This document discusses various methods of interpolation and numerical differentiation using divided differences and Newton's formulas. It introduces Lagrange interpolation for both equal and unequal intervals. Inverse interpolation and Newton's divided difference interpolation are also covered. Forward and backward difference formulas are presented for interpolation with equal intervals. Numerical differentiation can be performed by taking derivatives of the interpolation polynomial or using forward difference formulas to estimate derivatives at the data points.
2. INTERPOLATION FOR UNEQUAL INTERVALS
LAGRANGE’S INTERPOLATING
FORMULA
Let y = f(x) be the given function
Let 𝑦0, 𝑦1, … 𝑦𝑛 be the (𝑛 + 1) points of the
given function corresponding to 𝑥0, 𝑥1, … 𝑥 𝑛
The polynomial 𝑦 = 𝑓 𝑥 can be written as
𝑓 𝑥 =
𝑥−𝑥1 𝑥−𝑥2 ··· 𝑥−𝑥 𝑛
𝑥0−𝑥1 𝑥0−𝑥2 ··· 𝑥0−𝑥 𝑛
𝑦0 +
𝑥 − 𝑥0 𝑥 − 𝑥2 ··· 𝑥 − 𝑥 𝑛
𝑥1 − 𝑥0 𝑥1 − 𝑥2 ··· 𝑥1 − 𝑥 𝑛
𝑦1 +
𝑥 − 𝑥0 𝑥 − 𝑥1 ··· 𝑥 − 𝑥 𝑛
𝑥2 − 𝑥0 𝑥2 − 𝑥1 ··· 𝑥2 − 𝑥 𝑛
𝑦2
+ ⋯ +
𝑥−𝑥0 𝑥−𝑥1 ··· 𝑥−𝑥 𝑛−1
𝑥 𝑛−𝑥0 𝑥 𝑛−𝑥1 ··· 𝑥 𝑛−𝑥 𝑛−1
𝑦 𝑛
3. INVERSE INTERPOLATION
It is the process of finding a value of x
for the corresponding value of y and
we use Lagrange’s interpolation
formula by taking the independent
variable as y and the dependent
variable as x. It is the inverse process
of direct interpolation in which we find
the values of y corresponding to a
value of x, not present in the table.
5. USE OF LAGRANGIAN
INTERPOLATION
It is a process of computing
intermediate values of a function from
a given set of tabular values of the
function.
6. DIVIDED DIFFERENCE
Let 𝑦 = 𝑓(𝑥) be the given function
which takes the values
f(𝑥0), 𝑓(𝑥1) … 𝑓(𝑥 𝑛) corresponding to
the arguments 𝑥0, 𝑥1, … 𝑥 𝑛
respectively, where the intervals
𝑥1 – 𝑥0 , 𝑥2 − 𝑥1 , … 𝑥 𝑛 – 𝑥 𝑛−1 need not
equal
8. PROPERTIES OF DIVIDED DIFFERENCES
1. The divided differences are
symmetrical in all their arguments, that
is, the value of any difference is
independent of the order of the
arguments.
2. The divided difference of the product
of a constant and a function is equal to
the product of the constant and the
divided difference of the function.
3. The operator Δ1 is linear
4. The 𝑛 𝑡ℎ
divided difference of a
polynomial of degree 𝑛 is a constant
9. NEWTON DIVIDED DIFFERENCE INTERPOLATION
If 𝑓(𝑥) is a polynomial of degree𝑛, and
𝑓(𝑥0), 𝑓(𝑥1), … 𝑓(𝑥 𝑛) are the
corresponding values of arguments
𝑥0, 𝑥1, … . 𝑥 𝑛 respectively, not
necessarly equally spaced.
Then 𝑓 𝑥 = 𝑓 𝑥0 + ( 𝑥 −
10. INTERPOLATION WITH EQUAL INTERVALS
FORWARD DIFFERENCE
If 𝑦0, 𝑦1, … 𝑦𝑛 are the values of 𝑦 = 𝑓(𝑥)
corresponding to the arguments 𝑥0, 𝑥1, … 𝑥 𝑛 ,
where 𝑥1 − 𝑥0 , 𝑥2 − 𝑥1, … 𝑥 𝑛 − 𝑥 𝑛−1 are equal
𝑖. 𝑒 𝑥𝑖 = 𝑥0 + 𝑖ℎ, 𝑖 = 0,1,2 … 𝑛
Define Δ𝑦0 = 𝑦1 − 𝑦0 , Δ𝑦1 = 𝑦2 − 𝑦1 … . Δ𝑦 𝑛−1 =
𝑦𝑛 − 𝑦 𝑛−1
And Δ2
𝑦0 = Δ𝑦1 − Δ𝑦0 , Δ2
𝑦1 = Δ𝑦2 −
Δ𝑦1 … . Δ2
𝑦 𝑛−1 = Δ𝑦𝑛 − Δ𝑦 𝑛−1
And so on
Here Δ is called Newton’s forward difference
operator
11. NEWTON’S FORWARD
DIFFERENCE FORMULA
If 𝑦0, 𝑦1, … 𝑦𝑛 are the values of 𝑦 =
𝑓(𝑥) corresponding to the arguments
𝑥0, 𝑥1, … 𝑥 𝑛
Then y = 𝑦0 + 𝑝∆𝑦0 +
𝑝 𝑝−1
2!
∆2
𝑦0 +
𝑝 𝑝−1 𝑝−2
3!
∆3
𝑦0 + ⋯
Where 𝑝 =
𝑥−𝑥0
ℎ
.
17. DERIVATIVES USING DIVIDED DIFFERENCE
Procedure
Step 1. By using Newton’s divided
difference formula find 𝑓(𝑥) in terms of
x
Step 2. Find the derivatives of 𝑓(𝑥)
18. DERIVATIVE - USING NEWTON’S FORWARD DIFFERENCE
FORMULA
The first derivative of y at 𝑥 = 𝑥0 + 𝑝ℎ (near beginning of the data)
is
𝑑𝑦
𝑑𝑥
=
𝑑𝑦
𝑑𝑝
𝑑𝑝
𝑑𝑥
𝑑𝑦
𝑑𝑥
=
1
ℎ
∆𝑦0 +
2𝑝−1
2!
∆2 𝑦0 +
3𝑝2−6𝑝+2
3!
∆3 𝑦0 + ⋯
The second derivative of y at 𝑥 = 𝑥0 + 𝑝ℎ is
𝑑2 𝑦
𝑑𝑥2 =
𝑑2 𝑦
𝑑𝑝2
𝑑𝑝
𝑑𝑥
2
𝑑2 𝑦
𝑑𝑥2 =
1
ℎ2 ∆2
𝑦0 + (𝑝 − 1)∆3
𝑦0 +
6𝑝2−18𝑝+11
12
∆4
𝑦0 + ⋯
The third derivative y at 𝑥 = 𝑥0 + 𝑝ℎ
𝑑3 𝑦
𝑑𝑥3 =
1
ℎ3 ∆3
𝑦0 +
12𝑝−18
12
∆4
𝑦0+. .
For tabular values, at 𝑥 = 𝑥0, (𝑝 = 0)
𝑑𝑦
𝑑𝑥
=
1
ℎ
∆𝑦0 −
∆2 𝑦0
2
+
∆3 𝑦0
3
−
∆4 𝑦0
4
+ ⋯
𝑑2 𝑦
𝑑𝑥2 =
1
ℎ2 ∆2
𝑦0 − ∆3
𝑦0 +
11
12
∆4
𝑦0 − ⋯
𝑑3 𝑦
𝑑𝑥3 =
1
ℎ3 ∆3 𝑦0 −
3
2
∆4 𝑦0+. .
19. DERIVATIVE - USING NEWTON’S BACKWARD DIFFERENCE
FORMULA
The first derivative of y at 𝑥 = 𝑥 𝑛 + 𝑝ℎ (near end of the data) is
𝑑𝑦
𝑑𝑥
=
𝑑𝑦
𝑑𝑝
𝑑𝑝
𝑑𝑥
𝑑𝑦
𝑑𝑥
=
1
ℎ
𝛻𝑦 𝑛 +
2𝑝+1
2!
𝛻2 𝑦 𝑛 +
3𝑝2+6𝑝+2
3!
𝛻3 𝑦 𝑛 + ⋯
The second derivative of y at 𝑥 = 𝑥 𝑛 + 𝑝ℎ is
𝑑2 𝑦
𝑑𝑥2 =
𝑑2 𝑦
𝑑𝑝2
𝑑𝑝
𝑑𝑥
2
𝑑2 𝑦
𝑑𝑥2 =
1
ℎ2 𝛻2
𝑦 𝑛 + (𝑝 + 1)𝛻3
𝑦 𝑛 +
6𝑝2+18𝑝+11
12
𝛻4
𝑦0 + ⋯
The third derivative of y at 𝑥 = 𝑥 𝑛 + 𝑝ℎ
𝑑3 𝑦
𝑑𝑥3 =
1
ℎ3 𝛻3
𝑦 𝑛 +
12𝑝+18
12
𝛻4
𝑦0+. .
For tabular values, at 𝑥 = 𝑥 𝑛, (𝑝 = 0)
𝑑𝑦
𝑑𝑥
=
1
ℎ
𝛻𝑦 𝑛 +
𝛻2 𝑦 𝑛
2
+
𝛻3 𝑦 𝑛
3
+ ⋯
𝑑2 𝑦
𝑑𝑥2 =
1
ℎ2 𝛻2
𝑦 𝑛 + 𝛻3
𝑦 𝑛 +
11
12
𝛻4
𝑦0 + ⋯
𝑑3 𝑦
𝑑𝑥3 =
1
ℎ3 𝛻3 𝑦 𝑛 +
3
2
𝛻4 𝑦0+. .
27. Errors in Trapezoidal rule of numerical
integration.
When evaluating 𝑎
𝑏
𝑓 𝑥 𝑑𝑥, the error
in the trapezoidal rule is
<
𝑏−𝑎 2
12
ℎ2
𝑀, where ℎ =
𝑏−𝑎
𝑛
, n is the
number of subintervals of (a, b),
And 𝑀 = 𝑚𝑎𝑥{|𝑦0
′′
|, |𝑦1
′′
|,· · ·
, |𝑦 𝑛−1
′′
|}, 𝑦𝑟′′ = 𝑓′′(𝑥 𝑟)
Error in Trapezoidal rule is of order
𝒉 𝟐
28. Errors in Simpson’s rule of numerical
integration
The error in Simpson’s rule is <
𝑏−𝑎
180
ℎ4
𝑀
where ℎ =
𝑏−𝑎
2𝑛
, 2n is the number of
subintervals of (a, b),
𝑀 = 𝑚𝑎𝑥 𝑦0
1
, 𝑦1
4
,· · ·