(28/48/50) Lambert W Function

[John Keith]

(04-02-2023 11:12 PM)Albert Chan Wrote:  f=0 guaranteed we get the correct branch, which is nice.
There is no need to use complicated W guesses.

lua> a, k = 3+4*I, 5
lua> x = 2*k*pi*I -- guess for W_k(a)
lua> repeat h=f(x)/df(x); x=x-h; print(x, I.abs(h)) until x == x+1e-6*h

(-1.815554248884793+30.714634540472343*I)       1.9462907525577031
(-1.817005891456528+30.71333413897562*I)        0.0019489253984594729
(-1.8170058918466274+30.713334137004896*I)      2.0089622485451688e-009
(-1.8170058918466274+30.713334137004896*I)      0

>>> from mpmath import *
>>> lambertw(3+4j, k=5)
mpc(real='-1.8170058918466274', imag='30.713334137004896')

Thanks for the John D. Cook link, I hadn't seen that one before.

My program gets exact or near-exact results for all branches, but it is slower for branches other than 0 and -1. I will try your code and see how it compares.