Newton's Interpolation Methods: P. Sam Johnson

Newton’s Interpolation Methods
P. Sam Johnson
February 7, 2020
P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 1 / 47

Overview
One of the basic ideas in Mathematics is that of a function and most

useful tool of numerical analysis is interpolation.
According to Thiele (a numerical analyst), “Interpolation is the art of

reading between the lines of the table.”
Broadly speaking, interpolation is the problem of obtaining the value of a

function for any given functional information about it.
Interpolation technique is used in various disciplines like economics,

business, population studies, price determination etc. It is used to fill in
the gaps in the statistical data for the sake of continuity of information.

Overview
The concept of interpolation is the selection of a function p(x) from a

given class of functions in such a way that the graph of
y = p(x)
passes through a finite set of given data points. The function p(x) is
known as the interpolating function or smoothing function.
If p(x) is a polynomial, then it is called the interpolating polynomial and

the process is called the polynomial interpolation.
Similarly, if p(x) is a finite trigonometric series, we have trigonometric

interpolation. But we restrict the interpolating function p(x) to
being a polynomial.

Overview
The study of interpolation is based on the calculus of finite differences.
Polynomial interpolation theory has a number of important uses. Its

primary uses is to furnish some mathematical tools that are used in
developing methods in the areas of approximation theory, numerical
integration, and the numerical solution of differential equations.
We discuss Newtons forward/backward formulae (for equally spaced

nodes) and error bounds in two lectures.

Introduction
A census of the population of the India is taken every 10 years. The

following table lists the population, in thousands of people, from 1951 to
2011.
Year 1951 1961 1971 1981 1991 2001 2011
Population 361,088 439,235 548,160 683,329 846,388 1,028,737 1,210,193
(in thousands)
In reviewing these data, we might ask whether they could be used to

provide a reasonable estimate of the population, say, in 1996, or even in
the year 2014. Predictions of this type can be obtained by using a function
that fits the given data.
This process is called interpolation / extrapolation.

Introduction
If y is a function of x, then the functional relation may be denoted by the

equation
y = f (x).
The forms of f (x) can, of course, be very diverse, but we consider f (x) as
a polynomial of the nth degree in x
y = a0 + a1 x + · · · + an x n (an 6= 0).
We call x as the independent variable and y as the dependent variable.

It is usual to call x as argument and y as function of the argument or
entry.
Since the polynomials are relatively simple to deal with, we interpolate to

the data by polynomials.

Weierstrass Approximation Theorem
If the value of x whose corresponding value y is to be estimated lies within

the given range of x, then it is a problem of interpolation. On the other
hand, if the value lies outside the range, then it is a problem of
extrapolation.
Thus for the theory of interpolation, it is not esssential that the functional
form of f (x) be known. The only information needed is the values of the
function given for some values of the argument.
In the method of interpolation, it is assumed that the function is capable

of being expressed as a polynomial. This assumption is based on
Weierstrass approximation theorem. That is, the existence of an
interpolating polynomial is supported by the theorem.

Weierstrass Approximation Theorem
Given any function, defined and continuous on a closed and bounded

interval, there exists a polynomial that is as “close” to the given function
as desired. This result is expressed precisely in the following theorem.
Theorem 1 (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 ∈ [a, b].

Why polynomials are important?
Weierstrass approximation
theorem is illustrated in the
figure.
In science and engineering,

polynomials arise everywhere.
An 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 they are also polynomials.
For these reasons, polynomials are often used for approximating

continuous functions. We introduce various interpolating polynomials
using the concepts of forward, backward and central differences.

Main Asssumption for Interpolation
There are no sudden jumps or falls in the values of the function from one
period to another. This assumption refers to the smoothness of f (x) i.e.,
the shape of the curve y = f (x) changes gradually over the period under
consideration.
For example, if the population figures are given for, 1931, 1951, 1961, 1971
and figures for 1941 are to be interpolated, we shall have to assume that
the year 1941 was not an exceptional year, such as that affected by
epidemics, war or other calamity or large scale immigration.

Rolle’s theorem
We recall Rolle’s theorem, which is useful in evaluating error bounds.
Theorem 2 (Rolle’s Theorem).

Let f be continuous on [a, b] and differentiable in (a, b). If f (a) = f (b),
then there is at least one point c ∈ (a, b) such that f 0 (c) = 0.

Generalized Rolle’s theorem
Theorem 3 (Generalized Rolle’s Theorem).
Let f be continuous on [a, b] and n times differentiable in (a, b). If f (x) is
zero at the n + 1 distinct numbers c0 , c1 , . . . , cn in [a, b], then a number c
in (a, b) exists with f (n) (c) = 0.

Error in Polynomial Interpolation
Let the function y (x), defined by the (n + 1) points
(xi , yi ), i = 0, 1, 2, . . . , n
be continuous and differentiable (n + 1) times, and let y (x) be

approximated by a polynomial pn (x) of degree not exceeding n such that
pn (xi ) = yi
for i = 0, 1, 2, . . . , n.

Using the polynomial pn (x) of degree n, we can obtain approximate values

of y (x) at some points other xi , 0 ≤ i ≤ n.
Since the expression y (x) − pn (x) vanishes for x = x0 , x1 , . . . , xn we put
y (x) − pn (x) = Lπn+1 (x) (1)
where
πn+1 (x) = (x − x0 )(x − x1 ) · · · (x − xn )
and L is to be determined such that the equation (1) holds for any
intermediate value of x 0 ∈ (x0 , xn ). Clearly
y (x 0 ) − pn (x 0 )
L= . (2)
πn+1 (x 0 )

We construct a function F (x) such that
F (x) = y (x) − pn (x) − Lπn+1 (x) (3)
where L is given by the equation (2) above.

It is clear that
F (x0 ) = F (x1 ) = · · · = F (xn ) = F (x 0 ) = 0
that is, F (x) vanishes (n + 2) times in the interval x0 ≤ x ≤ xn .

Consequently, by the repeated application of Rolle’s theorem, F 0 (x) must

vanish (n + 1) times, F 00 (x) must vanish n times, etc,. in the interval
x0 ≤ x ≤ xn . In particular, F (n+1) (x) must vanish once in the interval.
Let this point be given by x = ξ, x0 < ξ < xn . On differentiating the

equation (3) (n + 1) times with respect to x and putting x = ξ, we obtain
F (n+1) (ξ) = 0 = y (n+1) (ξ) − L(n + 1)!
so that
y (n+1) (ξ)
L= . (4)
(n + 1)!

Comparison of (2) and (4) yields the results
y (n+1) (ξ)
y (x 0 ) − pn (x 0 ) = πn+1 (x 0 ).
(n + 1)!
Dropping the prime on x 0 , we obtain, for some x0 < ξ < xn ,
(x − x0 )(x − x1 ) · · · (x − xn ) (n+1)
y (x) − pn (x) = y (ξ) (5)
(n + 1)!
which is the required expression for the error. Since y (x) is, generally,
unknown and hence we do not have any information concerning y (n+1) (x),
formula (5) is almost useless in practical computations.
On the other hand, it is extremely useful in theroetical work in

different branches of numerical analysis.

Newton’s Interpolation Formulae for Equally Spaced Points
Given the set of (n + 1) values,
(x0 , y0 ), (x1 , y1 ), (x2 , y2 ), . . . , (xn , yn ),
of x and y , it is required to find pn (x), a polynomial of the nth degree

such that y and pn (x) agree at the tabulated points.
Let the values of x be equidistant,
xi = x0 + ih, i = 0, 1, 2, . . . , n.
Since pn (x) is a polynomial of the nth degree, it may be written as
pn (x) = a0 +a1 (x −x0 )+a2 (x −x0 )(x −x1 )+· · ·+an (x −x0 )(x −x1 ) · · · (x −xn−1 ).

Polynomial Coefficients
Imposing the condition that y and pn (x) should agree at the set of
tabulated points, we obtain
a0 = y0
y1 − y0 ∆y0
a1 = =
x1 − x0 h
∆2 y 0
a2 =
h2 2!
∆3 y 0
a3 =
h3 3!
..
.
∆n y0
an = .
hn n!

Interpolating Polynomial
Therefore
∆y0 ∆2 y0
pn (x) = y0 + (x − x0 ) + 2 (x − x0 )(x − x1 ) + · · ·
h h 2!
∆n y0
··· + (x − x0 )(x − x1 ) · · · (x − xn−1 )
hn n!
is the polynomial of degree n agreeing with the (unknown) function y at
the tabulated points.

Remainder Term (Error) in Polynomial Interpolation
Theorem 4.
Let f (x) be a function defined in (a, b) and suppose that f (x) have n + 1
continuous derivatives on (a, b). If a ≤ x0 < x1 < · · · < xn ≤ b, then
(x − x0 )(x − x1 ) · · · (x − xn ) (n+1)
f (x) − pn (x) = f (ξ),
(n + 1)!
for some ξ between x and x0 depending on x0 , x1 , . . . , xn and f .

Newton’s Forward Difference Interpolation Formula
Setting x = x0 + ph and substituting for a0 , a1 , . . . , an , the above equation

becomes
p(p − 1) 2 p(p − 1)(p − 2) 3
pn (x) = y0 + p∆y0 + ∆ y0 + ∆ y0 + · · ·
2! 3!
p(p − 1)(p − 2) · · · (p − n + 1) n
··· + ∆ y0
n!
which is (Gregory)-Newton’s forward difference interpolation formula
and is useful for interpolation near the beginning of a set of tabular
values and is useful for extrapolating the values of y (to the left of y0 ).
The first two terms of Newton’s forward formula give the linear
interpolation while the first three terms give a parabolic interpolation
and so on.

Exercises
1. Prove that
(a) y2 = y0 + 2∆y0 + ∆2 y0 .
(b) y3 = y0 + 3∆y0 + 3∆2 y0 + ∆3 y0 .
2. Show that h i
(a) ∆(tan−1 x) = tan−1 h
1+hx+x 2 .
h i
4h
(b) ∆ log(1 + 4x) = log 1 + 1+4x .
3. Evaluate
(a) ∆3 (1 − x)(1 − 2x)(1 − 3x) if h = 1.
(b) ∆10 (1 − x)(1 − 2x)(1 − 3x) . . . (1 − 10x) taking h = 1.
(c) ∆10 [(1 − x)(1 − 2x 2 )(1 − 3x 3 )(1 − 4x 4 )] if h = 2.
(d) ∆10 [(1 − ax)(1 − bx 2 )(1 − cx 3 )(1 − dx 4 )].

Exercises
1. If f (x) and g (x) are any two functions of x, prove that
∆[f (x) g (x)] = f (x) ∆g (x) + g (x + h) ∆f (x).
Hence, evaluate ∆(x sin x).

2. Evaluate
(a) ∆n (abx+c ).
(b) ∆n [cos(ax + b)].
3. Prove that
(a) yn−2 = yn − 2∇yn + ∇2 yn .
(b) yn−3 = yn − 3∇yn + 3∇2 yn − ∇3 yn .

Exercises
1. Prove that
(a) (∆ − ∇)f (x) = ∆∇f (x).
(b) ∆∇ and ∆ − ∇ are equal operators. That is, ∆∇ = ∆ − ∇.
2. Prove that ∆[f (x − 1)∆g (x − 1)] = ∇[f (x)∆g (x)].
[ Hint : ∆[f (x) g (x)] = f (x) ∆g (x) + g (x + h) ∆f (x). ]

Shift Operator
Let us recall the following expression for the first forward difference of a
function f (x) for equally spaced values of x with step-length h:
∆f (x) = f (x + h) − f (x).
This may be rewritten as
f (x + h) = f (x) + ∆f (x)
= (1 + ∆)f (x).
We denote (1 + ∆) by E and call it as the (first order) shift operator.

Thus,
E =1+∆
Ef (x) = f (x + h).

Shift Operator
Since ∆ is a linear operator, it automatically follows that E is also a linear

operator.
We define E 2 f (x), E 3 f (x), E 4 f (x), . . . by
E 2 f (x) = E {Ef (x)} = Ef (x + h) = f (x + 2h)

E 3 f (x) = E {E 2 f (x)} = Ef (x + 2h) = f (x + 3h)
E 4 f (x) = E {E 3 f (x)} = Ef (x + 3h) = f (x + 4h)
and so on. In general, E n f (x) is defined by
E n f (x) = f (x + nh), n = 1, 2, 3, . . . .
E n is called the nth order shift order as it shifts the value of the function at
x to the value at x + nh. We can write the formula in an alternative form:
E n ym = ym+n .
Shift Operator
Exercise 5.
Let the function y = f (x) take the values y0 , y1 , . . . , yn corresponding to
the values x0 , x0 + h, . . . , x0 + nh of x. Suppose f (x) is a polynomial of
degree n and it is required to evaluate f (x) for x = x0 + ph, where p is a
any real number. Derive Newton’s forward difference interpolation
formula, by using shift operator E .
[ Hint : yp = f (x) = f (x0 + ph) = E p f (x0 ) = (1 + ∆)p y0 . ]

Exercises
1. Evaluate
(a) E (x 2 − 3x + 1).
(b) (E 2 + 1)(x 3 + 2x 2 − 5x).
(c) E −1 (x 5 − x 3 + 1).
(d) E −3 (x 7 + x 5 + x 3 + x + 1).
2. Evaluate
(a) E 3 e x and then E n e x , where x varies by a constant interval h.
(b) (E − 1)(E − 2)x.
(c) (E − 3)(E + 4)3x .
(d) (E + 2)(E − 1)(2x/h + x).
3. Expand the following :
(a) (2E − 3)(5∆ + 4)x 2 .
(b) ∆2 E −3 x 5 and prove that ∆3 y3 = ∆3 y6 .

Exercises
∆2 ∆2 sin(x+h)
1. Evaluate E sin(x + h) + E sin(x+h) .
2. Estimate the missing term in the following table :
x 0 1 2 3 4
y=f(x) 1 2 6 ? 51
3. If u0 = 5, u1 = 11, u2 = 22, u3 = 40, u5 = 140, then find u4 , given
that the general term is represented by a fourth-degree polynomial.
4. Find log710 and log11 x
10 from the data given f (x) = log10 for some values
of x.
x 6 7 8 9 10 11 12
f (x) = logx10 0.77815 ? 0.90309 0.95424 1.0000 ? 1.07918
[Hint : 5 values of x and the corresponding values of y = f (x) are given, so ∆5 f (x) = 0,
for all x. By taking x as x0 and x1 , we get two expressions in y0 , y1 , . . . , y6 .]

Forward Difference Table
The values inside the boxes of the following difference table are used in
deriving the Newton’s forward difference interpolation formula.
Value Value First Second Third Fourth

of of Difference Difference Difference Difference
x y = f (x) ∆f (x) ∆2 f (x) ∆3 f (x) ∆4 f (x)
x0 y0
∆y0
x0 + h y1 ∆2 y 0
∆y1 ∆3 y 0
x0 + 2h y2 ∆2 y 1 ∆4 y 0
∆y2 ∆3 y 1
x0 + 3h y3 ∆2 y 2
∆y3
x0 + 4h y4

Error in Newton’s Forward Difference Interpolation Formula
To find the error committed in replacing the function y (x) by means of

the polynomial pn (x), we obtain
(x − x0 )(x − x1 ) . . . (x − xn ) (n+1)
y (x) − pn (x) = y (ξ)
(n + 1)!
for some ξ ∈ (x0 , xn ).
The error in the Newton’s forward difference interpolation formula is
p(p − 1(p − 2) · · · (p − n) n+1 (n+1)

y (x) − pn (x) = h y (ξ)
(n + 1)!
for some ξ ∈ (x0 , xn ), and x = x0 + ph.

As remarked earlier we do not have any information concerning y (n+1) (x),

and therefore the above formula is useless in practice.
Neverthless, if y (n+1) (x) does not vary too rapidly in the interval, a useful
estimate of the derivative can be obtained in the following way. Expanding
y (x + h) by Taylor’s series, we obtain
h2 00
y (x + h) = y (x) + hy 0 (x) + y (x) + · · · .
2!
Neglecting the terms containing h2 and higher powers of h, this gives
y (x + h) − y (x) ∆y (x)
y 0 (x) ≈ =
h h

Writing y 0 (x) as Dy (x) where D ≡ d/dx, the differentiation operator,

the above equation gives the operator relations
1 1
D≡ ∆ and so D n+1 ≡ n+1 ∆n+1 .
h h
We thus obtain
1
y (n+1) (x) ≈ ∆n+1 y (x).
hn+1
Hence the equation can be written as (equally spaced nodes, x = x0 + ph)
p(p − 1)(p − 2) · · · (p − n) n+1

y (x) − pn (x) = ∆ y (ξ)
(n + 1)!
for some ξ ∈ (x0 , xn ), which is suitable for computation.

Newton’s Backward Difference Interpolation Formula
Suppose we assume pn (x) in the following form
pn (x) = a0 + a1 (x − xn ) + a2 (x − xn )(x − xn−1 ) + · · ·
· · · + an (x − xn )(x − xn−1 ) · · · (x − x1 )
and then impose the condition that y and pn (x) should agree at the
tabulated points xn , xn−1 , . . . , x2 , x1 , x0 , we obtain (after some
simplification)
p(p + 1) 2 p(p + 1) · · · (p + n − 1) n
pn (x) = yn + p∇yn + ∇ yn + · · · + ∇ yn
2! n!
where p = (x − xn )/h.
This is (Gregory)-Newton’s backward difference interpolation

formula and it uses tabular values to the left of yn . This formula is
therefore useful for interpolation near the end of the tabular values and is
useful for extrapolating values of y (to the right of yn ).
Backward Difference Table
The values inside the boxes of the following difference table are used in
deriving the Newton’s backward difference interpolation formula.
Value Value First Second Third Fourth

of of Difference Difference Difference Difference
x y = f (x) ∇f (x) ∇2 f (x) ∇3 f (x) ∇4 f (x)
x0 y0
∇y1
x0 + h y1 ∇2 y 2
∇y2 ∇3 y 3
x0 + 2h y2 ∇2 y 3 ∇4 y 4
∇y3 ∇3 y 4
x0 + 3h y3 ∇2 y 4
∇y4
x0 + 4h y4

Error in Newton’s Backward Difference Interpolation
Formula
It can be shown that the error in this formula may be written as
p(p + 1)(p + 2) · · · (p + n) n+1 (n+!)

y (x) − pn (x) = h y (ξ)
(n + 1)!
where x0 < ξ < xn and x = xn + ph.

Taylor’s Theorem
Theorem 6.
Let f (x) have n + 1 continuous derivatives on [a, b] for some n ≥ 0, and
let x, x0 ∈ [a, b]. Then f (x) = pn (x) + Rn (x)
where
n
X (x − x0 )k (k)
pn (x) = f (x0 ) (n-degree polynomial)
k!
k=0
and
(x − x0 )n+1 (n+1)
Rn (x) = f (ξ) (error term)
(n + 1)!
for some ξ between x and x0 .
Hence “ξ between x and x0 ” means that either x0 < ξ < x or x < ξ < x0
depending on the particular values of x and x0 involved.

Taylor Polynomials
Here pn (X ) is called the nth Taylor polynomial for f about x0 , and Rn is

called the remainder term (or truncation error) associated with pn (x).
Since the number ξ in the truncation error Rn depends on the value of x
at which the polynomial pn (x) is being evaluated, it is a function of the
variable x.
Taylor’s theorem simply ensures that such a function exists, and that its
value lies between x and x0 .
In fact, one of the common problems in numerical methods is to try to

determine a realistic bound for the value of f (n+1) (ξ) when x is within
some specified interval.

Taylor polynomials are not useful for interpolation
The Taylor polynomials are one of the fundamental building blocks of

numerical analysis.
The Taylor polynomials agree as closely as possible with a given function

at a specific point, but they concentrate their accuracy near that point.
A good interpolation needs to provide a relatively accurate

approximation over an entire interval, and Taylor polynomials do not
generally do this.

Example
Example 7.
Taylor polynomials of various degree for f (x) = 1/x about x0 = 1 are
n
X
pn (x) = (−1)k (x − 1)k .
k=0
When we approximate f (3) = 1/3 by pn (3) for larger values of n, the

approximations become increasingly inaccurate, as shown in the following
table.
n 0 1 2 3 4 5 6 7
pn (3) 1 -1 3 -5 11 -21 43 -85

Taylor polynomials are not appropriate for interpolation
Since the Taylor polynomials have the property that all the information
used in the approximation is concentrated at the single point x0 , it is not
uncommon for these polynomials to give inaccurate approximations as we
move away from x0 . This limits Taylor polynomial approximation to the
situation in which approximations are needed only at points close to x0 .
For ordinary computational purposes it is more efficient to use methods

that include information at various points.
The primary use of Taylor polynomials in numerical analysis is not for

approximation purposes, but for the derivation of numerical techniques
and for error estimation.
Since the Taylor polynomials are not appropriate for interpolation,

alternative methods are needed.

Interpolation with Unequal Intervals
The problem of determining a polynomial of degree one that passes

through the distinct points (x0 , y0 ) and (x1 , y1 ) is the same as
approximating a function f for which f (x0 ) = y0 and f (x1 ) = y1 by means
of a first-degree polynomial interpolating, or agreeing with, the values of f
at the given points.
We first define the functions

x − x1 x − x0
L0 (x) = and L1 (x) = ,
x0 − x1 x1 − x0
and then define
x − x1 x − x0
p1 (x) = L0 (x)y0 + L1 (x)y1 = y0 + y1 .
x0 − x1 x1 − x0
Since L0 (x0 ) = 1, L0 (x1 ) = 0, L1 (x0 ) = 0, and L1 (x1 ) = 1, we have
p1 (x0 ) = y0 and p1 (x1 ) = y1 . So p1 is the unique linear function passing
(x0 , y0 ) and (x1 , y1 ).
Exercises
Exercises 8.
1. The table gives the distance in nautical miles of the visible horizon for
the given heights (in feet) above the earth’s surface.
x 100 150 200 250 300 350 400
y = f (x) 10.63 13.03 15.04 16.81 18.42 19.90 21.27
Find the values of y when x = 160 and x = 410.
2. From the following table, estimate the number of students who
obtained marks between 40 and 45.
Marks 30-40 40-50 50-60 60-70 70-80
No. of Students 31 42 51 35 31
3. Find the cubic polynomial which takes the following values.
x 0 1 2 3
y = f (x) 1 2 1 10
Also compute f (4).

Exercises
Exercises 9.
4. In the table below, the values of y are consecutive terms of a series of
which 23.6 is the 6th term. Find the first and tenth terms of the
series.
x 3 4 5 6 7 8 9
y = f (x) 4.8 8.4 14.5 23.6 36.2 52.8 73.9
5. Using Newton’s forward interpolation formula, show that
n 2
X
3 n(n + 1)
k = .
2
k=1

An Observation of Central Differences
We derived Newton’s forward and backward interpolation formulae which

are applicable for interpolation near the beginning and end of tabulated
values.
The following formulae are based on central differences which are best
suited for interpolation near the middle of the table.
Gauss’s forward interpolation formula

Gauss’s backward interpolation formula
Stirling’s formula
Bessel’s formula
Laplace-Everett’s formula.
The coefficients in the above central difference formulae are smaller and
converge faster than those in Newton’s formulae.
References
1. Richard L. Burden and J. Douglas Faires, Numerical Analysis -

Theory and Applications, Cengage Learning, Singapore.
2. Kendall E. Atkinson, An Introduction to Numerical Analysis, Wiley
India.
3. David Kincaid and Ward Cheney, Numerical Analysis -
Mathematics of Scientific Computing, American Mathematical
Society, Providence, Rhode Island.
4. S.S. Sastry, Introductory Methods of Numerical Analysis, Fourth
Edition, Prentice-Hall, India.

Newton's Interpolation Methods: P. Sam Johnson

Uploaded by

Copyright:

Available Formats

Newton's Interpolation Methods: P. Sam Johnson

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Newton's Interpolation Methods: P. Sam Johnson

Uploaded by

Copyright:

Available Formats

Newton’s Interpolation Methods

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 1 / 47

One of the basic ideas in Mathematics is that of a function and most

According to Thiele (a numerical analyst), “Interpolation is the art of

Broadly speaking, interpolation is the problem of obtaining the value of a

Interpolation technique is used in various disciplines like economics,

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 2 / 47

The concept of interpolation is the selection of a function p(x) from a

If p(x) is a polynomial, then it is called the interpolating polynomial and

Similarly, if p(x) is a finite trigonometric series, we have trigonometric

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 3 / 47

The study of interpolation is based on the calculus of finite differences.

Polynomial interpolation theory has a number of important uses. Its

We discuss Newtons forward/backward formulae (for equally spaced

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 4 / 47

A census of the population of the India is taken every 10 years. The

Year 1951 1961 1971 1981 1991 2001 2011

Population 361,088 439,235 548,160 683,329 846,388 1,028,737 1,210,193

In reviewing these data, we might ask whether they could be used to

This process is called interpolation / extrapolation.

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 5 / 47

If y is a function of x, then the functional relation may be denoted by the

We call x as the independent variable and y as the dependent variable.

Since the polynomials are relatively simple to deal with, we interpolate to

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 6 / 47

If the value of x whose corresponding value y is to be estimated lies within

In the method of interpolation, it is assumed that the function is capable

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 7 / 47

Given any function, defined and continuous on a closed and bounded

Theorem 1 (Weierstrass Approximation Theorem).

|f (x) − p(x)| < ε, for all x ∈ [a, b].

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 8 / 47

In science and engineering,

An important reason for considering the class of polynomials in the

For these reasons, polynomials are often used for approximating

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 9 / 47

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 10 / 47

We recall Rolle’s theorem, which is useful in evaluating error bounds.

Theorem 2 (Rolle’s Theorem).

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 11 / 47

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 12 / 47

Let the function y (x), defined by the (n + 1) points

be continuous and differentiable (n + 1) times, and let y (x) be

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 13 / 47

Using the polynomial pn (x) of degree n, we can obtain approximate values

Since the expression y (x) − pn (x) vanishes for x = x0 , x1 , . . . , xn we put

y (x) − pn (x) = Lπn+1 (x) (1)

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 14 / 47

We construct a function F (x) such that

F (x) = y (x) − pn (x) − Lπn+1 (x) (3)

where L is given by the equation (2) above.

F (x0 ) = F (x1 ) = · · · = F (xn ) = F (x 0 ) = 0

that is, F (x) vanishes (n + 2) times in the interval x0 ≤ x ≤ xn .

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 15 / 47

Consequently, by the repeated application of Rolle’s theorem, F 0 (x) must

Let this point be given by x = ξ, x0 < ξ < xn . On differentiating the

F (n+1) (ξ) = 0 = y (n+1) (ξ) − L(n + 1)!

P. Sam Johnson (NITK) Newton’s Interpolation Methods February 7, 2020 16 / 47

Comparison of (2) and (4) yields the results