[GSoC-PortA] factor constraints

[Ross Bennett]

Brian,

Thanks for the note. In a recent commit (r2783) I added a block of code to
constrained_objective that adds to the penalty term for weights that
violate the factor exposure constraint.

In commit r2784 I added DEoptim and random portfolios to the factor
exposure testing script. The examples compare group constraints and
industry exposure constraints that are the equivalent. The group
constraints are satisfied based on the fn_map function, but the industry
exposure constraints are handled using the penalty in
constrained_objective. Because of this I am getting different optimal
weights which I think should be expected, but I'm not entirely sure this is
ok. Although the weights are different, they do satisfy the constraints.
The objective measures (mean in the examples) are close - I am getting
about 0.0077 using group constraints and anywhere from 0.0075 to 0.0076
using factor exposure constraints.

Let me know if you have any other suggestions or if there is a better way
to handle the factor exposure constraint for the global numeric solvers.

Thanks,
Ross


On Wed, Aug 14, 2013 at 5:27 AM, Brian G. Peterson <brian at braverock.com>wrote:

> I see several commits in the past few days about factor and exposure
> constraints.
>
> That's fantastic.
>
> Now we need to make sure they will work with the global numeric solvers
> too.
>
> For that to happen, we need to handle them in the constrained.objective
> function as well.
>
> I'm thinking that we can always check them and account for them in
> constrained.objective.  In all the 'special case' optimizations designed
> for LP/QP etc solvers, we'll be using a different objective function anyway.
>
> Regards,
>
> Brian
>
> --
> Brian G. Peterson
> http://braverock.com/brian/
> Ph: 773-459-4973
> IM: bgpbraverock
>
> ______________________________**_________________
> GSoC-PortA mailing list
> GSoC-PortA at lists.r-forge.r-**project.org<GSoC-PortA at lists.r-forge.r-project.org>
> http://lists.r-forge.r-**project.org/cgi-bin/mailman/**listinfo/gsoc-porta<http://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/gsoc-porta>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/gsoc-porta/attachments/20130814/6e561d1a/attachment.html>