Timeline for Diverging solution to coupled second order ODEs from NDSolve
Current License: CC BY-SA 3.0
20 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 4, 2019 at 7:00 | history | edited | J. M.'s missing motivation | edited tags | |
| Aug 6, 2017 at 0:19 | history | edited | bbgodfrey | edited tags | |
| Mar 28, 2017 at 17:51 | history | edited | bbgodfrey | CC BY-SA 3.0 | corrected earlier my earlier choice of tags |
| Mar 24, 2017 at 0:21 | history | edited | bbgodfrey | added tag | |
| Nov 29, 2015 at 0:49 | history | edited | bbgodfrey | added tag | |
| S Jun 12, 2015 at 7:38 | history | bounty ended | Funzies | ||
| S Jun 12, 2015 at 7:38 | history | notice removed | Funzies | ||
| Jun 10, 2015 at 9:48 | vote | accept | Funzies | ||
| Jun 6, 2015 at 16:15 | answer | added | bbgodfrey | timeline score: 13 | |
| Jun 5, 2015 at 15:46 | comment | added | Albert Retey | @xzczd : I'm not sure why you did see other behaviour before, but that is the usual rule: precision will never automatically increase, so if you start out with machine precision numbers only in very limited cases increasing WorkingPrecision higher than MachinePrecision will have any effect, and usually not the desired one (AFAIK). To be on the safe side I do use completely "exact" input if I play around with the precision stuff beyond machine precision and that so far has been free from any surprises... | |
| Jun 5, 2015 at 13:36 | comment | added | xzczd♦ | @AlbertRetey Very surprising! In this case changing 1.5 to 3/2 is necessary, or WorkingPrecision -> 32 won't work at all! I used to think that though the warning NDSolve::precw will be generated, as long as a higher WorkingPrecision is set, approximate numbers in the differential equation won't influence the result. (Actually I never saw this principle failed before!) | |
| Jun 5, 2015 at 9:52 | comment | added | Albert Retey | I think WorkingPrecision->30 should be good enough, and I forgot to mention that this is of course to be used as an option to NDSolve... | |
| Jun 5, 2015 at 9:37 | comment | added | Albert Retey | I think it might be just numeric precision loss. Try something like WorkingPrecision -> 50 which will improve the situation a lot. You will have to take care that all constants will be given as exact numbers, but that would only mean to change 1.5 to 3/2 for the value in the Manipulate... | |
| S Jun 5, 2015 at 7:38 | history | bounty started | Funzies | ||
| S Jun 5, 2015 at 7:38 | history | notice added | Funzies | Draw attention | |
| Apr 25, 2015 at 6:12 | history | tweeted | twitter.com/#!/StackMma/status/591847345404452864 | ||
| Apr 24, 2015 at 8:00 | comment | added | Funzies | @user21 I have updated the code, it should be copy-pastable now. | |
| Apr 24, 2015 at 8:00 | history | edited | Funzies | CC BY-SA 3.0 | Updated code |
| Apr 23, 2015 at 16:12 | comment | added | user21 | Could you rework your code a bit such that it can be copy&pasted? | |
| Apr 23, 2015 at 14:37 | history | asked | Funzies | CC BY-SA 3.0 |