Consistency & Convergence

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

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

Chapter Newton s Method

Queens College, CUNY, Department of Computer Science Numerical Methods CSCI 361 / 761 Spring 2018 Instructor: Dr. Sateesh Mane.

6.3.4 Modified Euler s method of integration

Numerical Solution of Ordinary Differential Equations

NUMERICAL DIFFERENTIATION

ORDINARY DIFFERENTIAL EQUATIONS EULER S METHOD

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

ME 501A Seminar in Engineering Analysis Page 1

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

Appendix B. The Finite Difference Scheme

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

Handout # 6 (MEEN 617) Numerical Integration to Find Time Response of SDOF mechanical system. and write EOM (1) as two first-order Eqs.

Errors for Linear Systems

Singular Value Decomposition: Theory and Applications

MATH 5630: Discrete Time-Space Model Hung Phan, UMass Lowell March 1, 2018

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

1 Derivation of Point-to-Plane Minimization

Lecture 2: Numerical Methods for Differentiations and Integrations

Example: (13320, 22140) =? Solution #1: The divisors of are 1, 2, 3, 4, 5, 6, 9, 10, 12, 15, 18, 20, 27, 30, 36, 41,

Lecture 12: Discrete Laplacian

Numerical Heat and Mass Transfer

Some modelling aspects for the Matlab implementation of MMA

MMA and GCMMA two methods for nonlinear optimization

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

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

Second Order Analysis

A new Approach for Solving Linear Ordinary Differential Equations

Formal solvers of the RT equation

Report on Image warping

Chapter 3 Differentiation and Integration

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k)

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

Numerical Simulation of One-Dimensional Wave Equation by Non-Polynomial Quintic Spline

Topic 5: Non-Linear Regression

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

Discretization. Consistency. Exact Solution Convergence x, t --> 0. Figure 5.1: Relation between consistency, stability, and convergence.

ρ some λ THE INVERSE POWER METHOD (or INVERSE ITERATION) , for , or (more usually) to

The Exact Formulation of the Inverse of the Tridiagonal Matrix for Solving the 1D Poisson Equation with the Finite Difference Method

Modelli Clamfim Equazioni differenziali 22 settembre 2016

Perron Vectors of an Irreducible Nonnegative Interval Matrix

Dynamic Systems on Graphs

The lower and upper bounds on Perron root of nonnegative irreducible matrices

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

An Algorithm to Solve the Inverse Kinematics Problem of a Robotic Manipulator Based on Rotation Vectors

Trees and Order Conditions

PARTICIPATION FACTOR IN MODAL ANALYSIS OF POWER SYSTEMS STABILITY

2 Finite difference basics

The equation of motion of a dynamical system is given by a set of differential equations. That is (1)

EEE 241: Linear Systems

A Functionally Fitted 3-stage ESDIRK Method Kazufumi Ozawa Akita Prefectural University Honjo Akita , Japan

2.3 Nilpotent endomorphisms

FTCS Solution to the Heat Equation

Suppose that there s a measured wndow of data fff k () ; :::; ff k g of a sze w, measured dscretely wth varable dscretzaton step. It s convenent to pl

Note 10. Modeling and Simulation of Dynamic Systems

PHYS 705: Classical Mechanics. Calculus of Variations II

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

ACTM State Calculus Competition Saturday April 30, 2011

Lecture 21: Numerical methods for pricing American type derivatives

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

1 GSW Iterative Techniques for y = Ax

Homework Assignment 3 Due in class, Thursday October 15

Lecture 13 APPROXIMATION OF SECOMD ORDER DERIVATIVES

Linear Approximation with Regularization and Moving Least Squares

Continuous Time Markov Chain

Additional Codes using Finite Difference Method. 1 HJB Equation for Consumption-Saving Problem Without Uncertainty

The Study of Teaching-learning-based Optimization Algorithm

Curve Fitting with the Least Square Method

Inductance Calculation for Conductors of Arbitrary Shape

Tracking with Kalman Filter

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

THE SMOOTH INDENTATION OF A CYLINDRICAL INDENTOR AND ANGLE-PLY LAMINATES

Flow Induced Vibration

Review of Taylor Series. Read Section 1.2

Modelli Clamfim Equazioni differenziali 7 ottobre 2013

NUMERICAL METHODS FOR FIRST ORDER ODEs

Solution of Linear System of Equations and Matrix Inversion Gauss Seidel Iteration Method

Chapter 8. Potential Energy and Conservation of Energy

Integrals and Invariants of Euler-Lagrange Equations

2.29 Numerical Fluid Mechanics Fall 2011 Lecture 12

Chapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise.

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

C4B Machine Learning Answers II. = σ(z) (1 σ(z)) 1 1 e z. e z = σ(1 σ) (1 + e z )

CHALMERS, GÖTEBORGS UNIVERSITET. SOLUTIONS to RE-EXAM for ARTIFICIAL NEURAL NETWORKS. COURSE CODES: FFR 135, FIM 720 GU, PhD

Generalized Linear Methods

2.29 Numerical Fluid Mechanics

[ ] λ λ λ. Multicollinearity. multicollinearity Ragnar Frisch (1934) perfect exact. collinearity. multicollinearity. exact

Supplementary Notes for Chapter 9 Mixture Thermodynamics

Norms, Condition Numbers, Eigenvalues and Eigenvectors

Application of B-Spline to Numerical Solution of a System of Singularly Perturbed Problems

Hidden Markov Models & The Multivariate Gaussian (10/26/04)

Common loop optimizations. Example to improve locality. Why Dependence Analysis. Data Dependence in Loops. Goal is to find best schedule:

APPROXIMATE PRICES OF BASKET AND ASIAN OPTIONS DUPONT OLIVIER. Premia 14

ECEN 667 Power System Stability Lecture 21: Modal Analysis

Correlation and Regression. Correlation 9.1. Correlation. Chapter 9

Chapter 3 Describing Data Using Numerical Measures

CIVL 8/7117 Chapter 10 - Isoparametric Formulation 42/56

Dynamical Systems and Information Theory

Transcription:

/9/007 CHE 374 Computatonal Methods n Engneerng Ordnary Dfferental Equatons Consstency, Convergence, Stablty, Stffness and Adaptve and Implct Methods ODE s n MATLAB, etc Consstency & Convergence Consstency Whether or not the numercal soluton gves the same soluton as the dfferental equaton Convergence Whether the soluton to the dfference equatons becomes closer and closer to the soluton of the de for smaller and smaller step szes

/9/007 STABILITY A DE s sad to be stable f small changes n the problem causes only small changes n the soluton as t Y Stable : comes closer Unstable : soluton separates as t ncreases For numercal methods, small changes n the problem are nevtable Truncaton errors Round-off errors y = -y y =y Stablty of Euler s Method Consder the followng problem y' = λ y, y(0) = β Exact soluton yx ( ) = βe λx Stable f h λ

/9/007 3 System of n equatons For a system of st order ODE's y' = Λy, wth constant matrx Λ The Soluton s: yx ( ) = λ x where λ are egenvalues of Λ and v are correspondng egenvectors Cve Step sze needed for stablty s determned by the largest egevalue h λ max Problem s stff f there s a large dsparty n the magntudes of the Postve egenvalues. Specal methods requred for stff problems Stffness An ntal value problem s sad to be stff f some terms n the soluton vector vary more rapdly wth x (or t) than others dy t = 000y+ 3000 000 e, y(0) = 0 dt Analytcalsol -000t y=3-0.998e.00 e t

/9/007 4 Stffness y 3.5 3.5.5 0.5 0 0 4 6 8 0 t Backward Euler (Implct) Euler s method can be consdered as a result of numercally ntegratng y = f(t,y) at the left of the nterval. We can derve another method by usng f(t,y) at t k+ y + = y + hf(t +, y + ) Implct methods have a much larger regon of stablty!!

/9/007 5 Stablty of Backward Euler y = y + hf( t, y ) + + + Implct because t evaluates the functon f wth argument y before we know t's value + STABILITY: ODE y'=-λy y = y + h( λ y ) + + ( + hλ) y = y + f y0 s the ntal condton, applcaton of ths equaton n tmes gves after n steps HOLDS for any h>0 (uncondtonally stable) yn = y0 + hλ + hλ n Adaptve Methods We have looked at methods whch employ a constant step sze. Ths can present a challenge to problems wth steep changes Algorthms whch automatcally adjust the step sze wll have a great advantage Adapt to soluton accordngly. Most of the RK methods mplemented n MATLAB are adaptve

/9/007 MultStep methods Sngle step methods utlze the nformaton at a sngle pont x, to predct a value of the dependent value y + at a future pont, x + Utlzes the prevous nformaton. Not popular now, complcated. Some MATLAB solvers for ODE 6

/9/007 Ode3 and ode45 (non-stff dfferental eqs.) solver Descrpton ode3 ODE3 Solve non-stff dfferental equatons, low order method. ode45 ODE45 Solve non-stff dfferental equatons, medum order method. Usage [T,Y] = ODE45(ODEFUN,TSPAN,Y0) wth TSPAN = [T0 TFINAL] ntegrates the system of dfferental equatons y' = f(t,y) from tme T0 to TFINAL wth ntal condtons Y0. Functon ODEFUN(T,Y) must return a column vector correspondng to f(t,y). Each row n the soluton array Y corresponds to a tme returned n the column vector T. To obtan solutons at specfc tmes T0,T,...,TFINAL (all ncreasng or all decreasng), use TSPAN = [T0 T... TFINAL]. Example [t,y]=ode45(@vdp,[0 0],[ 0]); plot(t,y(:,)); solves the system y' = vdp(t,y), usng the default relatve error tolerance e-3 and the default absolute tolerance of e-6 for each component, and plots the frst component of the soluton. Stff and moderately stff ODE s solver ode3s ode3t Descrpton ODE3s Solve stff dfferental equatons, low order method. ODE3T Solve moderately stff ODEs and DAEs, trapezodal rule. ode5s ODE5S Solve stff dfferental equatons and DAEs, varable order method. odeset Create/alter ODE OPTIONS structure. OPTIONS = ODESET('NAME',VALUE,'NAME',VALUE,...) RelTol, AbsTol, IntalStep - Suggested ntal step sze See matlab for more nformaton!! 7