Simpson's Rule with negative values

[amanno]

Hey guys,

Not claiming to be an expert on numerical methods here but I am doing some digital integration using simpson's rule that takes 10 samples a second to provide me with the "approximate" integral for that time period.

Right now I am currently taking the magnitude of the value I receive (before integrating it), it is possible that for some of my samples, within the 1 second time period, could be negative. What does this do to the result of simpsons rule? Will it still provide accurate results or is the error greater?

Thanks!

 

[HallsofIvy]

Whether or not there are negative function values does not change the accuracy of Simpson's. You can, in fact, add a fixed number to each value- adding that number times the length of the interval to the integral but not changing the accuracy.