Discrete Fourier Transform { } ( ) sin( ) Discrete Sine Transformation. n, n= 0,1,2,, when the function is odd, f (x) = f ( x) L L L N N.

Σχετικά έγγραφα
Estimators when the Correlation Coefficient. is Negative

Κεφάλαιο 3. Εξίσωση Καθαρής Συναγωγής Εξίσωση Καθαρής Συναγωγής Ρύπου

Markov Processes and Applications

Exam Statistics 6 th September 2017 Solution

Matrices and Determinants

Answers - Worksheet A ALGEBRA PMT. 1 a = 7 b = 11 c = 1 3. e = 0.1 f = 0.3 g = 2 h = 10 i = 3 j = d = k = 3 1. = 1 or 0.5 l =

Multi-dimensional Central Limit Theorem

IIT JEE (2013) (Trigonomtery 1) Solutions

Multi-dimensional Central Limit Theorem

CS 1675 Introduction to Machine Learning Lecture 7. Density estimation. Milos Hauskrecht 5329 Sennott Square

Presentation of complex number in Cartesian and polar coordinate system

Fourier Series. MATH 211, Calculus II. J. Robert Buchanan. Spring Department of Mathematics

1. For each of the following power series, find the interval of convergence and the radius of convergence:

Other Test Constructions: Likelihood Ratio & Bayes Tests

Probability and Random Processes (Part II)

Solve the difference equation

Quadratic Expressions

The Heisenberg Uncertainty Principle

CHAPTER 103 EVEN AND ODD FUNCTIONS AND HALF-RANGE FOURIER SERIES

CHAPTER 25 SOLVING EQUATIONS BY ITERATIVE METHODS

CT Correlation (2B) Young Won Lim 8/15/14

FREE VIBRATION OF A SINGLE-DEGREE-OF-FREEDOM SYSTEM Revision B


Last Lecture. Biostatistics Statistical Inference Lecture 19 Likelihood Ratio Test. Example of Hypothesis Testing.

Homework 3 Solutions

INTEGRATION OF THE NORMAL DISTRIBUTION CURVE

Homework for 1/27 Due 2/5

Inverse trigonometric functions & General Solution of Trigonometric Equations

New bounds for spherical two-distance sets and equiangular lines

P AND P. P : actual probability. P : risk neutral probability. Realtionship: mutual absolute continuity P P. For example:

n r f ( n-r ) () x g () r () x (1.1) = Σ g() x = Σ n f < -n+ r> g () r -n + r dx r dx n + ( -n,m) dx -n n+1 1 -n -1 + ( -n,n+1)

Phys460.nb Solution for the t-dependent Schrodinger s equation How did we find the solution? (not required)

Political Science 552

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

Boundary-Fitted Coordinates!

CHAPTER 101 FOURIER SERIES FOR PERIODIC FUNCTIONS OF PERIOD

Examples of Cost and Production Functions

SUPERPOSITION, MEASUREMENT, NORMALIZATION, EXPECTATION VALUES. Reading: QM course packet Ch 5 up to 5.6

ΜΕΡΟΣ ΙΙI ΜΟΡΙΑΚΟ ΒΑΡΟΣ ΠΟΛΥΜΕΡΩΝ

α & β spatial orbitals in

D Alembert s Solution to the Wave Equation

p n r

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

3.4 SUM AND DIFFERENCE FORMULAS. NOTE: cos(α+β) cos α + cos β cos(α-β) cos α -cos β

6.1. Dirac Equation. Hamiltonian. Dirac Eq.

Nowhere-zero flows Let be a digraph, Abelian group. A Γ-circulation in is a mapping : such that, where, and : tail in X, head in

Second Order Partial Differential Equations

Cytotoxicity of ionic liquids and precursor compounds towards human cell line HeLa

(6,5 μονάδες) Θέμα 1 ο. Τμήμα Πολιτικών Μηχανικών Σχολή Τεχνολογικών Εφαρμογών Διεθνές Πανεπιστήμιο Ελλάδος ΟΝΟΜΑΤΕΠΩΝΥΜΟ

Markov Processes and Applications

forms This gives Remark 1. How to remember the above formulas: Substituting these into the equation we obtain with

([28] Bao-Feng Feng (UTP-TX), ( ), [20], [16], [24]. 1 ([3], [17]) p t = 1 2 κ2 T + κ s N -259-

Statistical Inference I Locally most powerful tests

Universal Levenshtein Automata. Building and Properties

Differential equations

Homework 8 Model Solution Section

Chapter 6 BLM Answers

Errata (Includes critical corrections only for the 1 st & 2 nd reprint)

ANSWERSHEET (TOPIC = DIFFERENTIAL CALCULUS) COLLECTION #2. h 0 h h 0 h h 0 ( ) g k = g 0 + g 1 + g g 2009 =?

Exercises 10. Find a fundamental matrix of the given system of equations. Also find the fundamental matrix Φ(t) satisfying Φ(0) = I. 1.

Fourier Series. Fourier Series

Homework 4.1 Solutions Math 5110/6830

EE512: Error Control Coding

ECE Spring Prof. David R. Jackson ECE Dept. Notes 2

Finite Field Problems: Solutions

derivation of the Laplacian from rectangular to spherical coordinates

Math221: HW# 1 solutions

MET 4302 LECTURE 3A 23FEB18

Symplecticity of the Störmer-Verlet algorithm for coupling between the shallow water equations and horizontal vehicle motion

Fourier Analysis of Waves

8.323 Relativistic Quantum Field Theory I

Roman Witu la 1. Let ξ = exp(i2π/5). Then, the following formulas hold true [6]:

ΔΙΑΚΡΙΤΟΣ ΜΕΤΑΣΧΗΜΑΤΙΣΜΟΣ FOURIER - Discrete Fourier Transform - DFT -

Biorthogonal Wavelets and Filter Banks via PFFS. Multiresolution Analysis (MRA) subspaces V j, and wavelet subspaces W j. f X n f, τ n φ τ n φ.

ST5224: Advanced Statistical Theory II

1. If log x 2 y 2 = a, then dy / dx = x 2 + y 2 1] xy 2] y / x. 3] x / y 4] none of these

k A = [k, k]( )[a 1, a 2 ] = [ka 1,ka 2 ] 4For the division of two intervals of confidence in R +

Econ 2110: Fall 2008 Suggested Solutions to Problem Set 8 questions or comments to Dan Fetter 1

The Equivalence Theorem in Optimal Design

Part III - Pricing A Down-And-Out Call Option

Comparison of Evapotranspiration between Indigenous Vegetation and Invading Vegetation in a Bog

MATH423 String Theory Solutions 4. = 0 τ = f(s). (1) dτ ds = dxµ dτ f (s) (2) dτ 2 [f (s)] 2 + dxµ. dτ f (s) (3)

Three-Dimensional Experimental Kinematics

2. THEORY OF EQUATIONS. PREVIOUS EAMCET Bits.

Example Sheet 3 Solutions

= 0.927rad, t = 1.16ms

Mock Exam 7. 1 Hong Kong Educational Publishing Company. Section A 1. Reference: HKDSE Math M Q2 (a) (1 + kx) n 1M + 1A = (1) =

8.1 The Nature of Heteroskedasticity 8.2 Using the Least Squares Estimator 8.3 The Generalized Least Squares Estimator 8.

Lanczos and biorthogonalization methods for eigenvalues and eigenvectors of matrices


Lecture 2. Soundness and completeness of propositional logic

C.S. 430 Assignment 6, Sample Solutions

MATH 38061/MATH48061/MATH68061: MULTIVARIATE STATISTICS Solutions to Problems on Matrix Algebra

Lifting Entry (continued)

SOLVING CUBICS AND QUARTICS BY RADICALS

of the methanol-dimethylamine complex

(6,5 μονάδες) Θέμα 1 ο. Τμήμα Πολιτικών Μηχανικών Σχολή Τεχνολογικών Εφαρμογών Διεθνές Πανεπιστήμιο Ελλάδος ΟΝΟΜΑΤΕΠΩΝΥΜΟ

Eight Examples of Linear and Nonlinear Least Squares

The Simply Typed Lambda Calculus

Solution Set #2

Transcript:

Dscrete Fourer Trasform Refereces:. umercal Aalyss of Spectral Methods: Theory ad Applcatos, Davd Gottleb ad S.A. Orszag, Soc. for Idust. App. Math. 977.. umercal smulato of compressble flows wth smple boudares. I. Galerk (spectral) represetato, Stud. Appl Math., 9 (97). Dscrete Se Trasformato whe the fucto s odd, f (x) f (x) k,,,,, / x,,,,, f ( x ) a s k x f ( x) a s k x a s k x x / a f( x )s k x <proof> a f( x )s k x f ( x ) a s k x <proof> l cos + l f( x )s s ( k x ) a s( k x ) ( k x ) m m s a s k x k x m a δ a m m s ( ) s ( k x) δ m m l l l l l l l l l s s cos s cos s cos s cos + + + + l l l l 4l l s + s + s s + s s 5l l 6l 4l + s s + s s l ( ) ( ) ( ) + + s s l s + l s l + ( ) s l s l s + l + + { } l l l s + + s ( ) l s ( l + )

s ( ) s ( k x) δ m m l cos + l s ( ) s ( k x) δ m m l cos + l <proof> f m: s s cos cos ( ) ( kmx) { (( k km) x) (( k + km) x) } ( m) ( + m) cos cos <proof> f m: s ( ) s( ) cos( ) + ( ) + cos ( ) m m ( ) + + + + Dscrete Cose Trasformato whe the fucto s eve, f (x) f (x) f ( x) bcos( ) b b b eve odd / x,,,,, b ( b b + b), for, 4,6, b b + b b b, for,,5, f ( x) bcos( odd ) b f( x )cos k x b b + b m m m odd b b + b m m m+ odd m m m odd b b + b <proof> b f( x)cos( ) m m b f( x) bmcos b + bmcos m m m b + bm cos m m m cos cos cos + (for ) m m { ( ) } b b + bm k f ( x ) b cos k x ( m) m m m

<proof> b b b + m m m+ odd m ( )cos b f x k x m b bmcos cos ( m+ ) ( m) bm cos + cos m m+ m b + b m ( ) + ( ) m m k f ( x ) b cos k x Dscrete Fourer Trasformato f ( x) aexp k x,,,,, f ( x ) a exp k x a f x k x exp a f x k x <proof> exp f ( x ) a exp k x f mod(, ) exp otherwse <proof> f mod(, ), wrte l f x k x a k x k x ( ) exp( ) mexp( m ) exp( ) m exp cos( l) + s ( l) ( + ) a exp ( m) m m am m m δ a f mod(, ), wrte θ exp cos θ + s θ { } ( ) + l cos + l

Dscrete Fourer Trasformato a a f ( x ) a exp k x a f x k x : complex umber exp * a (complex cougate) ( data) f f (x) s real. oly data eeds storg the amout of computatos s o the order of. Fast Fourer Trasform (FFT) the radx- Cooley-Tukey algorthm exp a f x k x f ( x ) exp f ( x) exp + f ( x) exp eve odd M m m+ + M M ( ) f ( xm) exp f ( xm+ ) exp m M m M Fast Fourer Trasform (FFT) the radx- Cooley-Tukey algorthm m m a f x f x M M a M M ( m) exp + exp ( m+ ) exp m m E + exp O M E, O : DFT of sze M /,,,, M M ( M + ) m f ( xm ) exp m M ( M + ) ( M + ) m + exp f ( xm+ ) exp m M M+ M a Fast Fourer Trasform (FFT) the radx- Cooley-Tukey algorthm m E exp O M M f ( xm ) exp m M M+ M m exp f ( xm+ ) exp m M,,,, M K If, for stace, the amout of computatos becomes o the order of log. 4 8 4 6 4 8 4 4 8 4 6 4 8 4 4

Covoluto For stace : u u uv + + t x y Stadard fte dfferece schemes v.s. spectral methods Example: a perodc fucto of perod uxt (, ) a( t)exp vxt (, ) b( t)exp uxtvxt (, ) (, ) c( t)exp arbs r r + s c ~ the amout of computatos s o the order of fx a exp( x) a expwx w w Spectral method: dw f( x) a exp( x ) a exp wx + w+ dw Fte dfferece methods: f f( x ),,,,, x Frst dervatve Exact : f( x) a expwx f ( x) w a expwx w d order cetral dfferece: f f f f ( x) + f( x) a expwx w aexpwx+ expwx w aexpwx expwdx expwdx w dw spectral: f ( x) w a exp wx w+ dw spectral exact except w ± a wx w w exp s s( w) aexpwx w exact wa dx 5

4rd order cetral dfferece: f( x) a expwx w spectral f f ( x ) f 8f + 8f f + + 8s( w) s( w) aexp wx 6 w exact wa kη req' d ~ ωε w k ε tolerace ωε k η smallest legthscale compact fte dfferece schemes: ele S.K. (99) J. Comp. Phys.,6-4 Frst dervatve: β f +α f + f +α f +βf + + f f f f f f a + b + c 6h 4h h + + + w k Questo: What choces of values of parameters a,b,c,α,β possess the best accuracy (resoluto)? () () () (v) (v) a / 4/9 5/6 7/.94985 b /9 /5 /5.4 c -/8 /.4559 α /4 / /8 /.58 β /.959 order 4 6 8 4 6

Pseudo-spectral method alasg error trucated wave doma grd shft techque mxed method da Burger's equato: t x x u u u + ν uxt (, ) a( t)exp u ( x, t) b( t)exp exp exp exp dt Galerk: ( k x) + k b t ( k x) ν k a t ( k x) exp k x dx m da k b νk a dt + orthogoalty Covoluto Pseudo-spectral method uxt (, ) a( t)exp vxt (, ) b( t)exp uxtvxt (, ) (, ) c( t)exp c a b r s r r+ s ux aexp vx bexp wx ux vx c w( x )exp k x

Pseudo-spectral method Pseudo-spectral method c a exp k x b exp k x exp k x r r s s r s abexp ( k + k k ) x r s r s r s ( r+ s) Γ( rs,, ) exp f mod( r+ s, ) Γ ( rs,, ) otherwse r+ s ab r sexp r s for rs,, <, < r+ s < ab Γ r s r s ( r, s, ) Γ r, s, δ ( r+ s ) +δ ( r+ s ) +δ ( r+ s + ) Pseudo-spectral method r s r s (,, ) c a b Γ r s c a b + a b + a b r s r s r s rs, rs, rs, r+ s r+ s r+ s + c c + a b + a b r s r s rs, rs, r+ s r+ s + Pseudo-spectral method possble soluto: trucated wave doma < K < rs,, < K K < r+ s < K < Γ rs,, δ ( r+ s) choose K (alasg error)

Grd Shft Techque Cosder x x + / ( + ) x +Δx u u( x ) a ( t)exp k x ( + / ) exp exp u u x a k Δx k x smlarly w w v v( x ) b ( t)exp k x ( + / ) exp exp v v x b k Δx k x u v c uv exp( ) c + a b + a b r s r s rs, rs, r+ s r+ s + u v cˆ exp u v + / uv exp( + /) exp ˆ a exp k x b exp k x exp k x ( + / ) ( + / ) ( + / ) r r s s r s abexp k + k k x ( + / ) r s r s rs, ( k Δx ) exp cˆ u v k x c c ab ab r s r s rs, rs, r+ s r+ s + c c + a b + a b r s r s rs, rs, r+ s r+ s + - ab exp( ( k + k k ) Δx ) exp ( k + k - k ) x r s r s rs, ( r+ s) ( r+ s) kr + ks k Δ x ( r s ) f r+ s or + or c c + c ( ˆ ) 6 FFTs are requred to obta alasg-error-free coeffcets.

Exteded to D Smulatos: ux ( ) aexp k x ( ) vx bexp k x ( ) wx ux vx c exp k x ( ) c rs, r+ s a( r) b( s) (D covoluto) Pseudo-spectral Method: c u x v x k x exp ( ) (,, ),,,. x,, x, x, x Δ Δ Δ Pseudo-spectral Method: alasg errors c r rs, s a( r) exp( k r x ) b s k x k x exp( s ) exp ( ) a( r) b( s) exp ( k r + k s k ) x ( ) k k k x r s r + s + ( r + s ) exp ( k + k k) x exp ( r + s ) r s { } δ r + s +δ r + s + +δ r + s c a r b s r s Γ(,, ) rs, Γ δ + +δ + + +δ + ( r, s, ) { ( r s ) ( r s ) ( r s ) } 6 terms cotrbutg to alasg errors 4

Grd shft techque: x Δ x ± Δ x ± Δ x ± ( ( ), ( ), ( ) ) ( m ) 8 combatos (8 shfted grds) m,,,,8 ( m) m exp( ) u a k x ( m) m exp( ) v b k x c u v k x exp ( ) ( m) ( m) ( m) ( m) (4 FFTs) 8 c c 8 ( m) m Mxed techque: Two grds used oly: x ( Δx, Δx, Δx ) x + Δ x + Δ x + Δ x + ( ( ), ( ), ( ) ) c u x v x k x exp ( ) ( ) ( ) exp( + + + ) cˆ u x v x k x Δx k exp Δx k exp ( ) ( ) exp( + + + ) cˆ u x v x k x cˆ u x v x k x exp ( ) ( ) exp + + ( ) ( r s ) ( ) ( r s ) δ + + c a( r) b( s) δ r + s + + rs, δ + ( r s ) ( ) ( r s ) δ + cˆ a( r) b( s) δ r + s + rs, δ + c c c a r b s r s + ˆ + δ ( + σ) rs, 5

Trucated wave doma: + ˆ + δ ( + σ) σ c c c a r b s r s rs, : exactly oe zero compoet ad the other two are ether + or a total of alasg-error terms Doma D: < ± <, f, a r b s r s D The all alasg-error terms become zero. Proof : suppose σ, σ r + s or r + s or The, oe of the followg must be true: r s r s + + or r s r s + + + ( r + s ) ( r + s ) ± r r s s ± + ± + ± < Mxed techque: completely alasg-error free 6 FFTs ( grds) + trucated wave doma Doma D: largest scrbed spercal wave doma (also true for D flows) < ± <, + + < 9 6