Approximate integration

In summary, we have shown that the Trapezoidal rule and Midpoint rule can be combined to obtain Simpson's rule, and that the same holds true for their corresponding values at twice the number of intervals. This provides a way to prove the given equations involving these rules.
  • #1
DivGradCurl
372
0
I have two similar problems. I need to show that

1. [tex] \frac{1}{2} \left( T_n + M_n \right) = T_{2n} [/tex]

2. [tex] \frac{1}{3} T_n + \frac{2}{3} M_n = S_{2n} [/tex]

where:

[tex]T[/tex] is the Trapezoidal rule.

[tex]M[/tex] is the Midpoint rule.

[tex]S[/tex] is Simpson's rule.

This is what I've got:

[tex]T_n = \frac{b-a}{2n}\left[ f\left( x_0 \right) + 2f\left( x_1 \right) + 2f\left( x_2 \right) + \dots + 2f\left( x_{n-2} \right) + 2f\left( x_{n-1} \right) + f\left( x_{n} \right) \right][/tex]

[tex]T_{2n} = \frac{b-a}{4n}\left[ f\left( x_0 \right) + 2f\left( x_2 \right) + 2f\left( x_4 \right) + \dots + 2f\left( x_{2n-2} \right) + 2f\left( x_{2n-1} \right) + f\left( x_{2n} \right) \right][/tex]

[tex]M_n = \frac{b-a}{n}\left[ f\left( \bar{x}_1 \right) + f\left( \bar{x}_2 \right) + f\left( \bar{x}_3 \right) + \dots + f\left( \bar{x}_{n} \right) \right][/tex]

[tex]S_n = \frac{b-a}{3n}\left[ f\left( x_0 \right) + 4f\left( x_1 \right) + 2f\left( x_2 \right) + 4f\left( x_3 \right) + \dots + 4f\left( x_{n-3} \right) + 2f\left( x_{n-2} \right) + 4f\left( x_{n-1} \right) + f\left( x_{n} \right) \right][/tex]

[tex]S_{2n} = \frac{b-a}{6n}\left[ f\left( x_0 \right) + 4f\left( x_2 \right) + 2f\left( x_4 \right) + 4f\left( x_6 \right) + \dots + 4f\left( x_{2n-3} \right) + 2f\left( x_{2n-2} \right) + 4f\left( x_{2n-1} \right) + f\left( x_{2n} \right) \right][/tex]

The main difficulty I've had is to deal with the [tex]f\left( \bar{x}_{n} \right)[/tex] terms, since I can't see a way to combine them with the [tex]f\left( x_{n} \right)[/tex] ones.

Any help is highly appreciated.
 
Physics news on Phys.org
  • #2
Express your points [itex] x_0, x_1, ..., x_n [/itex] in terms of the lower limit and the step. So the step is:

[tex]h= \frac {b -a} n[/tex]

[tex] x_0 = a [/tex]
[tex] x_1 = a + h [/tex]
.
.
.
[tex]x_n = a + nh [/tex]

Can you do the same for the [itex] \bar {x}_n [/itex]

You may want to use different symbols for the evaluation points of the different methods. This will help you to deal with the fact that [itex] x_n [/itex] is not the same point for the different methods. Once you have worked out the actual values, in terms af a and h you should be home free.
 
  • #3
Thank you for the tips!

I think I've found the solution to both problems. Here's what I have:

[tex]T_n = \frac{h}{2}\left\{ f(a) + 2f(a+h) + 2f(a+2h) + \dots + 2f\left[ a + (n-2)h \right] + 2f\left[ a + (n-1)h \right] + f\left( a + nh \right) \right\}[/tex]

[tex]T_{2n} = \frac{h}{4}\left\{ f(a) + 2 f\left( a + \frac{h}{2} \right) + 2 f\left( a + h \right) + 2 f\left( a + \frac{3h}{2} \right) + \dots + 2 f\left[ a + \frac{(2n-1)h}{2} \right] + f\left( a + nh \right) \right\}[/tex]

[tex]M_n = h \left\{ f\left( a + \frac{h}{2} \right) + f\left( a + \frac{3h}{2} \right) + f\left( a + \frac{5h}{2} \right) + \dots + f\left[ a + \frac{(2n-3)h}{2} \right] + f\left[ a + \frac{(2n-2)h}{2} \right] + f\left[ a + \frac{(2n-1)h}{2} \right] \right\}[/tex]

[tex]S_n = \frac{h}{3}\left\{ f(a) + 4 f(a+h) + 2 f(a+2h) + 4 f(a+3h) + \dots + 4 f\left[ a+(n-3)h \right] + 2 f\left[ a+(n-2)h \right] + 4 f\left[ a+(n-1)h \right] + f(a+nh) \right\}[/tex]

[tex]S_{2n} = \frac{h}{6} \left\{ f(a) + 4 f\left( a + \frac{h}{2} \right) + 2f(a+h) + 4 f\left( a + \frac{3h}{2} \right) + \dots + 2 f\left[ a+\frac{(2n-2)h}{2} \right] + 4 f\left[ a+\frac{(2n-1)h}{2} \right] + f(a+nh) \right\}[/tex]

To make things a bit simpler, I write them in shorthand notation:

[tex]T_n = \frac{h}{2}\alpha[/tex]

[tex]T_{2n} = \frac{h}{4}\beta[/tex]

[tex]M_n = h \gamma[/tex]

[tex]S_n = \frac{h}{3}\theta[/tex]

[tex]S_{2n} = \frac{h}{6}\lambda[/tex]

So, we have:

1. [tex]\frac{1}{2}\left( T_n + M_n \right) = T_{2n}[/tex]

[tex]\frac{1}{2} \left( \frac{h}{2} \alpha + h \gamma \right) = \frac{h}{4}\beta[/tex]

[tex]\frac{h}{2} \left( \frac{1}{2} \alpha + \gamma \right) = \frac{h}{4}\beta[/tex]

[tex]\frac{1}{2}\alpha + \gamma = \frac{1}{2}\beta[/tex]

Thus, we find

[tex]\frac{1}{2} f(a) + f(a+h) + f(a+2h) + \dots + f\left[ a + (n-2)h \right] + f\left[ a + (n-1)h \right] + \frac{1}{2} f\left( a + nh \right)[/tex]​

[tex]+[/tex]​

[tex]f\left( a + \frac{h}{2} \right) + f\left( a + \frac{3h}{2} \right) + f\left( a + \frac{5h}{2} \right) + \dots + f\left[ a + \frac{(2n-1)h}{2} \right][/tex]​


[tex]=[/tex]​

[tex]\frac{1}{2} f(a) + f\left( a + \frac{h}{2} \right) + f(a+h) + f\left( a + \frac{3h}{2} \right) + \dots + f\left[ a + \frac{(2n-1)h}{2} \right] + \frac{1}{2} f\left( a + nh \right) [/tex]​

which verifies the first relationship.

2. [tex]\frac{1}{3}T_n + \frac{2}{3}M_n = S_{2n}[/tex]

[tex]\frac{1}{3}\left( \frac{h}{2}\alpha \right) + \frac{2}{3}(h\gamma ) = \frac{h}{6}\lambda [/tex]

[tex]\frac{h}{6}\alpha + \frac{2h}{3}\gamma = \frac{h}{6}\lambda [/tex]

[tex]\frac{h}{6} \left( \alpha + 4\gamma \right) = \frac{h}{6}\lambda [/tex]

[tex]\alpha + 4\gamma = \lambda [/tex]

Thus, we find

[tex]f(a) + 2f(a+h) + 2f(a+2h) + \dots + 2f\left[ a + (n-2)h \right] + 2f\left[ a + (n-1)h \right] + f\left( a + nh \right)[/tex]​

[tex]+[/tex]​

[tex]4f\left( a + \frac{h}{2} \right) + 4f\left( a + \frac{3h}{2} \right) + 4f\left( a + \frac{5h}{2} \right) + \dots + 4f\left[ a + \frac{(2n-3)h}{2} \right] + 4f\left[ a + \frac{(2n-2)h}{2} \right] + 4f\left[ a + \frac{(2n-1)h}{2} \right][/tex]​

[tex]=[/tex]​

[tex]f(a) + 4 f\left( a + \frac{h}{2} \right) + 2f(a+h) + 4 f\left( a + \frac{3h}{2} \right) + \dots + 2 f\left[ a+\frac{(2n-2)h}{2} \right] + 4 f\left[ a+\frac{(2n-1)h}{2} \right] + f(a+nh)[/tex]​

which verifies the second relationship.
 

What is approximate integration?

Approximate integration is a method used to estimate the value of a definite integral when the integrand cannot be easily integrated analytically. It involves dividing the interval of integration into smaller subintervals and using numerical techniques to calculate the area under the curve.

What are the advantages of approximate integration?

Approximate integration allows for the estimation of the value of a definite integral without having to find an exact solution. It also allows for the integration of functions that cannot be easily integrated analytically, and can provide accurate results even for complex functions.

What are the limitations of approximate integration?

Approximate integration can introduce errors, especially when using simpler numerical techniques. It also requires a certain level of mathematical knowledge and understanding of the function being integrated.

What are some common numerical techniques used in approximate integration?

Some common numerical techniques used in approximate integration include the trapezoidal rule, Simpson's rule, and the midpoint rule. These methods involve dividing the interval of integration into smaller subintervals and using a formula to calculate the area under the curve.

How can I determine the accuracy of an approximate integration?

The accuracy of an approximate integration can be determined by comparing the result to an exact solution, if one is available. Additionally, using smaller subintervals and more advanced numerical techniques can improve the accuracy of the approximation.

Similar threads

  • Introductory Physics Homework Help
Replies
7
Views
674
Replies
1
Views
141
  • Introductory Physics Homework Help
Replies
5
Views
468
Replies
4
Views
808
  • Introductory Physics Homework Help
Replies
2
Views
540
  • Introductory Physics Homework Help
3
Replies
95
Views
4K
  • Introductory Physics Homework Help
Replies
28
Views
363
  • Introductory Physics Homework Help
Replies
2
Views
449
  • Introductory Physics Homework Help
Replies
13
Views
790
  • Introductory Physics Homework Help
Replies
6
Views
569
Back
Top