Solving Nonlinear Equation

Similar documents
SAMPLE QUESTIONS FOR FINAL EXAM. (1) (2) (3) (4) Find the following using the definition of the Riemann integral: (2x + 1)dx

Chapter 5: Inner Product Spaces

Determining the sample size

Lecture 5: Span, linear independence, bases, and dimension

Sampling Distribution And Central Limit Theorem

LECTURE 13: Cross-validation

By Gilberto E. Urroz, September 2004

FIBONACCI NUMBERS: AN APPLICATION OF LINEAR ALGEBRA. 1. Powers of a matrix

AP Calculus AB 2006 Scoring Guidelines Form B

S. Tanny MAT 344 Spring be the minimum number of moves required.

Soving Recurrence Relations

1. C. The formula for the confidence interval for a population mean is: x t, which was

.04. This means $1000 is multiplied by 1.02 five times, once for each of the remaining sixmonth

In nite Sequences. Dr. Philippe B. Laval Kennesaw State University. October 9, 2008

Normal Distribution.

Infinite Sequences and Series

WHEN IS THE (CO)SINE OF A RATIONAL ANGLE EQUAL TO A RATIONAL NUMBER?

CHAPTER 7: Central Limit Theorem: CLT for Averages (Means)

1 Computing the Standard Deviation of Sample Means

Confidence Intervals

Solving Logarithms and Exponential Equations

Convexity, Inequalities, and Norms

Systems Design Project: Indoor Location of Wireless Devices

Here are a couple of warnings to my students who may be here to get a copy of what happened on a day that you missed.

1 Correlation and Regression Analysis

CHAPTER 3 THE TIME VALUE OF MONEY

Theorems About Power Series

Hypothesis testing. Null and alternative hypotheses

BENEFIT-COST ANALYSIS Financial and Economic Appraisal using Spreadsheets

Properties of MLE: consistency, asymptotic normality. Fisher information.

THE REGRESSION MODEL IN MATRIX FORM. For simple linear regression, meaning one predictor, the model is. for i = 1, 2, 3,, n

NATIONAL SENIOR CERTIFICATE GRADE 12

Chapter 7: Confidence Interval and Sample Size

Overview on S-Box Design Principles

5: Introduction to Estimation

Lecture 4: Cauchy sequences, Bolzano-Weierstrass, and the Squeeze theorem

CS103A Handout 23 Winter 2002 February 22, 2002 Solving Recurrence Relations

Z-TEST / Z-STATISTIC: used to test hypotheses about. µ when the population standard deviation is unknown

Lecture 13. Lecturer: Jonathan Kelner Scribe: Jonathan Pines (2009)

Confidence Intervals for One Mean

Trigonometric Form of a Complex Number. The Complex Plane. axis. ( 2, 1) or 2 i FIGURE The absolute value of the complex number z a bi is

Repeating Decimals are decimal numbers that have number(s) after the decimal point that repeat in a pattern.

Section 11.3: The Integral Test

Maximum Likelihood Estimators.

CHAPTER 11 Financial mathematics

TO: Users of the ACTEX Review Seminar on DVD for SOA Exam MLC

CME 302: NUMERICAL LINEAR ALGEBRA FALL 2005/06 LECTURE 8

NEW HIGH PERFORMANCE COMPUTATIONAL METHODS FOR MORTGAGES AND ANNUITIES. Yuri Shestopaloff,

Sequences and Series

2-3 The Remainder and Factor Theorems

3. If x and y are real numbers, what is the simplified radical form

Modified Line Search Method for Global Optimization

CHAPTER 3 DIGITAL CODING OF SIGNALS


Vladimir N. Burkov, Dmitri A. Novikov MODELS AND METHODS OF MULTIPROJECTS MANAGEMENT

Heat (or Diffusion) equation in 1D*

Confidence Intervals. CI for a population mean (σ is known and n > 30 or the variable is normally distributed in the.

Math C067 Sampling Distributions

Lecture 16: Address decoding

Time Value of Money. First some technical stuff. HP10B II users

Question 2: How is a loan amortized?

2-3-4 Trees and Red- Black Trees

Problem Solving with Mathematical Software Packages 1

4.3. The Integral and Comparison Tests

Learning objectives. Duc K. Nguyen - Corporate Finance 21/10/2014

Our aim is to show that under reasonable assumptions a given 2π-periodic function f can be represented as convergent series

SECTION 1.5 : SUMMATION NOTATION + WORK WITH SEQUENCES

Asymptotic Growth of Functions

Time Value of Money, NPV and IRR equation solving with the TI-86

Best of security and convenience

Partial Di erential Equations

Section 8.3 : De Moivre s Theorem and Applications

Chapter 5 Unit 1. IET 350 Engineering Economics. Learning Objectives Chapter 5. Learning Objectives Unit 1. Annual Amount and Gradient Functions

Chapter XIV: Fundamentals of Probability and Statistics *

The following example will help us understand The Sampling Distribution of the Mean. C1 C2 C3 C4 C5 50 miles 84 miles 38 miles 120 miles 48 miles

where: T = number of years of cash flow in investment's life n = the year in which the cash flow X n i = IRR = the internal rate of return

5 Boolean Decision Trees (February 11)

A Recursive Formula for Moments of a Binomial Distribution

Elementary Theory of Russian Roulette

How To Solve The Homewor Problem Beautifully

Chapter 7 Methods of Finding Estimators

GCE Further Mathematics (6360) Further Pure Unit 2 (MFP2) Textbook. Version: 1.4

5.4 Amortization. Question 1: How do you find the present value of an annuity? Question 2: How is a loan amortized?

Factoring x n 1: cyclotomic and Aurifeuillian polynomials Paul Garrett <garrett@math.umn.edu>

Integer Factorization Algorithms

AP Calculus BC 2003 Scoring Guidelines Form B

Topic 5: Confidence Intervals (Chapter 9)

THE ARITHMETIC OF INTEGERS. - multiplication, exponentiation, division, addition, and subtraction

Overview. Learning Objectives. Point Estimate. Estimation. Estimating the Value of a Parameter Using Confidence Intervals

Complex Numbers. where x represents a root of Equation 1. Note that the ± sign tells us that quadratic equations will have

1. MATHEMATICAL INDUCTION

(VCP-310)

Escola Federal de Engenharia de Itajubá


CS103X: Discrete Structures Homework 4 Solutions

Case Study. Normal and t Distributions. Density Plot. Normal Distributions

An Efficient Polynomial Approximation of the Normal Distribution Function & Its Inverse Function


Multi-server Optimal Bandwidth Monitoring for QoS based Multimedia Delivery Anup Basu, Irene Cheng and Yinzhe Yu

Math 113 HW #11 Solutions

Transcription:

Solvi Noliear Equatio oot r

Noliear Equatios Give uctio, we id value or which Solutio is root o equatio, or zero o uctio So problem is kow as root idi or zero idi Numerical Methods We-Chieh Li

Noliear Equatios: two cases Sile oliear equatio i oe ukow, where Solutio is scalar or which = System o coupled oliear equatios i ukows, where Solutio is vector or which all compoets o are zero simultaeously, = Numerical Methods We-Chieh Li 3

Eamples: Noliear Equatios Eample o -D oliear equatio 3 si e or which =.364 is oe approimate solutio Eample o system o oliear equatios i two dimesios 4 e or which = [-.8,.8] is oe approimate solutio vector Numerical Methods We-Chieh Li 4

Multiplicity I = = = = m- = but m, the root has multiplicity m I m = =,, the is simple root Numerical Methods We-Chieh Li 5

Iterval Halvi Bisectio Bisectio method beis with iitial bracket ad repeatedly halves its leth util solutio has bee isolated as accurately as desired while b-a > tol, m = a + b-/; I a*m <, b = m; else a = m; a m b ed; ed; http://www.cse.uiuc.edu/iem/oliear_eqs/bisectio/ Numerical Methods We-Chieh Li 6

Bisectio cot. Simple ad uarateed to work i is cotiuous i [a, b] [a, b] brackets a root Needed iteratios to achieve a speciied accuracy is kow i advace Error ater iteratios < b - a / Slow to covere Good or iitial uess or other root idi alorithms Fidi the iitial bracket may be a problem i is ot ive eplicitly Numerical Methods We-Chieh Li 7

Use raphi to assist root idi Set the iitial bracket Detect multiple roots Numerical Methods We-Chieh Li 8

Ca we id a root i a better way? Bisectio oly utilizes uctio values We ca id a root with ewer iteratios i other iormatio is used liear approimatio o Secat lie secat method Taet lie Newto s method Polyomial approimatio o Muller s method Numerical Methods We-Chieh Li 9

Numerical Methods We-Chieh Li Secat Method Approimate a uctio by a straiht lie Compute the itersectio o the lie ad -ais oot r

Secat Method cot. Update edpoits i, swap with epeat Numerical Methods We-Chieh Li oot r

Eample = 3 + si ep Fid the root i [, ] http://www.cse.uiuc.edu/iem/oliear_eqs/secat/.5.5 -.5 -.5.5 Numerical Methods We-Chieh Li

Method o False Positio Problem o secat method emedy Always bracket a root i the iterval [, ] How to do this? Numerical Methods We-Chieh Li 3

Numerical Methods We-Chieh Li 4 Newto s method Better approimatio usi the irst derivative ' ta ' '

Iterpretatio o Newto s method Trucated Taylor series h h ' is a liear uctio o h approimati ear eplace oliear uctio by this uctio, whose zero is h = - / Zeros o oriial uctio ad liear approimatio are ot idetical, so repeat process ' Numerical Methods We-Chieh Li 5

Eample: Newto s method http://www.cse.uiuc.edu/iem/oliear_eqs/newto/ Numerical Methods We-Chieh Li 6

Numerical Methods We-Chieh Li 7 Compariso o Secat ad Newto s methods ' Secat method Newto s method

Pros ad Cos o Newto s method Pros eiciet Cos Need to kow the derivative uctio Numerical Methods We-Chieh Li 8

Whe will Newto s method ot covere? ' passi maimum/miimum = 6, loop Numerical Methods We-Chieh Li 9

Muller s method Istead o liear approimatio, Muller s method uses quadratic approimate Evaluatio o derivatives are ot required See the tetbook or details Numerical Methods We-Chieh Li

Fied-poit Iteratio Method earrae = ito a equivalet orm = = = I r is a root o, the r = r-r = r=r I iterative orm, Also called uctio iteratio For ive equatio =, there may be may equivalet ied-poit problems = with dieret choice o Will the method always covere? Numerical Methods We-Chieh Li

Eample: Fied-poit Iteratio 3 3 3 Numerical Methods We-Chieh Li

Eample: Fied-poit Iteratio cot. 3 Divere! 3 3 Numerical Methods We-Chieh Li 3

Coverece ate For eeral iterative methods, deie error at iteratio by e = where is approimate solutio ad is true solutio For methods that maitai iterval kow to cotai solutio, rather tha speciic approimate value or solutio, take error to be leth o iterval cotaii solutio Numerical Methods We-Chieh Li 4

Coverece ate cot. Sequece coveres with rate r i lim or some iite ozero costat C Some cases o iterest r=: liear C< r>: superliear r=: quadratic e e r C Numerical Methods We-Chieh Li 5

Coverece ate o Bisectio Leth o iterval cotaii solutio reduced by hal at each iteratio Liearly coveret r = C =.5 Numerical Methods We-Chieh Li 6

Coverece o Fied-poit Iteratio I = ad <, the there is a iterval cotaii such that iteratio + = coveres to i started withi that iterval I >, the iterative scheme diveres Numerical Methods We-Chieh Li 7

Numerical Methods We-Chieh Li 8 Proo o Coveret Coditio e e ' ' Mea Value Theorem, C e e r lim Fied-poit iteratio is liearly coveret C '

Coverece o Newto s Method epreset Newto s method i ied-poit iteratio orm ' Coditio or coverece " ' [ ' ] Numerical Methods We-Chieh Li 9

Numerical Methods We-Chieh Li 3 Coverece rate o Newto s method " ' ' ] ' [ " ' " " e e C e e r lim Newto s method is quadratically coveret! ecall that

Questio rom the last class Is the iitial solutio importat or the coverece o ied-poit iteratio? Numerical Methods We-Chieh Li 3

Aswer to the coverece problem o ied-poit Iteratio < is a ecessary coditio i >, the ied-poit iteratio will divere eve i the iitial coditio is very close to a root sice the iteratio will evetually reach the reio causi diverece Iitial solutio is importat but less critical The alorithm may ot covere i the iitial solutio is ar rom the true solutio ecall the coditios that Newto s method does ot covere? Numerical Methods We-Chieh Li 3

Eample: Newto s Method or Fidi = 3 + + 5 Comple oots Numerical Methods We-Chieh Li 33

Start Newto s method with a comple value 3 5 ' 3 4 i 5 i ' 3 i 5 i i.48638. 4587i ' 3 i.44839. 3665i ' Numerical Methods We-Chieh Li 34

Numerical Methods We-Chieh Li 35 Newto s Methods or Multiple oots Quadratically coveret or simple root, Liearly coveret or multiple roots as ' ] ' [ " ' ] ' [ " '

emedies or Multiple oots with Newto s method I has a root o multiplicity k at =, we ca actor out - k rom to et k Q With a slihtly modiied Newto s method ' It ca be proved that ' k ad Newto s method still coveres quadratically k k Numerical Methods We-Chieh Li 36

emedies or Multiple oots with Newto s method cot. I practice, we do t kow k i advace emedies Try ad error Delate /-s where s is a approimate Be wared that a idetermiate orm at = is created Numerical Methods We-Chieh Li 37

Systems o Noliear Equatios Solvi systems o oliear equatios is much more diicult tha scalar case because Wide variety o behavior is possible, so determii eistece ad umber o solutios or ood starti uess is much more comple I eeral, there is o simple way to uaratee coverece to desired solutio or to bracket solutio to produce absolutely sae method Computatioal overhead icreases rapidly with dimesio o problem Numerical Methods We-Chieh Li 38

Eample: Systems i Two Dimesios From Michael T. Heath Numerical Methods We-Chieh Li 39

Numerical Methods We-Chieh Li 4 Newto s Method I dimesios, Newto s method has orm where J is Jacobia matri o I practice, we do ot eplicitly ivert J, but istead solve liear system or Newto step s, the take as et iterate J s J s

Numerical Methods We-Chieh Li 4 Eample: Newto s Method 4,, e e J T.7 83...783 3.4 J s J.83..783 3.4 s T.798] [.43 s

Numerical Methods We-Chieh Li 4 Eample: Newto s Method 4,, e e J T.798.43 4.965 4 8.653 e e.73 3.4596.86 J s J 4.965 4 8.653.73 3.4596.86 e e s T.79637] [.469 s

Fied-Poit Iteratio Fied-poit problem or vector such that Correspodi ied-poit iteratio is Coveres i starts close eouh to solutio J : is to id J : Jacobia matri : true solutio A : maimum comple orm o the eievalues o a matri A Numerical Methods We-Chieh Li 43

Fied-Poit Iteratio cot. Coverece rate is ormally : liear, with costat C ive by J I J the coverece rate is at least quadratic, e.., Newto s method Numerical Methods We-Chieh Li 44

Numerical Methods We-Chieh Li 45 Eample: Fied-poit Iteratio 4,, e 4 l T.7. -.7358756888.555387438 -.79388579.398634834 -.79746479957.48743976 -.796466478.463985 -.796466783

Net Moday Itroductio to Matlab/Octave Matlab oers a studet versio with discouted price Octave is a shareware that has similar eviromet as Matlab! You are ecouraed to bri your laptop with Octave or Matlab istalled et Moday Numerical Methods We-Chieh Li 47