integration accuracy - Wrong numerical answer for simple integration of 1/ln(x) - Messages
#1 Posted: 5/22/2017 6:28:13 PM
The integral of 1/(ln(x) with limits of x from 2 to 1000 does not agree with mathlab or maxima --
The numerical value of the following integration is off -- should be 176.56 and Smath gives 177.1
Any idea what could be the problem?
Thanks
Joey Ooi
#2 Posted: 5/22/2017 6:50:40 PM
Hello [userlink]Joey Ooi[/userlink],
You have to set a better accuracy in the global options (tools >> options >> calculation)
![2017-05-22 23_48_28-SMath Studio - [Page1_].png](/en-US/file/4wWzTR/2017-05-22-23_48_28-SMath-Studio---_Page1___png)
You have to set a better accuracy in the global options (tools >> options >> calculation)
If you like my plugins please consider to support the program buying a license; for personal contributions to me: paypal.me/dcprojects
#3 Posted: 5/22/2017 9:32:54 PM
Thank you for your quick and accurate response Davide.
It was very helpful indeed.
Best regards
Joey
It was very helpful indeed.
Best regards
Joey
#4 Posted: 5/23/2017 7:53:55 AM
Any idea what could be the problem?
Smath Simpson iterative integrator is known to range from
exact [sometimes] to poor. Increasing accuracy is fine in
this example but will not converge in other instances like
in the Fourier Quantum double integral. You can rescue
your example using the attached Romberg.
Mathematica online integrator link added.
Jean
Integrate Mathematica.sm
#5 Posted: 5/23/2017 8:37:48 AM
#6 Posted: 5/23/2017 8:53:58 AM
... like this
#7 Posted: 5/23/2017 2:32:33 PM
How many numerical integrators have been imagined ?
1+1+1+... Even Mathcad aren't so easy to deal with.
This visit is worth the time spent.
Jean
Fourier Quantum CoC Model.sm
1+1+1+... Even Mathcad aren't so easy to deal with.
This visit is worth the time spent.
Jean
Fourier Quantum CoC Model.sm
-
New Posts
-
No New Posts