generalized Lambert



 

--Barton

On 30/09/2011 8:11 PM, Barton Willis wrote:
> My generalized Lambert code doesn't do numerical (binary64 or big float) evaluation.
> If you are looking for something fun to do, the project is yours.
>
> --Barton
I'm interested in taking this on. ?I have some numerical code working - 
just a matter of providing a starting guess on the correct branch - 
although some more testing is in order.

?? ? David

Great--numerical code will help me test my code that decides when
generalized_lambert_w(k, x*exp(x)) = x. So far, I've only tried the
k=0 & k=-1 cases.

On the tiny chance you haven't seen it, page 16 of
http://www.cs.uwaterloo.ca/research/tr/1993/03/W.pdf is helpful. Also,
the picture

  wxdraw2d( x_voxel = 60, y_voxel = 60, fill_color = navy, region((x*sin(y)+y*cos(y)) > 0,x,-9,9,y,-9,9));

has helped me (and looks cool too).

--Barton