echu
commented
11 years ago okay; sounds great. i'll take a look at it. by the way, i changed the underlying solver to ecos instead of cvxopt. furthermore, ecos takes numpy/scipy arrays and sparse matrices as arguments and that's what prob2socp outputs.
if you want to call CVXOPT's solver, you'll have to convert the arrays to cvxopt matrix and the scipy matrices to cvxopt spmatrices manually before calling the solver. just fyi.
On Sat, Sep 7, 2013 at 2:47 PM, Peter H. Li notifications@github.comwrote:
Here's my attempt at formulation. The constraints started working after pulling, but then I realized I had them wrong...
They need to be simple elementwise constraints on the values of these vectors in parallel. I wasn't sure if that was notatable so I made up this _i notation to indicate the ith subscript of each vector.
Let me know if that makes sense...
You can merge this Pull Request by running
git pull https://github.com/chinasaur/qcml cbp/1
Or view, comment on, or merge it at:
https://github.com/cvxgrp/qcml/pull/24 Commit Summary
- Starting to formulate CBP
- Merge commit 'c1ce1b87df53cced27e717dd50151cca2d2013e2' into cbp/1
- Moved
- Filled in remaining constraints
- Fixed constraints using made up notation
- Didn't mean to leave this here...
- Or this
- Finished "more natural" comment
File Changes
- A examples/cbp.pyhttps://github.com/cvxgrp/qcml/pull/24/files#diff-0(54)
Patch Links:
Here's my attempt at formulation. The constraints started working after pulling, but then I realized I had them wrong...
They need to be simple elementwise constraints on the values of these vectors in parallel. I wasn't sure if that was notatable so I made up this _i notation to indicate the ith subscript of each vector.
Let me know if that makes sense...