Error Bound of The Approximation Lagrange Method
Uploaded by
ScribdTranslationsError Bound of The Approximation Lagrange Method
Uploaded by
ScribdTranslationsThe steps to follow to construct the interpolation polynomial of the points {(x0, and0),
(x1, and1) (xn, andn}, developed by E. Waring, but called Lagrange's formula, are
the following:
1. We form n polynomials of degree n that meet the following requirements:
P0that it is canceled at the points x1, x2,
xnand that for x = x0take the value and0.
To achieve this, we propose the formula P0λ0(x-x1)
(x-x n).
Indeed P0(xi= 0, for i = 1, 2, ..., n and as is intended that P0(x0) = y0,
the value of λ0will result
P1that is canceled at the points x0, x2, , xnand that for x = x1take the value and1.
Similarly, we propose the formula P1(x) = λ0(x-x0) (x-x2(x-xn).
Here it results P0(xi) = 0 for i = 0, 2, 3, ..., n as intended, and as
I asked for P1(x1) = y1, the value of λ1will result
And so on, we create a polynomial P.iof degree n that is annulled for
all except xi, where it takes the value of yiuntil the last one is canceled in
all points except on x, where it takes the value yn.
2. Adding all the obtained polynomials, it results in a polynomial of degree to.
sumo n, in the form:
that meets condition Pn(xi) = yIfor i = 0, 1, n. (Verify it!!)
If we call
then the polynomial can be expressed in a more compact form, as
indicate in:
In this formula, the expressions Ln,j(x) can be called directly Lj(x) if not
There are doubts regarding his degree.
When making an approximation of the value of a function at a point, it is
it is good to know a bound of the error made. In some cases, it can be,
others no.
Error bound of the approximation using interpolation polynomials
The following theorem provides a formula to calculate the remainder or error bound in which
one incurs when approximating the value of a function at a point using the polynomial of
interpolation.
Theorem
Sean x0, x1, xndistinct real numbers in the interval [a, b], and let
f Cn+1Then for every x in the interval [a, b], with P the
The interpolation polynomial given by the expression (L.3) exists a
number (x) in the interval (a, b) that satisfies:
The specific use of this error formula is limited to functions whose derivatives
they have known quotas. But it is a very important theoretical result, given that the
Lagrange polynomials are often used to derive differentiation formulas.
numerical and integration methods. The error bounds of these techniques are derived from
the Lagrange error formula.
Example 1:
A table of values of the function f(x) = sin(x) is desired in the interval
[0, Assuming that the number of decimal places of each value is d ≥ 8 and h
the size of the step between two consecutive values of x. What should the value of h be
so that linear interpolation yields an absolute value of at most 10-6?
Let's consider the points x0, x1, x2, , xnof the interval [0, 1], where it will be evaluated the
given function. If x is between xjy xj+1, it is fulfilled, according to the formula:
Therefore, if x is in the interval [0, 1], it results:
If we define the function g(x) = (x-jh)(x-(j+1)h), and calculate its extrema, we get:
Therefore, as expressed in it results:
How is it intended for the error to be less than 10?-3, then proposing that the limit
obtained in sea is less than this value, remains
Solving for h from this expression results in h < 0.0975047
Since n = (1-0)/h must be an integer, an appropriate choice would be to take h = 0.05
http://www.frsn.utn.edu.ar/gie/an/iac/21_lagrange.html