You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the only type of transformation used for handling infinite bounds is $u \mapsto a + t/(1-t)$ and $u \mapsto t/(1-t)^2$. It could be nice to eventually support other transformations.
In some of the integrals I've dealt with in my work, I've found that transformations like
which typically gave better results when applying Gauss-Legendre quadrature afterwards. These integrals I dealt with had issues with oscillations and singularities, etc., so the currently used transform is still a good default. Another useful transform is $t \mapsto (2/\pi)\arctan(t)$, giving
Currently the only type of transformation used for handling infinite bounds is$u \mapsto a + t/(1-t)$ and $u \mapsto t/(1-t)^2$ . It could be nice to eventually support other transformations.
In some of the integrals I've dealt with in my work, I've found that transformations like
and
could be useful, giving the (rather complicated..) results
which typically gave better results when applying Gauss-Legendre quadrature afterwards. These integrals I dealt with had issues with oscillations and singularities, etc., so the currently used transform is still a good default. Another useful transform is$t \mapsto (2/\pi)\arctan(t)$ , giving
The text was updated successfully, but these errors were encountered: