Introduction to Numerical Analysis

(Numerical Integration)

MA 214, Spring 2023-24.

MA 214 - NA Spring 2023-24 1 / 40

Numerical Integration

MA 214 - NA Spring 2023-24 2 / 40

Numerical Integration

The problem is to evaluate the number

∫ b
I(f ) = f (x)dx.
a

MA 214 - NA Spring 2023-24 3 / 40

Numerical Integration

The problem is to evaluate the number

∫ b
I(f ) = f (x)dx.
a

The process of approximating the value of I(f ) is usually referred to as

numerical integration or quadrature rule.

MA 214 - NA Spring 2023-24 3 / 40

Numerical Integration

The problem is to evaluate the number

∫ b
I(f ) = f (x)dx.
a

The process of approximating the value of I(f ) is usually referred to as

numerical integration or quadrature rule.

IDEA

f (x) ≈ Simpler function.

MA 214 - NA Spring 2023-24 3 / 40

Numerical Integration
The problem is to evaluate the number
∫ b
I(f ) = f (x)dx.
a

The process of approximating the value of I(f ) is usually referred to as

numerical integration or quadrature rule.
IDEA

∫b ∫b
f (x) dx ≈ Simpler function dx.
a a
MA 214 - NA Spring 2023-24 3 / 40
Interpolating Polynomial (Recall)

Theorem
For a given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

MA 214 - NA Spring 2023-24 4 / 40

Interpolating Polynomial (Recall)

Theorem
For a given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

there ALWAYS exists an interpolating polynomial

MA 214 - NA Spring 2023-24 4 / 40

Interpolating Polynomial (Recall)

Theorem
For a given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

there ALWAYS exists an interpolating polynomial

and

this interpolating polynomial is UNIQUE

MA 214 - NA Spring 2023-24 4 / 40
Lagrange form of Interpolating Polynomial
Formula 1: Lagrange Form
Theorem (Lagrange form of Interpolating Polynomial)
Given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

Lagrange form of interpolating polynomial is given by

∑
n ∏
n
(x − xi )
pn (x) = yk lk (x), lk (x) = .
(xk − xi )
k=0 i=0
i6=k
MA 214 - NA Spring 2023-24 5 / 40
Newton form of Interpolating Polynomial (contd.)
Theorem (Newton form of Interpolating Polynomial)
Given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

MA 214 - NA Spring 2023-24 6 / 40

Newton form of Interpolating Polynomial (contd.)
Theorem (Newton form of Interpolating Polynomial)
Given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

The Newton form of interpolating polynomial pn (x) can be written

as
∏
pn (x) = A0 + A1 (x − x0 ) +A2 (x − x0 )(x − x1 ) + A3 2i=0 (x − xi )
∏
+ · · · + An n−1i=0 (x − xi )

MA 214 - NA Spring 2023-24 6 / 40

Newton form of Interpolating Polynomial (contd.)
Theorem (Newton form of Interpolating Polynomial)
Given data set
x x0 x1 x2 x3 · · · xn
,
y y0 y1 y2 y3 · · · yn

The Newton form of interpolating polynomial pn (x) can be written

as
∏
pn (x) = A0 + A1 (x − x0 ) +A2 (x − x0 )(x − x1 ) + A3 2i=0 (x − xi )
∏
+ · · · + An n−1i=0 (x − xi )

where A0 , A1 , · · · , An are constants obtained to satisfy interpolation

conditions.
MA 214 - NA Spring 2023-24 6 / 40
Newton’s Divided Differences
General Data Set
x x0 x1 x2 x3 · · · xn
.
f(x) f(x0 ) f(x1 ) f(x2 ) f(x3 ) · · · f(xn )

MA 214 - NA Spring 2023-24 7 / 40

Newton’s Divided Differences
General Data Set
x x0 x1 x2 x3 · · · xn
.
f(x) f(x0 ) f(x1 ) f(x2 ) f(x3 ) · · · f(xn )
Newton form of interpolating polynomial
∏
pn (x) = A0 + A1 (x − x0 ) +A2 (x − x0 )(x − x1 ) + A3 2i=0 (x − xi )
∏
+ · · · + An n−1
i=0 (x − xi ),
A0 = f (x0 )

MA 214 - NA Spring 2023-24 7 / 40

Newton’s Divided Differences
General Data Set
x x0 x1 x2 x3 · · · xn
.
f(x) f(x0 ) f(x1 ) f(x2 ) f(x3 ) · · · f(xn )
Newton form of interpolating polynomial
∏
pn (x) = A0 + A1 (x − x0 ) +A2 (x − x0 )(x − x1 ) + A3 2i=0 (x − xi )
∏
+ · · · + An n−1
i=0 (x − xi ),
A0 = f (x0 )

f (x1 ) − p0 (x1 )
A1 =
(x1 − x0 )
MA 214 - NA Spring 2023-24 7 / 40
Newton’s Divided Differences
General Data Set
x x0 x1 x2 x3 · · · xn
.
f(x) f(x0 ) f(x1 ) f(x2 ) f(x3 ) · · · f(xn )
Newton form of interpolating polynomial
∏
pn (x) = A0 + A1 (x − x0 ) +A2 (x − x0 )(x − x1 ) + A3 2i=0 (x − xi )
∏
+ · · · + An n−1
i=0 (x − xi ),
A0 = f (x0 )

f (x1 ) − p0 (x1 )
A1 =
(x1 − x0 )
MA 214 - NA Spring 2023-24 7 / 40
Newton’s Divided Differences

General Data Set

x x0 x1 x2 x3 · · · xn
.
f(x) f(x0 ) f(x1 ) f(x2 ) f(x3 ) · · · f(xn )

Newton form of interpolating polynomial

In new notation, we write the interpolating polynomial as :

∑
n ∏
k−1
pn (x) = f [x0 ] + f [x0 , x1 , · · · , xk ] (x − xi )
k=1 i=0

MA 214 - NA Spring 2023-24 8 / 40

We then showed that :
f [x1 , x2 , · · · , xn ] − f [x0 , x1 , · · · , xn−1 ]
f [x0 , x1 , · · · , xn ] =
xn − x0

MA 214 - NA Spring 2023-24 9 / 40

We then showed that :
f [x1 , x2 , · · · , xn ] − f [x0 , x1 , · · · , xn−1 ]
f [x0 , x1 , · · · , xn ] =
xn − x0

This ratio is called nth -order divided difference.

Definition (Divided Differences)
Let x0 , x1 , · · · , xn be distinct nodes.

MA 214 - NA Spring 2023-24 9 / 40

We then showed that :
f [x1 , x2 , · · · , xn ] − f [x0 , x1 , · · · , xn−1 ]
f [x0 , x1 , · · · , xn ] =
xn − x0

This ratio is called nth -order divided difference.

Definition (Divided Differences)
Let x0 , x1 , · · · , xn be distinct nodes.
Let pn (x) be the polynomial interpolating a function f at the nodes
x0 , x1 , · · · , xn .

MA 214 - NA Spring 2023-24 9 / 40

We then showed that :
f [x1 , x2 , · · · , xn ] − f [x0 , x1 , · · · , xn−1 ]
f [x0 , x1 , · · · , xn ] =
xn − x0

This ratio is called nth -order divided difference.

Definition (Divided Differences)
Let x0 , x1 , · · · , xn be distinct nodes.
Let pn (x) be the polynomial interpolating a function f at the nodes
x0 , x1 , · · · , xn .
The coefficient of xn in the polynomial pn (x)

MA 214 - NA Spring 2023-24 9 / 40

We then showed that :
f [x1 , x2 , · · · , xn ] − f [x0 , x1 , · · · , xn−1 ]
f [x0 , x1 , · · · , xn ] =
xn − x0

This ratio is called nth -order divided difference.

Definition (Divided Differences)
Let x0 , x1 , · · · , xn be distinct nodes.
Let pn (x) be the polynomial interpolating a function f at the nodes
x0 , x1 , · · · , xn .
The coefficient of xn in the polynomial pn (x), is denoted by
f [x0 , x1 , · · · , xn ]

MA 214 - NA Spring 2023-24 9 / 40

We then showed that :
f [x1 , x2 , · · · , xn ] − f [x0 , x1 , · · · , xn−1 ]
f [x0 , x1 , · · · , xn ] =
xn − x0

This ratio is called nth -order divided difference.

Definition (Divided Differences)
Let x0 , x1 , · · · , xn be distinct nodes.
Let pn (x) be the polynomial interpolating a function f at the nodes
x0 , x1 , · · · , xn .
The coefficient of xn in the polynomial pn (x), is denoted by
f [x0 , x1 , · · · , xn ], is called an nth divided difference of f.

MA 214 - NA Spring 2023-24 9 / 40

Numerical Integration

The problem is to evaluate the number

∫ b
I(f ) = f (x)dx.
a

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn

MA 214 - NA Spring 2023-24 10 / 40

Numerical Integration

The problem is to evaluate the number

∫ b
I(f ) = f (x)dx.
a

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn Construct pn (x)

MA 214 - NA Spring 2023-24 10 / 40

Numerical Integration

The problem is to evaluate the number

∫ b
I(f ) = f (x)dx.
a

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn Construct pn (x)

I(f ) ≈ I(pn )

MA 214 - NA Spring 2023-24 10 / 40

Numerical Integration

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn Construct pn (x)

I(f ) ≈ I(pn )
The approximation is written as
( n )
∑
I(f ) ≈ I(pn ) = I f (xi )li (x)
i=0

MA 214 - NA Spring 2023-24 11 / 40

Numerical Integration

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn Construct pn (x)

I(f ) ≈ I(pn )
The approximation is written as
( n )
∑ ∑
n
I(f ) ≈ I(pn ) = I f (xi )li (x) = f (xi )I(li )
i=0 i=0

MA 214 - NA Spring 2023-24 11 / 40

Numerical Integration

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn Construct pn (x)

I(f ) ≈ I(pn )
The approximation is written as
( n )
∑ ∑
n
I(f ) ≈ I(pn ) = I f (xi )li (x) = f (xi )I(li )
i=0 i=0
= f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn ,
MA 214 - NA Spring 2023-24 11 / 40
Numerical Integration

Approximation

Given n + 1 nodes x0 , x1 , · · · , xn Construct pn (x)

I(f ) ≈ I(pn )
The approximation is written as
( n )
∑ ∑
n
I(f ) ≈ I(pn ) = I f (xi )li (x) = f (xi )I(li )
i=0 i=0
= f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn , wi = I(li )
MA 214 - NA Spring 2023-24 11 / 40
Quadrature Rule 1: Rectangle Rule

MA 214 - NA Spring 2023-24 12 / 40

Numerical Integration: Rectangle Rule

I(f ) ≈ I(pn ) = f (x0 )w0 +f (x1 )w1 + · · · + f (xn )wn ,

MA 214 - NA Spring 2023-24 13 / 40

Numerical Integration: Rectangle Rule

I(f ) ≈ I(pn ) = f (x0 )w0 +f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 0.

MA 214 - NA Spring 2023-24 13 / 40

Numerical Integration: Rectangle Rule

I(f ) ≈ I(pn ) = f (x0 )w0 +f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 0.
Then p0 (x) = f (x0 ), and therefore
I(p0 ) = (b − a)f (x0 ).

MA 214 - NA Spring 2023-24 13 / 40

Numerical Integration: Rectangle Rule

I(f ) ≈ I(pn ) = f (x0 )w0 +f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 0.
Then p0 (x) = f (x0 ), and therefore
I(p0 ) = (b − a)f (x0 ).
If x0 = a, then this approximation becomes
I(f ) ≈ IR (f ) := (b − a)f (a)
and is called rectangle rule.

MA 214 - NA Spring 2023-24 13 / 40

Numerical Integration: Rectangle Rule

I(f ) ≈ I(pn ) = f (x0 )w0 +f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 0.
Then p0 (x) = f (x0 ), and therefore
I(p0 ) = (b − a)f (x0 ).
If x0 = a, then this approximation becomes
I(f ) ≈ IR (f ) := (b − a)f (a)
y f (x)
and is called rectangle rule.

I(f)

a b x
MA 214 - NA Spring 2023-24 13 / 40
Numerical Integration: Rectangle Rule

I(f ) ≈ I(pn ) = f (x0 )w0 +f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 0.
Then p0 (x) = f (x0 ), and therefore
I(p0 ) = (b − a)f (x0 ).
If x0 = a, then this approximation becomes
I(f ) ≈ IR (f ) := (b − a)f (a)
y f (x)
and is called rectangle rule.

IRI(f)
(f )

a b x
MA 214 - NA Spring 2023-24 14 / 40
Numerical Integration: Rectangle Rule (contd.)

Mathematical Error

MER (f ) := I(f ) − I(p0 ).

MA 214 - NA Spring 2023-24 15 / 40

Numerical Integration: Rectangle Rule (contd.)

Mathematical Error

MER (f ) := I(f ) − I(p0 ).

Theorem (Error in Rectangle Rule)

Let f ∈ C1 [a, b].

MA 214 - NA Spring 2023-24 15 / 40

Numerical Integration: Rectangle Rule (contd.)

Mathematical Error

MER (f ) := I(f ) − I(p0 ).

Theorem (Error in Rectangle Rule)

Let f ∈ C1 [a, b].The mathematical error MER (f ) of the rectangle rule
takes the form
f 0 (η)(b − a)2
MER (f ) = ,
2
for some η ∈ (a, b).
MA 214 - NA Spring 2023-24 15 / 40
Numerical Integration: Rectangle Rule (contd.)

Proof

For each x ∈ (a, b], from the linear interpolating polynomial for f at the
nodes a and x, we can write

f (x) = p0 (x) + f [a, x](x − a).

MA 214 - NA Spring 2023-24 16 / 40

Numerical Integration: Rectangle Rule (contd.)

Proof

For each x ∈ (a, b], from the linear interpolating polynomial for f at the
nodes a and x, we can write

f (x) = p0 (x) + f [a, x](x − a).

Therefore, the mathematical error in the rectangle rule is given by

∫b
MER (f ) = I(f ) − IR (f ) = f [a, x](x − a) dx.
a

MA 214 - NA Spring 2023-24 16 / 40

Numerical Integration: Rectangle Rule (contd.)
Proof

Therefore, the mathematical error in the rectangle rule is given by

∫b
MER (f ) = I(f ) − IR (f ) = f [a, x](x − a) dx.
a

Using mean-value theorem for integrals, we get

∫b
MER (f ) = f [a, ξ] (x − a) dx,
a

for some ξ ∈ (a, b).

MA 214 - NA Spring 2023-24 17 / 40
Numerical Integration: Rectangle Rule (contd.)
Proof

Using mean-value theorem for integrals, we get

∫b
MER (f ) = f [a, ξ] (x − a) dx,
a

for some ξ ∈ (a, b).

By mean value theorem for derivatives, f [a, ξ] = f 0 (η) for some
η ∈ (a, ξ). Thus, we get (for some η ∈ (a, b))
f 0 (η)(b − a)2
MER (f ) = .
2
MA 214 - NA Spring 2023-24 18 / 40
Quadrature Rule 1(a): Mid-Point Rule

MA 214 - NA Spring 2023-24 19 / 40

Numerical Integration: Mid-Point Rule

If x0 = (a + b)/2, we get
( )
a+b
I(f ) ≈ IM (f ) := (b − a)f
2
and is called the mid-point rule.

MA 214 - NA Spring 2023-24 20 / 40

Numerical Integration: Mid-Point Rule

If x0 = (a + b)/2, we get
( )
a+b
I(f ) ≈ IM (f ) := (b − a)f
2
and is called the mid-point rule.

Mathematical error is left as an exercise

MA 214 - NA Spring 2023-24 20 / 40

Numerical Integration: Composite Rectangle Rule

To improve on the approximation

MA 214 - NA Spring 2023-24 21 / 40

Numerical Integration: Composite Rectangle Rule

To improve on the approximation

break the interval [a, b] into smaller subintervals and

MA 214 - NA Spring 2023-24 21 / 40

Numerical Integration: Composite Rectangle Rule

To improve on the approximation

break the interval [a, b] into smaller subintervals and
apply the rectangle rule on each subinterval.

MA 214 - NA Spring 2023-24 21 / 40

Numerical Integration: Composite Rectangle Rule

To improve on the approximation

break the interval [a, b] into smaller subintervals and
apply the rectangle rule on each subinterval.
We will derive a general formula for this.

MA 214 - NA Spring 2023-24 21 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n

MA 214 - NA Spring 2023-24 22 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.

MA 214 - NA Spring 2023-24 22 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.
Then, we get ∫ b
I(f ) = f (x)dx
a

MA 214 - NA Spring 2023-24 22 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.
Then, we get ∫ b
I(f ) = f (x)dx
a
∫ xn
= f (x)dx
x0
MA 214 - NA Spring 2023-24 22 / 40
Numerical Integration: Composite Rectangle Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.
Then, we get ∫ b
I(f ) = f (x)dx
a
∫ xn n−1 ∫
∑ xj+1
= f (x)dx = f (x)dx.
x0 j=0 xj
MA 214 - NA Spring 2023-24 22 / 40
Numerical Integration: Composite Rectangle Rule (contd.)
∫ b
I(f ) = f (x)dx
a
∫ xn n−1 ∫
∑ xj+1
= f (x)dx = f (x)dx.
x0 j=0 xj

Using Rectangle rule on the subinterval [xj , xj+1 ], we get

MA 214 - NA Spring 2023-24 23 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
∫ b
I(f ) = f (x)dx
a
∫ xn n−1 ∫
∑ xj+1
= f (x)dx = f (x)dx.
x0 j=0 xj

Using Rectangle rule on the subinterval [xj , xj+1 ], we get

∫ xj+1
f (x)dx ≈ hf (xj ), j = 0, 1, · · · , n.
xj

MA 214 - NA Spring 2023-24 23 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
∫ b
I(f ) = f (x)dx
a
∫ xn n−1 ∫
∑ xj+1
= f (x)dx = f (x)dx.
x0 j=0 xj

Using Rectangle rule on the subinterval [xj , xj+1 ], we get

∫ xj+1
f (x)dx ≈ hf (xj ), j = 0, 1, · · · , n.
xj

Substituting this in the above equation, we get

I(f ) ≈ hf (x0 ) + hf (x1 ) + · · · + hf (xn ).

MA 214 - NA Spring 2023-24 23 / 40

Numerical Integration: Composite Rectangle Rule (contd.)
∫ b
I(f ) = f (x)dx
a
∫ xn n−1 ∫
∑ xj+1
= f (x)dx = f (x)dx.
x0 j=0 xj

Using Rectangle rule on the subinterval [xj , xj+1 ], we get

∫ xj+1
f (x)dx ≈ hf (xj ), j = 0, 1, · · · , n.
xj

Substituting this in the above equation, we get

I(f ) ≈ hf (x0 ) + hf (x1 ) + · · · + hf (xn ).
This rule is called the Composite Rectangle rule.
MA 214 - NA Spring 2023-24 23 / 40
Quadrature Rule 2: Trapezoidal Rule

MA 214 - NA Spring 2023-24 24 / 40

Numerical Integration: Trapezoidal Rule

I(f) ≈ I(pn ) = f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn ,

MA 214 - NA Spring 2023-24 25 / 40

Numerical Integration: Trapezoidal Rule

I(f) ≈ I(pn ) = f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 1.

MA 214 - NA Spring 2023-24 25 / 40

Numerical Integration: Trapezoidal Rule

I(f) ≈ I(pn ) = f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 1.
Then
p1 (x) = f (x0 ) + f [x0 , x1 ](x − x0 ),

MA 214 - NA Spring 2023-24 25 / 40

Numerical Integration: Trapezoidal Rule

I(f) ≈ I(pn ) = f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 1.
Then
p1 (x) = f (x0 ) + f [x0 , x1 ](x − x0 ),
and therefore
∫ b
I(f) ≈ IT (f) := (f (x0 ) + f [x0 , x1 ](x − x0 )) dx.
a

MA 214 - NA Spring 2023-24 25 / 40

Numerical Integration: Trapezoidal Rule

I(f) ≈ I(pn ) = f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 1.
Then
p1 (x) = f (x0 ) + f [x0 , x1 ](x − x0 ),
and therefore
∫ b
I(f) ≈ IT (f) := (f (x0 ) + f [x0 , x1 ](x − x0 )) dx.
a
Taking x0 = a and x1 = b, we get
( )
f (a) + f (b)
IT (f) = (b − a)
2

MA 214 - NA Spring 2023-24 25 / 40

Numerical Integration: Trapezoidal Rule

I(f) ≈ I(pn ) = f (x0 )w0 + f (x1 )w1 + · · · + f (xn )wn , i.e. TAKE n = 1.
Then
p1 (x) = f (x0 ) + f [x0 , x1 ](x − x0 ),
and therefore
∫ b
I(f) ≈ IT (f) := (f (x0 ) + f [x0 , x1 ](x − x0 )) dx.
a
Taking x0 = a and x1 = b, we get
( )
f (a) + f (b)
IT (f) = (b − a)
2
and is called the Trapezoidal Rule.
MA 214 - NA Spring 2023-24 25 / 40
Numerical Integration: Trapezoidal Rule (contd.)

y f (x)

I(f)

a b x

MA 214 - NA Spring 2023-24 26 / 40

Numerical Integration: Trapezoidal Rule (contd.)

y
f (x)

IT (f)

a b x

MA 214 - NA Spring 2023-24 27 / 40

Numerical Integration: Trapezoidal Rule (contd.)

Mathematical Error

MET (f) := I(f) − I(p1 ).

MA 214 - NA Spring 2023-24 28 / 40

Numerical Integration: Trapezoidal Rule (contd.)

Mathematical Error

MET (f) := I(f) − I(p1 ).

Theorem (Error in Trapezoidal Rule)

Let f ∈ C2 [a, b].

MA 214 - NA Spring 2023-24 28 / 40

Numerical Integration: Trapezoidal Rule (contd.)

Mathematical Error

MET (f) := I(f) − I(p1 ).

Theorem (Error in Trapezoidal Rule)

Let f ∈ C2 [a, b].The mathematical error MET (f) of the trapezoidal rule
takes the form
f 00 (η)(b − a)3
MET (f) = − ,
12
for some η ∈ (a, b).
MA 214 - NA Spring 2023-24 28 / 40
Numerical Integration: Trapezoidal Rule (contd.)

Proof

We have
f (x) = f (a) + f [a, b](x − a) + f [a, b, x](x − a)(x − b)

MA 214 - NA Spring 2023-24 29 / 40

Numerical Integration: Trapezoidal Rule (contd.)

Proof

We have
f (x) = f (a) + f [a, b](x − a) + f [a, b, x](x − a)(x − b)
= p1 (x) + f [a, b, x](x − a)(x − b)

MA 214 - NA Spring 2023-24 29 / 40

Numerical Integration: Trapezoidal Rule (contd.)

Proof

We have
f (x) = f (a) + f [a, b](x − a) + f [a, b, x](x − a)(x − b)
= p1 (x) + f [a, b, x](x − a)(x − b)
Integrating over the interval [a, b], we get
∫b
I(f) = IT (f) + f [a, b, x](x − a)(x − b)dx.
a

MA 214 - NA Spring 2023-24 29 / 40

Numerical Integration: Trapezoidal Rule (contd.)

Proof

f (x) = p1 (x) + f [a, b, x](x − a)(x − b)

Integrating over the interval [a, b], we get
∫b
I(f) = IT (f) + f [a, b, x](x − a)(x − b)dx.
a
Therefore the mathematical error is given by
∫b
MET (f) = I(f) − IT (f) = f [a, b, x](x − a)(x − b)dx.
a
MA 214 - NA Spring 2023-24 30 / 40
Numerical Integration: Trapezoidal Rule (contd.)

The function

x 7−→ f [a, b, x] is continuous and we have (x − a)(x − b) < 0, x ∈ (a, b)

MA 214 - NA Spring 2023-24 31 / 40

Numerical Integration: Trapezoidal Rule (contd.)

The function

x 7−→ f [a, b, x] is continuous and we have (x − a)(x − b) < 0, x ∈ (a, b)

Mean-value theorem for integrals then gives

∫b
MET (f) = f [a, b, η] (x − a)(x − b)dx,
a

for some η ∈ (a, b).

MA 214 - NA Spring 2023-24 31 / 40

Numerical Integration: Trapezoidal Rule (contd.)

We use the formula

f (n+1) (ξx )
f [x0 , x1 , · · · , xn , x] =
(n + 1)!

for some ξx ∈ (a, b).

MA 214 - NA Spring 2023-24 32 / 40

Numerical Integration: Trapezoidal Rule (contd.)

We use the formula

f (n+1) (ξx )
f [x0 , x1 , · · · , xn , x] =
(n + 1)!

for some ξx ∈ (a, b).

Using this formula in the mathematical error
∫b
MET (f) = f [a, b, η] (x − a)(x − b)dx,
a
we get the desired formula after a simple integration.
MA 214 - NA Spring 2023-24 32 / 40
Numerical Integration: Trapezoidal Rule (contd.)
Example: Approximate the integral
∫ 1
1
I= dx,
0 1 + x

MA 214 - NA Spring 2023-24 33 / 40

Numerical Integration: Trapezoidal Rule (contd.)
Example: Approximate the integral
∫ 1
1
I= dx,
0 1 + x

The true value is I = log(2) ≈ 0.693147.

MA 214 - NA Spring 2023-24 33 / 40

Numerical Integration: Trapezoidal Rule (contd.)
Example: Approximate the integral
∫ 1
1
I= dx,
0 1 + x

The true value is I = log(2) ≈ 0.693147.

Using the trapezoidal rule, we get
( )
1 1 3
IT (f) = 1+ = = 0.75.
2 2 4

MA 214 - NA Spring 2023-24 33 / 40

Numerical Integration: Trapezoidal Rule (contd.)
Example: Approximate the integral
∫ 1
1
I= dx,
0 1 + x

The true value is I = log(2) ≈ 0.693147.

Using the trapezoidal rule, we get
( )
1 1 3
IT (f) = 1+ = = 0.75.
2 2 4

⇒ MET (f ) ≈ −0.0569.

MA 214 - NA Spring 2023-24 33 / 40

Numerical Integration: Trapezoidal Rule (contd.)

⇒ MET (f ) ≈ −0.0569.
Recall
f 00 (η)(b − a)3
MET (f) = − ,
12

MA 214 - NA Spring 2023-24 34 / 40

Numerical Integration: Trapezoidal Rule (contd.)

⇒ MET (f ) ≈ −0.0569.
Recall
f 00 (η)(b − a)3
MET (f) = − ,
12
Using the error formula, we get the bounds for MET (f) as
1 1
− ≈ −0.1667 < MET (f) < − ≈ −0.02083
6 48
which clearly holds in the present case.
MA 214 - NA Spring 2023-24 34 / 40
Quadrature Rule 2(a): Composite Trapezoidal Rule

MA 214 - NA Spring 2023-24 35 / 40

Numerical Integration: Composite Trapezoidal Rule

To improve on the approximation, especially, when f (x) is not a nearly

linear function on [a, b],

MA 214 - NA Spring 2023-24 36 / 40

Numerical Integration: Composite Trapezoidal Rule

To improve on the approximation, especially, when f (x) is not a nearly

linear function on [a, b],
break the interval [a, b] into smaller subintervals and

MA 214 - NA Spring 2023-24 36 / 40

Numerical Integration: Composite Trapezoidal Rule

To improve on the approximation, especially, when f (x) is not a nearly

linear function on [a, b],
break the interval [a, b] into smaller subintervals and
apply the Trepezoidal rule on each subinterval.

MA 214 - NA Spring 2023-24 36 / 40

Numerical Integration: Composite Trapezoidal Rule

To improve on the approximation, especially, when f (x) is not a nearly

linear function on [a, b],
break the interval [a, b] into smaller subintervals and
apply the Trepezoidal rule on each subinterval.
We will derive a general formula for this.

MA 214 - NA Spring 2023-24 36 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n

MA 214 - NA Spring 2023-24 37 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.

MA 214 - NA Spring 2023-24 37 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.
Then, we get ∫ ∫
b xn
I(f) = f (x)dx = f (x)dx
a x0

MA 214 - NA Spring 2023-24 37 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Let us subdivide the interval [a, b] into n equal subintervals of length
b−a
h=
n
with endpoints of the subintervals as
xj = a + jh, j = 0, 1, · · · , n.
Then, we get ∫ ∫
b xn
I(f) = f (x)dx = f (x)dx
a x0
∑ ∫ xj+1
n−1
= f (x)dx.
j=0 xj
MA 214 - NA Spring 2023-24 37 / 40
Numerical Integration: Composite Trapezoidal Rule (contd.)
Then, we get
∫ b n−1 ∫ xj+1
∑
I(f ) = f (x)dx = f (x)dx.
a j=0 xj

Using Trapezoidal rule on the subinterval [xj , xj+1 ], we get

MA 214 - NA Spring 2023-24 38 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Then, we get
∫ b n−1 ∫ xj+1
∑
I(f ) = f (x)dx = f (x)dx.
a j=0 xj

Using Trapezoidal rule on the subinterval [xj , xj+1 ], we get

∫ xj+1 ( )
f (xj ) + f (xj+1 )
f (x)dx ≈ h , j = 0, 1, · · · , n.
xj 2

MA 214 - NA Spring 2023-24 38 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Then, we get
∫ b n−1 ∫ xj+1
∑
I(f ) = f (x)dx = f (x)dx.
a j=0 xj

Using Trapezoidal rule on the subinterval [xj , xj+1 ], we get

∫ xj+1 ( )
f (xj ) + f (xj+1 )
f (x)dx ≈ h , j = 0, 1, · · · , n.
xj 2
Substituting this in the above equation, we get
[ ] [ ] [ ]
f (x0 ) + f (x1 ) f (x1 ) + f (x2 ) f (xn−1 ) + f (xn )
I(f ) ≈ h +h +· · ·+h .
2 2 2
MA 214 - NA Spring 2023-24 38 / 40
Numerical Integration: Composite Trapezoidal Rule (contd.)

Then, we get
∫ b n−1 ∫
∑ xj+1
I(f) = f (x)dx = f (x)dx.
a j=0 xj

Using Trapezoidal rule on the subinterval [xj , xj+1 ], we get

[ ]
n 1 1
IT (f ) := h f (x0 ) + f (x1 ) + f (x2 ) + · · · + f (xn−1 ) + f (xn ) .
2 2

This rule is called the Composite Trapezoidal rule.

MA 214 - NA Spring 2023-24 39 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Example: Using composite trapezoidal rule with n = 2,

MA 214 - NA Spring 2023-24 40 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Example: Using composite trapezoidal rule with n = 2, let us
approximate the integral
∫ 1
1
I= dx.
0 1 + x

MA 214 - NA Spring 2023-24 40 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Example: Using composite trapezoidal rule with n = 2, let us
approximate the integral
∫ 1
1
I= dx.
0 1 + x

The true value is I = log(2) ≈ 0.693147.

MA 214 - NA Spring 2023-24 40 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Example: Using composite trapezoidal rule with n = 2, let us
approximate the integral
∫ 1
1
I= dx.
0 1 + x

The true value is I = log(2) ≈ 0.693147.

Now, the composite trapezoidal rule with x0 = 0, x1 = 1/2 and x2 = 1
gives
IT2 (f) ≈ 0.70833.

MA 214 - NA Spring 2023-24 40 / 40

Numerical Integration: Composite Trapezoidal Rule (contd.)
Example: Using composite trapezoidal rule with n = 2, let us
approximate the integral
∫ 1
1
I= dx.
0 1 + x

The true value is I = log(2) ≈ 0.693147.

Now, the composite trapezoidal rule with x0 = 0, x1 = 1/2 and x2 = 1
gives
IT2 (f) ≈ 0.70833.
Thus the error is −0.0152.
MA 214 - NA Spring 2023-24 40 / 40
Numerical Integration: Composite Trapezoidal Rule (contd.)
Example: Using composite trapezoidal rule with n = 2, let us
approximate the integral
∫ 1
1
I= dx.
0 1 + x

The true value is I = log(2) ≈ 0.693147.

Now, the composite trapezoidal rule with x0 = 0, x1 = 1/2 and x2 = 1
gives
IT2 (f) ≈ 0.70833.
Thus the error is −0.0152. Recall trapezoidal rule gave −0.0569.
MA 214 - NA Spring 2023-24 40 / 40