( ) [ ( k) ( k) ( x) ( ) ( ) ( ) [ ] ξ [ ] [ ] [ ] ( )( ) i ( ) ( )( ) 2! ( ) = ( ) 3 Interpolation. Polynomial Approximation.

Similar documents
Nice plotting of proteins II

2.29 Numerical Fluid Mechanics

1 Introduction We consider a class of singularly perturbed two point singular boundary value problems of the form: k x with boundary conditions

Polynomial Regression Models

ME 501A Seminar in Engineering Analysis Page 1

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

Lecture 13 APPROXIMATION OF SECOMD ORDER DERIVATIVES

APPENDIX A Some Linear Algebra

Least squares cubic splines without B-splines S.K. Lucas

ALGORITHM FOR THE CALCULATION OF THE TWO VARIABLES CUBIC SPLINE FUNCTION

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

Convexity preserving interpolation by splines of arbitrary degree

The Finite Element Method: A Short Introduction

PART 8. Partial Differential Equations PDEs

Chapter 4: Root Finding

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

Report on Image warping

OPTIMISATION. Introduction Single Variable Unconstrained Optimisation Multivariable Unconstrained Optimisation Linear Programming

Société de Calcul Mathématique SA

Solution for singularly perturbed problems via cubic spline in tension

Lecture 2 Solution of Nonlinear Equations ( Root Finding Problems )

Formal solvers of the RT equation

TR/95 February Splines G. H. BEHFOROOZ* & N. PAPAMICHAEL

Visualization of Data Subject to Positive Constraints

Solutions to exam in SF1811 Optimization, Jan 14, 2015

NUMERICAL DIFFERENTIATION

Math1110 (Spring 2009) Prelim 3 - Solutions

Lecture 21: Numerical methods for pricing American type derivatives

Chapter 3 Differentiation and Integration

Lecture 10 Support Vector Machines II

Ballot Paths Avoiding Depth Zero Patterns

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

Finite Difference Method

p 1 c 2 + p 2 c 2 + p 3 c p m c 2

Section 3.6 Complex Zeros

Weighted Fifth Degree Polynomial Spline

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space.

Which Separator? Spring 1

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family

Maejo International Journal of Science and Technology

8/25/17. Data Modeling. Data Modeling. Data Modeling. Patrice Koehl Department of Biological Sciences National University of Singapore

Complex Numbers, Signals, and Circuits

More metrics on cartesian products

However, since P is a symmetric idempotent matrix, of P are either 0 or 1 [Eigen-values

Cubic Trigonometric B-Spline Applied to Linear Two-Point Boundary Value Problems of Order Two

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS

Lecture 10: Euler s Equations for Multivariable

Generalized Linear Methods

U.C. Berkeley CS294: Beyond Worst-Case Analysis Luca Trevisan September 5, 2017

: Numerical Analysis Topic 2: Solution of Nonlinear Equations Lectures 5-11:

Upwind schemes for the wave equation in second-order form

Shape preserving third and fifth degrees polynomial splines

is the calculated value of the dependent variable at point i. The best parameters have values that minimize the squares of the errors

CIS526: Machine Learning Lecture 3 (Sept 16, 2003) Linear Regression. Preparation help: Xiaoying Huang. x 1 θ 1 output... θ M x M

Linear Approximation with Regularization and Moving Least Squares

For all questions, answer choice E) NOTA" means none of the above answers is correct.

Modeling curves. Graphs: y = ax+b, y = sin(x) Implicit ax + by + c = 0, x 2 +y 2 =r 2 Parametric:

Kernel Methods and SVMs Extension

A new Approach for Solving Linear Ordinary Differential Equations

PHYS 705: Classical Mechanics. Calculus of Variations II

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

Lecture Notes on Linear Regression

A Piecewise Rational Quintic Hermite Interpolant for Use in CAGD

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

MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS

Numerical Methods Solution of Nonlinear Equations

Mathematical Economics MEMF e ME. Filomena Garcia. Topic 2 Calculus

Developement of an interpolation routine for gridded EOP data

Curve Fitting with the Least Square Method

Maximum Likelihood Estimation of Binary Dependent Variables Models: Probit and Logit. 1. General Formulation of Binary Dependent Variables Models

Feb 14: Spatial analysis of data fields

Georgia Tech PHYS 6124 Mathematical Methods of Physics I

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens

An efficient algorithm for multivariate Maclaurin Newton transformation

Army Ants Tunneling for Classical Simulations

Chapter 2 Transformations and Expectations. , and define f

The univariate Bernstein-Bézier form

Appendix B. The Finite Difference Scheme

Lecture 3: Dual problems and Kernels

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

Bezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0

Modelli Clamfim Equazioni differenziali 7 ottobre 2013

Single Variable Optimization

A NOTE ON CES FUNCTIONS Drago Bergholt, BI Norwegian Business School 2011

Instituto Tecnológico de Aeronáutica FINITE ELEMENTS I. Class notes AE-245

MACHINE APPLIED MACHINE LEARNING LEARNING. Gaussian Mixture Regression

Vector Norms. Chapter 7 Iterative Techniques in Matrix Algebra. Cauchy-Bunyakovsky-Schwarz Inequality for Sums. Distances. Convergence.

Computational Astrophysics

Finite Differences, Interpolation, and Numerical Differentiation

Module 2. Random Processes. Version 2 ECE IIT, Kharagpur

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.

Chapter 13: Multiple Regression

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

e i is a random error

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

Difference Equations

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

MMA and GCMMA two methods for nonlinear optimization

Module 3: Element Properties Lecture 1: Natural Coordinates

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 8 Luca Trevisan February 17, 2016

Transcription:

3 Interpolaton {( y } Gven:,,,,,, [ ] Fnd: y for some Mn, Ma Polynomal Appromaton Theorem (Weerstrass Appromaton Theorem --- estence ε [ ab] f( P( <ε [ a, b]. If f Cab, and gven >, then there ests a polynomal P defned on, wth the property that for all {( y } Gven:,,,,,, smplest method: pecewse lnear nterpolaton Soluton: fnd the unque polynomal of degree, P (, passng through all these ( ponts. P a a a a Defnton: Lagrange nterpolatng polynomal L, ( ( ( ( ( ( ( ( ( ( ( ( ( polynomal of degree f j L, ( j δ j f j Consequently, j j j j P y L, P y L y δ y, for all j ξ Theorem If,,, are dstrnct numbers n ab, and f C ab,, then for each a, b, a, b such that ( f ( ξ ( ( (! f P pecewse lnear nterpolaton ~ use P ( between two adjacent ponts,.e. for [, ] Lagrange error ( ( f ξ! ( f ma Ma ( ( ma 4 f ~ O( Δ

th Polynomal appromaton Lagrange error [ ab] [ ] Consder,, ( f ( ξ (! ( dependng on the choces of nodes ( unformly spaced nodes: constant;,,,, T ( th ( Chebychev nodes: roots of the Chebychev polynomal T cos cos π cos π,,,, The Chebychev nodes mnmze the mamum absolute value of ( ( all sets of nodes a b Mn Ma evll s method --- computng hgher-order polynomal based on lower-order polynomals Defne Pm, m,, m the polynomal of degree passng through the ponts,,,,,, 3 { m } { } e.g.,,4 ( m ym ( m ym ( m ym P,,4 ( the polynomal of degree passng through ( y ( y ( y the ponts,,,,, 4 4

Construct the th order polynomal usng the ( th polynomals P,,, ( ( j P,,, j, j, ( ( P,,,,, ( ( j P y P y P, P y P P,,, P y P P P 3 3,3,,3,,,3 P y P P P P 4 4 3,4,3,4,,3,4,,,3,4 P,,, ( ( j P,,, j, j, ( ( P,,,,, ( ( j Because P y for j,, j, j, P y fo r,,,, ( ( ( j j y y P,,, ( y for, j,,, ( P ( ( j y ( P,,,,, j ( P,,, j, j, ( j ( ( j y y j j j j,,, j yj P ewton form of nterpolaton polynomal Gven: Fnd: P ( Wrte: P ( a ( a a a a 3 a ( 3( ( ( ( ( ( P a a a a a a a a3 a ( a ( { ( [ a a3( a ( ( ]} a a ( ( a a P ( ( a a 3 a ( nested multplcaton ( a a (

ewton form of nterpolaton polynomal a a P ( a ( a 3 a ( ( a a ( STEP: compute all a,,,, STEP : b a ( ( b a a a b b a b b a b,,,,, ( P b STEP: compute all a,,,, Defne ewton s dvded dfference : f [, ] f y [,, ] f f [ ] f [,,, ] f [, ] [, ] f f [,, ] [,,, ] f f ( 3( ( ( ( ( ( P a a a a a ( P a y f ( ( P a a y f a f [ ] f ( P a a a y [, ]( ( ( f f a f ( ( [, ]( a f f f f [ ] f a f [, ] f a [, ] f [, ] f [,, ] f [, ] f f f f [,,, ] a f y y y y 3 3 a [, ] f [, ] f [, ] f 3 ( P a b a a [,, ] f a f [,, 3] f [,,, ] b a b,,,,, 3 a 3 P b

Hermte nterpolaton Gven: f and f,,,, Fnd: a polynomal passng all the ponts wth the gven slops Theorem: If f C a, b and,,, are dstnct, the unque polynomal of least degree agreeng f and f at,,, s the polynomalof degree gven by Theorem: If,, then, f C a b ξ a b ( f ( ξ (! f ( H ( H ( s called the Hermte polynomal. ˆ H f H f H j, j j, j j j { ( } H L L, j j, j j, j ( ˆ, j j, j H L ewton s form of Hermte polynomal { } { } Defne ˆ ˆ,.e. ˆ,,,,,, ( ˆ ( ˆ ( ˆ ( ˆ ( ˆ ( ˆ H a a a a3 a a a a a 3 4 ( ˆ { ( ˆ { ( ˆ { ( ˆ ( ( ˆ } a a a a a a [ ab] Gven:, Want: f H? STEP : fnd a,,,,, STEP : b a b a b for,,,, b H

ewton s form of Hermte polynomal { } { } Defne ˆ ˆ,.e. ˆ,,,,,, [ ˆ ] [ ˆ ] Defne f f f ˆ : ˆ ˆ ˆ ˆ ˆ ˆ ˆ : st : 3 4 5 f f f f f f Defne [ ˆ, ˆ ] f ( f ( [ ˆ, ˆ ] f f f [ ˆ ˆ ˆ ] Defne f,,, [ ˆ, ˆ,, ˆ ] [ ˆ, ˆ,, ˆ ] f f for 3 ˆ ˆ nd : 3rd : f DD f DD f DD DD DD DD [ ˆ, ˆ,, ˆ ] a f DD: ewton dvded dfference Cubc splne nterpolaton ~ pecewse cubc polynomals ( Gven:, y f,,,, Use a cubc polynomal to ft for each subnterval, S ( for [, ] hgh-order polynomals > oscllatng nature Alternatve choce: pecewse polynomal appromaton Pecewse lnear nterpolaton: dsadvantage: dscontnuous frst dervatves at nodes n general Pecewse cubc nterpolaton: Goal: contnuous st and nd dervatves at nodes C : S y S y S y S y C : S S C : S S 6 constrants 8 degrees of freedom

Cubc Splne Interpolaton: Let a < < < < b. A cubc splne nterpolant S for f s a functon that satsfes the followng condtons: ( S s a cubc polynomal on the nterval,, denoted by S, for,,,, ( S y and S y, for,,,, ( S S, for,,,, ( v S S, for,,,, 4 degrees of freedom! boundary condtons: ( a free or natural splne: S S ( b clamped splne: S f and S f for,,,, ( ( ( 3 wrte S a b c d 3 S b c d 6 S c d defne h for,,, : C S a y S a b h c h d h y 3 for,,, C : for,,, 3 S b c d 3 S b c d S b c h 3d h S b b c h 3d h b C : for,,, 6 S c d 6 S c d c 6d h c

Summary: ( ( y a,,, ( 3 ( y a bh ch dh,,, ( ( b ch 3 dh b,,, ( ( v c 6d h c,,, From ( v : d d c ( v Substtute ( v nto ( : b b a, c ( v Substtute ( v( v nto ( : c c a, c For,,, 3 3 h c h h c h c a a a a h h a S y a b h c h d h 3 c S c 6d h ( ( { c } ~ equatons for unnowns boundary condtons: ( a free or natural splne: S S c c ( b f (, 3 ( b clamped splne: S f and S f S S b c h d h f ( a free or natural splne: c c c h ( h h h c r h ( h h h c r h- ( h- h h h- ( h- h- h r - c 3 3 r a a a a h h ~ trdagonal matr ~ ( ( ( 3 ( f a a h h c c ( a clamped splne: f a a h h c c h 3 h 3 h ( h h h h ( h h h h- ( h- h h h- ( h- h- h- h 3 h 3 r f a a h r f a a h 3 ~ trdagonal matr ~

Tragonal lnear system: A r a a b a a a3 b a3 a33 a34 a, a, a, a, a, a, b a, a, b Gven: Want: an appromaton of f ( Step : fnd the subnterval that belongs to, say, ( ( ( f ( 3 Step : compute S a b c d v f C a b Ma f M S 4 Theorem: Let, wth. If s the unque a b clamped cubc splne nterpolant to f wth respect to the nodes a < < < b. Then 5M Ma f ( S( Ma ( 4 j j a b 384 j cf.. natural splne: 4th order too. Boundary condtons for splne nterpolatons ( atural splne ( f ( f ( Parabolc runout f f f f ( clamped splne b Ths choce mnmzes the value of. a mplyng f f (parabolc endng curves f, f, or f nown f, f, or f nown f d (v Ft a cubc polynomal C( to the four endng ponts. f C f C (v Perodc boundary condton f f f f h f f h f f f f

Tenson splne nterpolaton Let a < < < < b. A tenson splne nterpolant S for f s a functon that satsfes the followng condtons: ( S s a cubc polynomal on the nterval,, denoted by S, for,,,, ( S y and S y, for,,,, ( S S, for,,,, ( v S S, for,,,, ( v s a soluton of τ. v S S S ( v s a soluton of τ. v S S S a specal case τ cubc polynomals cubc splne! τ another specal case appromately lnear fttng (stll C ( v S for, τ S S y C S y S z C S z nown unnown Tenson splne nterpolaton { snh[ τ( ] snh[ τ( ]} τ snh ( τ S z z h ( τ ( ( τ ( y z h y z h C S S : α z β β z α z γ γ ( h α h τ snh τ β τcosh τh snh τh h γ τ ( y y h for,,, { } ( equatons for unnowns z S,,,,, Two dmensonal nterpolatons Cartesan Product and Grd Gven: f f, y, M, j j j Soluton: D nterpolaton > two-layer D nterpolatons

Frst, nterpolate n one drecton M (, (, M, LM, ( f y f y L L j, j, j( y j (, (, f y f y L y M ( ( ( y y ( yj y j f, y f, y L y f L y ( ( j, j j, j j j M (, f y f L yl j, j M, j Irregular Grd Gven: f f, y, for,,,, Possble soluton: a polynomal to ft and nterpolate. Π m (R the set consstng of all real-coeffcent polynomals wth two varables and y of degree at most m. r If P (, y Π ( R, P (, y C y m m m rs r s m ( m ( m ( m degrees of freedom 3 Wsh f y, P y, m want f, y f P, y for,,,, m unque choce f m m s eample: 3 m m or m (, P y C C C y (, P y C C C y f for,, y C f y C f y C f eample: 6 m m or m (, P y C C C y C C y C y y y y C f y y y C f y y y C f 3 y3 3 3y3 y C 3 f3 4 y C 4 4 4y4 y 4 f4 C f5 5 y5 5 5y5 y 5 Matr nvertable? ot always! Theorem: Interpolaton of arbtray data by the subspace Π(R s possble on a set of (m(m/ nodes f these nodes le on lnes L, L,,L m n such a way that for each L contans eactly ( nodes.

Shepard nterpolaton method Wrte P(,y represents a pont n the R space. Q s another pont. Let Φ(P,Q be a real-valued functon on R R whch satsfes Φ(P,Q f and only f P Q ( PQ P Q ( ( y y e.g. Φ, p q p q Defne D Lagrange nterpolaton functon as or wrtten as L L,, ( P (, y Φ Φ ( PP, ( P, P Φ Φ (, y ;, y (, y ;, y L ( P ( Pj, P Φ δ Φ ( P, P, j j (, (, (, f y Py f L y,, P, y f L, y f δ f j j j j j j References: Kncad & Cheny Lancaster & Salausas Curve & Surface Fttng (98