The Spiral of Theodorus, Numerical Analysis, and Special Functions

Σχετικά έγγραφα
Computing the Macdonald function for complex orders

Απόκριση σε Μοναδιαία Ωστική Δύναμη (Unit Impulse) Απόκριση σε Δυνάμεις Αυθαίρετα Μεταβαλλόμενες με το Χρόνο. Απόστολος Σ.

The circle theorem and related theorems for Gauss-type quadrature rules

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

DiracDelta. Notations. Primary definition. Specific values. General characteristics. Traditional name. Traditional notation

Section 9.2 Polar Equations and Graphs

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

SCHOOL OF MATHEMATICAL SCIENCES G11LMA Linear Mathematics Examination Solutions

Example Sheet 3 Solutions

Problem Set 9 Solutions. θ + 1. θ 2 + cotθ ( ) sinθ e iφ is an eigenfunction of the ˆ L 2 operator. / θ 2. φ 2. sin 2 θ φ 2. ( ) = e iφ. = e iφ cosθ.

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

Arithmetical applications of lagrangian interpolation. Tanguy Rivoal. Institut Fourier CNRS and Université de Grenoble 1

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

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

Homework 3 Solutions

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

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

An Inventory of Continuous Distributions

CHAPTER 101 FOURIER SERIES FOR PERIODIC FUNCTIONS OF PERIOD

CHAPTER 48 APPLICATIONS OF MATRICES AND DETERMINANTS

Section 8.3 Trigonometric Equations

Second Order RLC Filters

g-selberg integrals MV Conjecture An A 2 Selberg integral Summary Long Live the King Ole Warnaar Department of Mathematics Long Live the King

2 Composition. Invertible Mappings

Trigonometric Formula Sheet

Homework 8 Model Solution Section

w o = R 1 p. (1) R = p =. = 1

Integrals in cylindrical, spherical coordinates (Sect. 15.7)

Partial Differential Equations in Biology The boundary element method. March 26, 2013

Statistical Inference I Locally most powerful tests

Bessel functions. ν + 1 ; 1 = 0 for k = 0, 1, 2,..., n 1. Γ( n + k + 1) = ( 1) n J n (z). Γ(n + k + 1) k!

ECE 468: Digital Image Processing. Lecture 8

HOMEWORK 4 = G. In order to plot the stress versus the stretch we define a normalized stretch:

Math221: HW# 1 solutions

Exercises to Statistics of Material Fatigue No. 5

Inverse trigonometric functions & General Solution of Trigonometric Equations

CRASH COURSE IN PRECALCULUS

Second Order Partial Differential Equations

Mean-Variance Analysis

University of Illinois at Urbana-Champaign ECE 310: Digital Signal Processing

Solution Series 9. i=1 x i and i=1 x i.

6.1. Dirac Equation. Hamiltonian. Dirac Eq.

CHAPTER 25 SOLVING EQUATIONS BY ITERATIVE METHODS

Matrices and Determinants

A Two-Sided Laplace Inversion Algorithm with Computable Error Bounds and Its Applications in Financial Engineering

Ordinal Arithmetic: Addition, Multiplication, Exponentiation and Limit

Parametrized Surfaces

Lecture 34 Bootstrap confidence intervals

EE512: Error Control Coding

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

A summation formula ramified with hypergeometric function and involving recurrence relation

b. Use the parametrization from (a) to compute the area of S a as S a ds. Be sure to substitute for ds!

DETERMINATION OF DYNAMIC CHARACTERISTICS OF A 2DOF SYSTEM. by Zoran VARGA, Ms.C.E.

Tridiagonal matrices. Gérard MEURANT. October, 2008

DERIVATION OF MILES EQUATION FOR AN APPLIED FORCE Revision C

Discontinuous Hermite Collocation and Diagonally Implicit RK3 for a Brain Tumour Invasion Model

ST5224: Advanced Statistical Theory II

( y) Partial Differential Equations

ES440/ES911: CFD. Chapter 5. Solution of Linear Equation Systems

MATRIX INVERSE EIGENVALUE PROBLEM

Lecture 26: Circular domains

High order interpolation function for surface contact problem

The k-α-exponential Function

Overview. Transition Semantics. Configurations and the transition relation. Executions and computation

Variational Wavefunction for the Helium Atom

Foundations of fractional calculus

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

Survival Analysis: One-Sample Problem /Two-Sample Problem/Regression. Lu Tian and Richard Olshen Stanford University

( ) 2 and compare to M.

Areas and Lengths in Polar Coordinates

Finite Field Problems: Solutions

Areas and Lengths in Polar Coordinates

Probability and Random Processes (Part II)

Answer sheet: Third Midterm for Math 2339

Uniform Convergence of Fourier Series Michael Taylor

Section 7.6 Double and Half Angle Formulas

Pg The perimeter is P = 3x The area of a triangle is. where b is the base, h is the height. In our case b = x, then the area is

Additional Results for the Pareto/NBD Model

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

D Alembert s Solution to the Wave Equation

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

Evaluation of some non-elementary integrals of sine, cosine and exponential integrals type

Numerical Analysis FMN011

Jesse Maassen and Mark Lundstrom Purdue University November 25, 2013

SCITECH Volume 13, Issue 2 RESEARCH ORGANISATION Published online: March 29, 2018

Nachrichtentechnik I WS 2005/2006

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

Fundamentals of Signals, Systems and Filtering

Practice Exam 2. Conceptual Questions. 1. State a Basic identity and then verify it. (a) Identity: Solution: One identity is csc(θ) = 1

SPECIAL FUNCTIONS and POLYNOMIALS

The Pohozaev identity for the fractional Laplacian

4.6 Autoregressive Moving Average Model ARMA(1,1)

Reminders: linear functions

Laplace Expansion. Peter McCullagh. WHOA-PSI, St Louis August, Department of Statistics University of Chicago

Section 8.2 Graphs of Polar Equations

2. Let H 1 and H 2 be Hilbert spaces and let T : H 1 H 2 be a bounded linear operator. Prove that [T (H 1 )] = N (T ). (6p)

Space-Time Symmetries

Numerical Methods for Civil Engineers. Lecture 10 Ordinary Differential Equations. Ordinary Differential Equations. d x dx.

ΔΙΑΚΡΙΤΟΣ ΜΕΤΑΣΧΗΜΑΤΙΣΜΟΣ FOURIER - Discrete Fourier Transform - DFT -

ON NEGATIVE MOMENTS OF CERTAIN DISCRETE DISTRIBUTIONS

Transcript:

Theo p. / The Spiral of Theodorus, Numerical Analysis, and Special Functions Walter Gautschi wxg@cs.purdue.edu Purdue University

Theo p. 2/ Theodorus of ca. 46 399 B.C.

Theo p. 3/ spiral of Theodorus 6 5 4 3 2 2 3 4 5 6 4 2 2 4 6

Theo p. 3/ spiral of Theodorus 6 5 4 3 2 2 3 4 5 6 4 2 2 4 6 numerical analysis k 3/2 + k /2

Theo p. 3/ spiral of Theodorus 6 5 4 3 2 2 3 4 5 6 4 2 2 4 6 numerical analysis k 3/2 + k /2 =.862579229378695957743324666524252345

Theo p. 3/ spiral of Theodorus 6 5 4 3 2 2 3 4 5 6 4 2 2 4 6 numerical analysis k 3/2 + k /2 =.862579229378695957743324666524252345 special functions F (x) = e x2 x e t2 dt Dawson s integral

Theo p. 4/ discrete spiral of Theodorus (also known as Quadratwurzelschnecke ; Hlawka, 98) 2 T 5 T 4 T 3 T T T 2 T 8 T 7 T n 2 Tn+ 3 4 3 2 2 3 4 5

Theo p. 4/ discrete spiral of Theodorus (also known as Quadratwurzelschnecke ; Hlawka, 98) 2 T 5 T 4 T 3 T T T 2 T 8 T 7 T n 2 Tn+ 3 4 3 2 2 3 4 5 parametric representation T (α) C, α

Theo p. 4/ discrete spiral of Theodorus (also known as Quadratwurzelschnecke ; Hlawka, 98) 2 T 5 T 4 T 3 T T T 2 T 8 T 7 T n 2 Tn+ 3 4 3 2 2 3 4 5 parametric representation T (α) C, α defining properties T (n) = T n T n = n T n+ T n = n =,, 2,...

Theo p. 5/ problem: Interpolate the discrete Theodorus spiral by a smooth (or even analytic) spiral

Theo p. 5/ problem: Interpolate the discrete Theodorus spiral by a smooth (or even analytic) spiral solution (P.J. Davis, 993; inspired by Euler s work on the gamma function)

Theo p. 5/ problem: Interpolate the discrete Theodorus spiral by a smooth (or even analytic) spiral solution (P.J. Davis, 993; inspired by Euler s work on the gamma function) T (α) = + i/ k + i/ k + α, α (i = )

Theo p. 5/ problem: Interpolate the discrete Theodorus spiral by a smooth (or even analytic) spiral solution (P.J. Davis, 993; inspired by Euler s work on the gamma function) T (α) = + i/ k + i/ k + α, α (i = ) properties T (α) = α ( ) T (α + ) = + i α T (α)

Theo p. 5/ problem: Interpolate the discrete Theodorus spiral by a smooth (or even analytic) spiral solution (P.J. Davis, 993; inspired by Euler s work on the gamma function) T (α) = + i/ k + i/ k + α, α (i = ) properties T (α) = α ( ) T (α + ) = + i α T (α) heart of the spiral: T (α), α 2

Theo p. 5/ problem: Interpolate the discrete Theodorus spiral by a smooth (or even analytic) spiral solution (P.J. Davis, 993; inspired by Euler s work on the gamma function) T (α) = + i/ k + i/ k + α, α (i = ) properties T (α) = α ( ) T (α + ) = + i α T (α) heart of the spiral: T (α), α 2 (Gronau, 24) Davis s function is the unique solution of the above difference equation with T (α) and arg T (α) monotonically increasing, and T () =.

a little bit of number theory Theo p. 6/

Theo p. 6/ a little bit of number theory distribution of the angles ϕ n = T T T n+ = n sin k +, n =, 2, 3,...

Theo p. 6/ a little bit of number theory distribution of the angles ϕ n = T T T n+ = n sin k +, n =, 2, 3,... equidistribution (Hlawka, 979) The sequence {ϕ n } n= is equidistributed mod 2π

Theo p. 6/ a little bit of number theory distribution of the angles ϕ n = T T T n+ = n sin k +, n =, 2, 3,... equidistribution (Hlawka, 979) The sequence {ϕ n } n= is equidistributed mod 2π more generally ϕ n (α) = T (α)t T (α + n) = n sin k + α < α < 2, n =, 2, 3,...

Theo p. 6/ a little bit of number theory distribution of the angles ϕ n = T T T n+ = n sin k +, n =, 2, 3,... equidistribution (Hlawka, 979) The sequence {ϕ n } n= is equidistributed mod 2π more generally ϕ n (α) = T (α)t T (α + n) = n sin k + α < α < 2, n =, 2, 3,... The sequence {ϕ n (α)} n= is equidistributed mod 2π for any α with < α < 2 (Niederreiter, email Feb. 3, 29)

Theo p. 7/ logarithmic derivative of T (α) T (α) T (α) = + i/ k + α + i/ k d dα ( + i/ ) k + i/ k + α

Theo p. 7/ logarithmic derivative of T (α) T (α) T (α) = + i/ k + α + i/ k d dα ( + i/ ) k + i/ k + α = ( + i/ k + α ) i 2 (k + α ) 3/2 ( + i/ k + α ) 2

Theo p. 7/ logarithmic derivative of T (α) T (α) T (α) = + i/ k + α + i/ k d dα ( + i/ ) k + i/ k + α = ( + i/ k + α ) i 2 (k + α ) 3/2 ( + i/ k + α ) 2 = i 2 (k + α )( k + α + i)

Theo p. 7/ logarithmic derivative of T (α) T (α) T (α) = + i/ k + α + i/ k d dα ( + i/ ) k + i/ k + α = ( + i/ k + α ) i 2 (k + α ) 3/2 ( + i/ k + α ) 2 = i 2 = i 2 (k + α )( k + α + i) k + α i (k + α )(k + α)

Theo p. 8/ logarithmic derivative of T (α) (cont ) = 2 (k + α )(k + α) + i 2 (k + α ) 3/2 + (k + α ) /2

Theo p. 8/ logarithmic derivative of T (α) (cont ) = 2 (k + α )(k + α) + i 2 = 2 (k + α ) 3/2 + (k + α ) /2 ( k + α ) + i k + α 2 U(α)

Theo p. 8/ logarithmic derivative of T (α) (cont ) = 2 (k + α )(k + α) + i 2 where = 2 (k + α ) 3/2 + (k + α ) /2 ( k + α ) + i k + α 2 U(α) = 2α + i 2 U(α) U(α) = (k + α ) 3/2 + (k + α ) /2

T (α) T (α) = 2α + i 2 U(α) Theo p. 9/

Theo p. 9/ integrate from to α> T (α) T (α) = 2α + i 2 U(α) ln T (α) = ln(α /2 ) + i 2 α U(α)dα

Theo p. 9/ integrate from to α> T (α) T (α) = 2α + i 2 U(α) ln T (α) = ln(α /2 ) + i 2 polar representation of T (α) T (α) = ( i α exp 2 α α U(α)dα ) U(α)dα, α >

Theo p. 9/ integrate from to α> T (α) T (α) = 2α + i 2 U(α) ln T (α) = ln(α /2 ) + i 2 polar representation of T (α) T (α) = ( i α exp 2 α α U(α)dα ) U(α)dα, α > since T () = + i U(), the slope of the tangent vector to the 2 2 spiral at α = is U() = (Theodorus constant) k 3/2 + k /2

Theo p. 9/ integrate from to α> T (α) T (α) = 2α + i 2 U(α) ln T (α) = ln(α /2 ) + i 2 polar representation of T (α) T (α) = ( i α exp 2 α α U(α)dα ) U(α)dα, α > since T () = + i U(), the slope of the tangent vector to the 2 2 spiral at α = is U() = (Theodorus constant) k 3/2 + k /2 numerical analysis and special functions: compute and identify U(α) = U(α)dα for < α < 2, (k+α ) 3/2 +(k+α ) /2 α

Theo p. / first digression summation by integration (G. & Milovanović, 985) s = a k, a k = (Lf)(k)

Theo p. / first digression summation by integration (G. & Milovanović, 985) s = a k, a k = (Lf)(k) s = (Lf)(k) = e kt f(t)dt

Theo p. / first digression summation by integration (G. & Milovanović, 985) s = a k, a k = (Lf)(k) = s = (Lf)(k) = ( ) e kt f(t)dt = e kt f(t)dt t e t f(t) t dt

Theo p. / first digression summation by integration (G. & Milovanović, 985) s = a k, a k = (Lf)(k) Thus = s = (Lf)(k) = ( ) e kt f(t)dt = a k = ε(t) = t e t f(t) t ε(t)dt, e kt f(t)dt t e t f(t) t f = L a Bose Einstein distribution dt

Theo p. / Theodorus: a k = k /2 = k 3/2 + k/2 k + convolution theorem for Laplace transform Lg Lh = Lg h, (g h)(t) = t g(τ)h(t τ)dτ

Theo p. / Theodorus: a k = k /2 = k 3/2 + k/2 k + convolution theorem for Laplace transform Lg Lh = Lg h, (g h)(t) = t g(τ)h(t τ)dτ application to a k k /2 = k + ( L t /2 π ) (k) = ( L e t) (k)

Theo p. / Theodorus: a k = k /2 = k 3/2 + k/2 k + convolution theorem for Laplace transform Lg Lh = Lg h, (g h)(t) = t g(τ)h(t τ)dτ application to a k k /2 = k + ( L t /2 π ) (k) = ( L e t) (k) a k = ( L t /2 π ) (k) (L e t ) (k) = t (L π ) τ /2 e (t τ) dτ (k)

Theo p. 2/ Theodorus (cont ) f(t) = π e t t = 2 π e t t τ /2 e τ dτ e x2 dx = 2 π F ( t)

Theo p. 2/ Theodorus (cont ) f(t) = π e t t = 2 π e t t τ /2 e τ dτ e x2 dx = 2 π F ( t) k 3/2 + k /2 = f(t) t ε(t)dt = 2 π F ( t) t w(t)dt

Theo p. 2/ Theodorus (cont ) f(t) = π e t t = 2 π e t t τ /2 e τ dτ e x2 dx = 2 π F ( t) k 3/2 + k /2 = f(t) t ε(t)dt = 2 π F ( t) t w(t)dt w(t) = t /2 ε(t) = t/2 e t

Theo p. 3/ second digression Gaussian quadrature n-point quadrature formula g(t)w(t)dt = n λ (n) k g(τ (n) k ), g P 2n

Theo p. 3/ second digression Gaussian quadrature n-point quadrature formula g(t)w(t)dt = n λ (n) k g(τ (n) k ), g P 2n orthogonal polynomials (π k, π l ) =, k l, where (u, v) = u(t)v(t)w(t)dt

Theo p. 3/ second digression Gaussian quadrature n-point quadrature formula g(t)w(t)dt = n λ (n) k g(τ (n) k ), g P 2n orthogonal polynomials (π k, π l ) =, k l, where (u, v) = three-term recurrence relation u(t)v(t)w(t)dt π k+ (t) = (t α k )π k (t) β k π k (t), k =,, 2,... π (t) =, π (t) = where α k = α k (w) R, β k = β k (w) >, β = w(t)dt

Theo p. 4/ Jacobi matrix J n (w) = α β β α β 2 β 2 α 2......... β n β n α n

Theo p. 4/ Jacobi matrix J n (w) = α β β α β 2 β 2 α 2......... β n β n α n Gaussian nodes and weights (Golub & Welsch, 969) τ (n) k = eigenvalues of J n, λ (n) k = β v 2 k, v k, = first component of (normalized) eigenvector v k

Theo p. 4/ Jacobi matrix J n (w) = α β β α β 2 β 2 α 2......... β n β n α n Gaussian nodes and weights (Golub & Welsch, 969) τ (n) k = eigenvalues of J n, λ (n) k = β v 2 k, v k, = first component of (normalized) eigenvector v k moments of w µ k = t k w(t)dt, k =,,..., 2n

Theo p. 4/ Jacobi matrix J n (w) = α β β α β 2 β 2 α 2......... β n β n α n Gaussian nodes and weights (Golub & Welsch, 969) τ (n) k = eigenvalues of J n, λ (n) k = β v 2 k, v k, = first component of (normalized) eigenvector v k moments of w µ k = Chebyshev algorithm t k w(t)dt, k =,,..., 2n {µ k } 2n k= {α k, β k } n k=

Theo p. 5/ numerical results for Gaussian quadrature (in 5D-arithmetic) µ k = t k w(t)dt = t k+/2 e t dt = Γ(k + 3/2)ζ(k + 3/2)

Theo p. 5/ numerical results for Gaussian quadrature (in 5D-arithmetic) µ k = t k w(t)dt = t k+/2 e t dt = Γ(k + 3/2)ζ(k + 3/2) k 3/2 + k = 2 [F ( t)/ t ] w(t) /2 π s n = 2 n ( ) λ (n) π k F τ (n) k / τ (n) k

Theo p. 5/ numerical results for Gaussian quadrature (in 5D-arithmetic) µ k = t k w(t)dt = t k+/2 e t dt = Γ(k + 3/2)ζ(k + 3/2) k 3/2 + k = 2 [F ( t)/ t ] w(t) /2 π s n = 2 n ( ) λ (n) π k F τ (n) k / τ (n) k n s n 5.85997... 5.862579227... 25.862579229378689... 35.86257922937869595774... 45.8625792293786959577433246665235... 55.862579229378695957743324666524252 65.862579229378695957743324666524252 75.862579229378695957743324666524252

Theo p. 6/ computation and identification of U(α) and α U(α)dα (k + α ) /2 (k + α ) + = t ) (L τ /2 e (t τ) dτ (k + α ) π

Theo p. 6/ computation and identification of U(α) and α U(α)dα (k + α ) /2 (k + α ) + = t ) (L τ /2 e (t τ) dτ (k + α ) π applying the shift property for Laplace transform yields (Lg) (s + b) = ( L e bt g(t) ) (s) (k + α ) /2 (k + α ) + = t L (e αt π ) τ /2 e τ dτ (k)

Theo p. 6/ computation and identification of U(α) and α U(α)dα (k + α ) /2 (k + α ) + = t ) (L τ /2 e (t τ) dτ (k + α ) π applying the shift property for Laplace transform yields hence (Lg) (s + b) = ( L e bt g(t) ) (s) (k + α ) /2 (k + α ) + = t L (e αt π f(t) = π e αt t ) τ /2 e τ dτ (k) τ /2 e τ dτ = 2 π e (α )t F ( t)

Theo p. 7/ U(α) = f(t) t ε(t)dt = 2 e (α )t F ( t) π t w(t)dt w(t) = t/2 e t, α < 2

Theo p. 7/ U(α) = f(t) t ε(t)dt = 2 e (α )t F ( t) π t w(t)dt w(t) = t/2 e t, α < 2 identification of U(α) as a Laplace transform U(α) = (Lu) (α ), u(t) = 2 π F ( t) t w(t)

Theo p. 7/ U(α) = f(t) t ε(t)dt = 2 e (α )t F ( t) π t w(t)dt w(t) = t/2 e t, α < 2 identification of U(α) as a Laplace transform U(α) = (Lu) (α ), u(t) = 2 π F ( t) t w(t) the integral of U(α) α U(α)dα = 2(α ) π e (α )t (α )t F ( t) t w(t)dt

Theo p. 8/ Epilogue summation theory can be generalized to series s + = k ν r(k), s = ( ) k k ν r(k), < ν < where r(k) is a rational function

Theo p. 8/ Epilogue summation theory can be generalized to series s + = k ν r(k), s = ( ) k k ν r(k), < ν < where r(k) is a rational function regularizing transformation (J. Waldvogel, in preparation) α = r 2, r R

Theo p. 8/ Epilogue summation theory can be generalized to series s + = k ν r(k), s = ( ) k k ν r(k), < ν < where r(k) is a rational function regularizing transformation (J. Waldvogel, in preparation) then T (α) = T (r 2 ) = α = r 2, + i + i/r k=2 r R + i/ k + i/ r 2 + k is analytic for r C \ [i, i ] [ i, i ]

Theo p. 8/ Epilogue summation theory can be generalized to series s + = k ν r(k), s = ( ) k k ν r(k), < ν < where r(k) is a rational function regularizing transformation (J. Waldvogel, in preparation) then T (α) = T (r 2 ) = α = r 2, + i + i/r k=2 r R + i/ k + i/ r 2 + k is analytic for r C \ [i, i ] [ i, i ] r > : spiral of Theodorus

Theo p. 8/ Epilogue summation theory can be generalized to series s + = k ν r(k), s = ( ) k k ν r(k), < ν < where r(k) is a rational function regularizing transformation (J. Waldvogel, in preparation) then T (α) = T (r 2 ) = α = r 2, + i + i/r k=2 r R + i/ k + i/ r 2 + k is analytic for r C \ [i, i ] [ i, i ] r > : spiral of Theodorus r < : analytic continuation of the spiral

Theo p. 9/ twin-spiral of Theodorus 6 5 4 3 2 2 3 4 5 6 4 2 2 4 6