ME 310 Numerical Methods. Differentiation

Similar documents
Tests for Two Correlations

UNIVERSITY OF BOLTON WESTERN INTERNATIONAL COLLEGE, RAS AL KHAIMAH BENG(HONS) MECHANICAL ENGINEERING SEMESTER TWO EXAMINATION 2014/2015

2. Equlibrium and Efficiency

SIMPLE FIXED-POINT ITERATION

Spring 2018 Social Sciences 7418 University of Wisconsin-Madison. Transactions and Portfolio Crowding Out

Fall 2016 Social Sciences 7418 University of Wisconsin-Madison. Transactions and Portfolio Crowding Out

Numerical Analysis ECIV 3306 Chapter 6

Supplementary material for Non-conjugate Variational Message Passing for Multinomial and Binary Regression

CS 541 Algorithms and Programs. Exam 1 Solutions

Supplementary Material for Borrowing Information across Populations in Estimating Positive and Negative Predictive Values

Understanding Annuities. Some Algebraic Terminology.

Linear Combinations of Random Variables and Sampling (100 points)

MgtOp 215 Chapter 13 Dr. Ahn

Capability Analysis. Chapter 255. Introduction. Capability Analysis

Static (or Simultaneous- Move) Games of Complete Information

A Case Study for Optimal Dynamic Simulation Allocation in Ordinal Optimization 1

A Correction to: The Structure of General Equilibrium Shadow Pricing Rules for a Tax-Distorted Economy

Deconvolution and correlation-based interferometric redatuming by wavefield inversion

II. Random Variables. Variable Types. Variables Map Outcomes to Numbers

Economic Design of Short-Run CSP-1 Plan Under Linear Inspection Cost

S yi a bx i cx yi a bx i cx 2 i =0. yi a bx i cx 2 i xi =0. yi a bx i cx 2 i x

Multifactor Term Structure Models

Elton, Gruber, Brown, and Goetzmann. Modern Portfolio Theory and Investment Analysis, 7th Edition. Solutions to Text Problems: Chapter 9

Tests for Two Ordered Categorical Variables

ISyE 512 Chapter 9. CUSUM and EWMA Control Charts. Instructor: Prof. Kaibo Liu. Department of Industrial and Systems Engineering UW-Madison

15-451/651: Design & Analysis of Algorithms January 22, 2019 Lecture #3: Amortized Analysis last changed: January 18, 2019

A MODEL OF COMPETITION AMONG TELECOMMUNICATION SERVICE PROVIDERS BASED ON REPEATED GAME

Homework 1 Answers` Page 1 of 12

iii) pay F P 0,T = S 0 e δt when stock has dividend yield δ.

EDC Introduction

OCR Statistics 1 Working with data. Section 2: Measures of location

Parallel Prefix addition

Macroeconomic Theory and Policy

The convolution computation for Perfectly Matched Boundary Layer algorithm in finite differences

Explaining Movements of the Labor Share in the Korean Economy: Factor Substitution, Markups and Bargaining Power

Appendix for Solving Asset Pricing Models when the Price-Dividend Function is Analytic

ECO 209Y MACROECONOMIC THEORY AND POLICY LECTURE 8: THE OPEN ECONOMY WITH FIXED EXCHANGE RATES

Hardware-Software Cosynthesis of Multi-Mode Multi-Task Embedded Systems with Real-Time Constraints

CHAPTER 9 FUNCTIONAL FORMS OF REGRESSION MODELS

Evaluating Performance

Macroeconomic Theory and Policy

Notes on experimental uncertainties and their propagation

/ Computational Genomics. Normalization

CS 286r: Matching and Market Design Lecture 2 Combinatorial Markets, Walrasian Equilibrium, Tâtonnement

The Integration of the Israel Labour Force Survey with the National Insurance File

A Single-Product Inventory Model for Multiple Demand Classes 1

Note on Cubic Spline Valuation Methodology

4. Greek Letters, Value-at-Risk

Jean-Paul Murara, Västeras, 26-April Mälardalen University, Sweden. Pricing EO under 2-dim. B S PDE by. using the Crank-Nicolson Method

Time Domain Decomposition for European Options in Financial Modelling

The Direct Control and Penalty Methods for American Put Options

Chapter 5 Student Lecture Notes 5-1

Module Contact: Dr P Moffatt, ECO Copyright of the University of East Anglia Version 2

Lecture Note 2 Time Value of Money

Mode is the value which occurs most frequency. The mode may not exist, and even if it does, it may not be unique.

Financial mathematics

Task Force on quality of business survey data

Graphical Methods for Survival Distribution Fitting

Ch Rival Pure private goods (most retail goods) Non-Rival Impure public goods (internet service)

Appendix - Normally Distributed Admissible Choices are Optimal

Project Management Project Phases the S curve

Major: All Engineering Majors. Authors: Autar Kaw, Jai Paul

Centre for International Capital Markets

Random Variables. b 2.

Cyclic Scheduling in a Job shop with Multiple Assembly Firms

ECE 586GT: Problem Set 2: Problems and Solutions Uniqueness of Nash equilibria, zero sum games, evolutionary dynamics

Computation of the Compensating Variation within a Random Utility Model Using GAUSS Software

Alternatives to Shewhart Charts

Maximum Likelihood Estimation of Isotonic Normal Means with Unknown Variances*

A Bootstrap Confidence Limit for Process Capability Indices

An Application of Alternative Weighting Matrix Collapsing Approaches for Improving Sample Estimates

Production and Supply Chain Management Logistics. Paolo Detti Department of Information Engeneering and Mathematical Sciences University of Siena

Copyright 2017 by Taylor Enterprises, Inc., All Rights Reserved. Dr. Wayne A. Taylor

A General Control Variate Method

Fall 2017 Social Sciences 7418 University of Wisconsin-Madison Problem Set 3 Answers

Mathematical Modeling of Financial Derivative Pricing

CS54701: Information Retrieval

Homework 9: due Monday, 27 October, 2008

Chapter 3 Student Lecture Notes 3-1

Interval Estimation for a Linear Function of. Variances of Nonnormal Distributions. that Utilize the Kurtosis

Fast Valuation of Forward-Starting Basket Default. Swaps

ECONOMETRICS - FINAL EXAM, 3rd YEAR (GECO & GADE)

Data Mining Linear and Logistic Regression

Cracking VAR with kernels

Raising Food Prices and Welfare Change: A Simple Calibration. Xiaohua Yu

COMPARISON OF THE ANALYTICAL AND NUMERICAL SOLUTION OF A ONE-DIMENSIONAL NON-STATIONARY COOLING PROBLEM. László Könözsy 1, Mátyás Benke 2

Risk and Return: The Security Markets Line

The Institute of Chartered Accountants of Sri Lanka

Optimal Information Disclosure in Auctions: The Handicap Auction

1 Introducton Opton prcng teory as been te core of modern matematcal nance snce te dervaton of te famous Blac-Scoles (19) formula wc provdes a partal

University of Toronto November 9, 2006 ECO 209Y MACROECONOMIC THEORY. Term Test #1 L0101 L0201 L0401 L5101 MW MW 1-2 MW 2-3 W 6-8

University of Toronto November 9, 2006 ECO 209Y MACROECONOMIC THEORY. Term Test #1 L0101 L0201 L0401 L5101 MW MW 1-2 MW 2-3 W 6-8

Calibration Methods: Regression & Correlation. Calibration Methods: Regression & Correlation

3: Central Limit Theorem, Systematic Errors

On the uniqueness of stable marriage matchings

Basket options and implied correlations: a closed form approach

Scribe: Chris Berlind Date: Feb 1, 2010

Spurious Seasonal Patterns and Excess Smoothness in the BLS Local Area Unemployment Statistics

Creating a zero coupon curve by bootstrapping with cubic splines.

A New and Efficient Congestion Evaluation Model in Floorplanning: Wire Density Control with Twin Binary Trees

Transcription:

M 0 Numercal Metods fferentaton Tese presentatons are prepared by r. Cuneyt Sert Mecancal ngneerng epartment Mddle ast Tecncal Unversty Ankara, Turkey csert@metu.edu.tr Tey can not be used wtout te permsson of te autor

ervatve: Rate of cange of a dependent varable wt respect to an ndependent varable. f df dx X lm x0 x x tangent slope = f (x x x wc can be wrtten approxmately as a dfference equaton df dx X f x x x x f x x Numercal dfferentaton s consdered f te functon can not be dfferentated analytcally te functon s known at dscrete ponts only te dfferentaton s to be automated n an algortm.

Fnte vded fference Formulas usng TS Formulas for te frst dervatve Forward dfferencng (use st order TS of + around x. Call ts TS f(! were = x + - x f(! O( Backward dfferencng (use st order TS of - around x. Call ts TS f(! were = x x - f(! O( Centered dfferencng (use TS TS. But consder nd order terms also.!! O( O( O(

Forward and centered dfference formulas are frst order, O(, accurate. Tat s te error drops approxmately by a factor of as te step sze drops to /. Centered dfference formula s second order, O(. rror drops by a factor of 4 as drops to /. Centered dfference formula uses te same number of artmetc operatons as forward and backward formulas, and t offers better accuracy. Terefore t s more effcent. xample : Poston of a body movng n a stragt pat s sown below. Fnd ts velocty. t x v 0.0 0.00.50 0. 0.5 0. 0.47.5 Use forward dfferencng at t = 0.0 x(0. x(0.0 0.5 0.00 v(0.0.50 0. Use centered dfferencng at t = 0., 0., 0. and 0.4 0. 0.6 0.4 0.84 v(0. x(0. x(0. 0.6 0.5 0..5 0.5 0.98.40 Use backward dfferencng at t = 0.5 v(0.5 x(0.5 x(0.4 0.98 0.84 0..40 xercse : Complete te above table. 4

Hger order formulas for te frst dervatve To derve tem use proper combnatons of TS of +, -, +, - Forward dfferencng 4 O( Backward dfferencng 4 O( Centered dfferencng 8 8 O( 4 See pages 6-64 for even more ger order formulas. xercse : erve te above formulas. xercse : Solve te prevous example usng te above formulas. Note tat for t=0.0 and 0. forward dfferencng must be used. For t=0.4 and 0.5 backward dfferencng s sutable. Centered dfferencng can be used only for t=0. and 0.. 5

Formulas for te second dervatve Forward dfferencng f (x O( Backward dfferencng f (x O( Centered dfferencng f (x O( See pages 6-64 for formulas for te rd and 4 t dervatves. xercse 4: erve te above formulas. Use proper combnatons of TS of +, -, + and -. xercse 5: Use te table gven for te frst example to calculate te acceleraton of te partcle. Use te above formulas to take te second dervatve of tme. You can also take te frst dervatve of te prevously calculated veloctes. Compare and comment on te results. 6

How can we mprove te dervatve estmates? Use small values. Use ger order approxmatons. xercse 6: Te followng two tables are for te functon = e x. Tey use step szes of 0. and 0.. Calculate te frst dervatve of te functon usng O( estmates. Calculate true errors. Compare te results. x e t 0.4.49847 0.6.888 0.8.55409.0.7888 x e t 0.4.49847 0.5.6487 0.6.888 0.7.0757 0.8.55409 0.9.45960.0.7888 Anoter alternatve to mprove dervatve estmates s to use Rcardson xtrapolaton. It combnes two estmates obtaned wt dfferent values to get a better estmate. Later we wll use Rcardson xtrapolaton for ntegraton too. 7

8 : xact dervatve (usually not known : stmated dervatve usng =. : rror of te estmaton. = + : stmated dervatve usng =. : rror of te estmaton. = + If we used second-order dfferentaton to get and, tan = O(, = O( Substtute ts nto te frst equaton Solve for Combne ts wt Ts estmate s or order O( 4, obtaned by two O( estmates. A specal case of = / results n 4 Rcardson s xtrapolaton

xample 4: Use Rcardson xtrapolaton to estmate te frst dervatve of sn(x at x=p/4 usng step szes of =p/ and =p/6. Use centered dfferences of O(. Usng =p/ f( p/4 sn( p/4 p/ sn( p/4 ( p/ p/ 0.5847760 e t = 7. % Usng =p/6 f( p/4 sn( p/4 p/6 sn( p/4 ( p/6 p/6 0.67577 e t = 4.5 % Apply Rcardson xtrapolaton f( p/4 4 0.67577 0.5847760 0.70595 e t = 0.4 % Romberg Algortm Apply multple Rcardson xtrapolaton one after te oter untl te error falls below a specfed tolerance. xercse 7: In te above example use =p/ to calculate. Combne wt, and wt to get two O( 4 estmates. Tan combne tese two estmates to get an estmate of order O( 6. 9