Chapter 6: Techniques of Integration
Section 6.7: Numeric Methods
Example 6.7.6
Use the error bound in Table 6.7.1 to estimate the value of the partition n for which Simpson's rule makes an absolute error of no more than 10−4 when estimating λ, the value of the definite integral in Example 6.7.1.
What is the actual value of n for which Simpson's rule achieves this accuracy?
Solution
Mathematical Solution
If fx=1+sinxlnx+1, write h4180 maxx(|f″|) b−a, the error bound in Table 6.7.1, as b−a5180 n4maxx(|f″|) and solve the inequality 4−15180 n4M≤10−4 for n, where M=maxxf4≐1.0443. This gives n≐± 10.9, so the appropriate value is n=12. However, for n=10, Simpson's rule actually approximates λ with an error no worse than 10−4.
Maple Solution
Determine M, the maximum of the absolute value of the fourth derivative of the integrand over the interval of integration.
Determine M=maxxf4≐1.0443
Context Panel: Assign Function
fx=1+sinxlnx+1→assign as functionf
From Figure 6.7.6(a), a graph of |f4x| on 1,4, estimate:
M=maxxf4=f41≐1.0443
In fact, f41.0 = 1.044281148
Figure 6.7.6(a) Graph of fx on 1,4
With a=1,b=4,M≐1.0443, solve the inequality b−a5180 n4maxx(|f4|)≤10−4 for n.
Write the inequality b−a5180 n4maxx(|f4|).
Context Panel: Solve≻Solve
4−15180 n41.0443≤10−4→solven≤−10.89656871,10.89656871≤n
The appropriate choice of n is the first positive even integer greater than 10.9. Hence, n=12 guarantees that Simpson's rule will approximate λ with an error of no more than 10−4. From Example 6.7.1, take λ to be the number 5.078061188. To determine the actual value of n for which Simpson's rule approximates λ with the desired accuracy, use the ApproximateInt command as per Table 6.7.6(a).
Initialize
Tools≻Load Package: Student Calculus 1
Loading Student:-Calculus1
Define L as the actual value of the integral.
L≔5.078061188:
Use the ApproximateInt command and compare to λ
L−ApproximateIntfx,x=1..4.0,partition=8,method=simpson,partitiontype=normal
0.000140137
L−ApproximateIntfx,x=1..4.0,partition=10,method=simpson,partitiontype=normal
0.000057487
Table 6.7.6(a) The smallest value of n for which Simpson's rule approximates λ to within 10−4
By experiment, it is determined that n=10 is the smallest value of n for which Simpson's rule approximates λ with an error no worse than 10−4.
Note that for Simpson's rule, the ApproximateInt command defaults to the subinterval form of the partitiontype option. This form of the rule evaluates the integrand at 2 k+1 points, thereby obtaining a much greater accuracy. To keep the number of function evaluations to k+1, it is essential to include the normal form of the partitiontype option.
<< Previous Example Section 6.7 Next Example >>
© Maplesoft, a division of Waterloo Maple Inc., 2024. All rights reserved. This product is protected by copyright and distributed under licenses restricting its use, copying, distribution, and decompilation.
For more information on Maplesoft products and services, visit www.maplesoft.com
Download Help Document