Integration of 3-body encounter. Figure taken from

Slides:



Advertisements
Similar presentations
Differential Equations Brannan Copyright © 2010 by John Wiley & Sons, Inc. All rights reserved. Chapter 08: Series Solutions of Second Order Linear Equations.
Advertisements

SE301: Numerical Methods Topic 8 Ordinary Differential Equations (ODEs) Lecture KFUPM Read , 26-2, 27-1 CISE301_Topic8L8&9 KFUPM.
Computational Modeling for Engineering MECN 6040
CSE245: Computer-Aided Circuit Simulation and Verification Lecture Note 5 Numerical Integration Prof. Chung-Kuan Cheng 1.
Computational Methods in Physics PHYS 3437
Ordinary Differential Equations
2. Numerical differentiation. Approximate a derivative of a given function. Approximate a derivative of a function defined by discrete data at the discrete.
1cs542g-term Notes  Notes for last part of Oct 11 and all of Oct 12 lecture online now  Another extra class this Friday 1-2pm.
Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display. 1 Ordinary Differential Equations Equations which are.
CSE245: Computer-Aided Circuit Simulation and Verification Lecture Note 5 Numerical Integration Spring 2010 Prof. Chung-Kuan Cheng 1.
1 EE 616 Computer Aided Analysis of Electronic Networks Lecture 12 Instructor: Dr. J. A. Starzyk, Professor School of EECS Ohio University Athens, OH,
CSE245:Lec4 02/24/2003. Integration Method Problem formulation.
Numerical Integration CSE245 Lecture Notes. Content Introduction Linear Multistep Formulae Local Error and The Order of Integration Time Domain Solution.
Initial-Value Problems
1 EE 616 Computer Aided Analysis of Electronic Networks Lecture 12 Instructor: Dr. J. A. Starzyk, Professor School of EECS Ohio University Athens, OH,
Multistep methods previous methods use information at xi to predict yi+1 multistep methods use information from xi-1, xi-2, etc. leads to better results.
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Implicit ODE Solvers Daniel Baur ETH Zurich, Institut.
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL EQUATIONS
Chapter 16 Integration of Ordinary Differential Equations.
CISE301_Topic8L31 SE301: Numerical Methods Topic 8 Ordinary Differential Equations (ODEs) Lecture KFUPM (Term 101) Section 04 Read , 26-2,
Ordinary Differential Equations (ODEs)
Differential Equations and Boundary Value Problems
Numerical Solution of Ordinary Differential Equation
MATH 685/ CSI 700/ OR 682 Lecture Notes Lecture 10. Ordinary differential equations. Initial value problems.
1 Chapter 6 Numerical Methods for Ordinary Differential Equations.
ME751 Advanced Computational Multibody Dynamics Implicit Integration Methods BDF Methods Handling Second Order EOMs April 06, 2010 © Dan Negrut, 2010 ME751,
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Explicit ODE Solvers Daniel Baur ETH Zurich, Institut.
Professor Walter W. Olson Department of Mechanical, Industrial and Manufacturing Engineering University of Toledo Solving ODE.
Boyce/DiPrima 9th ed, Ch 8.4: Multistep Methods Elementary Differential Equations and Boundary Value Problems, 9th edition, by William E. Boyce and Richard.
Lecture 35 Numerical Analysis. Chapter 7 Ordinary Differential Equations.
Interpolation. Interpolation is important concept in numerical analysis. Quite often functions may not be available explicitly but only the values of.
EE3561_Unit 8Al-Dhaifallah14351 EE 3561 : Computational Methods Unit 8 Solution of Ordinary Differential Equations Lesson 3: Midpoint and Heun’s Predictor.
Review Taylor Series and Error Analysis Roots of Equations
Modelling & Simulation of Chemical Engineering Systems Department of Chemical Engineering King Saud University 501 هعم : تمثيل الأنظمة الهندسية على الحاسب.
6. Introduction to Spectral method. Finite difference method – approximate a function locally using lower order interpolating polynomials. Spectral method.
Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display. ~ Ordinary Differential Equations ~ Stiffness and Multistep.
7. Introduction to the numerical integration of PDE. As an example, we consider the following PDE with one variable; Finite difference method is one of.
Scientific Computing Multi-Step and Predictor-Corrector Methods.
CHAPTER 3 NUMERICAL METHODS
Large Timestep Issues Lecture 12 Alessandra Nardi Thanks to Prof. Sangiovanni, Prof. Newton, Prof. White, Deepak Ramaswamy, Michal Rewienski, and Karen.
5. Integration method for Hamiltonian system. In many of formulas (e.g. the classical RK4), the errors in conserved quantities (energy, angular momentum)
Numerical Analysis – Differential Equation
1 EE 616 Computer Aided Analysis of Electronic Networks Lecture 12 Instructor: Dr. J. A. Starzyk, Professor School of EECS Ohio University Athens, OH,
Please remember: When you me, do it to Please type “numerical-15” at the beginning of the subject line Do not reply to my gmail,
Chapter 5. Ordinary Differential Equation
Today’s class Ordinary Differential Equations Runge-Kutta Methods
Lecture 40 Numerical Analysis. Chapter 7 Ordinary Differential Equations.
ECE 576 – Power System Dynamics and Stability Prof. Tom Overbye Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign.
Ordinary Differential Equations
Sec 21: Generalizations of the Euler Method Consider a differential equation n = 10 estimate x = 0.5 n = 10 estimate x =50 Initial Value Problem Euler.
ACSL, POSTECH1 MATLAB 입문 CHAPTER 8 Numerical Calculus and Differential Equations.
1/14  5.2 Euler’s Method Compute the approximations of y(t) at a set of ( usually equally-spaced ) mesh points a = t 0 < t 1
Lecture 39 Numerical Analysis. Chapter 7 Ordinary Differential Equations.
This chapter is concerned with the problem in the form Chapter 6 focuses on how to find the numerical solutions of the given initial-value problems. Main.
Runge Kutta schemes Taylor series method Numeric solutions of ordinary differential equations.
CISE301_Topic8L71 CISE301: Numerical Methods Topic 8 Ordinary Differential Equations (ODEs) Lecture KFUPM (Term 101) Section 04 Read , 26-2,
Keywords (ordinary/partial) differencial equation ( 常 / 偏 ) 微分方程 difference equation 差分方程 initial-value problem 初值问题 convex 凸的 concave 凹的 perturbed problem.
Ordinary Differential Equations
ECE 576 – Power System Dynamics and Stability
CSE245: Computer-Aided Circuit Simulation and Verification
Class Notes 18: Numerical Methods (1/2)
Class Notes 19: Numerical Methods (2/2)
CSE 245: Computer Aided Circuit Simulation and Verification
Chapter 26.
CSE245: Computer-Aided Circuit Simulation and Verification
Numerical Analysis Lecture 37.
5.3 Higher-Order Taylor Methods
MATH 175: NUMERICAL ANALYSIS II
Presentation transcript:

Integration of 3-body encounter. Figure taken from

4. Integration of Ordinary Differential Equations (ODE). We consider ODE with one variable, Most of results for this single ODE can be applicable for the above n- coupled ODE.

Existence, uniqueness, and stability of solution for ODE, A quick look for

Euler’s method: Consider the initial value problem,

Error analysis for Euler’s method

Stability of Euler’s method

Taylor’s method:

Exc 4-1) has a solution x(t) = t ( 1+ ln t). Apply Euler’s method changing the h = 1/2 n, n=1 to 8, and estimate (a) the absolute error |y i - w i | at t = 6, and (b) the error ratio for the successive h = 1/2 n at the same instant t = 6, namely, Exc 4-2)Perform the same analysis as 4-1) using the second order and fourth order Taylor methods. Because computations of higher derivatives is cumbersome, higher-order formula which involves evaluation f(t,y) only is more convenient.

Summary of the key concept on numerical method for ODE. Global discretization error : Local truncation error  i : The amount that the solution of ODE fails to satisfy the the finite difference equation. ex.) One-step method. Definitions: (Consistency, Convergence and Stability)

One-step method. Theorem: (Convergence and stability of the one-step method.) One-step method is consistent if Remark: Above theorem says the one-step method is consistent ) convergent. It can be proved under the same conditions, the one-step method is convergent, consistent.

One-step method. 2 nd order Runge-Kutta method. Determine constants a 1, a 2,  2,  2, so that  becomes O(h 2 ) approximation of the O(h 2 ) Taylor method. Heun method. (One step Euler + Trapezoidal integration.) Modified Euler method. (Half step Euler + Midpoint integration.)

Classical 4 th order Runge-Kutta method. Exc 4-3) Using an algebraic computing software, show that the local truncation error of Classical 4 th -order Runge-Kutta method is O(h 4 ). Optimal RK2 method.

General s-stage Runge-Kutta method.

Remarks on General s-stage Runge-Kutta method. (1) For Explicit s-stage formula, it is not known in general what order O(h p ) of formula one can construct for each level s. Formulas with properties (1) Small local truncation error, (2) Coefficients to be rational numbers, (3) many zeros in a j,l, are more practical. Exc 4-4) Using the idea of Gauss-Legendre integration formula, derive 2- stage Runge-Kutta formula with 4 th order accuracy. (2) For Implicit s-stage formula, it is known that O(h 2s ) formula can be construct for each level s.

Error control: Estimate the local truncation error and make it smaller than a certain threshold value by changing step size. How to estimate the error ? 1) Just to make h ! h/2. This is fine, but inefficient. 2) Embedded formula. : for RK 8 th order formula.

Linear (m-step) Multistep method. Substitute the form f(t,y(t)) = p(t) + R(t) in the integral form of ODE, and integrate it to calculate b j. Implicit linear multistep method is derived from interpolating polynomial of order m. Adams method: a j = 0, for j = 2,.., m Derivation: Integrating the both side of ODE, Explicit linear multistep method from interpolating polynomial of order m-1

Explicit scheme is also called Adams-Bashforth, implicit Adams-Moulton. Explicit scheme may be efficient since the f(t i,w i ) of earlier steps are used. The starting values for w 0 (initial value), w 1, …, w m-1 are required for the m-step method. These are calculated from one-step method of the same order. For the implicit method, value at t i+1, w i+1, is calculated from an algebraic equation. It is iteratively solved using w i+1 of an explicit multistep method of the same order as an initial guess. Usually this iteration is done by direct substitution, and only one or two iteration is made. This procedure is called predictor – corrector schemes. For this scheme, the 4 th -order formula is the most popular.

Exc4-5) Assuming uniform discretization in t domain, derive linear 2-step, 3- step 4-step explicit formulas and 1-step, 2-step, 3-step implicit formulas. Compare the coefficients of error terms between implicit and explicit formulas of the same order. In the Adams method, Newton form of polynomial interpolation formula is used for changing the step size as well as the starting formula. (Krogh type formula.) The local truncation error estimation is often made by the difference between predictor value and corrector value, which can be used for control the step size h. Exc4-6) Report on Krogh type formula. Exc4-7) Report on the ODE solver that uses Richardson extrapolation.

Conditions for the coefficients a j, and b J to satisfy for having the local truncation error  i = O(h p ) are written Order of General Linear (m-step) Multistep method. Exc4-8) Derive these conditions. hint) Substitute dy/dt = f(t,y) to the equation for  i above, and expand y and y’ around t* = t i+1 - m.

Consistency: (  i ! 0, as h ! 0) is satisfied if the local truncation error  i is a at least O(h). Consistency Convergence, and Stability of General Linear Multistep method. Note: the starting values w 1, …, w m-1 are assumed to be converge.

Stability: Consider the case with f(t,y) = 0. Definition: (Root condition) The linear multistep method satisfies the Root condition if the zeros of the associated characteristic polynomial satisfy Theorem: (Stability) A linear multistep method is stable if and only if it satisfies the root condition. Definition: A stable multistep method is said to be strongly stable if = 1 is the only zero of P( ) with | | = 1, and to be weakly stable otherwise. Remark: A linear multistep method that satisfies the consistency condition will always have at least one zero with =1.

Convergence: Theorem: (Convergence) A linear multistep method is convergent, if and only if it is both consistent and stable

Absolute stability and stiff equations. Test problem: This problem models a system of linear ODEs, in which case represents an eigenvalue of the Jacobian associated with the r.h.s. When the asymptotic character of a numerical approximation w n, n ! 1 matches that of exact solution y(t), t ! 1, the numerical method is said to be absolute stable. One-step method: Consider the m th order Taylor methods. Definition: The region of absolute stability for a one-step method is the set

Multistep method: Consider the linear m-step multistep method. Definition: The region of absolute stability for a multistep method is the set Exc 4-9) Derive the roots  k for 2-step Adams-Bashforth method and 2-step Adams Moulton method, and show the region of absolute stability on the complex plane.

Stiffness ratio: Suppose { k } is the set of characteristic expoennts associated with a particular ODE, stiffness ratio is defined by Methods to efficiently compute a solution of stiff ODE are required to have regions of absolute stability as large as possible. Possibly whole z < 0 plane. Definition: (A-stable, Dahlquist) A numerical method is A-stable if it is absolutely stable for all h such that Re h  Some known results: Explicit Runge-Kutta methods are not A-stable. Explicit linear multistep methods are not A-stable. Order of A-stable implicit multistep methods is less than 2. Among the A-stable implicit multistep methods, the trapezoidal method has the smallest local truncation error. Any symmetric s-stage and O( h 2s ) order implicit Runge-Kutta method (implicit Gauss formula) are A-stable. Also Radau, Lobatto formulas.