Timeline for Strategies to solve an oscillatory integrand only known numerically
Current License: CC BY-SA 3.0
24 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 19, 2012 at 9:27 | answer | added | fpghost | timeline score: 1 | |
| Oct 19, 2012 at 8:08 | comment | added | fpghost | @celtschk thanks! | |
| Oct 19, 2012 at 8:07 | history | edited | fpghost | CC BY-SA 3.0 |
added 58 characters in body
|
| Oct 19, 2012 at 7:55 | comment | added | celtschk |
@fpghost: Hint: If you want people to be notified of your answers to their comments, prepend @ to the user name of the user you reply to (as I did with yours in this comment, although in this case it's not strictly necessary because question/answer authors always get notified about comments on their post).
|
|
| Oct 19, 2012 at 7:06 | history | edited | fpghost | CC BY-SA 3.0 |
rewrite of integrand to something maybe suitable for LevinRule
|
| Oct 19, 2012 at 7:01 | history | edited | fpghost | CC BY-SA 3.0 |
rewrite of integrand to something maybe suitable for LevinRule
|
| Oct 17, 2012 at 19:49 | comment | added | fpghost | splitting my integral ranges by hand seems to remove the slwcon errors, not sure if that just means each has a low error, but the total still has just has much as if I did it in one go. | |
| Oct 17, 2012 at 18:37 | comment | added | fpghost | No unfortunately not. | |
| Oct 17, 2012 at 18:30 | history | edited | fpghost | CC BY-SA 3.0 |
added 931 characters in body
|
| Oct 17, 2012 at 18:24 | history | edited | fpghost | CC BY-SA 3.0 |
added 931 characters in body
|
| Oct 17, 2012 at 18:10 | history | edited | fpghost | CC BY-SA 3.0 |
added 513 characters in body
|
| Oct 17, 2012 at 3:47 | answer | added | J. M.'s missing motivation♦ | timeline score: 6 | |
| Oct 17, 2012 at 2:36 | answer | added | Andrew Moylan | timeline score: 16 | |
| Oct 17, 2012 at 2:24 | comment | added | Andrew Moylan | Does the oscillatory part of your integrand satisfy a linear ODE? | |
| Oct 17, 2012 at 2:23 | comment | added | Andrew Moylan | "I can force it to do the integral by making the WorkingPrecision higher, but I think this is cheating if I don't believe my integrand any higher than 6 dp?" <-- Definitely. You will need the integrand to higher precision. | |
| Oct 16, 2012 at 21:08 | history | tweeted | twitter.com/#!/StackMma/status/258313492455628801 | ||
| Oct 16, 2012 at 19:04 | comment | added | fpghost | Searke: the highly oscillating part of my integrand isn't known symbolically, I only know it as the interpolating function obtained from NDSolve of my ODE. | |
| Oct 16, 2012 at 19:02 | comment | added | fpghost | celtschk: Not sure I can do that, the interpolating function solution to NDSolve is just one component that goes into make my integrand and integration is in a different variable than ODE. belisarius: how do I do that is there a Method->"FourierRule" type of thing? Vitaly: thanks I will take a look. | |
| Oct 16, 2012 at 18:41 | comment | added | Searke | A common symbolic trick is to rewrite the integral using integration by parts. Even if it is still symbolically unsolveable, you may be able to symbolically solve the highly oscillating component. | |
| Oct 16, 2012 at 17:47 | history | edited | rm -rf♦ | CC BY-SA 3.0 |
added 62 characters in body; edited title
|
| Oct 16, 2012 at 17:43 | comment | added | Vitaliy Kaurov | Have you seen this video ? It talks about similar problem and hybrid numeric-symbolic methods to address it. | |
| Oct 16, 2012 at 17:34 | comment | added | Dr. belisarius | Have you tried a Fourier on that ? | |
| Oct 16, 2012 at 17:34 | comment | added | celtschk | What about rewriting your differential equations to directly give the integral from NDSolve? | |
| Oct 16, 2012 at 17:28 | history | asked | fpghost | CC BY-SA 3.0 |