
Re: Terrible algebraic system
Posted:
Feb 23, 2012 5:29 AM


Ach...so, Waldek, would you suggest using some other approach? I was thinking about using a fixed point iteration, but under my conditions a>L>0, b>U>0:
a=(2/p)*(aL)^2/(a+b) b=(2/p)*(bU)^2/(a+b)
the 1norm of the first row of the Jacobian is already not less than 1. Any suggestions? I guess I'll just write a NetwonRaphson subroutine for systems and hold fingers crossed :)
Sergio

