Numerical Integration. Integration/Integration Techniques/Numerical Integration by M. Seppälä...

Post on 27-Mar-2015

301 views 4 download

Tags:

Transcript of Numerical Integration. Integration/Integration Techniques/Numerical Integration by M. Seppälä...

Numerical Integration

Integration/Integration Techniques/Numerical Integration by M. Seppälä

Definite Integrals

f t

k( ) Δxk

k=1

n

∑ D→ 0⏐ →⏐ ⏐ ⏐

f x( )dx

a

b

Integration/Integration Techniques/Numerical Integration by M. Seppälä

Riemann Sum S

D= f t

k( ) Δxk.

k=1

n

NUMERICAL INTEGRATION

Use decompositions of the type

D = a, a +

b −an

, a + 2b −a

n,K , a + n

b −an

⎝⎜⎞

⎠⎟.

General kth subinterval:

a + k −1( ) Δx, a + kΔx⎡

⎣⎤⎦, Δx =Δx

k=

b −an

.

Integration/Integration Techniques/Numerical Integration by M. Seppälä

Riemann Sum S

D= f t

k( ) Δx.k=1

n

RULES TO SELECT POINTS

Left Rule t

k=a + k −1( ) Δx

Δx =

b −an

Integration/Integration Techniques/Numerical Integration by M. Seppälä

Riemann Sum S

D= f t

k( ) Δx.k=1

n

RULES TO SELECT POINTS

Right Rule tk=a + kΔx

Δx =

b −an

Integration/Integration Techniques/Numerical Integration by M. Seppälä

Riemann Sum S

D= f t

k( ) Δx.k=1

n

RULES TO SELECT POINTS

Midpoint Rule t

k=a + k −

12

⎝⎜⎞

⎠⎟Δx

Δx =

b −an

Integration/Integration Techniques/Numerical Integration by M. Seppälä

RULES TO SELECT POINTS

Right Approximation

Left Approximation

f a + kΔx( ) Δx

k=1

n

∑RIGHT(n) =

f a + k −1( ) Δx( ) Δx

k=1

n

∑LEFT(n) =

Integration/Integration Techniques/Numerical Integration by M. Seppälä

RULES TO SELECT POINTS

Midpoint Approximation

MID(n) =

Integration/Integration Techniques/Numerical Integration by M. Seppälä

PROPERTIES

LEFT(n) ≤ f x( )dx ≤

a

b

If f is increasing,Property

RIGHT(n)

Integration/Integration Techniques/Numerical Integration by M. Seppälä

PROPERTIES

LEFT(n) = f(a + (k −1)Δx)Δxk=1

n

=Δx f a( ) + f a + Δx( ) +L + f a + n −1( )Δx( )( )

RIGHT(n) = f(a + kΔx)Δxk=1

n

=Δx f a + Δx( ) +L + f a + n −1( )Δx( ) + f b( )( ).

Integration/Integration Techniques/Numerical Integration by M. Seppälä

PROPERTIES

RIGHT(n) −LEFT(n)

=Δx f b( ) −f a( )( ) =b −a

nf b( ) −f a( )( ).

For any function, Property

Integration/Integration Techniques/Numerical Integration by M. Seppälä

PROPERTIES

If f is increasing,

RIGHT(n) − f x( )dx

a

b

∫ ≤b −a

nf b( ) −f a( )

Hence

Property

Integration/Integration Techniques/Numerical Integration by M. Seppälä

PROPERTIES

RIGHT(n) − f x( )dx

a

b

∫ ≤b −a

nf b( ) −f a( )

Property If f is increasing or decreasing:

LEFT(n) − f x( )dx

a

b

∫ ≤b −a

nf b( ) −f a( )

Integration/Integration Techniques/Numerical Integration by M. Seppälä

CONCAVITY

Recall

The graph of a function f is concave

up, if the graph lies above any of its

tangent line.

Integration/Integration Techniques/Numerical Integration by M. Seppälä

MIDPOINT APPROXIMATIONS

Midpoint Approximation

MID(n) =

Integration/Integration Techniques/Numerical Integration by M. Seppälä

MIDPOINT APPROXIMATIONS

The two blue areas on the left are the same.

The blue polygon in the middle is contained in the domain under the concave-up curve.

MID(n) ≤ f x( )dx

a

b

Integration/Integration Techniques/Numerical Integration by M. Seppälä

MIDPOINT APPROXIMATIONS

If the function f takes positive values, and if the

graph of f is concave-up

MID(n) ≤ f x( )dx

a

b

Integration/Integration Techniques/Numerical Integration by M. Seppälä

MIDPOINT APPROXIMATIONS

If the function f takes positive values, and if the

graph of f is concave-down

MID(n) ≥ f x( )dx

a

b

Integration/Integration Techniques/Numerical Integration by M. Seppälä

TRAPEZOIDAL APPROXIMATIONS

LEFT(n) rectangle

RIGHT(n) rectangle

TRAP(n) polygon

Integration/Integration Techniques/Numerical Integration by M. Seppälä

TRAPEZOIDAL APPROXIMATIONS

TRAP(n) polygon

If the function f takes positive values and is concave-up

f x( )dx

a

b

∫ ≤TRAP n( ).

Integration/Integration Techniques/Numerical Integration by M. Seppälä

COMPARING APPROXIMATIONS

Example

a b

fThe graph of a function f is increasing and concave up.

f x( )dx

a

b

Arrange the various numerical

approximations of the integral

into an increasing order.

Integration/Integration Techniques/Numerical Integration by M. Seppälä

COMPARING APPROXIMATIONS

Example

a b

fBecause f is increasing,

Because f is positive and concave-up,

Integration/Integration Techniques/Numerical Integration by M. Seppälä

COMPARING APPROXIMATIONS

Example

a b

f

LEFT n( ) ≤MID n( ) ≤TRAP n( ) ≤RIGHT n( )

Because f is increasing and concave-up,

Integration/Integration Techniques/Numerical Integration by M. Seppälä

COMPARING APPROXIMATIONS

Example

a b

f

LEFT n( ) ≤MID n( ) ≤ f x( )dxa

b

∫≤TRAP n( ) ≤RIGHT n( )

Because f is increasing and concave-up,

Integration/Integration Techniques/Numerical Integration by M. Seppälä

SUMMARY

Right Approximation

Left Approximation

f a + kΔx( ) Δx

k=1

n

∑RIGHT(n) =

f a + k −1( ) Δx( ) Δx

k=1

n

∑LEFT(n) =

Integration/Integration Techniques/Numerical Integration by M. Seppälä

SUMMARY

Midpoint Approximation

MID(n) =

Trapezoidal Approximation

TRAP n( ) =

LEFT n( ) +RIGHT n( )

2.

Integration/Integration Techniques/Numerical Integration by M. Seppälä

SIMPSON’S APPROXIMATION

In many cases, Simpson’s Approximation gives best results.

SIMPSON n( ) =

2MID n( ) + TRAP n( )

3.