cvxgrp / qcml

A Python parser for generating Python/C/Matlab solver interfaces
Other
42 stars 9 forks source link

Question: norm(x, p)? #43

Open steve3141 opened 10 years ago

steve3141 commented 10 years ago

Thanks for a brilliant package. I see that CVX (matlab) supports p-norms for values of p>=1 other than 1, 2, \infty. I'm wondering if there are any plans to support that in QCML?

echu commented 10 years ago

I don't have any plans to do this yet; is there a specific kind of norm you wanted to use?

On Fri, Jun 27, 2014 at 9:45 AM, steve3141 notifications@github.com wrote:

Thanks for a brilliant package. I see that CVX (matlab) supports p-norms for values of p>=1 other than 1, 2, \infty. I'm wondering if there are any plans to support that in QCML?

— Reply to this email directly or view it on GitHub https://github.com/cvxgrp/qcml/issues/43.

steve3141 commented 10 years ago

Thanks. I don't have any real requirement for other p-norms at the moment, I'm really just trying to compare CVX and QCML to use for future projects.

I dug into CVX a bit just now and I see that for p different from 1, 2, \infty it converts a p-norm into a semidefinite cone, thus generating an SDP not an SOCP. So it seems that wouldn't fit into QCML unless it expands from generating SOCP to generating SDP. (Or unless it uses an alternative transformation that produces an SOCP...)