I agree that having the equalities be automatically
translated into pairs of inequalities would be a nice
convenience.
But if the constraints are optional, what would the call
fmin_cobyla(f, vars, initx)
do?
Kostas
On 10/14/10 01:42 PM, Raymond Toy wrote:
> I was discussing some things about fmin_cobyla with David Billinghurst,
> but I think perhaps it's better suited to the mailing list.
>
> David wrote the documentation for fmin_cobyla (Thanks!) and mentioned
> that the equality constraint h(x) = 0 can be used with fmin_cobyla by
> adding the inequality constraints h(x)>= 0 and -h(x)>= 0, which
> fmin_coblya can handle.
>
> While the user can easily do this himself, I thought it would be nice if
> fmin_cobyla did this. It's not very hard. Also, the inequality
> constraints are optional, so it would be nice to make that parameter
> optional too.
>
> So, I propose fmin_coblya be changed from
>
> fmin_cobyla(f, vars, constraints, initx[, rhobeg, rhoend, iprint, maxfun])
>
> to
>
> fmin_cobyla(f, vars, initx[, ineq, eq, rhobeg, rhoend, iprint, maxfun])
>
> That is, the constraints parameter is now longer required and becomes
> the ineq keyword parameter, and add the eq keyword parameter.
>
> The output is also slightly changed. fmin_cobyla returns the value of
> each constraint. With equality constraints, fmin_cobyla appends the
> value of each equality constraint to the list of contraint values.
> (Perhaps this should be made a bit clearer with two lists returned: the
> values of the inequality constraints and the values of the equality
> constraints.)
>
> Ray
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima