1 Interpolation and Polynomial Approximation

Interpolation is a type of estimation, a method of constructing new data points

within the range of a discrete set of known data points.

Consider the following table:

Year 2000 2005 2010 2015 2020

Turkey Urban Population(in millions) 40,942 46,065 51,225 57,616 63,803

These data could be used to provide a reasonable estimate of the urban

population at some other years, say, in 2008 or in 2030.
One of the most useful and well-known classes of functions mapping the set
of real numbers into itself is the algebraic polynomials

Pn (x) = an xn + an−1 xn−1 + ... + a1 x + a0

where n is a nonnegative integer and a0 , ..., an are real constants. One reason
for their importance is that they uniformly approximate continuous functions.

Theorem: (Weierstrass Approximation Theorem)

Suppose that f is defined and continuous on [a, b]. For each ε > 0, there
exists a polynomial P (x), with the property that

|f (x) − P (x)| < ε

for all x in [a, b] .

Another important reason for considering the class of polynomials in the

approximation of functions is that the derivative and indefinite integral of a
polynomial are easy to determine and are also polynomials. For these reasons,
polynomials are often used for approximating continuous functions.

Definition: The interpolating polynomial for a set of n + 1 points

(x0 , f (x0 )), (x1 , f (x1 )), ..., (xn , f (xn ))

is the unique polynomial P (x) such that

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

QUESTION: How to construct a polynomial P (x) of degree at most n?

1
1.1 Lagrange Interpolating Polynomials
Suppose n = 1 and we have only two points (x0 , f (x0 )) and (x1 , f (x1 )). The
linear Lagrange interpolating polynomial through (x0 , f (x0 )) and (x1 , f (x1 ))
is
(x − x1 ) (x − x0 )
P (x) = L0 (x)f (x0 ) + L1 (x)f (x1 ) = f (x0 ) + f (x1 )
(x0 − x1 ) (x1 − x0 )
which implies that

P (x0 ) = 1.f (x0 ) + 0.f (x1 ) = f (x0 )

P (x1 ) = 0.f (x0 ) + 1.f (x1 ) = f (x1 )

Example 1: Determine the linear Lagrange interpolating polynomial that

passes through the points (2, 4) and (5, 1).

Solution: The linear Lagrange interpolating polynomial that passes through

the points (2, 4) and (5, 1) is

P (x) = L0 (x)f (x0 ) + L1 (x)f (x1 )

(x − 5) (x − 2)
= 4+ 1
(2 − 5) (5 − 2)
= −x + 6

In general, the nth Lagrange interpolating polynomial, is defined in

the following theorem.

Theorem: If x0 , x1 , ..., xn are n + 1 distinct numbers and f is a function

whose values are given at these numbers, then a unique polynomial P (x) of
degree at most n exists with

P (xk ) = f (xk ), k = 0, 1, 2, ..., n

This polynomial is given by

n
X
P (x) = Ln,0 (x)f (x0 ) + Ln,1 (x)f (x1 ) + ... + Ln,n (x)f (xn ) = Ln,k (x)f (xk )
k=0

where, for each k = 0, 1, ..., n,

(x − x0 ) (x − x1 ) ... (x − xk−1 ) (x − xk+1 ) ... (x − xn )
Ln,k (x) =
(xk − x0 ) (xk − x1 ) ... (xk − xk−1 ) (xk − xk+1 ) ... (xk − xn )
n
Y (x − xi )
=
i=0
(xk − xi )
i6=k

2
 NOTE: We will write Ln,k (x) simply as Lk (x) when there is no confusion
as to its degree.

Example 2:
(i) Use the nodes x0 = 2, x1 = 2.75, and x2 = 4 to find the second Lagrange
interpolating polynomial for f (x) = 1/x.
(ii) Use this polynomial to approximate f (3) = 1/3.

Solution:

(i) The second Lagrange interpolating polynomial for f (x) = 1/x is

2
X
P2 (x) = Lk (x)f (xk )
k=0
= L0 (x)f (x0 ) + L1 (x)f (x1 ) + L2 (x)f (x2 )
(x − x1 ) (x − x2 ) (x − x0 ) (x − x2 ) (x − x0 ) (x − x1 )
= f (x0 ) + f (x1 ) + f (x2 )
(x0 − x1 ) (x0 − x2 ) (x1 − x0 ) (x1 − x2 ) (x2 − x0 ) (x2 − x1 )
(x − 2.75) (x − 4) (x − 2) (x − 4) (x − 2) (x − 2.75)
= f (2) + f (2.75) + f (4)
(2 − 2.75) (2 − 4) (2.75 − 2) (2.75 − 4) (4 − 2) (4 − 2.75)
1 64 1
= (x − 2.75) (x − 4) − (x − 2) (x − 4) + (x − 2) (x − 2.75)
3 165 10
1 2 35 49
= x − x+
22 88 44

(ii)
1 2 35 49
f (3) ≈ P2 (3) = (3) − 3 + ≈ 0.32955
22 88 44

1
((The value of f (x) at x = 3 is 3 ≈ 0.3333)

Theorem: (Interpolation Error Theorem)

Suppose x0 , x1 , ..., xn are distinct numbers in the interval [a, b] and f ∈

C n+1 [a, b]. Then, for each x in [a, b], a number ξ(x) between x0 , x1 , ..., xn and
hence in (a, b), exists with

f (n+1) (ξ(x))
f (x) = P (x) + (x − x0 ) (x − x1 ) ... (x − xn )
(n + 1)!

where P (x) is the interpolating polynomial given as

n
X
P (x) = Ln,0 (x)f (x0 ) + Ln,1 (x)f (x1 ) + ... + Ln,n (x)f (xn ) = Ln,k (x)f (xk )
k=0

3
 Example: Compute the maximum absolute error for Example 2 for x ∈
[2, 4] .

Solution: We have f (x) = 1/x, so

f 0 (x) = −1/x2 , f 00 (x) = 2/x3 and f 000 (x) = −6/x4

Hence, the second Lagrange polynomial has the error form

f 000 (ξ(x)) 1
(x − x0 ) (x − x1 ) (x − x2 ) = − (x − 2) (x − 2.75) (x − 4)
3! (ξ(x))4
1 1 1
where ξ(x) in (2, 4) . The maximum value of (ξ(x))4 = 24 = 16 . Let’s define

35 2 49
g(x) = (x − 2) (x − 2.75) (x − 4) = x3 − x + x − 22
4 2
Since
 
d 35 2 49 35 49 1
x − x + x − 22 = 3x2 − x +
3
= (3x − 7)(2x − 7)
dx 4 2 2 2 2

the critical points are

7 7 25 7 7 9
x= with g( ) = and x = with g( ) = −
3 3 108 2 2 16
Also, from the endpoints
g(2) = g(4) = 0
Hence, the maximum error is

f 000 (ξ(x)) 1
(x − x0 ) (x − x1 ) (x − x2 ) = − (x − 2) (x − 2.75) (x − 4)
3! (ξ(x))4
1 9
≤ ≈ 0.0352
16 16

DİVİDED DİFFERENCES
(ALTERNATE ALGEBRAİC FORMULA FOR INTERPOLATION)

A major difficulty with the Lagrange Interpolation is that if the accuracy is

not good enough with polynomial of a certain degree, one needs to increase the
degree of the polynomial, and computations need to be started all over again.
Start with a table of values of a function f :

x x0 x1 x2 ... xn
f(x) f(x0 ) f(x1 ) f(x2 ) ... f(xn )

4
 The points x0 , x1 , ..., xn are assumed to be distinct.

Previously, we established that for each n = 0, 1, ..., there exists a unique

polynomial Pn such that

• The degree of Pn is at most n.

• Pn (xi ) = f (xi ) for i = 0, 1, ..., n.

The divided differences of f with respect to x0 , x1 , ..., xn are used to express

Pn (x) in the form

Pn (x) = a0 + a1 (x − x0 ) + a2 (x − x0 )(x − x1 ) + ... + an (x − x0 )...(x − xn−1 ) (1)

for appropriate constants a0 , a1 , ..., an .

Note that
a0 = Pn (x0 ) = f (x0 ) (2)
Similarly, when Pn (x) is evaluated at x1

f (x0 ) + a1 (x1 − x0 ) = Pn (x1 ) = f (x1 )

so
f (x1 ) − f (x0 )
a1 = (3)
x1 − x0
We now introduce the divided-difference notation. The zeroth divided
difference of the function f with respect to xi , denoted f [xi ], is simply
the value of f at xi :
f [xi ] = f (xi ) (4)
The remaining divided differences are defined recursively; the first divided
difference of f with respect to xi and xi+1 is denoted f [xi , xi+1 ] and defined
as
f [xi+1 ] − f [xi ]
f [xi , xi+1 ] = (5)
xi+1 − xi
The second divided difference, f [xi , xi+1 , xi+2 ] is defined as

f [xi+1 , xi+2 ] − f [xi , xi+1 ]

f [xi , xi+1 , xi+2 ] = (6)
xi+2 − xi

Similarly, after the (k − 1)st divided differences, the kth divided difference
relative to xi , xi+1 , xi+2 , ..., xi+k is

f [xi+1 , ..., xi+k−1 , xi+k ] − f [xi , xi+1 , ..., xi+k−1 ]

f [xi , xi+1 , ..., xi+k−1 , xi+k ] =
xi+k − xi
(7)

5
 The process ends with the single nth divided difference,

f [x1 , ..., xn ] − f [x0 , x1 , ..., xn−1 ]

f [x0 , x1 , ..., xn ] = (8)
xn − x0
Hence the interpolating polynomial in Eq. (1) is

Pn (x) = f [x0 ]+f [x0 , x1 ](x−x0 )+f [x0 , x1 , x2 ](x−x0 )(x−x1 )+...+f [x0 , x1 , ..., xn ](x−x0 )...(x−xn−1 )
(9)
So Pn (x) can be rewritten in a form called Newton’s Divided-Difference:
n
X
Pn (x) = f [x0 ] + f [x0 , x1 , ..., xk ](x − x0 )...(x − xk−1 ) (10)
k=1

Example 1 For the table write out the Newton-Form of the interpolating poly-
nomial.
x 1 -4 0
f(x) 3 13 -23

f [x0 ] = 3
f [x1 ] = 13
f [x2 ] = −23
f [x1 ] − f [x0 ] 13 − 3
f [x0 , x1 ] = = = −2
x1 − x0 −4 − 1
f [x2 ] − f [x1 ] −23 − 13
f [x1 , x2 ] = = = −9
x2 − x1 0+4
f [x1 , x2 ] − f [x0 , x1 ] −9 + 2
f [x0 , x1 , x2 ] = = =7
x2 − x0 0−1

Hence, from Eq. (9)

P2 (x) = 3 − 2 (x − 1) + 7 (x − 1) (x + 4)

Example 2 Construct a divided-difference diagram for the function f given in

the following table, and write out the Newton form of the interpolating polyno-
mial.
x 1 3/2 0 2
f(x) 3 13/4 3 5/3

6
 f [x0 ] = 3
13
f [x1 ] =
4
f [x2 ] = 3
5
f [x3 ] =
3
13
f [x1 ] − f [x0 ] 4 −3 1
f [x0 , x1 ] = = 3 =
x1 − x0 2 −1
2
f [x2 ] − f [x1 ] 3 − 13
4 1
f [x1 , x2 ] = = =
x2 − x1 0 − 32 6
5
f [x3 ] − f [x2 ] 3 −3 2
f [x2 , x3 ] = = =−
x3 − x2 2−0 3
1 1
f [x1 , x2 ] − f [x0 , x1 ] − 1
f [x0 , x1 , x2 ] = = 6 2 =
x2 − x0 0−1 3
f [x2 , x3 ] − f [x1 , x2 ] − 32 − 61 5
f [x1 , x2 , x3 ] = = 3 =−
x3 − x1 2− 2 3
f [x1 , x2 , x3 ] − f [x0 , x1 , x2 ] −5 − 1
f [x0 , x1 , x2 , x3 ] = = 3 3 = −2
x3 − x0 2−1
Thus, we obtain
   
1 1 3 3
P3 (x) = 3 + (x − 1) + (x − 1) x − − 2 (x − 1) x − x
2 3 2 2