Talk:Runge–Kutta–Fehlberg method
![]() | This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | ||||||||||
|
Errors in specification of truncation error
[edit]This article's specification for how to calculate truncation error makes no sense - it ignores the central idea of this method, namely comparing a fourth and fifth order estimate to approximate truncation error, and replaces it with a calculation based off of the derivatives and some seemingly arbitrary constants (CT(K)). I don't know where these constants come from, as I can't find them in the papers referenced. — Preceding unsigned comment added by 67.244.52.112 (talk) 19:01, 5 August 2022 (UTC)
- I was confused too, specifically because the 4th order coefficients don't seem to be used at all in the article. But It turns out that the CT coefficients are the difference between the 4th and 5th order coefficicents which is what you end up with if you subtract the 5th order estimate from the 4th order estimate. So those coefficients are and equivalent way to calculate the truncation error. This should probably be made clear in the article. 2620:0:E00:5529:78BD:18EE:B4BC:C056 (talk) 22:11, 31 July 2023 (UTC)
Untitled
[edit]The URL and title of this page contain 8-bit graphic characters; perhaps they should be changed to pure 7-bit ASCII for easier reading on systems not infected with Windows. 17:15, 8 March 2011 (UTC) —Preceding unsigned comment added by 136.160.250.253 (talk)
Removing paragraph on halving step size
[edit]I am removing the following paragraph from the article:
- The RKF45 solves the problem twice using step sizes h and h/2 and then compares the answers. If the two answers do not agree to a specified accuracy, the step size is reduced. If the answers agree to more significant digits than required, the step size is increased.
This is a description of how some adaptive methods work, but does not apply to RKF as described in the article, which uses a higher order method (rather than a halved step size) to get a more accurate solution and hence estimate the error. 131.111.17.186 (talk) 10:55, 23 October 2012 (UTC)
Assessment comment
[edit]The comment(s) below were originally left at Talk:Runge–Kutta–Fehlberg method/Comments, and are posted here for posterity. Following several discussions in past years, these subpages are now deprecated. The comments may be irrelevant or outdated; if so, please feel free to remove this section.
Add the coefficients. -- Jitse Niesen (talk) 11:59, 28 April 2007 (UTC) |
Last edited at 11:59, 28 April 2007 (UTC). Substituted at 02:34, 5 May 2016 (UTC)