-
-
Notifications
You must be signed in to change notification settings - Fork 26
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Underflow in stieltjes() when trying to compute recursion coefficients for very high degree #70
Comments
Hi there, would you be willing to post an MWE? I'll take a look then. :) |
Hi, Here is an example, almost directly pulled from the documentation:
This is of course trivial (these are just Jacobi polynomials) but it already breaks down. I am interested in more complicated weight functions ( x*cos(x a)^2 or even more complicated). |
I'll take a look. Sorry for the delay. :) I'll try the following (but go ahead and do it yourself): just call I'm not sure how the nature of the weight function influences the underflow --> |
just out of curiosity -- what's your use case anyway? |
Thanks for your help! I figured out how to call The use case is basically performing a change of basis to simplify numerical simulations of quantum many-body hamiltonians (see https://arxiv.org/abs/1006.4507). |
Mh, I guess reaching the numerical limit of the method. Have you tried the Do you really need so many coefficients anyway? Perhaps there is a way you can resort to a weight function to which there is an analytical solution to the three-term recurrence coefficients? But I'm sure you have thought about that. |
Hi there, isn't there an elegant way to compute the recurrence coefficients of the orthonormal polynomials directly from the recurrence coefficients of the monic orthogonal polynomials? I am almost positive. It must be in Gautschi's book on orthogonal polynomials. Here's the thing: if you had the recurrence coefficients of the orthonormal polynomials instead of the orthogonal monic ones, then you're fine. |
Hello,
I am trying to compute a high number of recursion coefficients (on the order of 500).
Using the OrthoPoly constructor as in the tutorial returns the error
I do not see a way to use `removeZeroWeights'.
I would be grateful for any help or tips in general about obtaining such high number of coefficients.
The text was updated successfully, but these errors were encountered: