THE COMBINED SHEPARD ABEL GONCHAROV UNIVARIATE OPERATOR

Similar documents
Lecture 4: Piecewise Cubic Interpolation

Jens Siebel (University of Applied Sciences Kaiserslautern) An Interactive Introduction to Complex Numbers

90 S.S. Drgomr nd (t b)du(t) =u()(b ) u(t)dt: If we dd the bove two equltes, we get (.) u()(b ) u(t)dt = p(; t)du(t) where p(; t) := for ll ; t [; b]:

International Journal of Pure and Applied Sciences and Technology

The Number of Rows which Equal Certain Row

Two Coefficients of the Dyson Product

4. Eccentric axial loading, cross-section core

A Family of Multivariate Abel Series Distributions. of Order k

UNIVERSITY OF IOANNINA DEPARTMENT OF ECONOMICS. M.Sc. in Economics MICROECONOMIC THEORY I. Problem Set II

DCDM BUSINESS SCHOOL NUMERICAL METHODS (COS 233-8) Solutions to Assignment 3. x f(x)

LOCAL FRACTIONAL LAPLACE SERIES EXPANSION METHOD FOR DIFFUSION EQUATION ARISING IN FRACTAL HEAT TRANSFER

Online Appendix to. Mandating Behavioral Conformity in Social Groups with Conformist Members

ON SIMPSON S INEQUALITY AND APPLICATIONS. 1. Introduction The following inequality is well known in the literature as Simpson s inequality : 2 1 f (4)

6 Roots of Equations: Open Methods

Research Article On the Upper Bounds of Eigenvalues for a Class of Systems of Ordinary Differential Equations with Higher Order

Math 497C Sep 17, Curves and Surfaces Fall 2004, PSU

The Schur-Cohn Algorithm

INTRODUCTION TO COMPLEX NUMBERS

Applied Statistics Qualifier Examination

Katholieke Universiteit Leuven Department of Computer Science

Lecture 36. Finite Element Methods

Chapter Newton-Raphson Method of Solving a Nonlinear Equation

Pyramid Algorithms for Barycentric Rational Interpolation

The combined Shepard-Lidstone bivariate operator

Convexity preserving interpolation by splines of arbitrary degree

Rank One Update And the Google Matrix by Al Bernstein Signal Science, LLC

Numerical Solution of Linear Fredholm Fuzzy Integral Equations by Modified Homotopy Perturbation Method

ψ ij has the eigenvalue

Effects of polarization on the reflected wave

Principle Component Analysis

FINITE NEUTROSOPHIC COMPLEX NUMBERS. W. B. Vasantha Kandasamy Florentin Smarandache

APPENDIX A Some Linear Algebra

On quasiperfect numbers

GAUSS ELIMINATION. Consider the following system of algebraic linear equations

Problem Do any of the following determine homomorphisms from GL n (C) to GL n (C)?

COMPLEX NUMBERS AND QUADRATIC EQUATIONS

CISE 301: Numerical Methods Lecture 5, Topic 4 Least Squares, Curve Fitting

Remember: Project Proposals are due April 11.

ANALOG CIRCUIT SIMULATION BY STATE VARIABLE METHOD

Solutions to exam in SF1811 Optimization, Jan 14, 2015

Study of Trapezoidal Fuzzy Linear System of Equations S. M. Bargir 1, *, M. S. Bapat 2, J. D. Yadav 3 1

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system

INTERPOLATION(1) ELM1222 Numerical Analysis. ELM1222 Numerical Analysis Dr Muharrem Mercimek

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

7.2 Volume. A cross section is the shape we get when cutting straight through an object.

NUMERICAL MODELLING OF A CILIUM USING AN INTEGRAL EQUATION

Computation of Fifth Degree of Spline Function Model by Using C++ Programming

8. INVERSE Z-TRANSFORM

Haddow s Experiment:

First day August 1, Problems and Solutions

Least squares. Václav Hlaváč. Czech Technical University in Prague

Neryškioji dichotominių testo klausimų ir socialinių rodiklių diferencijavimo savybių klasifikacija

Affine transformations and convexity

Dennis Bricker, 2001 Dept of Industrial Engineering The University of Iowa. MDP: Taxi page 1

Lecture notes. Fundamental inequalities: techniques and applications

ESCI 342 Atmospheric Dynamics I Lesson 1 Vectors and Vector Calculus

LAPLACE TRANSFORM SOLUTION OF THE PROBLEM OF TIME-FRACTIONAL HEAT CONDUCTION IN A TWO-LAYERED SLAB

6.6 The Marquardt Algorithm

Review of linear algebra. Nuno Vasconcelos UCSD

CHAPTER 4. Vector Spaces

A Theoretical Study on the Rank of the Integral Operators for Large- Scale Electrodynamic Analysis

CHOVER-TYPE LAWS OF THE ITERATED LOGARITHM FOR WEIGHTED SUMS OF ρ -MIXING SEQUENCES

HnUf> xk) = S0Jf(xk) (k = 1,..., «; j = 0,..., m - 1).

Perron Vectors of an Irreducible Nonnegative Interval Matrix

Uniqueness of Weak Solutions to the 3D Ginzburg- Landau Model for Superconductivity

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP

Quiz: Experimental Physics Lab-I

Many-Body Calculations of the Isotope Shift

Investigation phase in case of Bragg coupling

MSC: Primary 11A15, Secondary 11A07, 11E25 Keywords: Reciprocity law; octic residue; congruence; quartic Jacobi symbol

Self-complementing permutations of k-uniform hypergraphs

Lecture 12: Discrete Laplacian

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

Chapter Newton-Raphson Method of Solving a Nonlinear Equation

G. Herglotz s treatment of acceleration waves in his lectures on Mechanik der Kontinua applied to the jump waves of Christoffel

General viscosity iterative method for a sequence of quasi-nonexpansive mappings

Formulas for the Determinant

Randić Energy and Randić Estrada Index of a Graph

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009

Statistics and Probability Letters

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION

FUNDAMENTALS ON ALGEBRA MATRICES AND DETERMINANTS

Estimation of the composition of the liquid and vapor streams exiting a flash unit with a supercritical component

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

M-LINEAR CONNECTION ON THE SECOND ORDER REONOM BUNDLE

Physics 5153 Classical Mechanics. D Alembert s Principle and The Lagrangian-1

Ballot Paths Avoiding Depth Zero Patterns

VARIATION OF CONSTANT SUM CONSTRAINT FOR INTEGER MODEL WITH NON UNIFORM VARIABLES

COMPLEX NUMBER & QUADRATIC EQUATION

M/G/1/GD/ / System. ! Pollaczek-Khinchin (PK) Equation. ! Steady-state probabilities. ! Finding L, W q, W. ! π 0 = 1 ρ

Modelli Clamfim Equazioni differenziali 7 ottobre 2013

Partially Observable Systems. 1 Partially Observable Markov Decision Process (POMDP) Formalism

Numbers Related to Bernoulli-Goss Numbers

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification

The Minimum Universal Cost Flow in an Infeasible Flow Network

Linear Regression Analysis: Terminology and Notation

Reproducing Kernel Hilbert Space for. Penalized Regression Multi-Predictors: Case in Longitudinal Data

The Riemann Integral

Perfect Competition and the Nash Bargaining Solution

A New Method for Solving Fuzzy Volterra Integro-Differential Equations

Transcription:

REVUE D ANALYSE NUMÉRIQUE ET DE THÉORIE DE L APPROXIMATION Tome 32, N o 1, 2003, pp 11 20 THE COMBINED SHEPARD ABEL GONCHAROV UNIVARIATE OPERATOR TEODORA CĂTINAŞ Abstrct We extend the Sheprd opertor by combnng t wth the Abel- Gonchrov unvrte opertor n order to ncrese the degree of exctness nd to use some specfc functonls We study ths combned opertor nd gve some of ts propertes We ntroduce the correspondng nterpolton formul nd study ts remnder term MSC 2000 65D05 Keywords Sheprd-Abel-Gonchrov opertors, nterpolton formul, degree of exctness, remnder term 1 INTRODUCTION 11 The Sheprd unvrte opertor Recll frst some results regrdng the multvrte Sheprd opertor for the unvrte cse Let f be rel vlued functon defned on X R nd x X, = 0,, N, be some dstnct ponts The unvrte Sheprd opertor s defned by (1) (Sf) (x) = A (x) f (x ), where (2) A (x) = wth µ R (see, eg, [12) brycentrc form N j=0, j N j=0, j k x x j µ x x j µ, The bss functons A my be wrtten n A (x) = x x µ n x x k µ Ths work ws supported by CNCSIS Grnt no 8/139, 2003 Bbeş-Boly Unversty, Fculty of Mthemtcs nd Computer Scence, str M Kogălncenu 1, 3400 Cluj-Npoc, Romn, e-ml: tgule@mthubbclujro

12 Teodor Cătnş 2 It s esy to check tht nd (3) A (x v ) = δ v,, v = 0,, N, A (x) = 1 The mn propertes of the opertor S re: The nterpolton property (Sf) (x ) = f (x ), = 0,, N; The degree of exctness s gex (S) = 0 The gol when extendng the opertor S by combnng wth other opertors s to ncrese the degree of exctness nd to use other sets of functonls Let Λ := {λ = 0,, N} be set of functonls nd let P be the correspondng nterpolton opertor We consder tht Λ Λ re the subsets ssocted to the functonls λ, = 0,, N We hve N Λ = Λ nd Λ Λj, exceptng the cse Λ = {λ }, = 0,, N, when Λ Λj =, for j We ssocte the nterpolton opertor P to ech subset Λ, = 0,, N The opertor S P defned by (4) (S P f) (x) = A (x) (P f) (x) s the combned opertor of S nd P (see, eg, [12) Remrk 1 As noted n [12, f P, = 0,, N, re lner opertors, then S P s lner opertor Remrk 2 [12 Let P, = 0,, N, be some rbtrry lner opertors If gex (P ) = r, = 0,, N, then gex (S P ) = r m := mn {r 0,, r N } Assume tht Λ s set of Brkhoff type functonl, e, Λ B = { λ kj λ kj f = f (j) (x k ), j I k, k = 1,, N }, where I k {0, 1,, r k }, for r k N Denote r M = mx {r 1,, r N } Remrk 3 [12 If µ > r M then λ kj (S P f) = λ kj (f), j I k, k = 0,, N, where P s the nterpolton opertor correspondng to the set Λ B

3 The combned Sheprd Abel Gonchrov unvrte opertor 13 (5) In the proof of ths result the followng reltons re used: A (v) (x k ) = 0, v I k, k = 0,, N, k, A (v) (x ) = 0, v I, v 1, A (j) (x ) = 1 12 The Abel Gonchrov unvrte opertor Let n N,, b R, < b, nd f : [, b R be functon hvng the frst n dervtves f (), = 1, 2,, n Gven the nodes x [, b, 0 n, nd the vlues f () (x ), 0 n, we consder the Abel Gonchrov nterpolton problem of fndng polynoml P n f of degree n such tht (see, eg, [8 nd [10) (6) (P n f) () (x ) = f () (x ), 0 n The determnnt of ths lner system 1 x 0 x 2 0 x n 0 0 1! 2x 1 nx n 1 1 (7) D = 0 0 2! n(n 1)x n 2 2 = 1 1! 2!, 0 0 0 s lwys nonzero nd the problem (6) hs therefore unque soluton The Abel Gonchrov nterpolton polynoml P n f cn be wrtten n the form n (P n f)(x) = g k (x)f (k) (x k ), where g k, k = 0,, n re clled Gonchrov polynomls of degree k [9, determned by the condtons { g (s) (8) k (x s) = 0, f k s, (x) = 1 g (k) k Accordng to [8, [9 nd [10, we hve: g 0 (x) = 1, g 1 (x) = x x 0, g k (x) = x = 1 k! t1 tk 1 dt 1 dt 2 dt k x 0 x 1 x k 1 [ k 1 x k g j (x) ( ) k j x k j j, k = 2,, n j=0

14 Teodor Cătnş 4 Remrk 4 [10 When ll the nodes concde, then the problem (6) s Tylor nterpolton problem nd P n f tkes the form n (x x (P n f)(x) = 0 ) k k! f (k) (x 0 ) Regrdng the degree of exctness, we obtn the followng result Theorem 1 The Abel Gonchrov opertor P n hs the degree of exctness n, e, dex(p n ) = n Proof It s esly seen tht for the test functons e (x) = x, x [, b, we hve (P n e )(x) = e (x), = 0,, n, whle (P n e n1 )(x) = g 0 (x)x n1 0 g 1 (x)(n1)x n 1 g n (x)(n1) 2x n e n1 (x) We obtn the followng result regrdng the remnder R n f of the Abel Gonchrov nterpolton formul wth Theorem 2 If f H n1 [, b then (R n f)(x) = f = P n f R n f (9) ϕ n (x, s) = (x s)n ϕ n (x, s)f (n1) (s)ds, n g k (x) (x k s) n k (n k)! Proof From Theorem 1 we hve tht gex(p n ) = n By Peno s theorem we obtn wth For ll x [, b we hve (R n f)(x) = ϕ n (, s) = R n [ ( s) n ϕ n (x, s)f (n1) (s)ds, ϕ n (x, s) = (x s)n [ = ( s)n ( s) n P n n [ (xk s) g k (x) n (k), whch, fter some mmedte mnpultons, mples (9)

5 The combned Sheprd Abel Gonchrov unvrte opertor 15 2 THE COMBINED SHEPARD ABEL GONCHAROV UNIVARIATE OPERATOR In ths secton we shll ssume tht there exsts f () (x ), = 0,, N, on the set of N 1 prwse dstnct ponts x [, b, 0 N Let us consder the set of lner functonls of Abel Gonchrov type: Λ AG (f) := { λ (f) : λ (f) = f () (x ), = 0,, N } We ttch to ech node x, = 0,, N, set of nodes X,n, n N, n N, = 0,, N, defned by (10) X,n = {x, x 1,, x n } = {x υ : υ = 0,, n}, = 0,, N, where x Nk1 = x k, k = 0,, n We ssocte to ech set of nodes X,n, = 0,, N, the Abel Gonchrov nterpolton opertor, denoted P n, = 0,, N, correspondng to the set of functonls Λ AG The opertors P n, = 0,, N, exst nd re unque becuse the ponts of the sets X,n, = 0,, N, re prwse dstnct so the determnnt of the nterpolton system of the form (7) s lwys dfferent from zero We hve (11) (P n f) (k) (x k ) = f (k) (x k ), k n, 0 N Remrk 5 The set of lner functonl of Abel Gonchrov type, Λ AG, s ncluded n the set of lner functonl of Brkhoff type We notce tht n cse of the Abel Gonchrov nterpolton we hve the dvntge tht the determnnt of the nterpolton system of the form (7) s lwys dfferent from zero, thus the nterpolton polynoml lwys exsts nd s unque We consder the Abel Gonchrov polynomls of degree n, ssocted to the sets of nodes X,n, = 0,, N, nd the sets of lner functonls of Abel Gonchrov type gven by n (12) (P n f)(x) = g j (x)f (j ) (x j ), = 0,, N, j= wth the Gonchrov polynomls gven by g 0 (x) = 1, g 1 (x) = x x, [ g k (x) = 1 k! x k k 1 j=0 g j (x) ( k j ) x k j j, k 1 Theorem 3 The Abel Gonchrov opertors P n, = 0,, N, hve the degree of exctness n, e, (13) dex(p n ) = n, = 0,, N

16 Teodor Cătnş 6 The proof s obtned smlrly to tht of Theorem 1 Remrk 6 If we consder the sets X,n, = 0,, N, of the form (10) such tht ech of them hs n 1 elements, n N, then We denote by S AG n dex(p n ) = n, = 0,, N the Sheprd opertor of Abel Gonchrov type, gven by (S AG n f)(x) = A (x)(p n f)(x), where A, = 0,, N, re gven by (2) nd P n, = 0,, N, re gven by (12) We cll Sn AG the combned Sheprd Abel Gonchrov opertor Prtculr cse When ll the nodes concde, x 0 = = x N, we obtn the Sheprd opertor of Tylor type: n (Tnf)(x) = (P n (x x f)(x) = ) j j! f (j) (x ), = 0,, N, j=0 nd the combned Sheprd Tylor opertor gven by (ST n f)(x) = A (x)(tnf)(x) The mn propertes of the Sheprd Tylor opertor ST n re: For µ > n (ST n f) (j) (x ) = f (j) (x ), j = 0,, n, = 0,, N; gex(st n ) = n Theorem 4 The opertor S AG n s lner Proof For rbtrry h 1, h 2 : [, b R nd α, β R, one gets S AG n (αh 1 βh 2 )(x) = n = A (x) g j (x)(αh 1 βh 2 ) (j ) (x j ) = α = αs AG n j= n A (x) g j (x)h(j ) 1 (x j ) β j= (h 1 )(x) βsn AG (h 2 )(x), n A (x) g j (x)h 2 (j ) (x j ) whch shows the lnerty of S AG n j=

7 The combned Sheprd Abel Gonchrov unvrte opertor 17 Theorem 5 If µ > N, the opertor S AG n (14) (S AG n f) (k) (x k ) = f (k) (x k ), 0 k N hs the nterpolton property: Proof Tkng nto ccount tht µ > N, we hve k (Sn AG f) (k) ( (x k ) = k (ν) ν) A (x k )(P n f) (k ν) (x k ) By (5) nd (11) we obtn (14) ν=0 Theorem 6 Sn AG f = f, for ll f P n, where P n s the set of polynomls of degree t most n Proof From Theorem 2 we hve gex(sn AG ) = mn { gex(p n ) = 0,, N }, nd, tkng nto ccount (13), we obtn gex(s AG n ) = n Remrk 7 If the condtons of Remrk 6 re verfed, e, gex(p n ) = n, = 0,, N, then by Theorem 2 we hve gex(s AG n ) = mn {0,,N} n The Sheprd Abel Gonchrov nterpolton formul s f = Sn AG f Rn AG f, where Rn AG f denotes the remnder When ll the nodes concde we hve the Sheprd Tylor nterpolton formul nd the followng result s known Theorem 7 [4 If f H n1 [, b nd x 0 = x 1 = = x N, then where wth ϕ n (x, s) = 1 (R AG n f)(x) = { (x s) n ϕ n (x, s)f (n1) (s)ds, A (x) [ (x s) 0 n (x x ) (x x) } Theorem 8 If f H n1 [, b then (R AG n f)(x) = (15) ϕ n (x, s) = (x s)n ϕ n (x, s)f (n1) (s)ds, n A (x) g j (x) (x j s) n j (n j)! j=

18 Teodor Cătnş 8 Proof Theorem 6 mples gex(sn AG ) = n Applyng the Peno s theorem, we obtn wth ϕ n (, s) = R AG n For ll x [, b we hve (R AG n f)(x) = ϕ n (x, s) = (x s)n [ ( s) n = ( s)n ϕ n (x, s)f (n1) (s)ds, j= A ( )P n [ ( s) n n [ A (x) g j (x) (xj s) n (j ), nd fnlly (15) Prtculr cse We consder n = 1 We hve the correspondng Sheprd Abel Gonchrov opertor gven by (S AG 1 f)(x) = The nterpolton formul s A (x)(p 1 f) [ = A (x) g 0 (x)f(x ) g 1 (x)f (x 1 ) f = S AG 1 f R AG 1 f, where R1 AG f s the remnder, whch ccordng wth Theorem 8 hs the followng form: wth (R AG 1 f)(x) = ϕ 1 (x, s)f (s)ds, [ ϕ 1 (x, s) = (x s) A (x) g 0 (x)(x s) g 1 (x)(x 1 s) 0 Exmple 1 Consder f : [0, 4 R, f(x) = 3 sn πx 4, nd the nodes x =, = 0,, 4 Fgure 1 shows the Sheprd pproxmton correspondng to these dt, whle Fgure 2 shows the Sheprd Abel Gonchrov pproxmton The fgures were drwn usng Mtlb

9 The combned Sheprd Abel Gonchrov unvrte opertor 19 3 25 2 15 1 05 0 0 05 1 15 2 25 3 35 4 Fg 1: Sheprd nterpolton 35 3 25 2 15 1 05 0 05 0 05 1 15 2 25 3 35 4 Fg 2: Sheprd-Abel-Gonchrov nterpolton REFERENCES [1 Agrwl, A nd Wong, P J Y, Error Inequltes n Polynoml Interpolton nd ther Applctons, Kluwer Acdemc Publshers, Dordrecht, 1993 [2 Cheney, W nd Lght, W, A Course n Approxmton Theory, Brooks/Cole Publshng Compny, Pcfc Grove, 2000 [3 Comn, Gh, The remnder of certn Sheprd type nterpolton formuls, Stud Unv Bbeş Boly, Mthemtc, XXXII, no 4, pp 24 32, 1987 [4 Comn, Gh, Sheprd-Tylor nterpolton, Itnernt Semnr on Functonl Equtons, Approxmton nd Convexty, Cluj-Npoc, pp 5 14, 1988 [5 Comn, Gh nd Ţâmbule, L, A Sheprd-Tylor pproxmton formul, Stud Unv Bbeş Boly, Mthemtc, XXXIII, no 3, pp 65 73, 1988 [6 Comn, Gh nd Trîmbţş, R, Combned Sheprd unvrte opertors, Est Jurnl on Approxmtons, 7, no 4, pp 471 483, 2001

20 Teodor Cătnş 10 [7 Comn, Gh nd Trîmbţş, R, Unvrte Sheprd-Brkhoff nterpolton, Rev Anl Numér Théor Approx, 30, no 1, pp 15 24, 2001 [8 Dvs, Ph J, Interpolton nd Approxmton, Blsdell Publshng Compny, New York, 1963 [9 Dzhprdze, K nd Jnssen, R H P, A stochstc pproch to n nterpolton problem wth pplctons to Hellnger ntegrls nd rthmetc-geometrc men reltonshp, CWI Report, Amsterdm, 7, no 3, pp 245 258, 1994 [10 Sendov, B nd Andreev, A, Approxmton nd Interpolton Theory, n Hndbook of Numercl Anlyss, vol III, ed P G Crlet nd J L Lons, 1994 [11 Shyne, W, Refnements of the Peno kernel theorem, Numer Funct Anl Optmz, 20, nos 1 2, pp 147 161, 1999 [12 Stncu, D D, Comn, Gh, Agrtn, O nd Trîmbţş, R, Numercl Anlyss nd Approxmton Theory, vol I, Pres Unverstră Clujenă, 2001 (n Romnn) [13 Stncu, D D, Comn, Gh, nd Blg, P, Numercl Anlyss nd Approxmton Theory, vol II, Pres Unverstră Clujenă, 2002 (n Romnn) Receved by the edtors: November 9, 2002