[Coin-discuss] MOSEK interface

[Brady Hunsaker]

Hi Bo,

I agree that improved documentation for interface authors/maintainers
is important.  I hope to work on that before too long, but I may wait
until the OSI redesign is complete.

I ran into some of the same questions when I was working on the GLPK
interface.  I'll be happy to try to answer your questions as they come
up.  In some cases, there is no definitive answer yet.

getColSolution without setColSolution should be unambiguous.  

setColSolution was discussed a while back.  It's really more important
for interior-point solvers; not simplex solvers.  Many simplex solvers
can not or do not use the values given.  

I don't know much about MOSEK, but what I've read indicates that it
has both interior-point and simplex solvers.  For the simplex solvers,
if MOSEK does not have a natural way to use this information, then
don't use it.

If you are using a cache for data, then I recommend changing the
column values in the cache but not actually telling MOSEK.  Personally
I don't think that's the best approach, but I think that's the most
common approach to facilitate use of unitTest for testing.  At some
point we will discuss this issue again, but it should wait for the OSI
redesign and possibly for incorporation.

If this didn't answer your question, please clarify your concern and
I'll try again.

Brady

On Fri, Aug 01, 2003 at 06:06:28PM +0200, Bo Jensen wrote:

> As ealier noted I'am working on an interface to MOSEK, but have 
> difficulties with the documentation. What is the expected behaviour of 
> setColSoution and getColSolution in OSI ? The different solvers do quite 
> different things.
> Bo
> 
> 
> *************************************************************************
> Bo Jensen
> Email: bo.jensen at mosek.com
> 
> MOSEK ApS
> C/O Symbion Science Park
> Fruebjergvej 3, Boks 16
> DK-2100 Copenhagen O
> Denmark
> *************************************************************************
> 
> _______________________________________________
> Coin-discuss mailing list
> Coin-discuss at www-124.ibm.com
> http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-discuss
----------------
Brady Hunsaker
Georgia Institute of Technology
Program in Algorithms, Combinatorics, and Optimization
School of Industrial and Systems Engineering

E-mail address:   hunsaker at isye.gatech.edu