Solving scalar IVP s : Runge-Kutta Methods

Similar documents
Multistage Methods I: Runge-Kutta Methods

Fourth Order RK-Method

Euler s Method, cont d

Numerical Methods for Differential Equations

Consistency and Convergence

Applied Math for Engineers

Ordinary Differential Equations II

You may not use your books, notes; calculators are highly recommended.

Ordinary differential equation II

Differential Equations

PowerPoints organized by Dr. Michael R. Gustafson II, Duke University

The family of Runge Kutta methods with two intermediate evaluations is defined by

PowerPoints organized by Dr. Michael R. Gustafson II, Duke University

Part IB Numerical Analysis

Numerical solution of ODEs

ECE257 Numerical Methods and Scientific Computing. Ordinary Differential Equations

Chapter 5 Exercises. (a) Determine the best possible Lipschitz constant for this function over 2 u <. u (t) = log(u(t)), u(0) = 2.

Ordinary Differential Equations II

Numerical Methods for the Solution of Differential Equations

Numerical Methods for Differential Equations

Introduction to the Numerical Solution of IVP for ODE

Numerical Analysis II. Problem Sheet 9

Numerical Methods - Initial Value Problems for ODEs

Modeling & Simulation 2018 Lecture 12. Simulations

Numerical Differential Equations: IVP

Introduction to standard and non-standard Numerical Methods

2.29 Numerical Fluid Mechanics Fall 2011 Lecture 20

Do not turn over until you are told to do so by the Invigilator.

MTH 452/552 Homework 3

Chap. 20: Initial-Value Problems

NUMERICAL SOLUTION OF ODE IVPs. Overview

Do not turn over until you are told to do so by the Invigilator.

Differential Equations

Four Point Gauss Quadrature Runge Kuta Method Of Order 8 For Ordinary Differential Equations

Lecture IV: Time Discretization

Initial-Value Problems for ODEs. Introduction to Linear Multistep Methods

Review Higher Order methods Multistep methods Summary HIGHER ORDER METHODS. P.V. Johnson. School of Mathematics. Semester

Lecture Notes on Numerical Differential Equations: IVP

Mini project ODE, TANA22

What we ll do: Lecture 21. Ordinary Differential Equations (ODEs) Differential Equations. Ordinary Differential Equations

Ordinary Differential Equations

Lecture V: The game-engine loop & Time Integration

Lecture Notes to Accompany. Scientific Computing An Introductory Survey. by Michael T. Heath. Chapter 9

Initial value problems for ordinary differential equations

Section 7.4 Runge-Kutta Methods

Runge-Kutta Theory and Constraint Programming Julien Alexandre dit Sandretto Alexandre Chapoutot. Department U2IS ENSTA ParisTech SCAN Uppsala

Scientific Computing: An Introductory Survey

Scientific Computing with Case Studies SIAM Press, Lecture Notes for Unit V Solution of

Efficient Simulation of Physical System Models Using Inlined Implicit Runge Kutta Algorithms

Math 128A Spring 2003 Week 12 Solutions

Initial value problems for ordinary differential equations

A Class of an Implicit Stage-two Rational Runge-Kutta Method for Solution of Ordinary Differential Equations

CS520: numerical ODEs (Ch.2)

Solving Ordinary Differential equations

TABLE OF CONTENTS INTRODUCTION, APPROXIMATION & ERRORS 1. Chapter Introduction to numerical methods 1 Multiple-choice test 7 Problem set 9

ODE Runge-Kutta methods


The collocation method for ODEs: an introduction

Exponentially Fitted Error Correction Methods for Solving Initial Value Problems

8.1 Introduction. Consider the initial value problem (IVP):

Almost Runge-Kutta Methods Of Orders Up To Five

Chapter 6 - Ordinary Differential Equations

Ordinary Differential Equations

PowerPoints organized by Dr. Michael R. Gustafson II, Duke University

Math Numerical Analysis Homework #4 Due End of term. y = 2y t 3y2 t 3, 1 t 2, y(1) = 1. n=(b-a)/h+1; % the number of steps we need to take

Research Article Diagonally Implicit Block Backward Differentiation Formulas for Solving Ordinary Differential Equations

PRECONDITIONING AND PARALLEL IMPLEMENTATION OF IMPLICIT RUNGE-KUTTA METHODS.

Ordinary differential equations - Initial value problems

Integrating-Factor-Based 2-Additive Runge Kutta methods for Advection-Reaction-Diffusion Equations

CS 450 Numerical Analysis. Chapter 9: Initial Value Problems for Ordinary Differential Equations

A Brief Introduction to Numerical Methods for Differential Equations

Solving PDEs with PGI CUDA Fortran Part 4: Initial value problems for ordinary differential equations

Linear Multistep Methods

EXAMPLE OF ONE-STEP METHOD

1 Ordinary Differential Equations

551614:Advanced Mathematics for Mechatronics. Numerical solution for ODEs School of Mechanical Engineering

Efficiency of Runge-Kutta Methods in Solving Simple Harmonic Oscillators

Lecture 4: Numerical solution of ordinary differential equations

Variable Step Size Differential Equation Solvers

A NOTE ON EXPLICIT THREE-DERIVATIVE RUNGE-KUTTA METHODS (ThDRK)

Logistic Map, Euler & Runge-Kutta Method and Lotka-Volterra Equations

Stability regions of Runge-Kutta methods. Stephan Houben Eindhoven University of Technology

Exponential Integrators

2 Numerical Methods for Initial Value Problems

AN OVERVIEW. Numerical Methods for ODE Initial Value Problems. 1. One-step methods (Taylor series, Runge-Kutta)

Investigation on the Most Efficient Ways to Solve the Implicit Equations for Gauss Methods in the Constant Stepsize Setting

Study the Numerical Methods for Solving System of Equation

Numerical Analysis. A Comprehensive Introduction. H. R. Schwarz University of Zürich Switzerland. with a contribution by

Validated Explicit and Implicit Runge-Kutta Methods

Butcher tableau Can summarize an s + 1 stage Runge Kutta method using a triangular grid of coefficients

NUMERICAL ANALYSIS 2 - FINAL EXAM Summer Term 2006 Matrikelnummer:

Ordinary Differential Equations

Chapter 8. Numerical Solution of Ordinary Differential Equations. Module No. 1. Runge-Kutta Methods

Numerical Integration of SDEs: A Short Tutorial

Lecture 5: Single Step Methods

Ordinary differential equations. Phys 420/580 Lecture 8

The Milne error estimator for stiff problems

Advanced methods for ODEs and DAEs

[ ] is a vector of size p.

Runge Kutta Theory and Constraint Programming

Transcription:

Solving scalar IVP s : Runge-Kutta Methods Josh Engwer Texas Tech University March 7, NOTATION: h step size x n xt) t n+ t + h x n+ xt n+ ) xt + h) dx = ft, x) SCALAR IVP ASSUMED THROUGHOUT: dt xt ) = x EXPLICIT ν-stage RUNGE-KUTTA METHOD: F = hft n, x n ) F = hft n + c h, x n + a F F = hft n + c h, x n + a F + a F ) F ν = hft n + c ν h, x n + a ν F + a ν F + + a ν,ν F ν ) x n+ = x n + w F + w F + + w ν F ν ) TEST IVP: { x = λx x) = } c a c a a c ν a ν a ν a ν,ν w w w ν w ν x = e λt IMPLICIT ν-stage RUNGE-KUTTA METHOD: F = hft n + c h, x n + a F + a F + + a ν F ν ) F = hft n + c h, x n + a F + a F + + a ν F ν ) F = hft n + c h, x n + a F + a F + + a ν F ν ) F ν = hft n + c ν h, x n + a ν F + a ν F + + a νν F ν ) x n+ = x n + w F + w F + + w ν F ν ) c a a a ν c a a a ν c a a a ν c ν a ν a ν a νν w w w ν c A w T REMARKS: Like Taylor Series Methods, RK-Methods are single step methods and have a high-order truncation error Unlike Taylor Series Methods, RK-Methods do not take derivatives, which would require symbolic programming In the Butcher tableau form, matrix A is strictly lower triangular and c = in explicit RK-Methods Implicit RK-Methods are better suited for stiff IVP s than explicit RK-Methods A unique implicit A-stable ν-stage RK-Method of order ν exists for each positive integer ν A k th -order explicit RK-Method, when applied to Test IVP, gives the k th -order Taylor approximation to e hλ e z A k th -order implicit RK-Method, when applied to Test IVP, gives a k th -order Padé approximation to e hλ e z th -order explicit RK-Methods are ideal as higher-order explicit RK-Methods require more stages as shown here: Order desired 5 6 7 8 Stages required or 5 6 7 or 8 9 or Copyright Josh Engwer Revised March 7,

st -ORDER RUNGE-KUTTA METHODS: F = hft Explicit) Euler s Method : x n+ = x n + hft n, x n ) n, x n ) x n+ = x n + F F = hft Implicit Euler s Method : x n+ = x n + hft n+, x n+ ) n + h, x n + F ) x n+ = x n + F nd -ORDER RUNGE-KUTTA METHODS: Heun s Method : Midpoint Method : Ralston s Method : F = hft n, x n ) F = hft n + h, x n + F ) x n+ = x n + F + F ) Trapezoidal Method : Lobatto IIIC) : A-stable RK : F = hft n, x n ) F = hft n + h, x n + F ) x n+ = x n + F F = hft n, x n ) F = hft n + h, x n + F ) x n+ = x n + F + F ) F = hft n, x n ) F = hft n + h, x n + F + F ) x n+ = x n + F + F ) F = hft n, x n + F F ) F = hft n + h, x n + F + F ) x n+ = x n + F + F ) { F = hft n + h, x n + F ) x n+ = x n + F } rd -ORDER RUNGE-KUTTA METHODS: F = hft n + h, x n + 5 F F ) Radau IIA) : F = hft n + h, x n + F + F ) x n+ = x n + F + F ) 5 th -ORDER RUNGE-KUTTA METHODS: F = hft n, x n ) F = hft n + h, x n + F ) Classic RK : F = hft n + h, x n + F ) F = hft n + h, x n + F ) x n+ = x n + F 6 + F + F + F ) 6 6 F = hft n, x n + F 6 F + F 6 ) F Lobatto IIIC) : = hft n + h, x n + F 6 + 5 F F 6 5 ) F = hft n + h, x n + F 6 + F + F 6 6 ) x n+ = x n + F 6 6 + F + F ) 6 F = hft n + + ] h, x 6 n + F + + ] F 6 ) A-stable RK : F = hft n + ] h, x 6 n + ] F 6 + F ) x n+ = x n + F + F ) 6 + 6 6 6 6 6 ) + ) 6 Copyright Josh Engwer Revised March 7,

EXAMPLE: Show that Ralston s method is a nd -order Runge-Kutta method F = hft n, x n ) F = hft n + h, x n + F ) x n+ = x n + F + F { x = λx x) = F = hλx n F = hλ x n + ] hλx n) = hλx n + hλ) x n hλx n + ] hλ) x n }, which we know the analytic solution is x = e λt x n+ = x n + hλx n] + x n+ = + hλ +! ] hλ) x n = x n + hλx n) + hλx n) + hλ) x n Let z := hλ C Then : x n+ = + z +! ] z x n Kz)x n Notice that e z = Kz) + Oz ) ie Kz) is the nd -order Taylor approximation of e z ] Therefore, Ralston s method is of order = EXAMPLE: Show that Implicit Euler is a st -order Runge-Kutta method F = hft n + h, x n + F ) x x n+ = x n + F n+ = x n + hft n+, x n+ ) x = λx, which we know the analytic solution is x = e x) = λt x n+ = x n + hλx n + F ) = x n + hλx n+ hλ)x n+ = x n x n+ = Let z := hλ C Then : ) x n+ = x n Kz)x n z Notice that Kz) is the, )-Padé approximation of e z Therefore, Implicit Euler method is of order + = ) x n hλ Padé approximations to e z m\n z z + z + z 6 + z + z z 6 z + z + z + 6 + z + z + 6z + z z 6 z 6z + z Copyright Josh Engwer Revised March 7,

EXAMPLE: Show that A-stable RK is a nd -order Runge-Kutta method { F = hft n + h, x n + F } ) x n+ = x n + F First, observe that x n+ = x n + F F = x n+ x n ) F = hf t n + h, x n + ) x n+ x n ) x n+ = x n + hf t n + h, x n + ) x n+ x = λx, which we know the analytic solution is x = e x) = λt x n+ = x n + hλ x n + ) x n+ x n+ = x n + hλx n + hλx n+ ) hλ x n+ = + ) + hλ x n x n+ = hλ ) hλ x n Let z := hλ C Then : + x n+ = z z ) x n x n+ = ) + z x n Kz)x n z Notice that Kz) is the, )-Padé approximation to e z Therefore, A-stable RK method is of order + = Padé approximations to e z m\n z z + z + z 6 + z + z z 6 z + z + z + 6 + z + z + 6z + z z 6 z 6z + z Copyright Josh Engwer Revised March 7,

EXAMPLE: Show that Radau IIA) is a rd -order Runge-Kutta method F = hft n + h, x n + 5 5 F F ) F = hft n + h, x n + F + F ) x n+ = x n + F + F First, observe that F = hft n+, x n+ ) F = hf Now, x n+ = x n + F + F = x n + F + hft n+, x n+ ) F = x n+ x n hft n+, x n+ ) t n + h, x n + 5 F ) hft n+, x n+ ) F = x n+ x n hft n+, x n+ ) 5 F = 5 9 x n+ 5 9 x n 5 6 hft n+, x n+ ) x n + 5 F 5 F = x n + 9 x n+ 5 9 x n 5 ] 6 hft n+, x n+ ) hft n+, x n+ )] x n + 5 F F = 9 x n + 5 9 x n+ 9 hft n+, x n+ ) Hence, F = hf t n + h, 9 x n + 5 9 x n+ ) 9 hft n+, x n+ ) x n+ = x n + t hf n + h, 9 x n + 5 9 x n+ ) 9 hft n+, x n+ ) + hft n+, x n+ ) x = λx, which we know the analytic solution is x = e x) = λt x n+ = x n + t hf n + h, 9 x n + 5 9 x n+ ) 9 hft n+, x n+ ) + hft n+, x n+ ) ) x n+ = x n + hλ 9 x n + 5 9 x n+ 9 hλx n+ + hλx n+ x n+ = x n + hλx n + 5 hλx n+ 6 hλ) x n+ + hλx n+ x n+ = x n + hλx n + hλx n+ 6 hλ) x n+ hλ) + 6 ] hλ) x n+ = + ] hλ) + x n x n+ = hλ) ] hλ) + x n ] 6 hλ) 6 + z Let z := hλ C Then, x n+ = x 6 z + z n Kz)x n Notice that Kz) is the, )-Padé approximation to e z Therefore, Radau IIA) is of order + = Padé approximations to e z m\n z z + z + z 6 + z + z z 6 z + z + z + 6 + z + z + 6z + z z 6 z 6z + z Copyright Josh Engwer Revised March 7,

References ] A Ackleh, E Allen, R B Kearfott, P Seshaiyer, Numerical Analysis: Theory, Methods, and Practice, CRC Press, 9 ] K Burrage and J C Butcher, Stability Criteria for Implicit Runge-Kutta Methods, SIAM Journal on Numerical Analysis, Vol 6, No Feb, 979), pp 6-57 ] JC Butcher, Numerical Methods for Ordinary Differential Equations Wiley & Sons, ] J C Butcher 7) Runge-Kutta methods Scholarpedia, 9):7 5] D Kincaid, W Cheney, Numerical Analysis: Mathematics of Scientific Computing Brooks Cole, Pacific Grove, CA, rd Edition, Copyright Josh Engwer Revised March 7,