Register to reply

Error Bounds for Approximate Integration

by dekoi
Tags: approximate, bounds, error, integration
Share this thread:
dekoi
#1
Feb1-06, 07:34 PM
P: n/a
Does anyone know a trick for finding "K" in the error bound equations for approximate integration?

The approximate integrations we have learned so far are Midpoint Rule, Trapzoid Rule, and Simpson's Rule.

Thank you.
Phys.Org News Partner Science news on Phys.org
Security CTO to detail Android Fake ID flaw at Black Hat
Huge waves measured for first time in Arctic Ocean
Mysterious molecules in space
HallsofIvy
#2
Feb2-06, 07:47 AM
Math
Emeritus
Sci Advisor
Thanks
PF Gold
P: 39,348
The error estimates for each of those should be in the same section of your text as the derivation of the formula.

Googling on ' "trapezoidal rule" error ',etc. gave me this:

http://archives.math.utk.edu/visual....us/4/approx.2/
dekoi
#3
Feb2-06, 01:37 PM
P: n/a
I will take a look at the website later.


I was aware that error bounds are in my textbook-- I'm not that ignorant.

Anyway,
I was wondering about some "tricks" because my textbook does not have that-- it simply states what K is in each case, but does not go through the process of finding it.

HallsofIvy
#4
Feb2-06, 03:00 PM
Math
Emeritus
Sci Advisor
Thanks
PF Gold
P: 39,348
Error Bounds for Approximate Integration

My apology- I misread your first post and though you were asking for the error formulas rather than 'finding "K" in the error bound equations'. However, now I have to point out that not every text writes the formulas in exactly the same way- some I am not sure what "K" represents. The formulas on the page do not use "K". I suspect, however, that it is what the formulas I use call "M"- an upper bound on
in the case of the trapezoidal rule, first derivative of the integrand
in the case of the mid-point rule, the second derivative of the integrand
in the case of Simpson's rule, the fourth derivative of the integrand

There is no rule for determining those- remember these are "approximations"- obviously you can't find the exact value of the error! If you could, you could just add it on to get an exact value for the integral!

Essentially, you find the correct derivative and extimate how large it can be in that interval. For sine or cosine, for example, you know the value is never larger than 1. For an increasing function like ex, use the right endpoint.
dekoi
#5
Feb2-06, 03:49 PM
P: n/a
Yes, K represents the upper bound.

Thank you for your help.


Register to reply

Related Discussions
Error bounds with approximated M value Calculus 1
Error bounds Calculus 1
Numerical Analysis: Taylor Polynomials, Error, Bounds Calculus & Beyond Homework 0
Approximate integration Introductory Physics Homework 2
Error bounds - Simpson, Trap, and Midpoint Calculus 5