[SciPy-user] Semidefinite programming in scipy ?

[Zhiwen Chong]

Hello Nils,

On 4-Jul-05, at 3:45 PM, Nils Wagner wrote:
> Actually, I am interested in eigenvalue optimization.
> Are you aware of freely available software for this
> topic ?

Unfortunately I know very little about eigenvalue optimization. If  
you can express your problem as an SDP, then you can use any number  
of SDP solvers. Most accept the SDPA input format (which looks fairly  
simple).
http://infohost.nmt.edu/~sdplib/FORMAT

Here's a list of current SDP solvers: (scroll down a little)
http://orion.math.uwaterloo.ca/~hwolkowi/henry/software/readme.html#sdp

The guys working in the building across from mine are maintaining  
SeDuMi.
Blurb: SeDuMi is a software package to solve optimization problems  
over symmetric cones. This includes linear, quadratic, second order  
conic and semidefinite optimization, and any combination of these. Free.
http://sedumi.mcmaster.ca/

Sorry I couldn't be of more help. Perhaps someone else on this list  
knows more about this subject.


Zhiwen

Side note: I took a peek at scipy.optimize; I don't think you'll find  
an SDP optimizer there. Most of the functions there are your standard  
line searches and basic algorithms like Newton, Trust Region,  
Conjugate Gradient, variable metric BFGS  and some DFO-type  
algorithms etc.