ODA BULTUM UNIVERSITY

INSTITUTE OF TECHNOLOGY

POLYNOMIAL IN-
TERPOLATION AND
APPROXIMATION
 Outline for Group 3
• Introduction
• Class of common Approximation
• Linear Interpolation
• Lagrangian Polynomials
 Introduction
• Interpolation, the computing of values for a tabulated function
at points not in the table, is historically a most important task.
• Many famous mathematicians have their names associated
with procedures for interpolation: Gauss, Newton, Bessel,
etc.
• The need to interpolate began with the early studies of astro-
nomy when the motion of heavenly bodies was to be determ-
ined from. periodic observations.
• Interpolation methods are the basic for many others procedures
that we will study: numerical differentiation and integration,
solution methods for ordinary and partial differential equa-
tions, etc.
 .... Cont’s
• These methods demonstrate some important theory
about polynomials and the accuracy of numerical
methods.
• Interpolating with polynomials also serves as an ex-
cellent introduction to some techniques for drawing
smooth curves.
• In practice there is always need to approximate func-
tions to facilitate the numerical solution.
• Here , below, are given some of the cases in which
there is need for approximations.
 .... Cont's
1. At the end of a certain measurement we may obtain (yi) correspond-
ing to a pre fixed xi (an dependent variable) relative to a certain
physical phenomena (a function). We might then need to establish an
analytical relation between the yi and the xi which sufficiently de-
scribe the phenomenon.
2. Suppose we want to operate (say integrate) on a function, ƒ(x).
However, ƒ(x) could be an expression which is difficult to operate
on. Then, we need to approximate ƒ(x), in the interval of interest,
with another simpler function, ƒ°(x).
3. We might need to determine the solution of a differential equitation.
However, as it is usually difficult to obtain an exact solution, we
might then satisfy ourselves with an approximate solution ƒ°(x),
(which is in a pre- fixed, say a polynomial form), such that ƒ°(x),
satisfies the equation not at every point in the domain of interest, but
at a finite number of points.
.... Cont‘s
Class of common Approximation
Function
.... Cont's
 .... Cont's
• Rational functions are to simulate singularities (poles) or to represent non-
periodic phenomena over infinite interval. This is usually used to simulate
phenomena with exponential behaviour. Also, there are a piecewise linear
polynomials and are used as an alternative to the class of polynomials or
class of trigonometric polynomials, specially when the use of polynomials
with high degree are required to achieve a certain desired precision, ε.
• There are a variety of alternative forms for expressing an interpolating
polynomial, between the amount of most popular and useful forms, we are
presenting the:
– Linear interpolation
– Lagrange interpolating polynomials
– Neville’s method
– Divided Differences (Newton’s Divided-difference)
– Spline interpolation
– Least- square approximation
Linear interpolation
.... Cont’s
Lagrangian Polynomials
.... Cont's
 .... Cont's
• This procedure is awkward, especially if we want a new
polynomial that is also made to fit at the point (5.6 ; 51.7),
or if we want to see what difference it would make to use
a quadratic instead of cubic.
• Furthermore, this technique leads to an ill-conditioned
system of equations.
• The Lagrangian polynomial is perhaps the simplest way
to exhibit the existence of a polynomial for interpolation
with unevenly spaced data.
• Data where the x-values are not equispaced often occur
as the result of experimental observations or when histor-
ical data are examined.
.... Cont's
.... Cont's
 .... Cont's
• Lagrangian polynomials summary characteristics:
1. Very useful for interpolation with unevenly spaced
datathis technique can leads to an ill conditioned sys-
tem of equations.
2. Lagrangian polynomials is accuracy to interpolate the
xvalves enclosed inside the points used in its construc-
tion, nevertheless when the x-values are outside of in-
terval used to determine the polynomial you can obtain
undesired results.
3. The arithmetic in this method (Lagrangian) is tedious,
but write a computer program that implements the
method is not hard to do
4. If the function that generates the xi , fi values is “s-
mooth”, a low-degree polynomial should work satisfact-
orily, on the other hand, a “rough” function can be expec-
ted to have larger errors when interpolated.
5. The interpolation error is smaller if x is centered within
the xi, because this makes the product of the (x- xi)
terms smaller.
6. In Lagrangian polynomial technique we don’t know
which degree of polynomial is better to use. If the de-
gree is too low, the interpolating polynomial does not
give good estimated of f(x). If the degree is too high, un-
desirable oscillations in polynomials values can occurs.
.... Cont's
.... Cont’s
.... Cont's
THANKS FOR HAVING SUCH
A BEAUTIFUL TIME