[SciPy-User] Lagrange Multipliers in optimize.slsqp

[Guilherme P. de Freitas]

I was wondering if it is worth outputting the multipliers at all...
The low accuracy in a very simple example (see my previous post) makes
me wonder how innacurate the estimated multipliers can be in a more
complex problem.

On Thu, Mar 4, 2010 at 5:55 PM, Rob Falck <robfalck at gmail.com> wrote:
> Any opinions on how this change should be implemented?  I'd prefer to
> keep the code clean, and the solution below is simple, but I'd rather
> not break current implementations of SLSQP that people have employed,
> as inserting a new element into the output likely would.  I'm leaning
> towards allowing both boolean or integer types for the full_output
> argument.  0 or 1 would work as True and False do now, and 2 would
> return the Lagrange multipliers as well as everything else.



-- 
Guilherme P. de Freitas
http://www.gpfreitas.com