Starting algorithms for partitioned Runge-Kutta methods: the pair Lobatto IIIA-IIIB
|
|
|
- Jodie Summers
- 10 years ago
- Views:
Transcription
1 Monografías de la Real Academia de Ciencias de Zaragoza. : 49 8, 4). Starting algorithms for partitioned Runge-Kutta methods: the pair Lobatto IIIA-IIIB I. Higueras and T. Roldán Departamento de Matemática e Informática. Universidad Pública de Navarra. 3 Pamplona, Spain. Abstract Partitioned Runge-Kutta PRK) methods are suitable for the numerical integration of separable Hamiltonian systems. For implicit PRK methods, the computational effort may be dominated by the cost of solving the non-linear systems. For these non-linear systems, good starting values minimize both the risk of a failed iteration and the effort required for convergence. In this paper we consider the Lobatto IIIA-IIIB pair. Applying the theory developed by the authors for PRK methods, we obtain optimum predictors for this pair. Numerical experiments show the efficiency of these predictors. Key words and expressions: Starting algorithms, stage value predictors, partitioned Runge-Kutta methods, Hamiltonian systems. MSC: L, L, L8, H. Introduction We consider partitioned differential equations of the form y = fy, z) yx ) = y, z = gy, z) zx ) = z, ) where f : IR l IR m IR l and g : IR l IR m IR m are sufficiently smooth functions. An example of this type of systems are Hamiltonian systems p = H q, ) q = 49 H p, 3)
2 where the Hamiltonian function H has the special structure Hp, q) = Tp) + V q). In this case, the Hamiltonian system )-3) is of the form p q = fq) = gp) with f = q V, g = p T. For the numerical integration of ) we consider partitioned Runge-Kutta PRK) methods. The idea of the PRK methods is to take two different Runge-Kutta RK) methods, A, b, Â,ˆb), and to treat the variables y with the first method, A, b), and the z variables with the second one, Â,ˆb). In this way, the numerical solution from t n, y n, z n ) to t n+, y n+, z n+ ) with the s-stage PRK method A, b, Â,ˆb) is given by y n+ = y n + h b t I l )fy n+, Z n+ ), z n+ = z n + h ˆb t I m )gy n+, Z n+ ), where the internal stages vectors Y n+, Z n+ are obtained from the system Y n+ = e y n + h A I l )fy n+, Z n+ ), 4) Z n+ = e z n + h  I m)gy n+, Z n+ ). ) As usual, we have denoted by the Kronecker product and by e =,..., ) t. For implicit PRK methods, the computational effort may be dominated by the cost of solving the non-linear systems 4)-). These non-linear systems are solved with an iterative method that requires starting values Y ) n+, Z ) n+, and it is well known that these values must be as accurate as possible, because in other case, the number of iterations in each step may be too high or even worse, the convergence may fail. A common way to proceed is to take as starting values the last computed numerical solution, i.e., Y ) n+ = e y n, Z ) n+ = e z n. We will refer to them as the trivial predictors. However it is also possible to involve the last computed internal stages, Y n, Z n, and the approximation y n, and consider starting algorithms of the form Y ) n+ = b y n + B I l )Y n, ) Z ) n+ = c z n + C I m )Z n, 7) where b, c IR s, B and C are s s matrices which have to be determined. Several kinds of starting algorithms for RK methods, also called stage value predictors, have been studied by different authors [], [], [4], [7], [4], [3], [], [], [], []). The
3 starting algorithms considered for example in [], [] or [7] are of the form )-7), whereas the ones considered for example in [], [4] or [] also involve function evaluations of the internal stages. Although in this paper we only deal with Hamiltonian systems without restrictions ODEs), we are also interested in Hamiltonian systems with restrictions DAEs). For ODE systems we can use any of the stage value predictors mentioned in the previous paragraph, but this is not the case for DAEs. Predictors involving function evaluations cannot be used for this type of systems [], [8]). This is the reason why we use predictors of the form )-7). These predictors have already been studied for PRK methods in [9]. In this paper we consider the PRK methods Lobatto IIIA-IIIB, that have been proved to be suitable ones for Hamiltonian systems [], [8]) and construct good starting algorithms for them. This is done in Section. The numerical experiments done in Section 3, show the efficiency of these starting algorithms. Starting Algorithms for Lobatto IIIA-IIIB Methods We consider the PRK methods A, b, Â,ˆb), where A, b) is the Lobatto IIIA method and Â,ˆb) is the Lobatto IIB method. We will refer to this PRK method as the Lobatto IIIA-IIB method. Observe that these methods share the same nodes, that is to say c = ĉ. Furthermore, the s-stage Lobatto IIIA method satisfies Cs), and thus Ac q = c q /q, q =,..., s; whereas the s-stage Lobatto IIIB method satisfies Cs ), and therefore Âĉ q = ĉ q /q, q =,..., s Figure : The pair Lobatto IIIA-IIIB for s = Figure : The pair Lobatto IIIA-IIIB for s = 4
4 In [9] the general order conditions for predictors of the form )-7) have been obtained. Fortunately, these conditions are reduced considerably if the method PRK satisfies some simplifying assumptions, as it occurs with Lobatto IIIA-IIIB methods. For s 3, the s-stage Lobatto IIIA and IIIB methods satisfy at least C3) and C) respectively. In this case, the general order conditions in [9] for the variable y are reduced to the following ones Consistency: b + B e = e, Order : B c = e + r c, Order : B A c = eb t c + rae + rc), 8) Order 3: B Ac = eb t c + rae + rc), B AÂc = ebt Âc + raeb t c + r AÂe + rc), whereas the order conditions for the variable z are reduced to Consistency: c + C e = e, Order : C c = e + r c, Order : C  c = eˆb t c + râe + rc), 9) Order 3: C Âc = eˆb t c + râe + rc), C  c = eˆb t Âc + râeˆb t c + r  e + rc). The parameter r = h n+ /h n, where h n+ denotes the current stepsize, h n+ = t n+ t n, and h n denotes the last stepsize, h n = t n t n. For s = 3, if we take into account the number of parameters available and the number of order conditions, we get that the maximum order achieved is for both variables, y and z. It is still possible to impose one of the two conditions of order 3, but not both of them. For s 4, the s-stage Lobatto IIIA and IIIB methods satisfy at least C4) and C) respectively. Both of them satisfy b t c = ˆb t c = /. In this case, in 8) and 9) the two conditions of order 3 are equivalent. Hence for s = 4, it is possible to achieve order 3 for both variables, y and z. In order to simplify the implementation, it is interesting to initialize the variables y and z with the same coefficients. This situation corresponds to the case b = c and B = C in )-7).
5 In that case, for s 3, up to order, the joint order conditions 8) and 9) are reduced to the following ones Consistency: b + B e = e, Order : B c = e + r c, Order : B A c = eb t c + rae + rc), B Â c = eˆb t c + râe + rc). Now it is easy to conclude that there exists a unique joint predictor of order for the case s = 3. This predictor is given by the following coefficients r r b = r3 + r), B = r + 3r) r + r) + 3r + r ). ) r + r) r + 3r) 4r + r) + 3r + r For s 4, up to order 3, the joint order conditions 8) and 9) are reduced to the following ones Consistency: b + B e = e, Order : B c = e + r c, Order : B A c = eb t c + rae + rc), Order 3: B Ac = eb t c + rae + rc), B Âc = eˆb t c + râe + rc). We immediately obtain that for s = 4 there exists a unique joint predictor of order 3. This predictor is given by the following coefficients r 3 ) + + r ) ) r r 3 b = + ) r ) ), ) r r 3 r 3 r 9 r 3 + r 3 b b b 3 b 4 B =, ) b 3 b 3 b 33 b 34 b 4 b 4 b 43 b 44 3
6 where b = + ) r 3 + ) r ) r 3, ) ) + 3 r 9 + r b = ) r 3, b 3 = r + 3 ) r + ) r 3, b 4 = ) r + 3 ) r + ) r 3, b 3 = + + ) r ) r b 3 = r ) r + + ) r 3, ) ) 3 r 9 r b 33 = + b 34 = ) r ) r ) r 3, + ) r 3, + ) r 3, b 4 = + r + r + 4r 3, b 4 = + ) r b 43 = + ) r b 44 = + r + r + r ) r + 3 ) r + r 3, r 3, For Lobatto IIIA, the first stage of the variable y is explicit. As Y n, = y n and Y n,s = y n, in ) and )-) we get Y ) = y n, and therefore, as expected, the first stage of the variable y is not initialized. Observe that this is not the case for Lobatto IIIB method. In this case the first stage is implicit and the coefficients in ) and )-) are necessary to obtain a predictor. 3 Numerical Experiments In this section we test the predictor constructed in the previous section for the 3-stage pair Lobatto IIIA-IIIB. We have selected an academic problem and the well known restricted three-body problem. In both cases we have proceeded in the same way. The stopping criteria for the Newton iterations is Y k) Y k) TOL. 4
7 The problems have been integrated for different values of T OL and for different stepsizes. In Tables - we show the average number of iterations per step needed when the trivial predictor is used, and the ones when the optimum predictor given by ) is used. For each value of TOL and h, the values on the left correspond to the trivial predictor, whereas the values on the right correspond to the optimum predictor given by ). 3. Problem The problem is given by y = 4z + t) + t y) = 3) z = y t z+t) z) = The exact solution is yt) = sint + t, zt) = cost t. We have integrated this problem for t [, ]. h TOL.E-3.E-.E-7.E-./../.9 3./..E-3./../../..E-3./../../..E-3.898/../../.9 Table : Iterations per step. Problem It can be seen that except in one case, the number of iterations is lower with the predictor constructed in this paper. The exception is for h =.E-3 and TOL =.E-7 where the number of iterations coincides. We remark that in this case, the numerical solution obtained with the optimum predictor is more accurate. 3. The restricted three-body problem We have considered the restricted three-body problem from [3] x = v x x) = x y = v y y) = y z = v z z) = z v x = v y + x [ ] x+µ µ x µ + µ r 3 v r 3 x ) = v x v y = v x + y [ ] µ + µ r 3 r y 3 vy ) = v y v z = [ ] µ + µ r 3 r z 3 vz ) = v z 4)
8 where µ = µ, and r = x + µ ) + y + z r = x µ ) + y + z. We have integrated this problem for t [, ] with different values of the parameter µ and different initial conditions. Case I We take µ =.8 and the initial values.4,,,,, ). The numerical results are shown in Table. h TOL.E-3.E-.E-7.E-./.84.4/.3 3.9/.43.E-3.8/.3.3/.8.874/.87.E-3./..3/.49./..E-3.93/../..77/.938 Table : Iterations per step: µ =.8 Case II We take µ =.9 and the initial values.4,,,,.99,.). The numerical results are shown in Table 3. h TOL.E-3.E-.E-7.E-./..94/.4.4/.74.E-3./.3.49/.3.9/.3.E-3.4/../..9/..E-3.9/../.3.4/.37 Table 3: Iterations per step: µ =.9 Case III We take µ =.9994 and the initial values.74,,,,.43, ). The numerical results are shown in Table 4. h TOL.E-.E-7.E-9.E-./../../..E-3./../../..E-3./../../..E-3./../../. Table 4: Iterations per step: µ =.9994
9 In Tables -4 we again see that the number of iterations needed when we use the predictors constructed in this paper is lower than the ones needed when the trivial predictor is used. We remark that the reduction in the number of iterations when solving the non-linear systems implies a reduction in the total CPU time needed for the numerical integration, and hence it may be essential when long time integrations are done. Acknowledgments This research has been supported by the Ministerio de Ciencia y Tecnología, Project BFM-88. References [] Calvo, M., Laburta, P. and Montijano, J. I., 3: Starting algorithms for Gauss Runge-Kutta methods for Hamiltonian systems, Computers and Mathematics with Applications 4, 4 4. [] Calvo, M., Laburta, P. and Montijano, J. I., 3: Two-step high order starting values for implicit Runge-Kutta methods, Advances in Computational Mathematics 9, 4 4. [3] Calvo, M. P., : High order starting iterates for implicit Runge-Kutta methods. An improvement for variable-step symplectic integrators, IMA Journal of Numerical Analysis, 3. [4] González Pinto, S., Montijano, J. I. and Rodríguez, S. P., : On the starting algorithms for fully implicit Runge-Kutta methods, BIT 4, [] González Pinto, S., Montijano, J. I. and Rodríguez, S. P., 3: Stabilized starting algorithms for collocation Runge-Kutta methods, Computers and Mathematics with Applications 4, [] González Pinto, S., Montijano, J. I. and Rodríguez, S. P., 3: Variable order starting algorithms for implicit Runge-Kutta methods on stiff problems, Applied Numerical Mathematics 44, [7] Higueras, I. and Roldán, T., : Starting algorithms for some DIRK methods, Numerical Algorithms 3, [8] Higueras, I. and Roldán, T., 3: IRK methods for index and 3 DAEs: Starting algorithms, BIT 43,
10 [9] Higueras, I. and Roldán, T., 3: Stage value predictors for additive Runge Kutta methods, submitted for publication. [] Jay, L., 99: Symplectic partitioned Runge-Kutta methods for constrained Hamiltonian systems, SIAM Journal on Numerical Analysis 33, [] Jay, L., 998: Structure preservation for constrained dynamics with super partitioned additive Runge-Kutta methods, SIAM Journal on Scientific Computing, [] Laburta, M., 997: Starting algorithms for IRK methods, Journal of Computational and Applied Mathematics 83, [3] Murray, C. D. and Dermott S. F., 999: Solar System Dynamics, Cambridge University Press, Cambridge. [4] Olsson, H. and Söderlind, G., 999: Stage value predictors and efficient Newton iterations in implicit Runge-Kutta methods, SIAM Journal on Scientific Computing, 8. [] Roldán, T. and Higueras, I., 999: IRK methods for DAE: Starting algorithms, Journal of Computational and Applied Mathematics, [] Sand, J., 99, Methods for starting iterations schemes for implicit Runge-Kutta formulae, Institute of Mathematical Applied Conference Series, New Series vol. 39, Oxford University Press, New York,. [7] Sanz Serna, J. and Calvo, M., 994: Numerical Hamiltonian Problems, Chapman & Hall, London. [8] Sun, G., 993: Symplectic partitioned Runge Kutta methods, Journal of Computational Mathematics,
AN INTRODUCTION TO NUMERICAL METHODS AND ANALYSIS
AN INTRODUCTION TO NUMERICAL METHODS AND ANALYSIS Revised Edition James Epperson Mathematical Reviews BICENTENNIAL 0, 1 8 0 7 z ewiley wu 2007 r71 BICENTENNIAL WILEY-INTERSCIENCE A John Wiley & Sons, Inc.,
Extra-solar massive planets with small semi-major axes?
Monografías de la Real Academia de Ciencias de Zaragoza. 25: 115 120, (2004). Extra-solar massive planets with small semi-major axes? S. Fernández, D. Giuliodori and M. A. Nicotra Observatorio Astronómico.
General Framework for an Iterative Solution of Ax b. Jacobi s Method
2.6 Iterative Solutions of Linear Systems 143 2.6 Iterative Solutions of Linear Systems Consistent linear systems in real life are solved in one of two ways: by direct calculation (using a matrix factorization,
Symmetric planar non collinear relative equilibria for the Lennard Jones potential 3 body problem with two equal masses
Monografías de la Real Academia de Ciencias de Zaragoza. 25: 93 114, (2004). Symmetric planar non collinear relative equilibria for the Lennard Jones potential 3 body problem with two equal masses M. Corbera,
University of Lille I PC first year list of exercises n 7. Review
University of Lille I PC first year list of exercises n 7 Review Exercise Solve the following systems in 4 different ways (by substitution, by the Gauss method, by inverting the matrix of coefficients
(Quasi-)Newton methods
(Quasi-)Newton methods 1 Introduction 1.1 Newton method Newton method is a method to find the zeros of a differentiable non-linear function g, x such that g(x) = 0, where g : R n R n. Given a starting
Families of symmetric periodic orbits in the three body problem and the figure eight
Monografías de la Real Academia de Ciencias de Zaragoza. 25: 229 24, (24). Families of symmetric periodic orbits in the three body problem and the figure eight F. J. Muñoz-Almaraz, J. Galán and E. Freire
LEVERRIER-FADEEV ALGORITHM AND CLASSICAL ORTHOGONAL POLYNOMIALS
This is a reprint from Revista de la Academia Colombiana de Ciencias Vol 8 (106 (004, 39-47 LEVERRIER-FADEEV ALGORITHM AND CLASSICAL ORTHOGONAL POLYNOMIALS by J Hernández, F Marcellán & C Rodríguez LEVERRIER-FADEEV
General Theory of Differential Equations Sections 2.8, 3.1-3.2, 4.1
A B I L E N E C H R I S T I A N U N I V E R S I T Y Department of Mathematics General Theory of Differential Equations Sections 2.8, 3.1-3.2, 4.1 Dr. John Ehrke Department of Mathematics Fall 2012 Questions
Numerical Analysis An Introduction
Walter Gautschi Numerical Analysis An Introduction 1997 Birkhauser Boston Basel Berlin CONTENTS PREFACE xi CHAPTER 0. PROLOGUE 1 0.1. Overview 1 0.2. Numerical analysis software 3 0.3. Textbooks and monographs
Implicit Runge-Kutta Methods for Orbit Propagation
Implicit Runge-Kutta Methods for Orbit Propagation Jeffrey M. Aristoff and Aubrey B. Poore Numerica Corporation, 4850 Hahns Peak Drive, Suite 200, Loveland, Colorado, 80538, USA Accurate and efficient
Numerical methods for American options
Lecture 9 Numerical methods for American options Lecture Notes by Andrzej Palczewski Computational Finance p. 1 American options The holder of an American option has the right to exercise it at any moment
Modern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh
Modern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh Peter Richtárik Week 3 Randomized Coordinate Descent With Arbitrary Sampling January 27, 2016 1 / 30 The Problem
SIXTY STUDY QUESTIONS TO THE COURSE NUMERISK BEHANDLING AV DIFFERENTIALEKVATIONER I
Lennart Edsberg, Nada, KTH Autumn 2008 SIXTY STUDY QUESTIONS TO THE COURSE NUMERISK BEHANDLING AV DIFFERENTIALEKVATIONER I Parameter values and functions occurring in the questions belowwill be exchanged
CS 294-73 Software Engineering for Scientific Computing. http://www.cs.berkeley.edu/~colella/cs294fall2013. Lecture 16: Particle Methods; Homework #4
CS 294-73 Software Engineering for Scientific Computing http://www.cs.berkeley.edu/~colella/cs294fall2013 Lecture 16: Particle Methods; Homework #4 Discretizing Time-Dependent Problems From here on in,
Homework 2 Solutions
Homework Solutions Igor Yanovsky Math 5B TA Section 5.3, Problem b: Use Taylor s method of order two to approximate the solution for the following initial-value problem: y = + t y, t 3, y =, with h = 0.5.
A simple application of the implicit function theorem
Boletín de la Asociación Matemática Venezolana, Vol. XIX, No. 1 (2012) 71 DIVULGACIÓN MATEMÁTICA A simple application of the implicit function theorem Germán Lozada-Cruz Abstract. In this note we show
Is there chaos in Copenhagen problem?
Monografías de la Real Academia de Ciencias de Zaragoza 30, 43 50, (2006). Is there chaos in Copenhagen problem? Roberto Barrio, Fernando Blesa and Sergio Serrano GME, Universidad de Zaragoza Abstract
Central configuration in the planar n + 1 body problem with generalized forces.
Monografías de la Real Academia de Ciencias de Zaragoza. 28: 1 8, (2006). Central configuration in the planar n + 1 body problem with generalized forces. M. Arribas, A. Elipe Grupo de Mecánica Espacial.
G.A. Pavliotis. Department of Mathematics. Imperial College London
EE1 MATHEMATICS NUMERICAL METHODS G.A. Pavliotis Department of Mathematics Imperial College London 1. Numerical solution of nonlinear equations (iterative processes). 2. Numerical evaluation of integrals.
3.2 Sources, Sinks, Saddles, and Spirals
3.2. Sources, Sinks, Saddles, and Spirals 6 3.2 Sources, Sinks, Saddles, and Spirals The pictures in this section show solutions to Ay 00 C By 0 C Cy D 0. These are linear equations with constant coefficients
A new continuous dependence result for impulsive retarded functional differential equations
CADERNOS DE MATEMÁTICA 11, 37 47 May (2010) ARTIGO NÚMERO SMA#324 A new continuous dependence result for impulsive retarded functional differential equations M. Federson * Instituto de Ciências Matemáticas
ON COMPLETELY CONTINUOUS INTEGRATION OPERATORS OF A VECTOR MEASURE. 1. Introduction
ON COMPLETELY CONTINUOUS INTEGRATION OPERATORS OF A VECTOR MEASURE J.M. CALABUIG, J. RODRÍGUEZ, AND E.A. SÁNCHEZ-PÉREZ Abstract. Let m be a vector measure taking values in a Banach space X. We prove that
DEFORMATION OF DIRAC STRUCTURES ALONG ISOTROPIC SUBBUNDLES. and MARCO ZAMBON
Vol. 65 (2010) REPORTS ON MATHEMATICAL PHYSICS No. 2 DEFORMATION OF DIRAC STRUCTURES ALONG ISOTROPIC SUBBUNDLES IVÁN CALVO Laboratorio Nacional de Fusión, Asociación EURATOM-CIEMAT, E-28040 Madrid, Spain
We shall turn our attention to solving linear systems of equations. Ax = b
59 Linear Algebra We shall turn our attention to solving linear systems of equations Ax = b where A R m n, x R n, and b R m. We already saw examples of methods that required the solution of a linear system
Roots of Polynomials
Roots of Polynomials (Com S 477/577 Notes) Yan-Bin Jia Sep 24, 2015 A direct corollary of the fundamental theorem of algebra is that p(x) can be factorized over the complex domain into a product a n (x
1 Short Introduction to Time Series
ECONOMICS 7344, Spring 202 Bent E. Sørensen January 24, 202 Short Introduction to Time Series A time series is a collection of stochastic variables x,.., x t,.., x T indexed by an integer value t. The
24. The Branch and Bound Method
24. The Branch and Bound Method It has serious practical consequences if it is known that a combinatorial problem is NP-complete. Then one can conclude according to the present state of science that no
10.2 ITERATIVE METHODS FOR SOLVING LINEAR SYSTEMS. The Jacobi Method
578 CHAPTER 1 NUMERICAL METHODS 1. ITERATIVE METHODS FOR SOLVING LINEAR SYSTEMS As a numerical technique, Gaussian elimination is rather unusual because it is direct. That is, a solution is obtained after
Practical Guide to the Simplex Method of Linear Programming
Practical Guide to the Simplex Method of Linear Programming Marcel Oliver Revised: April, 0 The basic steps of the simplex algorithm Step : Write the linear programming problem in standard form Linear
ON GALOIS REALIZATIONS OF THE 2-COVERABLE SYMMETRIC AND ALTERNATING GROUPS
ON GALOIS REALIZATIONS OF THE 2-COVERABLE SYMMETRIC AND ALTERNATING GROUPS DANIEL RABAYEV AND JACK SONN Abstract. Let f(x) be a monic polynomial in Z[x] with no rational roots but with roots in Q p for
Numerical Analysis Introduction. Student Audience. Prerequisites. Technology.
Numerical Analysis Douglas Faires, Youngstown State University, (Chair, 2012-2013) Elizabeth Yanik, Emporia State University, (Chair, 2013-2015) Graeme Fairweather, Executive Editor, Mathematical Reviews,
Lecture 7: Finding Lyapunov Functions 1
Massachusetts Institute of Technology Department of Electrical Engineering and Computer Science 6.243j (Fall 2003): DYNAMICS OF NONLINEAR SYSTEMS by A. Megretski Lecture 7: Finding Lyapunov Functions 1
GEC320 COURSE COMPACT. Four hours per week for 15 weeks (60 hours)
GEC320 COURSE COMPACT Course Course code: GEC 320 Course title: Course status: Course Duration Numerical Methods (2 units) Compulsory Four hours per week for 15 weeks (60 hours) Lecturer Data Name: Engr.
9. Particular Solutions of Non-homogeneous second order equations Undetermined Coefficients
September 29, 201 9-1 9. Particular Solutions of Non-homogeneous second order equations Undetermined Coefficients We have seen that in order to find the general solution to the second order differential
ORDINARY DIFFERENTIAL EQUATIONS
ORDINARY DIFFERENTIAL EQUATIONS GABRIEL NAGY Mathematics Department, Michigan State University, East Lansing, MI, 48824. SEPTEMBER 4, 25 Summary. This is an introduction to ordinary differential equations.
Lecture Notes on Polynomials
Lecture Notes on Polynomials Arne Jensen Department of Mathematical Sciences Aalborg University c 008 Introduction These lecture notes give a very short introduction to polynomials with real and complex
Reference: Introduction to Partial Differential Equations by G. Folland, 1995, Chap. 3.
5 Potential Theory Reference: Introduction to Partial Differential Equations by G. Folland, 995, Chap. 3. 5. Problems of Interest. In what follows, we consider Ω an open, bounded subset of R n with C 2
Heriot-Watt University. M.Sc. in Actuarial Science. Life Insurance Mathematics I. Tutorial 5
1 Heriot-Watt University M.Sc. in Actuarial Science Life Insurance Mathematics I Tutorial 5 1. Consider the illness-death model in Figure 1. A life age takes out a policy with a term of n years that pays
FOR ACCELERATION OF ITERATIVE ALGORITHMS
DISTRIBUTED MINIMAL RESIDUAL (DMR) METHOD Q/ 2% FOR ACCELERATION OF ITERATIVE ALGORITHMS Seungsoo Lee and George S. Dulikravich Pennsylvania State University University Park, Pennsylvania, U.S.A. ABSTRACT
MATH1231 Algebra, 2015 Chapter 7: Linear maps
MATH1231 Algebra, 2015 Chapter 7: Linear maps A/Prof. Daniel Chan School of Mathematics and Statistics University of New South Wales [email protected] Daniel Chan (UNSW) MATH1231 Algebra 1 / 43 Chapter
Marketing Mix Modelling and Big Data P. M Cain
1) Introduction Marketing Mix Modelling and Big Data P. M Cain Big data is generally defined in terms of the volume and variety of structured and unstructured information. Whereas structured data is stored
NOTES ON LINEAR TRANSFORMATIONS
NOTES ON LINEAR TRANSFORMATIONS Definition 1. Let V and W be vector spaces. A function T : V W is a linear transformation from V to W if the following two properties hold. i T v + v = T v + T v for all
Adaptive Online Gradient Descent
Adaptive Online Gradient Descent Peter L Bartlett Division of Computer Science Department of Statistics UC Berkeley Berkeley, CA 94709 bartlett@csberkeleyedu Elad Hazan IBM Almaden Research Center 650
PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 4: LINEAR MODELS FOR CLASSIFICATION
PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 4: LINEAR MODELS FOR CLASSIFICATION Introduction In the previous chapter, we explored a class of regression models having particularly simple analytical
Overview of Violations of the Basic Assumptions in the Classical Normal Linear Regression Model
Overview of Violations of the Basic Assumptions in the Classical Normal Linear Regression Model 1 September 004 A. Introduction and assumptions The classical normal linear regression model can be written
Numerical simulation of the manoeuvrability of an underwater vehicle
XXI Congreso de Ecuaciones Diferenciales y Aplicaciones XI Congreso de Matemática Aplicada Ciudad Real, 21-25 septiembre 29 (pp. 1 8) Numerical simulation of the manoeuvrability of an underwater vehicle
A Learning Based Method for Super-Resolution of Low Resolution Images
A Learning Based Method for Super-Resolution of Low Resolution Images Emre Ugur June 1, 2004 [email protected] Abstract The main objective of this project is the study of a learning based method
ABSTRACT FOR THE 1ST INTERNATIONAL WORKSHOP ON HIGH ORDER CFD METHODS
1 ABSTRACT FOR THE 1ST INTERNATIONAL WORKSHOP ON HIGH ORDER CFD METHODS Sreenivas Varadan a, Kentaro Hara b, Eric Johnsen a, Bram Van Leer b a. Department of Mechanical Engineering, University of Michigan,
The Matrix Elements of a 3 3 Orthogonal Matrix Revisited
Physics 116A Winter 2011 The Matrix Elements of a 3 3 Orthogonal Matrix Revisited 1. Introduction In a class handout entitled, Three-Dimensional Proper and Improper Rotation Matrices, I provided a derivation
INTERPOLATION. Interpolation is a process of finding a formula (often a polynomial) whose graph will pass through a given set of points (x, y).
INTERPOLATION Interpolation is a process of finding a formula (often a polynomial) whose graph will pass through a given set of points (x, y). As an example, consider defining and x 0 =0, x 1 = π 4, x
Advanced CFD Methods 1
Advanced CFD Methods 1 Prof. Patrick Jenny, FS 2014 Date: 15.08.14, Time: 13:00, Student: Federico Danieli Summary The exam took place in Prof. Jenny s office, with his assistant taking notes on the answers.
Continued Fractions and the Euclidean Algorithm
Continued Fractions and the Euclidean Algorithm Lecture notes prepared for MATH 326, Spring 997 Department of Mathematics and Statistics University at Albany William F Hammond Table of Contents Introduction
BINOMIAL OPTIONS PRICING MODEL. Mark Ioffe. Abstract
BINOMIAL OPTIONS PRICING MODEL Mark Ioffe Abstract Binomial option pricing model is a widespread numerical method of calculating price of American options. In terms of applied mathematics this is simple
Lecture Notes to Accompany. Scientific Computing An Introductory Survey. by Michael T. Heath. Chapter 10
Lecture Notes to Accompany Scientific Computing An Introductory Survey Second Edition by Michael T. Heath Chapter 10 Boundary Value Problems for Ordinary Differential Equations Copyright c 2001. Reproduction
Valuation of American Options
Valuation of American Options Among the seminal contributions to the mathematics of finance is the paper F. Black and M. Scholes, The pricing of options and corporate liabilities, Journal of Political
The equivalence of logistic regression and maximum entropy models
The equivalence of logistic regression and maximum entropy models John Mount September 23, 20 Abstract As our colleague so aptly demonstrated ( http://www.win-vector.com/blog/20/09/the-simplerderivation-of-logistic-regression/
Instituto de Engenharia de Sistemas e Computadores de Coimbra Institute of Systems Engineering and Computers INESC Coimbra
Instituto de Engenharia de Sistemas e Computadores de Coimbra Institute of Systems Engineering and Computers INESC Coimbra João Clímaco and Marta Pascoal A new method to detere unsupported non-doated solutions
DATA ANALYSIS II. Matrix Algorithms
DATA ANALYSIS II Matrix Algorithms Similarity Matrix Given a dataset D = {x i }, i=1,..,n consisting of n points in R d, let A denote the n n symmetric similarity matrix between the points, given as where
Index tracking UNDER TRANSACTION COSTS:
MARKE REVIEWS Index tracking UNDER RANSACION COSS: rebalancing passive portfolios by Reinhold Hafner, Ansgar Puetz and Ralf Werner, RiskLab GmbH Portfolio managers must be able to estimate transaction
Chebyshev Expansions
Chapter 3 Chebyshev Expansions The best is the cheapest. Benjamin Franklin 3.1 Introduction In Chapter, approximations were considered consisting of expansions around a specific value of the variable (finite
Exam Introduction Mathematical Finance and Insurance
Exam Introduction Mathematical Finance and Insurance Date: January 8, 2013. Duration: 3 hours. This is a closed-book exam. The exam does not use scrap cards. Simple calculators are allowed. The questions
Computing divisors and common multiples of quasi-linear ordinary differential equations
Computing divisors and common multiples of quasi-linear ordinary differential equations Dima Grigoriev CNRS, Mathématiques, Université de Lille Villeneuve d Ascq, 59655, France [email protected]
METHODOLOGICAL CONSIDERATIONS OF DRIVE SYSTEM SIMULATION, WHEN COUPLING FINITE ELEMENT MACHINE MODELS WITH THE CIRCUIT SIMULATOR MODELS OF CONVERTERS.
SEDM 24 June 16th - 18th, CPRI (Italy) METHODOLOGICL CONSIDERTIONS OF DRIVE SYSTEM SIMULTION, WHEN COUPLING FINITE ELEMENT MCHINE MODELS WITH THE CIRCUIT SIMULTOR MODELS OF CONVERTERS. Áron Szûcs BB Electrical
Row Ideals and Fibers of Morphisms
Michigan Math. J. 57 (2008) Row Ideals and Fibers of Morphisms David Eisenbud & Bernd Ulrich Affectionately dedicated to Mel Hochster, who has been an inspiration to us for many years, on the occasion
OPTIMAl PREMIUM CONTROl IN A NON-liFE INSURANCE BUSINESS
ONDERZOEKSRAPPORT NR 8904 OPTIMAl PREMIUM CONTROl IN A NON-liFE INSURANCE BUSINESS BY M. VANDEBROEK & J. DHAENE D/1989/2376/5 1 IN A OPTIMAl PREMIUM CONTROl NON-liFE INSURANCE BUSINESS By Martina Vandebroek
TESLA Report 2003-03
TESLA Report 23-3 A multigrid based 3D space-charge routine in the tracking code GPT Gisela Pöplau, Ursula van Rienen, Marieke de Loos and Bas van der Geer Institute of General Electrical Engineering,
Monotonicity Hints. Abstract
Monotonicity Hints Joseph Sill Computation and Neural Systems program California Institute of Technology email: [email protected] Yaser S. Abu-Mostafa EE and CS Deptartments California Institute of Technology
Max-Min Representation of Piecewise Linear Functions
Beiträge zur Algebra und Geometrie Contributions to Algebra and Geometry Volume 43 (2002), No. 1, 297-302. Max-Min Representation of Piecewise Linear Functions Sergei Ovchinnikov Mathematics Department,
A Novel Fourier Transform B-spline Method for Option Pricing*
A Novel Fourier Transform B-spline Method for Option Pricing* Paper available from SSRN: http://ssrn.com/abstract=2269370 Gareth G. Haslip, FIA PhD Cass Business School, City University London October
Optimal batch scheduling in DVB-S2 satellite networks
Optimal batch scheduling in DVB-S2 satellite networks G. T. Peeters B. Van Houdt and C. Blondia University of Antwerp Department of Mathematics and Computer Science Performance Analysis of Telecommunication
H.Calculating Normal Vectors
Appendix H H.Calculating Normal Vectors This appendix describes how to calculate normal vectors for surfaces. You need to define normals to use the OpenGL lighting facility, which is described in Chapter
Convex Programming Tools for Disjunctive Programs
Convex Programming Tools for Disjunctive Programs João Soares, Departamento de Matemática, Universidade de Coimbra, Portugal Abstract A Disjunctive Program (DP) is a mathematical program whose feasible
IMPROVED NETWORK PARAMETER ERROR IDENTIFICATION USING MULTIPLE MEASUREMENT SCANS
IMPROVED NETWORK PARAMETER ERROR IDENTIFICATION USING MULTIPLE MEASUREMENT SCANS Liuxi Zhang and Ali Abur Department of Electrical and Computer Engineering Northeastern University Boston, MA, USA [email protected]
Figure 2.1: Center of mass of four points.
Chapter 2 Bézier curves are named after their inventor, Dr. Pierre Bézier. Bézier was an engineer with the Renault car company and set out in the early 196 s to develop a curve formulation which would
U.C. Berkeley CS276: Cryptography Handout 0.1 Luca Trevisan January, 2009. Notes on Algebra
U.C. Berkeley CS276: Cryptography Handout 0.1 Luca Trevisan January, 2009 Notes on Algebra These notes contain as little theory as possible, and most results are stated without proof. Any introductory
COMBINED NEURAL NETWORKS FOR TIME SERIES ANALYSIS
COMBINED NEURAL NETWORKS FOR TIME SERIES ANALYSIS Iris Ginzburg and David Horn School of Physics and Astronomy Raymond and Beverly Sackler Faculty of Exact Science Tel-Aviv University Tel-A viv 96678,
Cyclotomic Extensions
Chapter 7 Cyclotomic Extensions A cyclotomic extension Q(ζ n ) of the rationals is formed by adjoining a primitive n th root of unity ζ n. In this chapter, we will find an integral basis and calculate
On the degrees of freedom of lattice electrodynamics
arxiv:hep-lat/0408005v2 3 Jan 2005 On the degrees of freedom of lattice electrodynamics Bo He and F. L. Teixeira ElectroScience Laboratory and Department of Electrical Engineering, The Ohio State University,
Solving Mass Balances using Matrix Algebra
Page: 1 Alex Doll, P.Eng, Alex G Doll Consulting Ltd. http://www.agdconsulting.ca Abstract Matrix Algebra, also known as linear algebra, is well suited to solving material balance problems encountered
COUNTING INDEPENDENT SETS IN SOME CLASSES OF (ALMOST) REGULAR GRAPHS
COUNTING INDEPENDENT SETS IN SOME CLASSES OF (ALMOST) REGULAR GRAPHS Alexander Burstein Department of Mathematics Howard University Washington, DC 259, USA [email protected] Sergey Kitaev Mathematics
FACTORING POLYNOMIALS IN THE RING OF FORMAL POWER SERIES OVER Z
FACTORING POLYNOMIALS IN THE RING OF FORMAL POWER SERIES OVER Z DANIEL BIRMAJER, JUAN B GIL, AND MICHAEL WEINER Abstract We consider polynomials with integer coefficients and discuss their factorization
A three point formula for finding roots of equations by the method of least squares
A three point formula for finding roots of equations by the method of least squares Ababu Teklemariam Tiruneh 1 ; William N. Ndlela 1 ; Stanley J. Nkambule 1 1 Lecturer, Department of Environmental Health
Numerical Methods for Engineers
Steven C. Chapra Berger Chair in Computing and Engineering Tufts University RaymondP. Canale Professor Emeritus of Civil Engineering University of Michigan Numerical Methods for Engineers With Software
