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
Those FinderAlgorithm implementations that use Newton's method (at any point during their lifetime) do not correctly reflect the relative error that the user requested. This is due to the user request not being passed down to the invocation of scipy.optimize.newton. This should be fixed!
Together with this fix we should implement the following feature to make algorithms using Newton's method more robust: Instead of an arbitrary number of sample points (NewtonGridAlgorithm) or an arbitrary measure for sufficiently close starting points (SimpleArgumentNewton) we should employ algorithms that adapt dynamically. In particular this could mean
successive refinement of the starting point grid until further refinement finds no additional zeros
starting Newton's algorithm as soon as a simple root has been detected - if convergence cannot yet be achieved, go for further subdivision
The text was updated successfully, but these errors were encountered:
As part of this issue we should replace all invocations of Newton's method with a facade that provides a unified interface. In the future his would enable us to replace the scipy.optimize.newton implementation with different implementations or even completely different algorithms.
Those
FinderAlgorithm
implementations that use Newton's method (at any point during their lifetime) do not correctly reflect the relative error that the user requested. This is due to the user request not being passed down to the invocation ofscipy.optimize.newton
. This should be fixed!Together with this fix we should implement the following feature to make algorithms using Newton's method more robust: Instead of an arbitrary number of sample points (
NewtonGridAlgorithm
) or an arbitrary measure for sufficiently close starting points (SimpleArgumentNewton
) we should employ algorithms that adapt dynamically. In particular this could meanThe text was updated successfully, but these errors were encountered: