bamos
commented
5 years ago Hi @Benjiou - just a quick response here, let me know if there's anything else!
1) if extra terms in the objective still make your problem a QP it's probably the easiest (although not necessarily the most efficient) to re-formulate your problem so that it's in standard form. It may be more efficient to implement a solver for your problem
2) You should be able to do this with \hat z = argmax f(z) = argmin -f(z)
Hi @bamos ,
Thank you for this promising work and clean code if you don't mind.
I have some questions related to the solver.
Is it possible to :
1) Add extra terms in the objective function and variable constraints ? https://github.com/locuslab/qpth/blob/master/qpth/qp.py#L36
A) extra terms in the objective function
\hat z = argmin_z 1/2 z^T Q z + p^T z+......+....+......B) Add more variables constraints : for instance add
y_, F_, K_, n_toQ_, p_, G_, h_, A_, b_2) Is this constraint solver can deal with maximum solution rather than minimum ?
\hat z = argmax_z 1/2 z^T Q z + p^T z+......+....+......Thank you a lot for your answer