Scribd
Upload
5 views16 pages

L-6 Error

Uploaded by

rudeperson731
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
5 views16 pages

L-6 Error

Uploaded by

rudeperson731
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

MA 204 Numerical Methods

Dr. Debopriya Mukherjee


Lecture-6

January 23, 2024


Contents

Solution of a nonlinear equation, bisection and secant


methods, Newton’s method, rate of convergence.
Contents

Solution of a nonlinear equation, bisection and secant


methods, Newton’s method, rate of convergence.
Interpolation by polynomials, divided differences, error of the
interpolating polynomial, piecewise linear and cubic spline
interpolation.
Errors in Polynomial Interpolation

Given a function f (x) on x ∈ [a, b], and a set of distinct points


xi ∈ [a, b], i = 0, 1, · · · , n. Let Pn (x) ∈ Pn s.t.

Pn (xi ) = f (xi ), i = 0, 1, 2, · · · , n.
Errors in Polynomial Interpolation

Given a function f (x) on x ∈ [a, b], and a set of distinct points


xi ∈ [a, b], i = 0, 1, · · · , n. Let Pn (x) ∈ Pn s.t.

Pn (xi ) = f (xi ), i = 0, 1, 2, · · · , n.

Error function: e(x) = f (x) − Pn (x), x ∈ [a, b].


Theorem 1
There exists some value ξ ∈ (a, b), such that
n
1 Y
e(x) = f (n+1) (ξ) (x − xi ), for all x ∈ [a, b]. (1)
(n + 1)!
i=0
Proof. If f ∈ Pn , then, by Uniqueness Theorem of polynomial
interpolation, we must have f (x) = Pn (x). Then, e(x) = 0 and
the proof is trivial.
Proof. If f ∈ Pn , then, by Uniqueness Theorem of polynomial
interpolation, we must have f (x) = Pn (x). Then, e(x) = 0 and
the proof is trivial.
Now assume, f ∈ / Pn . If x = xi for some i, we have

e(xi ) = f (xi ) − Pn (xi ) = 0,

and the result holds.


Proof. If f ∈ Pn , then, by Uniqueness Theorem of polynomial
interpolation, we must have f (x) = Pn (x). Then, e(x) = 0 and
the proof is trivial.
Now assume, f ∈ / Pn . If x = xi for some i, we have

e(xi ) = f (xi ) − Pn (xi ) = 0,

and the result holds.


Now consider x 6= xi for any i.
n
Y
W (x) = (x − xi ) ∈ Pn+1 ,
i=0
Proof. If f ∈ Pn , then, by Uniqueness Theorem of polynomial
interpolation, we must have f (x) = Pn (x). Then, e(x) = 0 and
the proof is trivial.
Now assume, f ∈ / Pn . If x = xi for some i, we have

e(xi ) = f (xi ) − Pn (xi ) = 0,

and the result holds.


Now consider x 6= xi for any i.
n
Y
W (x) = (x − xi ) ∈ Pn+1 ,
i=0

it holds

W (xi ) = 0, W (x) = x n+1 + · · · , W (n+1) = (n + 1)!


Fix an x such that a ≤ x ≤ b and x 6= xi for any i. We define a
constant
f (x) − Pn (x)
c= ,
W (x)
Fix an x such that a ≤ x ≤ b and x 6= xi for any i. We define a
constant
f (x) − Pn (x)
c= ,
W (x)
and another function

ϕ(y ) = f (y ) − Pn (y ) − cW (y ).
Fix an x such that a ≤ x ≤ b and x 6= xi for any i. We define a
constant
f (x) − Pn (x)
c= ,
W (x)
and another function

ϕ(y ) = f (y ) − Pn (y ) − cW (y ).

We find all the zeros for ϕ(y ). We see that xi0 s are zeros since

ϕ(xi ) = f (xi ) − Pn (xi ) − cW (xi ) = 0.

Also, x is a zero because

ϕ(x) = f (x) − Pn (x) − cW (x) = 0.


Here goes our deduction:
ϕ(x) has atleast (n + 2) zeros on [a, b].
ϕ0 (x) has atleast (n + 1) zeros on [a, b].
ϕ00 (x) has atleast n zeros on [a, b].
···
ϕ(n+1) (x) has atleast 1 zero on [a, b].
Call it ξ s.t. ϕ(n+1) (ξ) = 0. So, we have

ϕ(n+1) (ξ) = f (n+1) (ξ) − 0 − cW (n+1) (ξ) = 0.

Recall W n+1 = (n + 1)!, we have, for every y,

f (y ) − Pn (y )
f (n+1) (ξ) = cW (n+1) (ξ) = (n + 1)!
W (y )
n
1 Y
e(x) = f (x) − Pn (x) = f (n+1) (ξ) (x − xi ),
(n + 1)!
i=0

for some ξ ∈ [a, b].


Disadvantages of polynomial interpolation Pn (x)
n− times differentiable; We do not need such high
smoothness;
big error in certain intervals (esp. near the ends);
no convergence result;
heavy to compute for large n.

You might also like