A Singly Diagonally Implicit Runge-Kutta- Nyström Method with Reduced Phase-lag

Similar documents
A Singly Diagonally Implicit Runge-Kutta- Nyström Method for Solving Oscillatory Problems

Fourth Order Four-Stage Diagonally Implicit Runge-Kutta Method for Linear Ordinary Differential Equations ABSTRACT INTRODUCTION

A New Diagonally Implicit Runge-Kutta-Nyström Method for Periodic IVPs

Derivation of 3-Point Block Method Formula for Solving First Order Stiff Ordinary Differential Equations

CS473-Algorithms I. Lecture 12b. Dynamic Tables. CS 473 Lecture X 1

Reaction Time VS. Drug Percentage Subject Amount of Drug Times % Reaction Time in Seconds 1 Mary John Carl Sara William 5 4

Simple Linear Regression Analysis

On a Truncated Erlang Queuing System. with Bulk Arrivals, Balking and Reneging

ROOT-LOCUS ANALYSIS. Lecture 11: Root Locus Plot. Consider a general feedback control system with a variable gain K. Y ( s ) ( ) K

Trignometric Inequations and Fuzzy Information Theory

Collapsing to Sample and Remainder Means. Ed Stanek. In order to collapse the expanded random variables to weighted sample and remainder

1. Linear second-order circuits

Multiple Choice Test. Chapter Adequacy of Models for Regression

Cubic Nonpolynomial Spline Approach to the Solution of a Second Order Two-Point Boundary Value Problem

Linear Approximating to Integer Addition

European Journal of Mathematics and Computer Science Vol. 5 No. 2, 2018 ISSN

On Modified Interval Symmetric Single-Step Procedure ISS2-5D for the Simultaneous Inclusion of Polynomial Zeros

A Result of Convergence about Weighted Sum for Exchangeable Random Variable Sequence in the Errors-in-Variables Model

On the Interval Zoro Symmetric Single Step. Procedure IZSS1-5D for the Simultaneous. Bounding of Real Polynomial Zeros

On the energy of complement of regular line graphs

STK4011 and STK9011 Autumn 2016

16 Homework lecture 16

Third handout: On the Gini Index

EVALUATION OF FUNCTIONAL INTEGRALS BY MEANS OF A SERIES AND THE METHOD OF BOREL TRANSFORM

Numerical Analysis Formulae Booklet

PGE 310: Formulation and Solution in Geosystems Engineering. Dr. Balhoff. Interpolation

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions

European Journal of Mathematics and Computer Science Vol. 5 No. 2, 2018 ISSN

Bounds on the expected entropy and KL-divergence of sampled multinomial distributions. Brandon C. Roy

Nargozy T. Danayev*, Darkhan Zh. Akhmed-Zaki* THE USAGE OF MATHEMATICAL MLT MODEL FOR THE CALCULATION OF THERMAL FILTRATION

REVIEW OF SIMPLE LINEAR REGRESSION SIMPLE LINEAR REGRESSION

C-1: Aerodynamics of Airfoils 1 C-2: Aerodynamics of Airfoils 2 C-3: Panel Methods C-4: Thin Airfoil Theory

Functions of Random Variables

Journal of Mathematical Analysis and Applications

ANALYSIS ON THE NATURE OF THE BASIC EQUATIONS IN SYNERGETIC INTER-REPRESENTATION NETWORK

Lecture 07: Poles and Zeros

International Journal of Pure and Applied Sciences and Technology

Chapter 2 - Free Vibration of Multi-Degree-of-Freedom Systems - II

T-DOF PID Controller Design using Characteristic Ratio Assignment Method for Quadruple Tank Process

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines

Mean is only appropriate for interval or ratio scales, not ordinal or nominal.

A Conventional Approach for the Solution of the Fifth Order Boundary Value Problems Using Sixth Degree Spline Functions

BERNSTEIN COLLOCATION METHOD FOR SOLVING NONLINEAR DIFFERENTIAL EQUATIONS. Aysegul Akyuz Dascioglu and Nese Isler

Beam Warming Second-Order Upwind Method

Theory study about quarter-wave-stack dielectric mirrors

Quiz 1- Linear Regression Analysis (Based on Lectures 1-14)

( ) Thermal noise ktb (T is absolute temperature in kelvin, B is bandwidth, k is Boltzamann s constant) Shot noise

Compound Means and Fast Computation of Radicals

Econometric Methods. Review of Estimation

Mathematical Model of Dengue Fever with and without awareness in Host Population

Bayes Estimator for Exponential Distribution with Extension of Jeffery Prior Information

TESTS BASED ON MAXIMUM LIKELIHOOD

Generalized One-Step Third Derivative Implicit Hybrid Block Method for the Direct Solution of Second Order Ordinary Differential Equation

Stability For a stable numerical scheme, the errors in the initial condition will not grow unboundedly with time.

2006 Jamie Trahan, Autar Kaw, Kevin Martin University of South Florida United States of America

Point Estimation: definition of estimators

The Primitive Idempotents in

Lecture 3 Probability review (cont d)

L5 Polynomial / Spline Curves

Comparing Different Estimators of three Parameters for Transmuted Weibull Distribution

AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET

Application of Laplace Adomian Padé approximant to solve exponential stretching sheet problem in fluid mechanics

Newton s Power Flow algorithm

d dt d d dt dt Also recall that by Taylor series, / 2 (enables use of sin instead of cos-see p.27 of A&F) dsin

A class of cubic and quintic spline modified collocation methods for the solution of two-point boundary value problems.

Scheduling Jobs with a Common Due Date via Cooperative Game Theory

Ideal multigrades with trigonometric coefficients

X X X E[ ] E X E X. is the ()m n where the ( i,)th. j element is the mean of the ( i,)th., then

Analysis of Lagrange Interpolation Formula

2.28 The Wall Street Journal is probably referring to the average number of cubes used per glass measured for some population that they have chosen.

New soliton solutions for some important nonlinear partial differential equations using a generalized Bernoulli method

Fibonacci Identities as Binomial Sums

Bivariate Vieta-Fibonacci and Bivariate Vieta-Lucas Polynomials

The number of observed cases The number of parameters. ith case of the dichotomous dependent variable. the ith case of the jth parameter

A note on testing the covariance matrix for large dimension

best estimate (mean) for X uncertainty or error in the measurement (systematic, random or statistical) best

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions.

Chapter 9 Jordan Block Matrices

Multivariate Transformation of Variables and Maximum Likelihood Estimation

8 The independence problem

Lecture 12 APPROXIMATION OF FIRST ORDER DERIVATIVES

IRREDUCIBLE COVARIANT REPRESENTATIONS ASSOCIATED TO AN R-DISCRETE GROUPOID

1. a. Houston Chronicle, Des Moines Register, Chicago Tribune, Washington Post

Some Wgh Inequalities for Univalent Harmonic Analytic Functions

Initial-Value Problems for ODEs. numerical errors (round-off and truncation errors) Consider a perturbed system: dz dt

ENGI 4421 Propagation of Error Page 8-01

Algorithms Theory, Solution for Assignment 2

Feature Selection: Part 2. 1 Greedy Algorithms (continued from the last lecture)

1 Onto functions and bijections Applications to Counting

A nearly parametric solution to Selective Harmonic Elimination PWM Bao-Xin Shang 1, Shu-Gong Zhang 2, Na Lei 3 *, Jing-Yi Chen 4

means the first term, a2 means the term, etc. Infinite Sequences: follow the same pattern forever.

18.413: Error Correcting Codes Lab March 2, Lecture 8

( ) 2 2. Multi-Layer Refraction Problem Rafael Espericueta, Bakersfield College, November, 2006

Rademacher Complexity. Examples

10.2 Series. , we get. which is called an infinite series ( or just a series) and is denoted, for short, by the symbol. i i n

ESS Line Fitting

Lecture 25 Highlights Phys 402

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model

Overview of the weighting constants and the points where we evaluate the function for The Gaussian quadrature Project two

Reliability and Cost Analysis of a Series System Model Using Fuzzy Parametric Geometric Programming

Transcription:

Sgly Dagoally Implct Ruge-Kutta- Nytröm Method wth Reduced Phae-lag N. Seu, M. Sulema, F. Imal, ad M. Othma btract I th paper a gly dagoally mplct Ruge- Kutta-Nytröm (RKN) method developed for ecod-order ordary dfferetal equato wth perodcal oluto. he method ha algebrac order four ad phae-lag order eght at a cot of four fucto evaluato per tep. h ew method more accurate whe compared wth curret method of mlar type for the umercal tegrato of ecod-order dfferetal equato wth perodc oluto, ug cotat tep ze. Keyword Ruge-Kutta-Nytröm method; Dagoally mplct; Phae-lag; Ocllatory oluto. I. INRODUCION HIS paper deal wth umercal method for ecod-order ODE, whch the dervatve doe ot appear explctly, y f( x y) y( x ) y y( x ) y () 0 0 0 0 for whch t kow advace that ther oluto ocllatg. Such problem ofte are dfferet area of egeerg ad appled cece uch a celetal mechac, quatum mechac, elatodyamc, theoretcal phyc ad chemtry, ad electroc. -tage Ruge-Kutta-Nytröm (RKN) method for the umercal tegrato of the IVP gve by y yh b k y y h h bk y Maucrpt receved October 9, 00; reved Jauary 9, 0. h work wa upported part by the UPM Reearch Uverty Grat Scheme (RUGS) uder Grat 05-0-0-0900RU ad Fudametal Reearch Grat Scheme (FRGS) uder Grat 0-07-0-97FR N. Seu wth the Departmet of Mathematc ad Ittute for Mathematcal Reearch, Uvert Putra Malaya, 4400 UPM Serdag, Selagor, MLYSI. (correpodg author phoe: 60-8946-6848; fax: 60-8947958; e-mal: razak@math.upm.edu.my ). M. Sulema ad F. Imal are wth Departmet of Mathematc ad Ittute for Mathematcal Reearch, Uvert Putra Malaya, 4400 UPM Serdag, Selagor, MLYSI. (e-mal: mohamed@math.upm.edu.my, fudzah@math.upm.edu.my ). M. Othma wth the Ittute for Mathematcal Reearch, Uvert Putra Malaya, 4400 UPM Serdag, Selagor, MLYSI. (e-mal: mothma@fktm.upm.edu.my). () where y j j j k f x chy ch h a k he RKN parameter ajbj bj ad cj are aumed to be real ad the umber of tage of the method. Itroduce the - dmeoal vector cb ad b ad matrx, where c [ ccc ] b [ bb b ] b[ b b b ] [ a j ] repectvely. RKN method ca be dvded to two broad clae: explct ( a jk 0, k j ) ad mplct ( a jk 0, k > j). he latter cota the cla of dagoally mplct RKN (DIRKN) method for whch all the etre the dagoal of are equal. he RKN method above ca be expreed Butcher otato by the table of coeffcet c b b Geerally problem of the form () whch have perodc oluto ca be dvded to two clae. he frt cla cot of problem for whch the oluto perod kow a pror. he ecod cla cot of problem for whch the oluto perod tally ukow. Several umercal method of varou type have bee propoed for the tegrato of both clae of problem. See Stefel ad Bett [], va der Houwe ad Sommejer [], Gautch [6] ad other. Whe olvg () umercally, atteto ha to be gve to the algebrac order of the method ued, ce th the ma crtero for achevg hgh accuracy. herefore, t derable to have a lower tage RKN method wth maxmal order. h wll lee the computatoal cot. If t tally kow that the oluto of () of perodc ature the t eetal to coder phae-lag (or dpero) ad amplfcato (or dpato). hee are actually two type of trucato error. he frt the agle betwee the true ad the approxmated oluto, whle the ecod the ace from a tadard cyclc oluto. I th paper we wll derve a ew dagoally mplct RKN method wth three-tage fourth-order wth dpero of hgh order. umber of umercal method for th cla of problem of the explct ad mplct type have bee extevely developed. For example, va der Houwe ad Sommejer [], Smo, Dma ad Sderd, [5], ad Seu, Sulema

ad Imal [8] have developed explct RKN method of algebrac order up to fve wth dpero of hgh order for olvg ocllatory problem. For mplct RKN method, ee for example va der Houwe ad Sommejer [], Sharp, Fe ad Burrage [4] ad Imo, Otuta ad Ramamoha [7]. I th paper a dpero relato mpoed ad together wth algebrac coo to be olved explctly. I the followg ecto the cotructo of the ew four-tage fourth-order dagoally mplct RKN method decrbed. It coeffcet are gve ug the Butcher tableau otato. Fally, umercal tet o ecod order dfferetal equato problem poeg a ocllatory oluto are performed. II. NLYSIS OF PHSE-LG I th ecto we hall dcu the aaly of phae-lag for RKN method. he frt aaly of phae-lag wa carred out by Bura ad Ngro [0]. he followed by Gladwell ad homa [5] for the lear multtep method, ad for explct ad mplct Ruge-Kutta(-Nytrom) method by va der Houwe ad Sommejer [], []. he phae aaly ca be dvded two part; homogeeou ad homogeeou compoet. Followg va der Houwe ad Sommejer [], homogeeou phae error cotat tme, meawhle the homogeeou phae error are accumulated a creae. hu, from that pot of vew we wll cofe our aaly to the phae-lag of homogeeou compoet oly. he phae-lag aaly of the method () vetgated ug the homogeeou tet equato y ( ) y( t) () lteratvely the method () ca be wrtte a where y h ( ) f t ch Y y y h h b f( t ch Y) y y b y j j j Y y ch h a f( t chy ) By applyg the geeral method () to the tet equato () we obta the followg recurve relato a how by Papageorgou, Famel ad toura [4] y y D z h, hy hy Hb ( I H) e Hb ( I H) c DH ( ) Hb ( I H) e Hb ( I H) c where H z e () c ( c c m ). Here D(H) the tablty matx of the RKN method ad t charactertc polyomal tr( Dz ( )) det( Dz ( )) 0, (4) (5) the tablty polyomal of the RKN method. Solvg dfferece ytem (5), the computed oluto gve by y c co( ) (6) he exact oluto of () gve by yt ( ) co( z) (7) Eq. (6) ad (7) led u to the followg defto. Defto. (Phae-lag). pply the RKN method () to (). q he we defe the phae-lag ( z) z. If ( z) O( z ), the the RKN method ad to have phae-lag order q. doally, the quatty ( z) called r amplfcato error. If ( z) O( z ), the the RKN method ad to have dpato order r. Let u deote R( z ) trace( D) ad S( z ) det( D) From Defto, t follow that Let u deote Rz ( ) ( z) zco S( z ) S( z ) R( z ) ad S( z ) the followg form z z Rz ( ) ˆ ( + z ) z z Sz ( ), ˆ ( + z ) where ˆ dagoal elemet the Butcher tableau. Here the eceary coo for the fourth-order accurate dagoally mplct RKN method () to have hae-lag order eght term of ad gve by (8) (9) 6 4 4 (0) 60 8 4 4 4 4 6 0 () 45 40 Notce that the fourth-order method already dperve order four ad dpatve order fve. Furthermore dperve order eve ad dpatve order odd. III. CONSRUCION OF HE MEHOD I the followg we hall derve a four-tage fourth-order accurate dagoally mplct RKN method wth dperve order eght, by takg to accout the dpero relato

ecto II. he RKN parameter mut atfy the followg algebrac coo to fd fourth-order accuracy a gve Harer ad Waer []. order order order b () b b c () bc b c 6 6 (4) order 4 bc b c b ajcj 4 6 4 4. (5) For mot method the c eed to atfy c aj ( ) (6) j four-tage method of algebrac order four ( p 4 ) wth dperve order eght ( q 8 ) ad dpatve order fve ( r 5 ) ow codered. he coo ()-(6) ad dpero relato (0)-() formed thrtee olear equato wth etee varable to be olved. Now, from algebrac coo ()-(6) ad phae-lag relato of order x (0) ad lettg be a free parameter, the we olve t multaeouly. herefore the followg oluto of oeparameter famly obta c c c c a 6 6 6 6 a a4 a a4 a4 0 6 6 a a a a b b b 0b 4 44 4 (80 ) b 0( 4 4 88 7 ) 60 55 60 0 b4 5( 4 4 88 7 ) From the above oluto, we are gog to derve a method wth dpero of order eght. he eght order dpero relato () eed to be atfed ad th ca be wrtte term of RKN parameter whch correpod to the above famly of oluto yeld the followg equato 7 6 6 5 5 (5806080 4550 4550 490 967680 4 4 60480 8440 80640 4768 44856 976 94 75585 49 ) [0960( )] 00 ad olvg for yeld -0.755795, -0.08545609, 0.0479776, 0.684065,0.49098846, 0.684677664, ad -0.056647. Numercal reult how that choog -0.08545609 wll gve u more accurate cheme compared to the other ad we metoed here oe fourth-order (p=4) wth dperve order eght (q=8) method. For -0.08545609, the followg method wll be produced. h method wll be deoted by DIRKN4(4,8)NEW (ee able I) c able : : he DIRKN(4,6) method 6 6 0 6 6 0 0 6 6 0 b b4 4 0 0 where c =-0.7049006, b =0.957499, b 4 =0.604875, ad = =0.0454747 h method ha PLE BLE I HE DIRKN4(4,8)NEW MEHOD (5) (5) 66970 ad 670. able II compare the properte of our method wth the method derved by va der Houwe ad Sommejer [0], Sharp, Fe ad Burrage [4] ad Imo, Otuta ad Ramamoha [7]. IV. PROBLEM ESED I th ecto we ue our method to olve homogeeou ad homogeeou problem whoe exact oluto cot of a rapdly or/ad a lowly ocllatg fucto. For purpoe of llutrato, we wll compare our reult wth thoe derved by ug three method; DIRKN three-tage fourth-order derved by va der Houwe ad Sommejer [0] ad Imo, Otuta, Ramamoha [7], three-tage fourth-order dperve order x derved by Sharp, Fe ad Burrage [4] ad four-tage fourth-order derved by l-khaaweh, Imal, Sulema [].

BLE II SUMMRY OF HE CHRCERISIC OF HE FOURH-ORDER DIRKN MEHODS Problem (Homogeou) d y t () 00 yt ( ) y(0) y(0) Exact oluto y() t (0) t co(0) t Problem Method q d d y t 5 () yt ( ) t y(0) y(0) Exact oluto yt ( ) ( t) co( t) t Source : lle ad Wg [9] Problem (Ihomogeeou ytem) d y () t vy ( x) vft ( ) f( t) y(0) a f(0) y (0) f(0) d y () t vy () tvft () f() t y (0) f(0) y (0) va f(0) ( p ) Exact oluto y() t aco( vt) f() t y() t a( vt) f() t f () t 005t choe to be e ad parameter v ad a are 0 ad 0. repectvely. Source : Lambert ad Wato [7] DIRKN4(4,8)NEW 8 5 4.84 0 67 0 DIRKN(4,4)IMONI 4-75 0 DIRKN(4,4)HS 4 4 4 0 65 0 4 DIRKN(4,6)SHRP 6 0 0 85 0 DIRKNRaed 4 80 0. 0 Notato : q Dpero order, d Dpato cotat ( p ) Error coeffcet for y ( p) Error coeffcet for y ( p) 6 0 0 59 0 4 66 0 4 7 0 Problem 4 ( almot Perodc Orbt problem) d y t d y () t () y ( t) 000co( t) y (0) y (0) 0 y( t) 000( t) y(0) 0 y (0) 09995 Exact oluto y ( t ) co( t ) 0 0005 t ( t ), y ( t ) ( t ) 0 0005 t co( t ) Source : Stefel ad Bett [] he followg otato are ued able III-VI: DIRKN4(4,8)NEW : four-tage fourth-order dperve order eght method wth mall dpato cotat ad prcpal local trucato error derved th paper. DIRKN(4,4)IMONI : three-tage fourth-order derved by Imo, Otuta ad Ramamoha [7]. DIRKN(4,4)HS : three-tage fourth-order dperve order four derved by va der Houwe ad Sommejer [0]. DIRKN(4,6)SHRP : three-tage fourth-order dperve order x a Sharp, Fe ad Burrage [4]. DIRKN4(4,4)Raed : four-tage fourth-order drved by l-khaaweh, Imal, Sulema []. V. NUMERICL RESULS he reult for the four problem above are tabulated able III-VI. Oe meaure of the accuracy of a method to exame the Emax( ), the maxmum error whch defed by Emax( ) max y( t) y t0 where t t0 h? h able III-VI how the abolute maxmum error for DIRKN4(4,8)NEW, DIRKN(4,4)IMONI, DIRKN(4,4)HS, DIRKN(,6)SHRP ad DIRKN4(4,4)Raed method whe olvg Problem -4 wth three dfferet tep value. From umercal reult able III-VI, we oberved that the ew method more accurate compared wth DIRKN(4,4)IMONI, DIRKN(4,4)HS ad DIRKN4(4,4)Raed method whch do ot relate to the dpero order of the method. lo the ew method more accurate compared wth DIRKN(4,6)SHRP method becaue the ew method ha dpero order eght whch the hghet ad alo the dpato cotat for our method maller tha the DIRKN(4,6)SHRP method (ee able II).

BLE III COMPRING OUR RESULS WIH HE MEHODS IN HE LIERURE FOR PROBLEM h Method =00 =000 =4000 0.005 DIRKN4(4,8)NEW.494(-9).0464(-7) 7.766(-7) DIRKN(4,4)IMONI.5646(-).46(-) 4.7069(-) DIRKN(4,4)HS.56(-7).689(-6) 5.84(-6) DIRKN(4,6)SHRP.050(-7).09(-6).0(-5) DIRKN4(4,4)Raed 9.774(-6) 9.904(-5).79(-4) 0.005 DIRKN4(4,8)NEW.54(-9).56(-8) 5.047(-7) DIRKN(4,4)IMONI.0(-).8480(-) 5.6(-) DIRKN(4,4)HS 6.6977(-7) 6.6966(-6).78(-5) DIRKN(4,6)SHRP.5569(-6).564(-5).055(-4) DIRKN4(4,4)Raed.48(-4).4849(-) 5.99(-) 0.0 DIRKN4(4,8)NEW 4.5984(-8) 4.05(-7).875664(-6) DIRKN(4,4)IMONI 5.9680(-) 4.6(-) 9.8605(-) DIRKN(4,4)HS.05(-5).6(-4).95597(-) DIRKN(4,6) SHRP.4(-4).448(-).6707(-) DIRKN4(4,4)Raed.699(-).786(-) 9.56865(-) BLE IV COMPRING OUR RESULS WIH HE MEHODS IN HE LIERURE FOR PROBLEM h Method =00 =000 =4000 0.065 DIRKN4(4,8)NEW.468(-8).764(-8).8789(-8) DIRKN(4,4)IMONI 5.96(-) 5.(-).004(-) DIRKN(4,4)HS 6.80(-7) 6.86(-6).794(-5) DIRKN(4,6) SHRP 4.007(-6) 4.06(-5).649(-4) DIRKN4(4,4)Raed 5.8594(-5) 5.8706(-4).509(-) 0.5 DIRKN4(4,8)NEW.475(-7) 5.87(-7).0986(-6) DIRKN(4,4)IMONI.04(-).00(-).69(-) DIRKN(4,4)HS.087(-5).090(-4) 4.85(-4) DIRKN(4,6)SHRP.006(-4).98(-) 5.657(-) DIRKN4(4,4)Raed 8.070(-4) 8.09(-).9(-) 0.5 DIRKN4(4,8)NEW 5.8948(-6).77(-5).5775(-5) DIRKN(4,4)IMONI.94(-).868(-) 6.958(-) DIRKN(4,4)HS.7(-4).7444(-) 7.0007(-) DIRKN(4,6) SHRP 4.480(-) 4.644(-).950(-) DIRKN4(4,4)Raed.897(-).969(-) 5.6(-) BLE V COMPRING OUR RESULS WIH HE MEHODS IN HE LIERURE FOR PROBLEM h Method =00 =000 =4000 0.005 DIRKN4(4,8)NEW 8.6798(-0).090(-8).509(-7) DIRKN(4,4)IMONI 5.9756(-) 4.600(-) 9.50(-) DIRKN(4,4)HS.9675(-7).9897(-6).608(-5) DIRKN(4,6)SHRP.8995(-6).9004(-5) 7.6048(-5) DIRKN4(4,4)Raed.9(-5).9(-4).650(-) 0.005 DIRKN4(4,8)NEW.64(-8) 9.49(-8) 4.04(-7) DIRKN(4,4)IMONI.7(-) 7.005(-) 9.90(-) DIRKN(4,4)HS 6.468(-6) 6.496(-5).544(-4) DIRKN(4,6)SHRP 6.59(-5) 6.776(-4).498(-) DIRKN4(4,4)Raed 4.66(-4) 4.6689(-).8754(-) 0.0 DIRKN4(4,8)NEW 5.54(-7).456(-6).84(-5) DIRKN(4,4)IMONI.9988(-) 9.040(-).000(-) DIRKN(4,4)HS.04(-4).056(-) 4.0589(-) DIRKN(4,6) SHRP.089(-).85(-).66(-) DIRKN4(4,4)Raed 7.506(-) 7.7409(-).57(-) BLE VI COMPRING OUR RESULS WIH HE MEHODS IN HE LIERURE FOR PROBLEM 4 h Method =00 =000 =4000 0.065 DIRKN4(4,8)NEW.495(-8).748(-8) 4.6060(-8) DIRKN(4,4)IMONI.998(-) 4.09(-).08(-) DIRKN(4,4)HS 5.605(-7) 5.808(-6).06(-5) DIRKN(4,6) SHRP.498(-6).68(-5).9990(-4) DIRKN4(4,4)Raed 4.8(-5) 4.777(-4).486(-) 0.5 DIRKN4(4,8)NEW.578(-7) 4.899(-7).0(-6) DIRKN(4,4)IMONI 7.90(-) 7.67(-).76(-) DIRKN(4,4)HS 7.6595(-6) 7.9664(-5) 4.794(-4) DIRKN(4,6)SHRP 9.794(-5) 9.749(-4) 5.67(-) DIRKN4(4,4)Raed 5.69(-4) 5.8856(-).0(-) 0.5 DIRKN4(4,8)NEW 5.84(-6).0(-5).604(-5) DIRKN(4,4)IMONI.995(-).655(-) 6.7067(-) DIRKN(4,4)HS.0(-4).75(-) 7.0099(-) DIRKN(4,6) SHRP.09(-).895(-).996(-) DIRKN4(4,4)Raed 9.0479(-) 9.459(-) 5.0(-) Notato :.45(-4) mea 45 0 4 VI. CONCLUSION I th paper we have derved dagoally mplct four-tage fourth-order ad dperve order eght wth mall dpato cotat ad prcpal local trucato error. We have alo performed varou umercal tet. From the reult tabulated able III-VI, we coclude that the ew method more accurate for tegratg ecod-order equato poeg a ocllatory oluto whe compared to the curret DIRKN method derved by va der Houwe ad Sommejer [0], Sharp, Fe ad Burrage [4], Imo, Otuta ad Ramamoha [7] ad l-khaaweh, Imal, Sulema []. REFERENCES [] D. I. Okubor ad R. D. Skeel, Caocal Ruge-Kutta-Nytröm method of order fve ad x, J. Comput. ppl. Math., vol. 5, pp. 75-8, 994. [] E. Harer ad G. Waer, heory for Nytrom Method, Numer. Math., vol. 5, pp. 8-400, 975. [] E. Stefel ad D.G. Bett, Stablzato of Cowell method, Numer. Math, vol., pp. 54-75, 969. [4] G. Papageorgou, I. h. Famel ad Ch. toura, P-table gle dagoally mplct Ruge-Kutta-Nytröm method, Numercal lgorthm, vol. 7, pp. 45-5, 998. [5] I. Gladwell ad R. homa, Dampg ad phae aaly of ome method for olvg ecod order ordary dfferetal equato, Iterat. J. Numer. Method Egrg., vol. 9, pp. 495-50, 98. [6] J.C. Butcher, Numercal Method for Ordary Dfferetal Equato, Wley & So Ltd., Eglad, 00. [7] J. D. Lambert ad I.. Wato, Symmetrc multtep method for perodc tal-value problem, J. It. Math pplc, vol. 8, pp. 89-0, 976. [8] J. D. Lambert, Numercal Method for ordary Dfferetal Sytem-he Ital Value Problem, Wley & So Ltd., Eglad, 99. [9] J. R. Dormad, Numercal Method for Dfferetal Equato, CRC Pre, Ic, Florda, 996. [0] L. Bura ad L. Ngro, oe-tep method for drect tegrato of tructural dyamc equato, Iterat. J. Numer. Method Egrg, vol. 5, pp. 685-699, 980. [] M. M. Chawla ad S. R Sharma, Iterval of perodcty ad abolute tablty of explct Nytrom method for y f( x y), BI, vol., pp. 455-469, 98. [] P. J. va der Houwe ad B. P. Sommejer, Explct Ruge-Kutta(-

Nytröm) method wth reduced phae error for computg ocllatg oluto, SIM J. Numer. al., vol. 4, o., pp. 595-67, 987. [] P. J. va der Houwe ad B. P. Sommejer, Dagoally mplct Ruge- Kutta(-Nytröm) method for ocllatory problem, SIM J. Numer. al., vol. 6, o., pp. 44-49, 989. [4] P. W. Sharp, J. M. Fe ad K. Burrage, wo-tage ad hree-tage Dagoally Implct Ruge-Kutta-Nytröm Method of Order hree ad Four, J. Of Numercal aly, vol. 0, pp. 489-504, 990. [5]. E. Smo, E. Dma,.B. Sderd, Ruge-Kutta-Nytröm method for the tegrato of pecal ecod-order perodc tal-value problem, J. Comput. ppl. Mat., vol 5, pp. 7-6, 994. [6] W. Gautch, Numercal tegrato of ordary dfferetal equato baed o trgoometrc polomal, Numer. Math. vol., pp. 8-97, 96. [7] S. O. Imo, F. O. Otuta ad. R. Ramamoha, Embedded mplct Ruge-Kutta-Nytröm method for olvg ecod-order dfferetal equato, Iteratoal Joural of Computer Mathematc, vol. 8, o., pp. 777-784, 006 [8] N. Seu, M. Sulema ad F. Imal. (009, Jue). embedded explct Ruge Kutta Nytröm method for olvg ocllatory problem. 80(). valable: tack.op.org/physcr/80/05005 [9] R. C. lle, Jr. ad G. M. Wg, varat mbeddg algorthm for the oluto of homogeeou lear two-pot boudary value problem, J. Computer Phyc, vol. 4, pp. 40-58, 974. [0] B.P. Sommejer, ote o a dagoally mplct Ruge-Kutta- Nytr om method, J. Comp. ppl. Math., vol 9, pp. 95-99, 987. [] N. Seu, M. Sulema, F. Imal, ad M. Othma, New Dagoally Implct Ruge-Kutta-Nytröm Method for Perodc IVP, WSES raacto o Mathematc, vol. 9, pp. 679-688, 00. [] l-khaaweh, R.., Imal, F., Sulema, M., Embedded dagoally mplct Ruge-Kutta-Nytröm 4() par for olvg pecal ecod-order IVP. ppl. Math. Comp., vol. 90, 007, 80-84.