Timeline for Plotting highly oscillatory integrand
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 9, 2022 at 10:56 | comment | added | Roman | @Patrycja you interchanged the order of integration; if you go back to the order I suggest, it works. Try to integrate step by step (one dimension at a time) and fiddle with assumptions until it works. | |
| Nov 9, 2022 at 9:57 | vote | accept | Patrycja | ||
| Nov 9, 2022 at 9:51 | comment | added | Patrycja | Thank you very much for your input, the provided solution worked. If by any chance you have time to look at the post again, where I added a problem appearing after applying the solution, it would be much appreciated. Either way, thanks :) | |
| Nov 8, 2022 at 13:36 | comment | added | Roman | The analytic integration is done once-and-for-all (hence an immediate assignment was used) and every future evaluation of f or eA is instantaneous. Numerical integration must be done for every value of the parameters anew (hence you used a delayed assignment). So what you are saying is only true if you evaluate your eA once; but as you are evaluating it multiple times, analytic integration is much faster. | |
| Nov 8, 2022 at 13:23 | comment | added | Patrycja | The problem with the analytic integration is that it takes significantly longer to compute than with numerical integration even for small boundaries. I'm pretty sure I need to stick with NIntegrate | |
| Nov 8, 2022 at 13:13 | history | edited | Roman | CC BY-SA 4.0 | deleted 28 characters in body |
| Nov 8, 2022 at 12:49 | history | answered | Roman | CC BY-SA 4.0 |