Simpson's Integration Rule
Simpson's Integration Rule
com
and it is a natural progression from the mid-point integration rule2 and the trapezoidal integration rule3. The derivation of the method is based on approximating the function to be integrated by a piecewise quadratic function4 and integrating that function. The domain is divided into subdomains where and . The points with odd indices lie within the relevant subinterval, usually at the central points.
by a quadratic function that passes through the points and The method of approximation is illustrated in the following graph. The integral is represented by a set of sectional areas below the piecewise quadratic function, and the approximation to the integral is formed by the summation of the areas.
For the subdomain , let the width be and hence the integral of the quadratic function approximates the integral over that subdomain. It is found that the following approximation to the integral is derived:
Numerical Integration or Quadrature Mid-point Integration Rule 3 Trapezoidal Integration Rule 4 Piecewise Polynomial Interpolation
1 2
www.numerical-methods.com
The error in Simpsons rule is found to be approximately proportional to the cube of , or in the limit as , using O-notation5. Example In order to demonstrate the method, let us use Simpsons integration rule to find
The exact solution6 to the integral is =1.718282 . The numerical solution to the integral will first be found using two subdomains ( ), and we will then go on to double the number of subdomains in order to illustrate the pattern in the numerical error. An Excel spreadsheet7 is available in order to automate the method. n=2 The domain [0,1] is divided into two subdomains of equal width [0,0.5] and. The points and function values are listed in the following table. The application of the trapezoidal integration rule gives the following approximation to the integral:
4*2.117000017 +
)= 1.718319
Hence the numerical error is |1.718282-1.718319|=0.000037 (to six decimal places). n=4 Using the spreadsheet with 4 subdomains the result from the method is 1.718284 and the numerical error is 0.000002.