Numerical Solution of Ordinary Differential Equations

Similar documents
NUMERICAL DIFFERENTIATION

PART 8. Partial Differential Equations PDEs

Chapter 12. Ordinary Differential Equation Boundary Value (BV) Problems

ORDINARY DIFFERENTIAL EQUATIONS EULER S METHOD

One-sided finite-difference approximations suitable for use with Richardson extrapolation

6.3.4 Modified Euler s method of integration

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

Formal solvers of the RT equation

A Hybrid Variational Iteration Method for Blasius Equation

Resource Allocation and Decision Analysis (ECON 8010) Spring 2014 Foundations of Regression Analysis

2 Finite difference basics

Trees and Order Conditions

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Consistency & Convergence

Introduction to Vapor/Liquid Equilibrium, part 2. Raoult s Law:

A new Approach for Solving Linear Ordinary Differential Equations

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system

Numerical Heat and Mass Transfer

ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family

Numerical Solutions of a Generalized Nth Order Boundary Value Problems Using Power Series Approximation Method

Modelli Clamfim Equazione del Calore Lezione ottobre 2014

Chapter Newton s Method

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS

Kinematics in 2-Dimensions. Projectile Motion

Global Sensitivity. Tuesday 20 th February, 2018

6.3.7 Example with Runga Kutta 4 th order method

E91: Dynamics. E91: Dynamics. Numerical Integration & State Space Representation

Note 10. Modeling and Simulation of Dynamic Systems

New Method for Solving Poisson Equation. on Irregular Domains

Week3, Chapter 4. Position and Displacement. Motion in Two Dimensions. Instantaneous Velocity. Average Velocity

e i is a random error

Modelli Clamfim Equazioni differenziali 7 ottobre 2013

Appendix B. The Finite Difference Scheme

FTCS Solution to the Heat Equation

Lecture 12: Discrete Laplacian

Lecture 10 Support Vector Machines II

System in Weibull Distribution

EEE 241: Linear Systems

Outline. Review Numerical Approach. Schedule for April and May. Review Simple Methods. Review Notation and Order

Section 8.3 Polar Form of Complex Numbers

Numerical Methods. ME Mechanical Lab I. Mechanical Engineering ME Lab I

Chapter 4: Root Finding

Physics 5153 Classical Mechanics. D Alembert s Principle and The Lagrangian-1

ME 501A Seminar in Engineering Analysis Page 1

A NUMERICAL COMPARISON OF LANGRANGE AND KANE S METHODS OF AN ARM SEGMENT

Integrals and Invariants of Euler-Lagrange Equations

Lecture 13 APPROXIMATION OF SECOMD ORDER DERIVATIVES

Inductance Calculation for Conductors of Arbitrary Shape

OPTIMISATION. Introduction Single Variable Unconstrained Optimisation Multivariable Unconstrained Optimisation Linear Programming

Modelli Clamfim Equazioni differenziali 22 settembre 2016

Workshop: Approximating energies and wave functions Quantum aspects of physical chemistry

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Bézier curves. Michael S. Floater. September 10, These notes provide an introduction to Bézier curves. i=0

Salmon: Lectures on partial differential equations. Consider the general linear, second-order PDE in the form. ,x 2

Second Order Analysis

PHYS 705: Classical Mechanics. Calculus of Variations II

Statistics for Business and Economics

Grid Generation around a Cylinder by Complex Potential Functions

FREQUENCY DISTRIBUTIONS Page 1 of The idea of a frequency distribution for sets of observations will be introduced,

Thermal-Fluids I. Chapter 18 Transient heat conduction. Dr. Primal Fernando Ph: (850)

Statistics for Economics & Business

Asymptotics of the Solution of a Boundary Value. Problem for One-Characteristic Differential. Equation Degenerating into a Parabolic Equation

The Feynman path integral

PHYS 705: Classical Mechanics. Canonical Transformation II

SIMPLE LINEAR REGRESSION

x yi In chapter 14, we want to perform inference (i.e. calculate confidence intervals and perform tests of significance) in this setting.

Supplementary Notes for Chapter 9 Mixture Thermodynamics

Code_Aster. Identification of the model of Weibull

Kernel Methods and SVMs Extension

Bezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0

Canonical transformations

Implicit Integration Henyey Method

Polynomial Regression Models

Correlation and Regression. Correlation 9.1. Correlation. Chapter 9

DUE: WEDS FEB 21ST 2018

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION

Comparison of Second Order Numerical Techniques for Linear and Non-Linear ODEs

Some modelling aspects for the Matlab implementation of MMA

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA

4DVAR, according to the name, is a four-dimensional variational method.

Using the estimated penetrances to determine the range of the underlying genetic model in casecontrol

RELIABILITY ASSESSMENT

6. Stochastic processes (2)

12. The Hamilton-Jacobi Equation Michael Fowler

6. Stochastic processes (2)

APPENDIX 2 FITTING A STRAIGHT LINE TO OBSERVATIONS

Chapter 13: Multiple Regression

STAT 3014/3914. Semester 2 Applied Statistics Solution to Tutorial 13

Linear Regression Analysis: Terminology and Notation

ALGORITHM FOR THE CALCULATION OF THE TWO VARIABLES CUBIC SPLINE FUNCTION

Convexity preserving interpolation by splines of arbitrary degree

NAME and Section No.

Properties of Least Squares

Differentiating Gaussian Processes

Introduction to Statistical Methods

Transcription:

Numercal Methods (CENG 00) CHAPTER-VI Numercal Soluton of Ordnar Dfferental Equatons 6 Introducton Dfferental equatons are equatons composed of an unknown functon and ts dervatves The followng are examples of dfferental equatons: dv dt c = g v (6a) m d x dx c kx = 0 (6b) dt dt m Dfferental equatons can be classfed as ordnar dfferental equatons (ODE) or partal dfferental equatons (PDE) ODEs are equatons n whch the functon nvolves one ndependent varable PDE's nvolve functons two or more varables Both Eq(6a) and Eq(6b) are ordnar dfferental equatons Dfferental equatons are also classfed as to ther order For Eq(6a) s a frst-order equaton because the hghest dervatve s a frst dervatve A second order equaton ncludes a second dervatve as n Eq (6b) above Smlarl an n th order equaton ncludes an nth dervatve Ths chapter s devoted to solvng ordnar dfferental equatons of the form d dx ( x ) = f (6) One-step Runge-Kutta (RK) methods can be generall expressed as = φh (63) Accordng to ths equaton the slope estmate of φ s used to extrapolate from an old value to new value over a dstance h Ths formula can be appled step b step to compute values of The smplest approach s to use the dfferental equaton to estmate the slope n the form of frst dervatve at x Academc ear 008/9 Instructor: Zerhun A

Numercal Methods (CENG 00) 6 Euler's Method The frst dervatve provdes a drect estmate of the slope at x That s ( ) φ = f x (64) f ( x ) s the dfferental equaton evaluated at x and Ths estmate can be substtuted nto Eq (63): = f ( x )h (65) Ths formula s referred to as Euler's method Error Analss for Euler's Method As n an other numercal procedure the numercal soluton of ODEs nvolves two tpes of errors () Truncaton error and () Round-off errors The truncaton errors are composed of two parts The frst s a local truncaton error that results from the applcaton of the method over a sngle step The second s a propagated truncaton error that results from the approxmaton produced durng the prevous steps The sum of the two s the total or global truncaton error Insght nto the magntude and propertes of the truncaton error can be ganed b dervng Euler's method drectl from the Talor seres expanson Remember that the dfferental equaton to be solved wll be of the general form ( x ) = f (66) = d dx and x and are the ndependent and the dependent varables respectvel If the soluton has contnuous dervatves t can be represented b a Talor seres expanson about a startng value ( x ) ( n) =! n! n h h h Rn (67) h = x x and n R = the remander term defned b Academc ear 008/9 Instructor: Zerhun A

Numercal Methods (CENG 00) R n = ( n ) ( ξ ) ( n )! h n (68) ξ les some n the nterval from x to x developed b substttng Eq(66) nto Eq(67) and (68) to eld An alternatve form can be n ( h ) ( n ( x ) f ) ( x ) n ( h ) f = f ( x ) h h O (69)! n! O specfes that the local truncaton error s proportonal to the step sze rased to the ( n )th power Thus we can see that Euler's method corresponds to the Talor seres up to and ncludng the term f ( x )h The true local truncaton error E t n the Euler method s thus gven as ( x ) n ( h ) f E = t h O (60)! For suffcentl small h the errors n the terms n Eq(60) usuall decrease as the order ncreases and the result s often represented as E a ( x ) f = h (6)! or ( ) E a = O h (6) E a = the approxmate local truncaton error Academc ear 008/9 Instructor: Zerhun A 3

Numercal Methods (CENG 00) Example Usng Euler s method solve the followng ntal value problem 63 Runge-Kutta Methods Runge-Kutta (RK) methods acheve the accurac of a Talor seres approach wthout havng requrng the calculaton of hgher dervatves Man varatons exst but all can be cast n the generalzed form: = ( x h)h (63) φ ( x h) φ s called an ncrement functon whch can be nterpreted as a representatve slope over the nterval The ncrement functon can be wrtten n the general form as φ = a k a k a k n n (64) Academc ear 008/9 Instructor: Zerhun A 4

Numercal Methods (CENG 00) the a 's are constants and the k 's are ( ) k = f x (65a) ( x p h q k h) k = f (65b) ( x p h q k h) k3 = f (65c) k ( x p h q k h q k h) = f n n n n n (65d) n Notce that the k's are recurrence relatonshps Varous tpes of Runge-Kutta methods can be devsed b emplong dfferent numbers of terms n the ncrement functon as specfed b n Note that the frst-order RK method wth n = s n fact Euler's method Once n s chosen values for the a's p's and q's are evaluated b settng eqn 64 equal to terms n a Talor seres expanson Fourth-order Runge-Kutta Methods The most popular and relatvel accurate RK methods are the fourth order There are an nfnte number of versons The followng known as the classcal fourth-order RK method s the most commonl used form ( k k k k ) = 3 6 (66) k = f ( x ) k k 3 = f x h kh = f x h k h ( x h k h) k4 = f 3 4 Academc ear 008/9 Instructor: Zerhun A 5

Numercal Methods (CENG 00) NB The n th order RK methods have local errors of o(h n ) and global error of o(h n ) Example Usng 4 th order Runge-Kutta methods solve the followng ntal value problem Academc ear 008/9 Instructor: Zerhun A 6