A Numerical Approach To The Steepest Descent Method
|
|
- Piers Hensley
- 5 years ago
- Views:
Transcription
1 A Numerical Approach To The Steepest Descent Method K.U. Leuven, Division of Numerical and Applied Mathematics Isaac Newton Institute, Cambridge, Feb 15, 27
2 Outline 1 One-dimensional oscillatory integrals 2
3 Outline One-dimensional oscillatory integrals 1 One-dimensional oscillatory integrals 2
4 One-dimensional oscillatory integrals The model problem I := b f (x)e iωg(x) dx a with... ω: frequency parameter; f : amplitude; g: oscillator f, g: smooth real functions classical quadrature deteriorates rapidly as ω increases take fixed number of points per oscillation amount of operations scales linearly with ω new oscillatory quadrature methods asymptotic, Filon, Levin, numerical steepest descent
5 One-dimensional oscillatory integrals What determines the value of the integral? Example: (1 x 2 )e i ω x regions where the oscillations do not cancel: boundary points regions where the integrand is (locally) not oscillatory: stationary points: solutions to g (x) =
6 One-dimensional oscillatory integrals What is the size of the integral (asymptotically)? I = O(ω 1/(r+1) ) with r: the largest order of any stationary point ξ in [a, b] g (j) (ξ) =, j = 1,..., r; g (r+1) (ξ) Our goal: to find a decomposition of the integral b a f (x)eiωg(x) dx = F (a) + F (b) + i F (ξ i) with ξ i [a, b] the stationary points
7 Basic idea: select a new integration path assume f and g are analytic Cauchy s theorem: the value of I does not depend on the complex path taken C a b
8 The path of steepest descent (Cauchy(1827), Riemann (1863)) e iωg(x) = e iω(rg(x)+iig(x)) = e ωig(x) e iωrg(x) 1 The function e iωg(x) does not oscillate if Rg(x) is fixed 2 The function e iωg(x) decays exponentially fast if Ig(x) > new path at point a: h a (p) such that g(h a (p)) = g(a) + p i, p
9 Example 1: Fourier oscillator g(x) = x (f (x) = 1 x 2 ) g(h a (p)) = g(a) + pi h a (p) = a + pi 3 2 C a h a h b 1 1 b
10 Decomposition: I = b a f (x)eiωg(x) = F (a) F (b) F (a) = f (h a (p))e iωg(ha(p)) h a(p)dp = e iωg(a) f (h a (p))e ωp h a(p)dp = e iωg(a) f (a + pi)e ωp idp
11 One-dimensional oscillatory integrals A new integration path Example 2: Quadratic oscillator g(x) = x 2 (f (x) = 1 x 2 ) g(h a (p)) = g(a) + pi h a (p) = ± a 2 + pi
12 Decomposition: I = F 1 (a) F 1 (ξ) + F 2 (ξ) F 2 (b) 1 1 f (x)e iωx2 dx = e iω f (h 1,1 (p))e ωp h 1,1(p)dp + f (h,1 (p))e ωp h,1(p)dp f (h,2 (p))e ωp h,2(p)dp e iω f (h 1,2 (p))e ωp h 1,2(p)dp Numerical singularity of the path: h ξ (p) p 1/2, p
13 Example 3: Cubic oscillator g(x) = x 3 g(h a (p)) = g(a) + pi h a (p) = 3 a 3 + pi e i2π/3k, k =, 1, 2 C i I = F 1 (a) F 1 (ξ) + F 2 (ξ) F 2 (b) Numerical singularity of the path: h ξ (p) p 2/3, p
14 Example 3: Cubic oscillator g(x) = x 3 (and f (x) = 1 x 2 )
15 Implementation issue 1. How to evaluate F j? F j (x) = e iωg(x) f (h x (p))h x(p)e ωp dp = eiωg(x) ω f (h x ( q ω ))h x( q ω )e q dq if exponential decay: Gauss-Laguerre (w(q) = e q ) if singularity: generalized Gauss-Laguerre (w(q) = q α e q ) or, generalized Gauss-Hermite (w(q) = e qn )
16 F j (x) Q F [f, g, h x ] := eiωg(x) ω n w i f (h x (x i /ω))h x(x i /ω) i=1 F j (ξ) Q r F [f, g, h ξ,j] := eiωg(ξ) ω 1/(r+1) n i=1 absolute error: O(ω ( 2n 1)/(r+1) ) relative error: O(ω 2n/(r+1) ) w i f (h ξ,j (x r+1 i /ω)) h ξ,j r+1 (xi /ω) xi r Gaussian convergence rate 2n as a function of 1/ω (r = )
17 Example: x eiωx dx I Q F [f, g, h a ] Q F [f, g, h b ] ω \ n E 3 3.1E 5 1.9E 6 1.7E 7 2.1E E 4 1.1E 6 2.3E 8 7.5E 1 3.2E E 5 3.9E 8 2.1E 1 2.E E E 6 1.2E 9 1.7E E E 17 rate 3.1(3) 5.(5) 6.9(7) 8.9(9) 1.8(11)
18 Example: x+2 eiωx3 dx I Q[f, g, h a ] Q 2 F [f, g, h ξ,1] + Q 2 F [f, g, h ξ,2] Q F [f, g, h b ] ω \ n E 4 2.4E 6 1.3E 8 6.9E E E 5 7.1E 7 2.3E 9 6.4E E E 5 2.1E 7 4.2E 1 6.1E E E 6 6.7E 8 8.1E E E 16 rate 1.3 (3/3) 1.7 (5/3) 2.4 (7/3) 3.4 (9/3) 3.9 (11/3)
19 Implementation issue 2. How to determine the path? if inverse of g is available: h a (p) = g 1 (g(a) + pi) otherwise...compute h a (x i /ω) and h a(x i /ω) numerically Apply Newton-Raphson iteration to g(h a (p)) g(a) pi = initial guess by truncated Taylor series of g only very few iterations necessary Derivative: g(h a (p)) g(a) pi = g (h a (p))h a(p) i =
20 Example: x eiω(x2 +x+1) 1/3 dx I Q F [f, g, h a ] Q F [f, g, h b ] with 2 nd order Taylor path ω \ n E 2 2.7E 3 7.4E 4 2.4E 4 8.9E E 3 2.6E 4 4.6E 5 1.E 5 2.5E E 4 1.8E 5 1.7E 6 2.E 7 2.9E E 5 1.1E 6 4.E 8 2.1E 9 1.5E E 6 6.8E 8 7.7E 1 1.6E E 13 rate
21 Example: x eiω(x2 +x+1) 1/3 dx I Q F [f, g, h a ] Q F [f, g, h b ] with 2 nd order Taylor path, followed by 1 to 4 Newton iteration steps ω \ n E 2 2.4E 3 7.4E 4 2.5E 4 7.5E E 3 2.4E 4 4.4E 5 1.E 5 2.4E E 4 1.5E 5 1.2E 6 1.5E 7 2.3E E 5 6.1E 7 1.8E 8 8.7E 1 6.2E E 6 2.1E 8 1.8E 1 2.7E E E 7 6.7E 1 1.5E E E 17 rate
22 Relation to Levin type methods Levin method: write I = F (b)e iωg(b) F (a)e iωg(a) with F (x) the non-oscillatory solution to F (x) + iωg (x)f (x) = f (x) It can be verified that the exact solution is given by F (x) = f (h x (p))h x(p)e ωp dp Numerical steepest descent: evaluate the analytical solution to the Levin ODE numerically!
23 The Generalized Filon s method (Iserles and Nørsett) Idea: approximate f globally on [a, b] by Hermite interpolation Assume: f (x) N i=1 c iφ i (x) Then: I N i=1 w ic i with w i = b a φ i(x)e iωg(x) dx Iserles and Nørsett: convergence O(ω s 1/(r+1) ) interpolate derivatives of order... s 1 at corner points interpolate derivatives of order... (s 1)(r + 1) at stationary points computation of moments, e.g., by numerical steepest descent
24 Localized Filon s method Idea: apply local Hermite (or Taylor) approximation of f for each integral contribution F j Define: From: We have: with: F j [f ](x) := e iωg(x) f (h x (p))h x(p)e ωp dp f (z) d j i= f (i) (x) (z x)i i! F j [f ](x) d j i= w i,jf (i) (x) w i,j := F j [ (z x)i i! ](x)
25 A classical quadrature rule Convergence result: I l dj j= i= w i,jf (i) (x j ) Let d j = s 1 at a non-stationary point, and let d j = s(r + 1) 1 at a stationary point then our rule has an absolute error of O(ω s 1/(r+1) ) and a relative error of O(ω s ), asymptotically for large ω.
26 Example 1: x 2 e iω(x 1/2)2 dx Filon type 3 weights localized Filon type numerical steepest descent Filon type 5 weights Filon (3 evals): O(ω 3/2 ) Local Filon (3 evals): O(ω 3/2 ) NSD (4 evals): O(ω 5/2 ) Filon (5 evals): O(ω 2 ) ω
27 Example 2: The Hankel oscillator I H [f ] := For large arguments H ν (1) (z) b a f (x)h ν (1) (ωg 1 (x))e iωg2(x) dx, 2 πz ei(z 1 2 νπ 1/4π), π < argz < 2π, z. Approximate oscillator: g(x) = g 1 (x) + g 2 (x) Quadrature rule: I H [f ] Q H [f ] := L l= dl j= w H l,j f (j) (x l ).
28 1 cos(x 1)H(1) (ωx)eiω(x2 +x 3 x) dx. Two quadrature points: x = : a singularity and a stationary point of order 1 x = 1: a regular endpoint ω \ (d, d 1 ) (, ) (1, ) (2, ) (3, 1) 1 1.2E 3 2.8E 5 1.3E 6 2.6E E 4 8.6E 6 2.9E 7 4.1E E 4 2.6E 6 6.4E 8 6.2E E E 7 1.4E 8 9.7E 11 rate 1.23 (1.25) 1.73 (1.75) 2.2 (2.25) 2.68 (2.75)
29 Outline 1 One-dimensional oscillatory integrals 2
30 Multivariate oscillatory integrals Model form I n := S f (x)e iωg(x) dx Contributing points? corner points critical points: g = resonance points: g S
31 Multivariate oscillatory integrals Main approach: repeated one-dimensional integration deform onto path of steepest descent for inner variable I 1 (x) := b(x) a(x) f (x, y)e iωg(x,y) dy = F (x, a(x)) F (x, b(x)) the function F is evaluated only in points on the boundary F (x, a(x)) = e iωg(x,a(x)) u(x, p) f (x, u(x, p)) e ωp dp p oscillator of F (x, a(x)) is exactly known: g(x, a(x))!
32 Example 1 Rectangular domain in two dimensions I := b d a c f (x, y)e iω(x+y) dy dx Step 1: deform onto path of steepest descent for y I := b a Smooth function G is given by G(x, c)e iω(x+c) G(x, d)e iω(x+d) dx G(x, y) = f (x, y + ip)ie ωp dp
33 Example 1 (continued) Rectangular domain in two dimensions b a G(x, c)e iω(x+c) = G(a, c)e iω(a+c) G(b, c)e iω(b+c) Total decomposition I := F (a, c) F (b, c) F (a, d) + F (b, d) Contributions are given by non-oscillatory double integrals with exponential decay in both variables F (x, y) = e iω(x+y) f (x + ip, y + iq)i 2 e ω(p+q) dq dp
34 Example 1 (continued) Rectangular domain in two dimensions I := b d a c f (x, y)e iω(x+y) dy dx (a,d) (b,d) (a,c) (b,c) I := F (a, c) F (b, c) F (a, d) + F (b, d)
35 Example 2: smooth boundaries 1. A two-dimensional simplex I := b x a a f (x, y)e iω(x+y) dy dx Step 1: deform onto path of steepest descent for y I := b a G(x, a)e iω(x+a) G(x, x)e iω(x+x) dx oscillators are different Result: contributions of the three corner points
36 Example 2: smooth boundaries (continued) 2. More general boundaries I := = b d(x) a b a c(x) f (x, y)e iω(x+y) dy dx G(x, c(x))e iω(x+c(x)) G(x, d(x))e iω(x+d(x)) dx new oscillator x + c(x) may have stationary points! resonance points: stationary point of oscillator g(x, c(x)) evaluated along the boundary happens when g S
37 Example 2: smooth boundaries (continued) Fourier integral on a circle 1 1 x 2 I := 1 f (x, y)e iω(x+y) dy dx 1 x ( 2 ) ( ) = F 2, F 2 2, 2
38 Example 3: critical points A rectangular domain with a critical point I := b d g(x, y) = x 2 xy y 2 g(, ) = g x (x, y) = x = y/2 g y (x, y) = y = x/2 a c f (x, y)e iω(x2 xy y 2) dy dx
39 Example 3: critical points (continued) A rectangular domain with a critical point (a, a/2) (a,d) (d/2,d) (b,d) (,) (a,c) (c/2,c) (b, b/2) (b,c)
40 Example 3: critical points (continued) A rectangular domain with a critical point Step 1: deform onto path of steepest descent for y I := b a G 1 (x, c)e iωg(x,c) G 1 (x, x/2)e iωg(x, x/2) + G 2 (x, x/2)e iωg(x, x/2) G 2 (x, d)e iωg(x,d) dx g 11 (x) := g(x, c) = x 2 cx c 2 g 12 (x) := g(x, x/2) = 5 4 x 2
41 Example 3: critical points (continued) A rectangular domain with a critical point Step 2: deform onto path of steepest descent for x I = F 111 (a, c) F 111 (c/2, c) + F 112 (c/2, c) F 112 (b, c) F 121 (a, a/2) + F 121 (, ) F 122 (, ) + F 122 (b, b/2) + F 211 (a, a/2) F 211 (, ) + F 212 (, ) F 212 (b, b/2) F 221 (a, d) + F 221 (d/2, d) F 222 (d/2, d) + F 222 (b, d).
42 What can be proved Theorem 4.2 (Huybrechs and Vandewalle, 26) I n := f (x)e iωg(x) dx = s λ F λ (x λ ) + O(e ωd ), S size(λ)=2n Conditions are: analyticity of f, g in a complex neighbourhood of S piecewise analytic parameterisation of S stationary points may be degenerate two additional conditions exclude special cases
43 Additional conditions 1. Exclude curves of resonance points and critical points If for some λ we have g λ (x, y), y then the integral in y is not oscillatory. Solution: integration in y can be performed numerically on the real line
44 Additional conditions (continued) Example curve of resonance points on circular boundary Example: x 2 +y 2 <=1 f (x, y)eiω(x2 +y 2) dv
45 Additional conditions (continued) 2. Each lower-dimensional integral in x is either: an integral along a curve of stationary points (of the same order) in y an integral along a curve that has no stationary point in y In particular: this excludes critical points on the boundary Reason: existence of complex stationary points in y that may (or may not) be arbitrarily close to the real line
46 Additional conditions (continued) Example: integral on the part in the upper right halfplane (a, a/2) (a,d) (d/2,d) (b,d) (,) (a,c) (c/2,c) (b, b/2) (b,c)
47 Example 1: 3D balls and ellipsoids Sphere with Fourier oscillator No stationary points, two boundary points Z 1 Z q 1 x 2 Z q 1 x I 3 = x2 2 q 1 1 x 1 2 q 1 x 1 2 e x 1 +x2 2 x 3 (3x 3 + cos(x 2 ))e iω(x 1 +x 2 +x 3 ) dx 3 dx 2 dx 1. x2 2 Contributing points: g S ( 3/3, 3/3, 3/3) and ( 3/3, 3/3, 3/3)
48 Example 1: 3D balls and ellipsoids Cubature rule Convergence I 3 2 i=1 j k l w i,j,k,l j+k+l f x j 1 x k 2 x l 3 (x i ) ω \ d e 5 2.4e 6 1.2e e 6 3.6e 7 8.e e 7 5.2e 8 5.5e e 8 3.8e 9 2.7e e 9 5.2e 1 1.7e 12 rate 3. (2.5) 2.9 (3.) 4. (3.5)
49 Example 1: 3D balls and ellipsoids Generalisation to an ellipsoid 1 I 3 := 4π k2 (n(x) 2 1)e iω a x dx 3 dx 2 dx 1. E length scales R 1, R 2, R 3 along X, Y and Z axis oscillator: a x = a 1 x 1 + a 2 x 2 + a 3 x 3 two resonance points application: scattering of light due to propagation in an object with refractive index n(x) (Sigal Trattner)
50 Example 1: an ellipsoid (continued) Absolute errors R 1 = 1, R 2 = 2, R 3 = 1 tensor-product Gauss-Laguerre with m points per dimension ω \ m (4m 3 ) 1 (4) 2 (32) 3 (18) 4 (256) 1 2.2e 2 4.9e 3 3.8e 3 1.7e e 3 7.5e 5 7.8e 5 6.7e e 3 6.e 7 5.9e 7 5.9e e 6 3.8e 8 6.6e e e 6 2.4e 9 1.1e e 14 rate
51 Example 2: a degenerate stationary point A rectangular domain with a degenerate stationary point I 2 := x + y eiω(x3 +y 3) dy dx, (a,d) (b,d) (a,c) (b,c)
52 Example 2 (continued) Localised Filon-type quadrature rule nine quadrature points use d function values and derivatives at each point ω \ d e 3 2.5e 4 2.9e e 3 9.9e 5 9.e e 3 3.9e 5 2.8e e 4 1.6e 5 8.9e e 4 6.1e 6 2.9e 7 rate.98 (1.) 1.34 (1.33) 1.63 (1.66)
53 Concluding remarks 1 One-dimensional integrals two types of points: stationary points and endpoints integration along the path of steepest descent construction of Filon-type quadrature rules 2 Multi-dimensional integrals three types of points: corners, critical points, resonance points integration on a manifold of steepest descent construction of Filon-type cubature rules 3 Problems not treated here functions with complex poles or stationary points oscillatory integral equations
Complex Gaussian quadrature of oscillatory integrals
Complex Gaussian quadrature of oscillatory integrals Alfredo Deaño Daan Huybrechs. March 18, 28 Abstract We construct and analyze Gauss-type quadrature rules with complexvalued nodes and weights to approximate
More informationCalculation of Highly Oscillatory Integrals by Quadrature Method
Calculation of Highly Oscillatory Integrals by Quadrature Methods Krishna Thapa Texas A&M University, College Station, TX May 18, 211 Outline Purpose Why Filon Type Quadrature Methods Our Expectations
More informationA combined Filon/asymptotic quadrature method for highly oscillatory problems
BIT Numerical Mathematics (26)46:- c Springer 26. DOI:.7/s543---x A combined Filon/asymptotic quadrature method for highly oscillatory problems A. Asheim Abstract. Department of Mathematical Sciences,
More informationOn the quadrature of multivariate highly oscillatory integrals over non-polytope domains
On the quadrature of multivariate highly oscillatory integrals over non-polytope domains Sheehan Olver February 16, 2006 Abstract In this paper, we present a Levin-type method for approximating multivariate
More informationMATH 205C: STATIONARY PHASE LEMMA
MATH 205C: STATIONARY PHASE LEMMA For ω, consider an integral of the form I(ω) = e iωf(x) u(x) dx, where u Cc (R n ) complex valued, with support in a compact set K, and f C (R n ) real valued. Thus, I(ω)
More informationPreconditioned GMRES for oscillatory integrals
Report no. 08/9 Preconditioned GMRES for oscillatory integrals Sheehan Olver Oxford University Computing Laboratory, Wolfson Building, Parks Road, Oxford, UK Sheehan.Olver@sjc.ox.ac.uk Abstract None of
More informationTHE MAGNUS METHOD FOR SOLVING OSCILLATORY LIE-TYPE ORDINARY DIFFERENTIAL EQUATIONS
THE MAGNUS METHOD FOR SOLVING OSCILLATORY LIE-TYPE ORDINARY DIFFERENTIAL EQUATIONS MARIANNA KHANAMIRYAN Department of Applied Mathematics and Theoretical Physics, University of Cambridge, Wilberforce Road,
More informationHighly oscillatory quadrature and its applications
UNIVERSITY OF CAMBRIDGE CENTRE FOR MATHEMATICAL SCIENCES DEPARTMENT OF APPLIED MATHEMATICS & THEORETICAL PHYSICS Highly oscillatory quadrature and its applications Arieh Iserles & Syvert Nørsett Report
More informationNumerical Approximation of Highly Oscillatory Integrals
Numerical Approximation of Highly Oscillatory Integrals Sheehan Olver Trinity Hall University of Cambridge 14 June 2008 This dissertation is submitted for the degree of Doctor of Philosophy Abstract The
More informationyou expect to encounter difficulties when trying to solve A x = b? 4. A composite quadrature rule has error associated with it in the following form
Qualifying exam for numerical analysis (Spring 2017) Show your work for full credit. If you are unable to solve some part, attempt the subsequent parts. 1. Consider the following finite difference: f (0)
More informationNumerical Methods for Highly Oscillatory Problems
Andreas Asheim Numerical Methods for Highly Oscillatory Problems Doctoral thesis for the degree of philosophiae doctor Trondheim, May 2010 Norwegian University of Science and Technology Faculty of Information
More informationFixed point iteration and root finding
Fixed point iteration and root finding The sign function is defined as x > 0 sign(x) = 0 x = 0 x < 0. It can be evaluated via an iteration which is useful for some problems. One such iteration is given
More informationFinal Exam May 4, 2016
1 Math 425 / AMCS 525 Dr. DeTurck Final Exam May 4, 2016 You may use your book and notes on this exam. Show your work in the exam book. Work only the problems that correspond to the section that you prepared.
More information(f(x) P 3 (x)) dx. (a) The Lagrange formula for the error is given by
1. QUESTION (a) Given a nth degree Taylor polynomial P n (x) of a function f(x), expanded about x = x 0, write down the Lagrange formula for the truncation error, carefully defining all its elements. How
More information0.1 Problems to solve
0.1 Problems to solve Homework Set No. NEEP 547 Due September 0, 013 DLH Nonlinear Eqs. reducible to first order: 1. 5pts) Find the general solution to the differential equation: y = [ 1 + y ) ] 3/. 5pts)
More informationSection 3.4. Second Order Nonhomogeneous. The corresponding homogeneous equation
Section 3.4. Second Order Nonhomogeneous Equations y + p(x)y + q(x)y = f(x) (N) The corresponding homogeneous equation y + p(x)y + q(x)y = 0 (H) is called the reduced equation of (N). 1 General Results
More informationSeries Solutions of Linear ODEs
Chapter 2 Series Solutions of Linear ODEs This Chapter is concerned with solutions of linear Ordinary Differential Equations (ODE). We will start by reviewing some basic concepts and solution methods for
More informationLECTURE 16 GAUSS QUADRATURE In general for Newton-Cotes (equispaced interpolation points/ data points/ integration points/ nodes).
CE 025 - Lecture 6 LECTURE 6 GAUSS QUADRATURE In general for ewton-cotes (equispaced interpolation points/ data points/ integration points/ nodes). x E x S fx dx hw' o f o + w' f + + w' f + E 84 f 0 f
More informationCh. 03 Numerical Quadrature. Andrea Mignone Physics Department, University of Torino AA
Ch. 03 Numerical Quadrature Andrea Mignone Physics Department, University of Torino AA 2017-2018 Numerical Quadrature In numerical analysis quadrature refers to the computation of definite integrals. y
More informationChapter 4: Interpolation and Approximation. October 28, 2005
Chapter 4: Interpolation and Approximation October 28, 2005 Outline 1 2.4 Linear Interpolation 2 4.1 Lagrange Interpolation 3 4.2 Newton Interpolation and Divided Differences 4 4.3 Interpolation Error
More informationHighly Oscillatory Problems: Computation, Theory and Applications
Highly Oscillatory Problems: Computation, Theory and Applications Isaac Newton Institute for Mathematical Sciences Edited by 1 Rapid function approximation by modified Fourier series Daan Huybrechs and
More informationHigh Frequency Scattering by Convex Polygons Stephen Langdon
Bath, October 28th 2005 1 High Frequency Scattering by Convex Polygons Stephen Langdon University of Reading, UK Joint work with: Simon Chandler-Wilde Steve Arden Funded by: Leverhulme Trust University
More informationAn Efficient Adaptive Levin-type Method for Highly Oscillatory Integrals
Applied Mathematical Sciences, Vol. 8, 2014, no. 138, 6889-6908 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ams.2014.49720 An Efficient Adaptive Levin-type Method for Highly Oscillatory Integrals
More informationEAD 115. Numerical Solution of Engineering and Scientific Problems. David M. Rocke Department of Applied Science
EAD 115 Numerical Solution of Engineering and Scientific Problems David M. Rocke Department of Applied Science Taylor s Theorem Can often approximate a function by a polynomial The error in the approximation
More informationA Mathematical Trivium
A Mathematical Trivium V.I. Arnold 1991 1. Sketch the graph of the derivative and the graph of the integral of a function given by a freehand graph. 2. Find the limit lim x 0 sin tan x tan sin x arcsin
More informationLECTURE NOTES ELEMENTARY NUMERICAL METHODS. Eusebius Doedel
LECTURE NOTES on ELEMENTARY NUMERICAL METHODS Eusebius Doedel TABLE OF CONTENTS Vector and Matrix Norms 1 Banach Lemma 20 The Numerical Solution of Linear Systems 25 Gauss Elimination 25 Operation Count
More informationSection 3.4. Second Order Nonhomogeneous. The corresponding homogeneous equation. is called the reduced equation of (N).
Section 3.4. Second Order Nonhomogeneous Equations y + p(x)y + q(x)y = f(x) (N) The corresponding homogeneous equation y + p(x)y + q(x)y = 0 (H) is called the reduced equation of (N). 1 General Results
More informationAN EFFICIENT METHOD FOR COMPUTING HIGHLY OSCILLATORY PHYSICAL OPTICS INTEGRAL
Progress In Electromagnetics Research, Vol. 17, 11 57, 1 AN EFFICIENT METHOD FOR COMPUTING HIGHLY OSCILLATORY PHYSICAL OPTICS INTEGRAL Y. M. Wu 1, L. J. Jiang 1, and W. C. Chew, * 1 Department of Electric
More informationTMA4120, Matematikk 4K, Fall Date Section Topic HW Textbook problems Suppl. Answers. Sept 12 Aug 31/
TMA420, Matematikk 4K, Fall 206 LECTURE SCHEDULE AND ASSIGNMENTS Date Section Topic HW Textbook problems Suppl Answers Aug 22 6 Laplace transform 6:,7,2,2,22,23,25,26,4 A Sept 5 Aug 24/25 62-3 ODE, Heaviside
More informationA Survey of Computational High Frequency Wave Propagation II. Olof Runborg NADA, KTH
A Survey of Computational High Frequency Wave Propagation II Olof Runborg NADA, KTH High Frequency Wave Propagation CSCAMM, September 19-22, 2005 Numerical methods Direct methods Wave equation (time domain)
More informationAnton ARNOLD. with N. Ben Abdallah (Toulouse), J. Geier (Vienna), C. Negulescu (Marseille) TU Vienna Institute for Analysis and Scientific Computing
TECHNISCHE UNIVERSITÄT WIEN Asymptotically correct finite difference schemes for highly oscillatory ODEs Anton ARNOLD with N. Ben Abdallah (Toulouse, J. Geier (Vienna, C. Negulescu (Marseille TU Vienna
More informationn 1 f n 1 c 1 n+1 = c 1 n $ c 1 n 1. After taking logs, this becomes
Root finding: 1 a The points {x n+1, }, {x n, f n }, {x n 1, f n 1 } should be co-linear Say they lie on the line x + y = This gives the relations x n+1 + = x n +f n = x n 1 +f n 1 = Eliminating α and
More informationEngg. Math. II (Unit-IV) Numerical Analysis
Dr. Satish Shukla of 33 Engg. Math. II (Unit-IV) Numerical Analysis Syllabus. Interpolation and Curve Fitting: Introduction to Interpolation; Calculus of Finite Differences; Finite Difference and Divided
More informationCHAPTER 4 NUMERICAL INTEGRATION METHODS TO EVALUATE TRIPLE INTEGRALS USING GENERALIZED GAUSSIAN QUADRATURE
CHAPTER 4 NUMERICAL INTEGRATION METHODS TO EVALUATE TRIPLE INTEGRALS USING GENERALIZED GAUSSIAN QUADRATURE 4.1 Introduction Many applications in science and engineering require the solution of three dimensional
More informationLECTURE 5: THE METHOD OF STATIONARY PHASE
LECTURE 5: THE METHOD OF STATIONARY PHASE Some notions.. A crash course on Fourier transform For j =,, n, j = x j. D j = i j. For any multi-index α = (α,, α n ) N n. α = α + + α n. α! = α! α n!. x α =
More information2. Determine whether the following pair of functions are linearly dependent, or linearly independent:
Topics to be covered on the exam include: Recognizing, and verifying solutions to homogeneous second-order linear differential equations, and their corresponding Initial Value Problems Recognizing and
More informationSOLUTION OF ALGEBRAIC AND TRANSCENDENTAL EQUATIONS BISECTION METHOD
BISECTION METHOD If a function f(x) is continuous between a and b, and f(a) and f(b) are of opposite signs, then there exists at least one root between a and b. It is shown graphically as, Let f a be negative
More informationdx n a 1(x) dy
HIGHER ORDER DIFFERENTIAL EQUATIONS Theory of linear equations Initial-value and boundary-value problem nth-order initial value problem is Solve: a n (x) dn y dx n + a n 1(x) dn 1 y dx n 1 +... + a 1(x)
More informationJim Lambers MAT 460/560 Fall Semester Practice Final Exam
Jim Lambers MAT 460/560 Fall Semester 2009-10 Practice Final Exam 1. Let f(x) = sin 2x + cos 2x. (a) Write down the 2nd Taylor polynomial P 2 (x) of f(x) centered around x 0 = 0. (b) Write down the corresponding
More informationNumerical Analysis Preliminary Exam 10 am to 1 pm, August 20, 2018
Numerical Analysis Preliminary Exam 1 am to 1 pm, August 2, 218 Instructions. You have three hours to complete this exam. Submit solutions to four (and no more) of the following six problems. Please start
More information(0, 0), (1, ), (2, ), (3, ), (4, ), (5, ), (6, ).
1 Interpolation: The method of constructing new data points within the range of a finite set of known data points That is if (x i, y i ), i = 1, N are known, with y i the dependent variable and x i [x
More information1. Nonlinear Equations. This lecture note excerpted parts from Michael Heath and Max Gunzburger. f(x) = 0
Numerical Analysis 1 1. Nonlinear Equations This lecture note excerpted parts from Michael Heath and Max Gunzburger. Given function f, we seek value x for which where f : D R n R n is nonlinear. f(x) =
More informationBessel s Equation. MATH 365 Ordinary Differential Equations. J. Robert Buchanan. Fall Department of Mathematics
Bessel s Equation MATH 365 Ordinary Differential Equations J. Robert Buchanan Department of Mathematics Fall 2018 Background Bessel s equation of order ν has the form where ν is a constant. x 2 y + xy
More informationFourier transforms. c n e inπx. f (x) = Write same thing in an equivalent form, using n = 1, f (x) = l π
Fourier transforms We can imagine our periodic function having periodicity taken to the limits ± In this case, the function f (x) is not necessarily periodic, but we can still use Fourier transforms (related
More informationRoot Finding (and Optimisation)
Root Finding (and Optimisation) M.Sc. in Mathematical Modelling & Scientific Computing, Practical Numerical Analysis Michaelmas Term 2018, Lecture 4 Root Finding The idea of root finding is simple we want
More informationx x2 2 + x3 3 x4 3. Use the divided-difference method to find a polynomial of least degree that fits the values shown: (b)
Numerical Methods - PROBLEMS. The Taylor series, about the origin, for log( + x) is x x2 2 + x3 3 x4 4 + Find an upper bound on the magnitude of the truncation error on the interval x.5 when log( + x)
More informationOn the Computation of Confluent Hypergeometric Functions for Large Imaginary Part of Parameters b and z
On the Computation of Confluent Hypergeometric Functions for Large Imaginary Part of Parameters b and z Guillermo Navas-Palencia 1 and Argimiro Arratia 2 1 Numerical Algorithms Group Ltd, UK, and Dept.
More informationLecture 10 Polynomial interpolation
Lecture 10 Polynomial interpolation Weinan E 1,2 and Tiejun Li 2 1 Department of Mathematics, Princeton University, weinan@princeton.edu 2 School of Mathematical Sciences, Peking University, tieli@pku.edu.cn
More informationSection 3.4. Second Order Nonhomogeneous. The corresponding homogeneous equation. is called the reduced equation of (N).
Section 3.4. Second Order Nonhomogeneous Equations y + p(x)y + q(x)y = f(x) (N) The corresponding homogeneous equation y + p(x)y + q(x)y = 0 (H) is called the reduced equation of (N). 1 General Results
More informationThe kissing polynomials and their Hankel determinants
The kissing polynomials and their Hankel determinants Alfredo Deaño, Daan Huybrechs and Arieh Iserles April 28, 2015 Abstract We study a family of polynomials that are orthogonal with respect to the weight
More informationScientific Computing: Numerical Integration
Scientific Computing: Numerical Integration Aleksandar Donev Courant Institute, NYU 1 donev@courant.nyu.edu 1 Course MATH-GA.2043 or CSCI-GA.2112, Fall 2015 Nov 5th, 2015 A. Donev (Courant Institute) Lecture
More informationInverse problems Total Variation Regularization Mark van Kraaij Casa seminar 23 May 2007 Technische Universiteit Eindh ove n University of Technology
Inverse problems Total Variation Regularization Mark van Kraaij Casa seminar 23 May 27 Introduction Fredholm first kind integral equation of convolution type in one space dimension: g(x) = 1 k(x x )f(x
More informationMATH 220: MIDTERM OCTOBER 29, 2015
MATH 22: MIDTERM OCTOBER 29, 25 This is a closed book, closed notes, no electronic devices exam. There are 5 problems. Solve Problems -3 and one of Problems 4 and 5. Write your solutions to problems and
More informationNumerical Analysis Preliminary Exam 10.00am 1.00pm, January 19, 2018
Numerical Analysis Preliminary Exam 0.00am.00pm, January 9, 208 Instructions. You have three hours to complete this exam. Submit solutions to four (and no more) of the following six problems. Please start
More informationPARTIAL DIFFERENTIAL EQUATIONS MIDTERM
PARTIAL DIFFERENTIAL EQUATIONS MIDTERM ERIN PEARSE. For b =,,..., ), find the explicit fundamental solution to the heat equation u + b u u t = 0 in R n 0, ). ) Letting G be what you find, show u 0 x) =
More informationSuperposition of electromagnetic waves
Superposition of electromagnetic waves February 9, So far we have looked at properties of monochromatic plane waves. A more complete picture is found by looking at superpositions of many frequencies. Many
More informationScientific Computing: An Introductory Survey
Scientific Computing: An Introductory Survey Chapter 7 Interpolation Prof. Michael T. Heath Department of Computer Science University of Illinois at Urbana-Champaign Copyright c 2002. Reproduction permitted
More informationCUBATURE FORMULA FOR APPROXIMATE CALCULATION OF INTEGRALS OF TWO-DIMENSIONAL IRREGULAR HIGHLY OSCILLATING FUNCTIONS
U.P.B. Sci. Bull., Series A, Vol. 80, Iss. 3, 08 ISSN 3-707 CUBATURE FORMULA FOR APPROXIMATE CALCULATION OF INTEGRALS OF TWO-DIMENSIONAL IRREGULAR HIGHLY OSCILLATING FUNCTIONS Vitaliy MEZHUYEV, Oleg M.
More informationROOT FINDING REVIEW MICHELLE FENG
ROOT FINDING REVIEW MICHELLE FENG 1.1. Bisection Method. 1. Root Finding Methods (1) Very naive approach based on the Intermediate Value Theorem (2) You need to be looking in an interval with only one
More informationApplied Numerical Analysis (AE2220-I) R. Klees and R.P. Dwight
Applied Numerical Analysis (AE0-I) R. Klees and R.P. Dwight February 018 Contents 1 Preliminaries: Motivation, Computer arithmetic, Taylor series 1 1.1 Numerical Analysis Motivation..........................
More information2003 Mathematics. Advanced Higher. Finalised Marking Instructions
2003 Mathematics Advanced Higher Finalised Marking Instructions 2003 Mathematics Advanced Higher Section A Finalised Marking Instructions Advanced Higher 2003: Section A Solutions and marks A. (a) Given
More informationFurther Mathematics SAMPLE. Marking Scheme
Further Mathematics SAMPLE Marking Scheme This marking scheme has been prepared as a guide only to markers. This is not a set of model answers, or the exclusive answers to the questions, and there will
More informationOn the convergence rate of a modified Fourier series [PREPRINT]
On the convergence rate of a modified Fourier series [PREPRINT Sheehan Olver St John s College St Giles Oxford OX1 3JP United Kingdom Sheehan.Olver@sjc.ox.ac.uk 18 April 008 Abstract The rate of convergence
More informationc) xy 3 = cos(7x +5y), y 0 = y3 + 7 sin(7x +5y) 3xy sin(7x +5y) d) xe y = sin(xy), y 0 = ey + y cos(xy) x(e y cos(xy)) e) y = x ln(3x + 5), y 0
Some Math 35 review problems With answers 2/6/2005 The following problems are based heavily on problems written by Professor Stephen Greenfield for his Math 35 class in spring 2005. His willingness to
More informationMath 4263 Homework Set 1
Homework Set 1 1. Solve the following PDE/BVP 2. Solve the following PDE/BVP 2u t + 3u x = 0 u (x, 0) = sin (x) u x + e x u y = 0 u (0, y) = y 2 3. (a) Find the curves γ : t (x (t), y (t)) such that that
More informationCS 450 Numerical Analysis. Chapter 8: Numerical Integration and Differentiation
Lecture slides based on the textbook Scientific Computing: An Introductory Survey by Michael T. Heath, copyright c 2018 by the Society for Industrial and Applied Mathematics. http://www.siam.org/books/cl80
More informationFormulas to remember
Complex numbers Let z = x + iy be a complex number The conjugate z = x iy Formulas to remember The real part Re(z) = x = z+z The imaginary part Im(z) = y = z z i The norm z = zz = x + y The reciprocal
More informationLectures 9-10: Polynomial and piecewise polynomial interpolation
Lectures 9-1: Polynomial and piecewise polynomial interpolation Let f be a function, which is only known at the nodes x 1, x,, x n, ie, all we know about the function f are its values y j = f(x j ), j
More informationChapter 2 Interpolation
Chapter 2 Interpolation Experiments usually produce a discrete set of data points (x i, f i ) which represent the value of a function f (x) for a finite set of arguments {x 0...x n }. If additional data
More information3 2 6 Solve the initial value problem u ( t) 3. a- If A has eigenvalues λ =, λ = 1 and corresponding eigenvectors 1
Math Problem a- If A has eigenvalues λ =, λ = 1 and corresponding eigenvectors 1 3 6 Solve the initial value problem u ( t) = Au( t) with u (0) =. 3 1 u 1 =, u 1 3 = b- True or false and why 1. if A is
More informationDepartment of Applied Mathematics and Theoretical Physics. AMA 204 Numerical analysis. Exam Winter 2004
Department of Applied Mathematics and Theoretical Physics AMA 204 Numerical analysis Exam Winter 2004 The best six answers will be credited All questions carry equal marks Answer all parts of each question
More informationLECTURE-13 : GENERALIZED CAUCHY S THEOREM
LECTURE-3 : GENERALIZED CAUCHY S THEOREM VED V. DATAR The aim of this lecture to prove a general form of Cauchy s theorem applicable to multiply connected domains. We end with computations of some real
More informationMath 3313: Differential Equations Second-order ordinary differential equations
Math 3313: Differential Equations Second-order ordinary differential equations Thomas W. Carr Department of Mathematics Southern Methodist University Dallas, TX Outline Mass-spring & Newton s 2nd law Properties
More informationBessel Functions Michael Taylor. Lecture Notes for Math 524
Bessel Functions Michael Taylor Lecture Notes for Math 54 Contents 1. Introduction. Conversion to first order systems 3. The Bessel functions J ν 4. The Bessel functions Y ν 5. Relations between J ν and
More informationMassachusetts Institute of Technology Instrumentation Laboratory Cambridge, Massachusetts
Massachusetts Institute of Technology Instrumentation Laboratory Cambridge, Massachusetts Space Guidance Analysis Memo #9 To: From: SGA Distribution James E. Potter Date: November 0, 96 Subject: Error
More informationComputational High Frequency Wave Propagation
Computational High Frequency Wave Propagation Olof Runborg CSC, KTH Isaac Newton Institute Cambridge, February 2007 Olof Runborg (KTH) High-Frequency Waves INI, 2007 1 / 52 Outline 1 Introduction, background
More informationEigenvalues and eigenfunctions of the Laplacian. Andrew Hassell
Eigenvalues and eigenfunctions of the Laplacian Andrew Hassell 1 2 The setting In this talk I will consider the Laplace operator,, on various geometric spaces M. Here, M will be either a bounded Euclidean
More informationMultiple integrals: Sufficient conditions for a local minimum, Jacobi and Weierstrass-type conditions
Multiple integrals: Sufficient conditions for a local minimum, Jacobi and Weierstrass-type conditions March 6, 2013 Contents 1 Wea second variation 2 1.1 Formulas for variation........................
More informationEAD 115. Numerical Solution of Engineering and Scientific Problems. David M. Rocke Department of Applied Science
EAD 115 Numerical Solution of Engineering and Scientific Problems David M. Rocke Department of Applied Science Multidimensional Unconstrained Optimization Suppose we have a function f() of more than one
More informationNumerical Methods in Physics and Astrophysics
Kostas Kokkotas 2 October 20, 2014 2 http://www.tat.physik.uni-tuebingen.de/ kokkotas Kostas Kokkotas 3 TOPICS 1. Solving nonlinear equations 2. Solving linear systems of equations 3. Interpolation, approximation
More informationEcient computation of highly oscillatory integrals by using QTT tensor approximation
Ecient computation of highly oscillatory integrals by using QTT tensor approximation Boris Khoromskij Alexander Veit Abstract We propose a new method for the ecient approximation of a class of highly oscillatory
More informationNumerical Methods for Differential Equations Mathematical and Computational Tools
Numerical Methods for Differential Equations Mathematical and Computational Tools Gustaf Söderlind Numerical Analysis, Lund University Contents V4.16 Part 1. Vector norms, matrix norms and logarithmic
More informationNotes 14 The Steepest-Descent Method
ECE 638 Fall 17 David R. Jackson Notes 14 The Steepest-Descent Method Notes are adapted from ECE 6341 1 Steepest-Descent Method Complex Integral: Ω g z I f z e dz Ω = C The method was published by Peter
More informationLesson 3: Linear differential equations of the first order Solve each of the following differential equations by two methods.
Lesson 3: Linear differential equations of the first der Solve each of the following differential equations by two methods. Exercise 3.1. Solution. Method 1. It is clear that y + y = 3 e dx = e x is an
More informationSET-A U/2015/16/II/A
. n x dx 3 n n is equal to : 0 4. Let f n (x) = x n n, for each n. Then the series (A) 3 n n (B) n 3 n (n ) (C) n n (n ) (D) n 3 n (n ) f n is n (A) Uniformly convergent on [0, ] (B) Po intwise convergent
More informationAnalytical Mechanics: Elastic Deformation
Analytical Mechanics: Elastic Deformation Shinichi Hirai Dept. Robotics, Ritsumeikan Univ. Shinichi Hirai (Dept. Robotics, Ritsumeikan Univ.) Analytical Mechanics: Elastic Deformation 1 / 60 Agenda Agenda
More informationNumerical Optimization
Numerical Optimization Unit 2: Multivariable optimization problems Che-Rung Lee Scribe: February 28, 2011 (UNIT 2) Numerical Optimization February 28, 2011 1 / 17 Partial derivative of a two variable function
More informationIn this chapter we study elliptical PDEs. That is, PDEs of the form. 2 u = lots,
Chapter 8 Elliptic PDEs In this chapter we study elliptical PDEs. That is, PDEs of the form 2 u = lots, where lots means lower-order terms (u x, u y,..., u, f). Here are some ways to think about the physical
More informationx 3y 2z = 6 1.2) 2x 4y 3z = 8 3x + 6y + 8z = 5 x + 3y 2z + 5t = 4 1.5) 2x + 8y z + 9t = 9 3x + 5y 12z + 17t = 7
Linear Algebra and its Applications-Lab 1 1) Use Gaussian elimination to solve the following systems x 1 + x 2 2x 3 + 4x 4 = 5 1.1) 2x 1 + 2x 2 3x 3 + x 4 = 3 3x 1 + 3x 2 4x 3 2x 4 = 1 x + y + 2z = 4 1.4)
More informationEach is equal to CP 1 minus one point, which is the origin of the other: (C =) U 1 = CP 1 the line λ (1, 0) U 0
Algebraic Curves/Fall 2015 Aaron Bertram 1. Introduction. What is a complex curve? (Geometry) It s a Riemann surface, that is, a compact oriented twodimensional real manifold Σ with a complex structure.
More informationFast Numerical Methods for Stochastic Computations
Fast AreviewbyDongbinXiu May 16 th,2013 Outline Motivation 1 Motivation 2 3 4 5 Example: Burgers Equation Let us consider the Burger s equation: u t + uu x = νu xx, x [ 1, 1] u( 1) =1 u(1) = 1 Example:
More informationEdexcel GCE Further Pure Mathematics (FP1) Required Knowledge Information Sheet. Daniel Hammocks
Edexcel GCE Further Pure Mathematics (FP1) Required Knowledge Information Sheet FP1 Formulae Given in Mathematical Formulae and Statistical Tables Booklet Summations o =1 2 = 1 + 12 + 1 6 o =1 3 = 1 64
More informationCOMPLEX VARIABLES. Principles and Problem Sessions YJ? A K KAPOOR. University of Hyderabad, India. World Scientific NEW JERSEY LONDON
COMPLEX VARIABLES Principles and Problem Sessions A K KAPOOR University of Hyderabad, India NEW JERSEY LONDON YJ? World Scientific SINGAPORE BEIJING SHANGHAI HONG KONG TAIPEI CHENNAI CONTENTS Preface vii
More informationResearch Statement. James Bremer Department of Mathematics, University of California, Davis
Research Statement James Bremer Department of Mathematics, University of California, Davis Email: bremer@math.ucdavis.edu Webpage: https.math.ucdavis.edu/ bremer I work in the field of numerical analysis,
More information1. Method 1: bisection. The bisection methods starts from two points a 0 and b 0 such that
Chapter 4 Nonlinear equations 4.1 Root finding Consider the problem of solving any nonlinear relation g(x) = h(x) in the real variable x. We rephrase this problem as one of finding the zero (root) of a
More informationUNIVERSITY OF SOUTHAMPTON. A foreign language dictionary (paper version) is permitted provided it contains no notes, additions or annotations.
UNIVERSITY OF SOUTHAMPTON MATH055W SEMESTER EXAMINATION 03/4 MATHEMATICS FOR ELECTRONIC & ELECTRICAL ENGINEERING Duration: 0 min Solutions Only University approved calculators may be used. A foreign language
More informationFinal Year M.Sc., Degree Examinations
QP CODE 569 Page No Final Year MSc, Degree Examinations September / October 5 (Directorate of Distance Education) MATHEMATICS Paper PM 5: DPB 5: COMPLEX ANALYSIS Time: 3hrs] [Max Marks: 7/8 Instructions
More informationDifferential Equations
Differential Equations A differential equation (DE) is an equation which involves an unknown function f (x) as well as some of its derivatives. To solve a differential equation means to find the unknown
More informationDiscrete Projection Methods for Integral Equations
SUB Gttttingen 7 208 427 244 98 A 5141 Discrete Projection Methods for Integral Equations M.A. Golberg & C.S. Chen TM Computational Mechanics Publications Southampton UK and Boston USA Contents Sources
More informationy = x 3 and y = 2x 2 x. 2x 2 x = x 3 x 3 2x 2 + x = 0 x(x 2 2x + 1) = 0 x(x 1) 2 = 0 x = 0 and x = (x 3 (2x 2 x)) dx
Millersville University Name Answer Key Mathematics Department MATH 2, Calculus II, Final Examination May 4, 2, 8:AM-:AM Please answer the following questions. Your answers will be evaluated on their correctness,
More information