-
Notifications
You must be signed in to change notification settings - Fork 33
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
improving multidimentional integrals? #86
Comments
Hi @jsitarek, I already thought about this, this is why I left a At the moment you can basically choose between two methods: the There is actually a test function checking the two integration methods ( I just wanted to say that I already started to think about this 😉, but let's discuss more! |
Hi @cosimoNigro |
I think it would be good to improve how the integrals are being done in agnpy, for the moment integration is done with simple traperoid integration in each dimention one by one over a predefined grid. There is no check if the integration has converged, and the points are not selected more densely in the regions where the integrant is varying fast. It would be good to use a dedicated multidimentional integration function with a speficied accuracy.
Connected with it, for the case of BLR sometimes the integrant function goes singular. Since the default binning is 50 bins over 5 decades in distance when you select e.g. r = 0.1 * R_line one of the points falls on top of the BLR shell. Proper multidimentional integral should hopefully solve this automatically
The text was updated successfully, but these errors were encountered: