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

net/publication/279849033

Study of Numerical Accuracy of Runge-Kutta Second, Third and Fourth Order

Method

Article · June 2015

CITATIONS READS

9 5,174

3 authors:

Najmuddin Ahmad Shiv Charan

Integral University Shri Ramswaroop Memorial Group of Professional Colleges
47 PUBLICATIONS 127 CITATIONS 9 PUBLICATIONS 46 CITATIONS

SEE PROFILE SEE PROFILE

Vimal Pratap Singh

Integral University
14 PUBLICATIONS 34 CITATIONS

SEE PROFILE

All content following this page was uploaded by Najmuddin Ahmad on 07 July 2015.

The user has requested enhancement of the downloaded file.

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

Study of Numerical Accuracy of Runge-Kutta Second, Third

and Fourth Order Method
Najmuddin Ahmad, Shiv Charan,Vimal Partap Singh
Department of Mathematics, Integral University , Lucknow

Abstract : We have considered ten ordinary differential equation of first order with boundary condition . These
equations have been solved by Heun’s method , Runge – Kutta third order method and by Runge – Kutta fourth
order method in the interval [ 0, 1 ] by dividing the interval into 200 parts with the help of MATLAB by us. At each
point of the interval we have calculated the value of y and compared it with its exact value at that point . Error in the
value of y is the difference between calculated and exact values . Percentage error has also been calculated at each point
of the intervals . Comparison of the results of the of the Heun’s method , Runge – Kutta third order method and by
Runge – Kutta fourth order method shows that Runge – Kutta fourth order method is better in all the cases . Mean of
the results for all differential equations shows Runge – Kutta fourth order method is times better than Runge – Kutta
second order method ( Heun’s method ) and , Runge – Kutta third order method .

Introduction : In mathematics and computational science , Heun‟s method ( also called the modified
Euler‟s method or the explicit trapezoidal rule ), named after Karl L. W. M. Heun, is a numerical procedure
for solving ordinary differential equations (ODEs) with a given initial . It can be
seen as an extension of the Euler method into a two – stage second – order Runge – Kutta
.
Heun‟s Method is used to generate a numerical solution to an initial value problem of the form :
=f ( x,y ) with the boundary conditions y( ) = .
Heun „s Method was an improvement over the rather simple Euler Method , and that though it
uses Euler „s method as a basic , it attempting to compensate for the Euler Method „s failure to take the
curvature of the solution curve into account . Heun‟s Method is one of the simplest of a class of method
called predictor – Corrector algorithms. One of the most powerful predictor – Corrector algorithms of all
one which is so accurate , that most computer packages designed to find numerical solutions for differential
equations will use it by default – the fourth order Runge – Kutta .
The Runge – Kutta is a little hard to follow even when one only considers it from a
geometric point of view . In reality the formula was not originally derived in this fashion , but with a purely
analytical approach . After all , among other things , our geometric “ explanation “ doesn‟t even account for
the weights that were used.
Numerical ordinary differential is part of numerical analysis which studies the
numerical solution of ordinary differential equations (ODEs) .This field is also known under the name
numerical integration , but some people reserve this term for the computation of integrals . Many differential
equations cannot be solved analytically , in which case we have to satisfy ourselves with an approximation to
the solution . The algorithms studied here can be used to compute such an approximation . An alternative
method is to use techniques from calculus to obtain a series expansion of the solution .
Ordinary differential occur in many scientific disciplines , for instance in mechanics ,
chemistry , biology and economics . In addition , some methods in numerical partial differential equation
convert the partial differential equation into an ordinary differential equation , which must then be
.
We have restricted ourselves to first – order differential equations . However , a higher – order equation
can easily be converted to a system of first – order equations by introducing extra .

111 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

The proper numerical modeling method heavily depends on the situation , the available resources , and
the desired accuracy of the result . If only a quick estimate of a differential equation is required , the Euler
method may provide the simplest solution . If much higher accuracy is required , a fifth – order Runge – Kutta
method may be . In all numerical models , as the step size is decreased , the accuracy of the
model is increased . The tradeoff here is that smaller step sizes require more computation and therefore
increase the amount of time to obtain a solution. A balance between desired accuracy and time required for
producing an answer can be achieved by selecting an appropriate step size . One suggested algorithm for
selecting a suitable step size is to produce models using two different . the steps size can then
be systematically cut in half until the difference between both model is acceptably small ( effectively
creating an error tolerance ).

Objective : We have considered ten differential equations with initial condition . We have developed
MATLAB to solve them to get exact value and approximated value by Heun‟s method and Runge – Kutta
fourth order method . We have then compared the accuracy of the results obtained by Heun‟s method and by
Runge – Kutta fourth order method . We have also studied that whether the accuracy obtained by Runge –
Kutta fourth order method can be achieved by Heun‟s method by increasing the number of intervals .
Differential equations under study are given below –
1. dy/dx = x + y ; y = 1 when x = 0
2. dy/dx = y - x ; y = 2 when x = 0
3. dy/dx = - 2x ; y = 1 when x = 0
4. dy/dx = x + y + 1 ; y = 1 when x = 0
5. dy/dx = -x + y + 1 ; y = 1 when x = 0
6. dy/dx = (1 + x) x ; y = 1 when x = 0
7 . dy/dx = (1 + x) /2 ; y = 1 when x = 0
8. dy/dx = ; y = 1 when x = 0
9. dy/dx = - y ; y = 1 when x = 0
10. dy/dx = 1 + y ; y = 1 when x = 0
With the help of computer programs , we have calculated the value of y, exact value of y , difference
between calculated and exact value of y and percentage error in the value of y . We have calculated the values
of y in the interval [ 0 ,1 ] by assuming the step size of 0.005 . Percentage error in the value of y is defined as
= × 100
We have also performed multi linear regression ( MLR ) analysis between the value of x and the
percentage error in the value y with the help of CAChe Software of Fujitsu. MLR equation is of the form
= m * x + n , rCV˄2=p , r˄2 =q
where is the percentage error in the value of y , rCV˄2 is the cross – validation coefficient , r˄2 is
the regression coefficient , m, n, p, q  R .
Predictive power of the regression equation depends on the value of cross – validation and regression
coefficients . A regression equation is said to possess a good predictive power if the value of cross –
validation coefficient is greater than 0.2 and the value of regression coefficient is greater than 0.5 . As the
value of regression coefficient increase , the predictive power of the regression equation increase . The
maximum value of regression coefficient may by unity and in this case the regression equation possesses
100% predictive power .

Methodology :
Second order Runge – Kutta Method - the fourth order Runge – Kutta method is one of the standard
algorithm to solve differential equations . Before we give the algorithm of the fourth order Runge – Kutta

112 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

method we will drive the second order Runge – Kutta method . WE start with the original differential equation
and integrate it formally
= f(t,y) y(t) =

 = +

= +
We essentially changed the task at hand from performing a differentiation to an integration . To do this we
expend f( t ) in a second order Taylor series around the midpoint of the integration subinterval
f(t,y)f( , ) + (t- ) ………….. (1)

Yet since the integral of ( t - ) vanishes when evaluate about the midpoint , we automatically get
improved precision using only the first term in (1)

f(t,y)f( , )

  +hf( , )

= +hf( , )

This algorithm cannot be applied immediately since it requires a knowledge of

which is not in the scheme of things . We thus approximate with Euler‟s algorithm .

 + t = +

 + hf

 +

= hf

= hf
The second order Runge – Kutta algorithm (2) requires the known derivative function f at the endpoint
and midpoint of the interval , and the unknown function y at the previous point . Since we start with initial
condition , the algorithm is self starting . Note too that it is applicable with a general function f ( for example
nonlinear ) , and simple to .
In numerical analysis , the Runge – Kutta methods are an important family of implicit and explicit
iterative methods for the approximation of solution of ordinary differential
. These techniques were developed around 1900 by the German
mathematicians Runge and Kutta .
The Classical Fourth – Order Runge – Kutta Method - One member of family of Runge – Kutta
methods is so commonly used , that it is often referred to as “ RK4” or simply as “the Runge – Kutta
.
Let an initial value problem be specified as follows

113 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

=f(t,y), y = .
Then , the RK4 method for this problem is given by the following equation :
= +

where =f( , ), = f ,

= f , =f( , ) .
Thus , the next value ( ) is determined by the present value ( plus the product of the size of the
interval (h) and an estimated slope . The slope is a weighted average of slopes :
 is the slope at the beginning of the interval ;
 is the slope at the midpoint of the interval , using slope to determine the value of y at the point
using Euler‟s method ;
 is again the slope at the midpoint , but now using the slope to determine the y – value ;
 is the slope at the end of the interval , with its y – value determined using .
When the four slopes are averaged , more weight is given to the slope at the
slope =
the RK4 method is a fourth – order method , meaning that the error per step is on the order of ,
while the total accumulated error has order .
Above formulas are valid for both scalar and vector valued function ( y can be a vector ) .

Result and Discussions :

Solution of Differential equation = x + y with boundary condition x = 0 , y = 1 by Heun‘ s
method
Differential equation is given by
=x+y ; x = 0,y=1.
Exact solution of this differential equation is given by
x + y +1=2
We have solved this differential equation by Heun‟s method with the computer program given in the
methodology at different values in the interval [ 0 , 1 ] by talking the step size of 0.005 . Value of x,
calculated value of y , exact value of y , difference between calculated and exact value of y and percentage
error in the value of y are shown in the Table – 1 . This Table indicates that the percentage error in the value
of y increases as the value of x increases . The maximum percentage error is 0.000656703 at x = 1 . Graph
between calculated value of y by Heun‟s method and exact value of y is shown in the graph – 1 . As the
maximum error in the value of y is only 0.000022568 , the curves of calculated and exact values of y
coincide . Difference between calculated and exact values of y at different values of x for the differential
equation is shown in the Graph – 2 . The shape of this curve is similar to a parabola . Graph – 3 shows the
percentage error in the value of y at different values of x for the different equation .
We have also performed multilinear regression ( MLR ) analysis between the values of x and the
percentage error in the value y with the help of CAChe Software of Fujitsu . MLR equation is given below –
P = 0.0006562*x + 3.4307e – 005

114 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

rCV˄2 = 0.992974
r˄2 = 0.992978
where P is the percentage error in the value of y , rCV˄2 is the cross – validation coefficient , r˄2
is the regression coefficient .
Predictive power of the regression equation depends on the value of cross – validation and regression
coefficient . A regression equation is said to possess a good predictive power if the value of cross – validation
coefficient is greater than 0.2 and the value of regression coefficient is greater than 0.5 A the value of
regression coefficient increase the predictive power of the regression equation increases . The maximum value
of regression coefficient may by unity and in this case the regression equation possesses 100% predictive
power . Here , the value of regression coefficient is 0.992978 which indicates that the regression equation has
very good predictive power .

Conclusions :
We have calculated the percentage error at each point of the interval [ 0 , 1 ] by dividing it in 200 equal parts
for each differential equation using RUnge – Kutta second order method and Runge – Kutta fourth order
method . Average percentage error has been calculated by the formula .
Average percentage error = Sum of percentage error at each point of the interval / 200.
In each differential equation , the average percentage error in Runge – Kutta fourth order method is
very low as compared to Runge – Kutta second method . Accuracy of Runge – Kutta fourth order method
over Runge – Kutta second order method has been defined by the factor Ä define as
Ä = average percentage error in Runge – Kutta second order method /
average percentage error in Runge – Kutta fourth order method .
Table – 3 includes average percentage error in Runge – Kutta second order method ( Heun‟s
method ) , average percentage error in Runge – Kutta fourth order method and accuracy factor ( Ä ) for each
differential equation . A references to Table – 3 indicades that
 In differential equation = x + y ; y = 1 , x = 0 ; the average percentage error in Heun‟s method is
0.000364047245 , , the average percentage error in Runge – Kutta third order method is
0.00000045494603 and the average percentage error in Runge – Kutta fourth order method is
0.00000000045485.
 In differential equation = y - x ; y = 2 , x = 0 ; the average percentage error in Heun‟s method is
0.000112987945 , , the average percentage error in Runge – Kutta third order method is
0.00000014119891 and the average percentage error in Runge – Kutta fourth order method is
0.00000000014121.
In differential equation = -2x ; y = 1 , x = 0 ; the average percentage error in Heun‟s method is
0.000149559645 , , the average percentage error in Runge – Kutta third order method is
0.000000838112311 and the average percentage error in Runge – Kutta fourth order method is
0.00000000029570 .
 In differential equation = x + y + 1 ; y = 1 , x = 0 ; the average percentage error in Heun‟s
method is 0.000392652030 , , the average percentage error in Runge – Kutta third order method is
0.00000049069346 and the average percentage error in Runge – Kutta fourth order method is
0.00000000049055 .
 In differential equation =- x + y + 1 ; y = 1 , x = 0 ; the average percentage error in Heun‟s
method is 0.000158251915 , , the average percentage error in Runge – Kutta third order method is
0.00000019776422 and the average percentage error in Runge – Kutta fourth order method is
0.00000000019779.

115 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

 In differential equation = ( 1 + x )x ; y = 1 , x = 0 ; the average percentage error in Heun‟s

method is 0.004678271690 , the average percentage error in Runge – Kutta third order method is
0.000057437713051 and the average percentage error in Runge – Kutta fourth order method is
0.00005742673978 .
 In differential equation = ( 1 + x ) /2 ; y = 1 , x = 0 ; the average percentage error in Heun‟s
method is 0.001440964555 , , the average percentage error in Runge – Kutta third order method is
0.00000311677763 and the average percentage error in Runge – Kutta fourth order method is
0.00000000833301.
 In differential equation = ; y = -1 , x = 0 ; the average percentage error in Heun‟s method
is 0.000243209175 , , the average percentage error in Runge – Kutta third order method is
0.00000033590607 and the average percentage error in Runge – Kutta fourth order method is
0.00000000034946.
 In differential equation = - y ; y = 1 , x = 0 ; the average percentage error in Heun‟s method is
0.000210160330 , , the average percentage error in Runge – Kutta third order method is
0.00000026276778 and the average percentage error in Runge – Kutta fourth order method is
0.00000000026280.
 In differential equation = 1 + y ; y = 1 , x = 0 ; the average percentage error in Heun‟s method is
0.000530911520 , , the average percentage error in Runge – Kutta third order method is
0.00000035735117 and the average percentage error in Runge – Kutta fourth order method is
0.00000000066374.
Table – 1 : Average Percentage Error in Runge – Kutta Second Order Method ( Heun’s method ) , Average
Percentage Error in Runge – Kutta Fourth Order Method and Factor ( Ä ) for Each Differential Equation .
S . Differential Average Percentage Average Percetage Average Percetage
No. Equation Error in Heun’s Metod Error in Runge – Kutta Error in Runge – Kutta
third Order Method third Order Method
1 dy/dx = x + y ; 0.000364047245
y=1, x=0 0.00000045494603 0.00000000045485
2 dy/dx = y - x ;
y=2, x=0 0.000112987945 0.00000014119891 0.00000000014121
3 dy/dx = - 2x ;
0.000149559645 0.00000083811231 0.00000000029570
y=1, x=0
4 dy/dx = x+y+1 ;
y=1, x=0 0.000392652030 0.00000049069346 0.00000000049055
5 dy / dx = -x+y+1;
y=1, x=0 0.000158251915 0.00000019776422 0.00000000019779
6 (1 + x) x ;
0.004678271690 0.00005743771305 0.00005742673978
y=1, x=0
7 (1 + x) /2;
y=1, x=0 0.001440964555 0.00000311677763 0.00000000833301
8 dy/dx = ;
0.000243209175 0.00000033590607 0.00000000034946
y = -1 , x = 0
9 dy / dx = - y ;
y=1, x=0 0.000210160330 0.00000026276778 0.00000000026280
10 dy / dx = 1+ y;
y = 1, x = 0 0.000530911520 0.00000035735117 0.00000000066374

116 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

References :
1. Bogacki, P. and Shampine , L. F. (1996) : Efficient Runge – Kutta ( 4 ,5 ) Pair Source :Computers and
Mathematics with Application , Vol. 32, No. 6 , pp 15-28, Compendex.
2. Butcher. J. C. (1965) : A Modified Multistep Method for the Numerical Integration of Ordinary Differential
Equation . Journal of the ACM , Vol. 12 (1) ,pp 124-135.
3. Calvo , M.; Montijano, J. I. and Randeg, L. (1990) : New Continuous Extensions for Fifth – Order RK Formulas .
Rev. Acad. Cienc. Zaragoza (2) 45, pp 69-81, Math. Sci. Net.
4. Cockburn. B. and Shu, C.W. ( 1989) : TVB Runge – Kutta Local Projection Discontinuous Galerkinnite
Element Method for Conservation Laws . ΙΙ .General Framework . Math. Comp., Vol. 52 No. 186, pp 411-435 .
5. Cotiu, A . (1962) : Delimitation of the error in Fehlberg‟s Procedure of Numerical Integration of first Order
Differential Equations . ( Romanian ) . Studia Univ. Babes – Bolyai Ser. Math. Phys. Vol. 7 , No. 2, pp 37 – 43 ,
Math. Sci. Net.
6. Daniel Okunbor and Robert D. Skeel ( 1992) : An Explicit Runge – Kutta – Nystrom Method is Canonical If and
Only If Its Adjoint is Explicit . SIAM Journal on Numerical Analysis , Vol. 29, No.2 , pp 521 – 527 , Jstor .
Dekker , K. and Verwer, J. G. ( 1984) : Stability of Runge – Kutta Methods for Nonlinear Differential Equations ,
Vol. 2 of CWI Monografhas . North – Holland Publishing Co. ( Amsterdam ) . England , R . (1968 ) : Error
Estimates for Runge – Kutta Solutions to Systems of Ordinary Differential Equations . Research and
Development Department , Pressed Steel Fisher Ltd., Cowley , Oxford , U . K . Ernst Hairer , Syvert Paul
Norsett and Gerhard Wanner (1993 ) : Solving Ordinary Differential Equations Ι : Nonstiff Problems , Second
Edition , Springer Verlag, Berlin , ISBN 3 – 540 – 56670 – 8 .
7. Fredebeul , C. ; Kornmair , D. and Muller , M. W. ( 2002 ) : Multiple Order Double Output Runge – Kutta
Fehlberg Formulae : Strategies for Efficient Application. Journal of Computational and Applied Mathematics ,
Vol. 144, No. 1 – 2 , pp 187 – 196 .
8. George D . Byrne and Robert J . Lambert ( 1966 ) : Pseudo – Runge – Kutta Methods Involving two points .
Journal of the ACM , Vol. 13 ( 1 ) , pp 114 – 123.
9. George , E . Forsythe ; Michael A. Malcolm and Cleve B . ( 1977 ) Moler Computer Methods for
Mathematical Computations . Englewood Cliffs , NJ : Prentice – Hall .Gerisch , A. and Weiner R . ( 2003 ) : The
Positivity of Low – Order Explicit Runge Kutta Schemes Applied in Splitting Methods. Comput . Math.Appl.,
Vol. 45 No. 1 – 3 , pp 53 – 67 .
10. Hairer , E . ; Narsett , S. P. and Wanner , G. ( 1993 ) : Solving Ordinary Differential Equations. Ι. Nonstiff
problems , Vol. 8 of Springer Series in Computational Mathematics . Springer – Verlag ( Berlin ) , Second Ed.
11. Hundsdofer , W . and Verwer , J. G. ( 2003 ) : Numerical Solution of Timedependent Advection – Diffusion –
Reaction Equations , Vol. 33 of Springer Series in Computational Mathematics . Springer ( Berlin ) .
12. Jain , R. K. and Jain , M. K. ( 1971 ) : Optimum Runge – Kutta Fehlberg Methods for First Order Differential
Equations . J. Inst. Math. Appl., Vol. 8, pp 386 – 396, Math. Sci. Net. John , W. ( 1958 ) : Carr , ΙΙΙ , Error
Bounds for the Runge – Kutta Single – Step Integration Process , Journal of the ACM ( JACM ) , Vol. 5 No. 1 ,
pp 39 – 44.
13. Kaw , Autar ; Kalu , Egwu ( 2008 ) : Numerical Methods with Applications ( 1 st Ed. ) , www.autarkaw.com.
14. Kendall , E. Atkinson ( 1998 ) : An Introduction to Numerical Analysis . John Wiley & Sons .
15. Kharab, A. (1994 ) : Spreadsheet Solution to an initial Value Problem using the Runge – Kutta – Fehlberg
Method . Computer Application in Engineering Education , Vol. 2 , No. 2 , pp 129 – 134 , Compendex .
16. Kraaijevanger J. F. B. M. ( 1991 ) : Contractivity of Runge – Kutta Methods . BIT , Vol. 31, No. 3, pp 482 – 528 .
17. Ling , Xiang; Tu, Shan – Tung ; Gong and Jain – Ming ( 2000 ) : Application of Runge – Kutta – Merson
Algorithm for Creep Damage Analysis . International Journal of Pressure Vessels and Piping , Vol. 77 , No. 5, pp
243 -248, Compendex .
18. Murugesan , K; Dhayabaran , D. Paul; Amiritharaj, E. C. Henry ; Evans , David J.( 2002 ) : A Fourth Order
Embedded Runge – Kutta RKACeM( 4 , 4 ) Method Based on Arithmetic and Centroidal Means with Error
Control . Int. J. Comput. Math., Vol. 79, No. 2 , pp 247 – 269 , Math. Sci. Net.
19. Numerical Methods in Physics , Chemistry and Engineering . Gottlieb S., Shu C. W. ( 1998 ) : Total Variation
Diminishing Runge – Kutta Schemes . Math. Comp., Vol. 67, No. 221, pp 73 – 85 .
20. Pimenov , V. G. ( 2001 ) : General Linear Methods for the Numerical Solution of Functional – Differential
Equations , Differential Equations , Vol. 37, No. 1 .
21. Procedures of Runge – Kutta – Fehlberg Type with Error of Order 0 , for the Approximation of First
Order Integrodifferential Equations of Volterra Type . ( Romanian ) Micula , Maria Stud. Cerc. Mat., 25
(1973) , pp 33 – 40, Math. Sci. Net.

117 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

 International Journal of Computer & Mathematical Sciences
IJCMS
ISSN 2347 – 8527
Volume 4, Issue 6
June 2015

22. Runge – Kutta – Fehlberg Type Procedure on Two Nodes for Numerical Integration of Systems Differential
Equations . Dumitras , Daria Elena Automat . Comput. Appl. Math., Vol. 2, pp 139 – 143 , Math. Sci. Net.
23. Ruuth , S. J. ( 2004 ) : Global optimization of Explicit Strong – Stability preserving Runge – Kutta Methods.
Tech. Rep., Department of Mathematics Simon Fraser University .
24. Sideridis , A. B. and Simos , T. E. ( 1992 ) : A Low – Order Embedded Runge – Kutta Method for Periodic Initial
Value Problems. J. Comput . Appl. Math., Vol. 44. No. 2, pp 235 – 244, Math. Sci. Net.
25. Simos, T. E. , ( 1993 ) : A Runge – Kutta Fehlberg Method with Phase – Leg of Order Infinity for initial –
Value Problem with Oscillating Solution . Computers and Mathematics with Applicatons, Vol. 25 , No. 6 pp 95 –
101.
26. Simos , T. E. , ( 1995 ) : Modified Runge – Kutta _ Fehlberg Methods for Periodic Initial – Value Problems .
Japan J . Indust. Appl. Math., Vol. 12, No. 1 , pp 109-122 , Math. Sci , Net.
27. Spiteri , R. J. and Ruuth, S. J. ( 2003 ) : Non – Linear Evolution using Optimal Fourth – Order Strong – Stability
– Preserving Runge - Kutta Methods. Math. Comput. Simulation , Vol. 62, No. 1 – 2 , pp 125 – 135 .
28. Strogatz , S. H. ( 1994 ) : Nonlinear Dynamics and Chaos with Applications in Physics , Biology , Chemistry and
Engineering . Perseus Books ( Reading , M . A. ) .
29. Suchitra Gupta ( 1985 ) : An Adaptive Boundary Value Runge – Kutta Solver for First Order Boundary Value
Problems SIAM Journal on Numerical Analysis, Vol. 22, No. 1., pp 114 – 126 .
30. Tavernini , L. ( 1971 ) : One – step Methods for the Numerical Solution of Volterra Functional Differential
Equations , SIAM J . Numer . Anal., Vol. 8, No. 4.
31. Test results on Initial Value Methods for Non – Stiff Ordinary Differential Equations W. H. Enright ; T. E. Hull
SIAM Journal on Numerical Analysis, Vol. 13, No. 6. ( Dec., 1976 ) , pp 944 – 961 , Jstor .
32. The Runge – Kutta Theory in a Nutshell Peter Albrecht SIAM Journal on Numerical Analysis , Vol. 33 , No. 5.
( Oct., 1996 ) , pp 1712 – 1735, Jstor .
33. Toro , E. F. ( 1999 ) : Riemann Solvers and Numerical Methods for Fluid Dynamics. A Practical Introducing .
Springer – Verlag ( Berlin ) , Second Ed.
34. Verner , J. H. ( 1990 ) : A Contrast of Some Runge – Kutta Formula Pairs SIAM Journal on Numerical Analysis
, Vol. 27 , No. 5 , pp 1332 – 1344 , Jstor .
35. Verner , J. H. ( 1991 ) : Some Runge – Kutta Formula Pairs . SIAM J. Numer . Anal. Vol. 28 , No. 2 , pp 496 –
511 , Math . Sci. Net.
36. Verner , J. H. ( 1996 ) : Higher – Order Explicit Runge – Kutta Pairs with Low Stage Order Applied
Numerical Mathematics , Vol. 22 , No. 1 – 3 , pp 345 – 357 , Compendex . William B. Gruttke ( 1970 ) :
Pseudo – Runge – Kutta Methods of the Fifth order . Journal of the ACM , Vol. 17 ( 4 ) , pp 613 – 628 .

118 Najmuddin Ahmad, Shiv Charan, Vimal Partap Singh

View publication stats