146 0.00059029 4.30484E-06 1.37552E-07 1.6 0.33335566 0.048386616 0.001254086 7.42895E-05 0.027579499 0.000730806 4.60236E-06 1.5911E-07 1.8 0.387022514 0.059557718 0.001493059 9.05732E-05 0.035230798 0.000889484 4.5988E-06 1.78296E-07 2.0 0.439687446 0.07241732 0.001746858 0.00010895 0.044256394 0.001067261 4.14598E-06 1.93005E-07 2.2 0.488551845 0.087166601 0.002010337 0.000129548 0.054764582 0.001264553 3.04482E-06 2.00312E-07 2.4 0.529682123 0.104013298 0.002275345 0.00015245 0.066820707 0.00148097 1.03004E-06 1.96219E-07 2.6 0.557655359 0.123164798 0.002529647 0.000177668 0.080418451 0.001714918 -2.24942E-06 1.75318E-07 2.8 0.565105866 0.14481787 0.002755517 0.000205102 0.095439695 0.001963076 -7.25541E-06 1.30368E-07 3.0 0.542146641 0.16914387 0.002927905 0.0002345 0.111599159 0.002219688 -1.45924E-05 5.17436E-08 3.2 0.475633024 0.196267946 0.003012063 0.000265386 0.128368873 0.002475627 -2.5048E-05 -7.32639E-08

Truncation error of Euler’s method is more than the other methods, whereas the truncation Error of Runge kutta order 4 together with Richardson extrapolation have less amount of error and it is more accurate than the others.

Example 2: 𝑦 = 𝑦 − 𝑥 , 0 ≤ 𝑥 ≤ 3.2 , 𝑦(0) = 2

The solution of the given IVP using Euler’s, Runge 2, Runge 3 and Runge 4 is as follows using tables and graphs.

Table 3: Solution of given IVP with Euler’s, Rung 2, Runge 3 and Runge 4 methods.

X- values Euler Rung 2 Rung 3 Rung 4 Exact

0 2 2 2 2 2

0.2 2.4 2.42 2.421333333 2.4214 2.421402758

0.4 2.84 2.8884 2.891655111 2.89181796 2.891824698 0.6 3.328 3.415848 3.421808109 3.422106456 3.4221188 0.8 3.8736 4.01533456 4.025034971 4.025520826 4.025540928 1.0 4.48832 4.702708163 4.717509377 4.718251137 4.718281828 1.2 5.185984 5.497303959 5.518984786 5.520071938 5.520116923 1.4 5.9831808 6.42271083 6.453586752 6.455135865 6.455199967 1.6 6.89981696 7.507707213 7.55078062 7.552942946 7.553032424 1.8 7.959780352 8.7874028 8.846553397 8.849524514 8.849647464 2.0 9.191736422 10.30463142 10.38485722 10.38888924 10.3890561 2.2 10.63008371 12.11165033 12.21937228 12.22478932 12.2250135 2.4 12.31610045 14.2722134 14.41566001 14.42287768 14.42317638 2.6 14.29932054 16.86410035 17.05379276 17.06334279 17.06373804 2.8 16.63918465 19.98220242 20.23156556 20.24412689 20.24464677 3.0 19.40702157 23.74228696 24.06841873 24.08485658 24.08553692 3.2 22.68842589 28.28559009 28.71022875 28.73164383 28.7325302

Runge kutta 4 th order method is more accurate than Euler’s, Runge 2, and Runge 3 methods.

(8)

Table: 4 Combination of Richardson methods with Euler’s, Runge 2, Runge 3 and Runge 4 methods.

X-Values E&RICH R 2&RICH R3&RICH R4&RICH Exact

0 2 2 2 2 2

0.2 2.42 2.421366667 2.421401937 2.421402742 2.421402758 0.4 2.8882 2.891736068 2.89182269 2.891824659 2.891824698 0.6 3.415122 3.421955568 3.422115121 3.422118729 3.4221188 0.8 4.01357762 4.025273713 4.025534935 4.025540812 4.025540928 1.0 4.69916492 4.717871741 4.718272675 4.718281651 4.718281828 1.2 5.490872753 5.519512768 5.520103502 5.520116663 5.520116923 1.4 6.411815872 6.454334656 6.455180837 6.455199597 6.455199967 1.6 7.490129013 7.5518184 7.553005714 7.553031908 7.553032424 1.8 8.760054275 8.847970862 8.849610752 8.849646755 8.849647464 2.0 10.26326348 10.38676932 10.38900626 10.38905514 10.3890561 2.2 12.05046617 12.22192575 12.22494652 12.22501221 12.2250135 2.4 14.1833649 14.41904171 14.42308711 14.42317466 14.42317638 2.6 16.73703254 17.05824008 17.06361988 17.06373575 17.06373804 2.8 19.80280258 20.23737949 20.24449131 20.24464377 20.24464677 3.0 23.49178296 24.07598021 24.08533342 24.085533 24.08553692 3.2 27.9391276 28.72001902 28.73226498 28.73252508 28.7325302

The solution of the given IVP is expresed graphically as follows.

Figure 3: The combined effect of Richardson together with Euler’s and Runge kutta methods As we can see table 4 and figure 3 Combination of Richardson extrapolation with runge kutta 4 th order method is the most accurate method and the result of this method is almost similar with the exact value. This indicates that instead of using Rung kutta 4 th order method only it is better to combine together with Richardson Extrapolation method.

In general the amount of error obtained using Euler > (Euler with Richardson) >

Rung 2 > Rung 3 >

0 0.5 1 1.5 2 2.5 3 3.5

0 5 10 15 20 25 30

x-value

dy /d x= y- x

Euler & rich

R2 and Rich

R3 and Rich

R4 and Rich

Exact Value

(9)

(Rung 2 with Richardson) >Runge 3 >Rung 4 > (Rung 3 with Richardson) > (Rung 4 with Richardson) The amount of error obtained at each method is as follows.

Table 5: Truncation error of the give IVP.

x- values

Euler Rung 2 Rung 3 Rung 4 E&RICH R 2&RICH R3&RICH R4&RICH

0 0 0 0 0 0 0 0 0

0.2 0.021402758 0.001402758 6.94248E-05 2.75816E-06 0.001402758 3.60915E-05 8.21652E-07 1.59194E-08 0.4 0.051824698 0.003424698 0.000169587 6.73764E-06 0.003624698 8.86301E-05 2.00772E-06 3.88885E-08 0.6 0.0941188 0.0062708 0.000310691 1.2344E-05 0.0069968 0.000163232 3.67944E-06 7.12487E-08 0.8 0.151940928 0.010206368 0.000505958 2.01027E-05 0.011963308 0.000267216 5.99385E-06 1.16033E-07 1.0 0.229961828 0.015573665 0.000772451 3.06919E-05 0.019116908 0.000410087 9.15381E-06 1.77155E-07 1.2 0.334132923 0.022812964 0.001132137 4.49845E-05 0.029244169 0.000604155 1.34205E-05 2.59657E-07 1.4 0.472019167 0.032489137 0.001613215 6.41015E-05 0.043384095 0.000865311 1.91294E-05 3.70007E-07 1.6 0.653215464 0.045325212 0.002251804 8.94784E-05 0.062903412 0.001214024 2.67103E-05 5.16495E-07 1.8 0.889867112 0.062244665 0.003094067 0.00012295 0.089593189 0.001676603 3.67128E-05 7.09714E-07 2.0 1.197319677 0.084424684 0.004198883 0.000166857 0.125792623 0.002286782 4.9838E-05 9.63174E-07 2.2 1.594929793 0.113363173 0.00564122 0.00022418 0.174547329 0.003087745 6.69791E-05 1.29408E-06 2.4 2.107075932 0.150962982 0.00751637 0.000298705 0.239811477 0.004134672 8.92716E-05 1.72431E-06 2.6 2.764417497 0.199637689 0.009945275 0.000395243 0.326705497 0.00549796 0.000118157 2.28161E-06 2.8 3.605462126 0.262444348 0.013081214 0.000519884 0.441844195 0.007267282 0.000155464 3.00116E-06 3.0 4.678515349 0.343249968 0.017118189 0.000680344 0.59375396 0.009556714 0.000203507 3.9275E-06 3.2 6.044104308 0.446940111 0.02230145 0.000886371 0.793402596 0.01251118 0.000265212 5.11693E-06

The amount of error occur at Euler’s, Runge 2, Runge 3 and Runge 4 methods before combination with Richardson extrapolation is as follows graphically.

Figure 4: The amount of truncation error by Euler’s, Runge 2, Runge 3 and Rung 4

The total amount of Error occur by using Euler’s method riches 7 units and Runge 2, Runge 3 and Runge 4 methods also have much amount of error compared with respective combination methods.

The amount of error occurs when Euler’s, Runge 2, Runge 3 and Runge 4 combined with Richardson extrapolation methods are as follows graphically.

0 0.5 1 1.5 2 2.5 3 3.5

0 1 2 3 4 5 6 7

x-value

Am ou nt o f E rro r

Euler

Rung 2

Rrung 3

Rung 4

(10)

Figure 5: The truncation error of Richardson with Euler’s, Runge 2, Runge 3 and Rung 4 methods.

When the numerical methods combined with Richardson Extrapolation method the amount of error occurred highly reduced. Thus the amount of error occur at Euler’s methods with Richardson method is not more than 0.8 unites. The amount of error obtained using Runge 2, Runge 3 and Runge 4 together with Richardson extrapolation is less than 0.1 unites.

Particularly the amount of error when we use Runge kutta 4 th order method together with Richardson extrapolation overlapped to the x-axis (Error approximately zero).

5. DISCUSSION OF THE RESULTS

The accuracy of numerical methods like Euler, runge kutta order 2, runge kutta order 3, runge kutta order4 compared with the combination of Richardson extrapolation together with Euler method, runge kutta order 2, runge kutta order 3, runge kutta order 4 relative to the exact value. we compared among them through MATLAB Program.

CONCLUSION

The results of numerical methods using MATLAB indicates that Richardson Extrapolation together with Runge kutta 4 th order method is more accurate than the other methods. Thus, for the solution of IVP, we should Use higher order approximations together with Richardson Extrapolation. The use of computational tools, such as MATLAB software, is extremely necessary both to solve the problems proposed, especially in numerical calculation, and to illustrate in a didactic way the graphical representation of the results obtained by the methods.

To increase the accuracy of IVPs we have to take small step size or combine the numerical methods together with Richardson extrapolation method but if we use small step size with Richardson extrapolation the result of the IVP have better accuracy.

0 0.5 1 1.5 2 2.5 3 3.5

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

x-value

Am ou nt o f e rro r

Euler and rich

R2 and Rich

R3 and Rich

R4 and Rich

(11)

REFERENCES

1. Mayers, S. E. An Introduction to Numerical Analysis. New York: Cambridg University Press (2003).

2. Gaunt J, The deferred approach to the limit. Philosophical Transactions of the Royal society of London, Seriea (1987).

3. Istvan F., H. Efficient implimentation of stable Richardson Extrapolation Algorizem.

Science Direct, Camwa (2010).

4. Istv A., O. Z. The convergence of explicit Rung-kutta methods combind with Richardson Etrapolation. Institut of Mathematics Cr, Parague (2012).

5. Chapra, S. Numerical methods for Engineers fifth edition. New York san Francisco:

McGraw Hill (2006).

6. C. Senthilnathan. Numerical solutions of initial value problems (IVP) for ordinary differential equations (ode) with Euler and higher order of Runge kutta methods using MATLAB. International Journal of Engineering Science Invention (IJESI) (2018).

7. Lative Z., I. F. Stability of Richardson extrapolation applied together with θ-method, Science Direct (2010).

8. Burden, F. Numerical methods 3rd Edition. Youngston State: Cole Publishing (June 18, 2002).

9. Hoffman, J. Numerical methods for Engineers and scientists second edition. New York,

Basel: McGRAW- Hill (1992).

References

Related documents

If we calculate the efficiency of a DMU by DEA models on the assumption that it is constant returns to scale in different times, we reach the conclusion that the amounts of

Somewhat lower values were recorded in the volleyball players (, ml/kg/min), while the lowest values were obtained in non-athletes (, ml/kg/min) whose VO max/BW

Differences between values of giant cells in both groups 2 and 3 (that had re- ceived respectively 10mg/kg and 25mg/kg of terfenadine) and the control group were statisti-

Nili Ahmadabadi, Effect of modified heat treatments on the microstructure and mechanical properties of a low alloy high strength steel, Materials Science and Technology ,

On this basis, the present study attempted to investigate the prediction of water level of piezometers of double-curvature arch dam using feed-forward multi- layer artificial

Radio modem, cellular, and also satellite communication with an automatic channels choice of communication in the conditions of the powerful electromagnetic fields having

Fiserv supports all of the payment channels, methods and services your customers demand, with an integrated multi-channel strategy that will help you accelerate your transition

Notwithstanding this, in the case of non-monetarily sovereign countries, fiscal austerity can eventually curtail rather than stimulate private investments through