RE: [Fwd: help for glpk]

[Meketon, Marc]

There are other fast solvers.  Two open source instances are CBC and SCIP.  And 
there are faster commercial solvers (Gurobi, Xpress, Ilog/Cplex).  And more.

However, if you want to get a 5-10 times speedup, is it because you intend to 
solve many of these types of problems?  If so, consider:

- Using multiple processes at the same time (GLPK by itself isn't 
multi-threaded, but you could execute glpsol many times)

- Using 'warm start'.  If the problems only differs by the objective function, 
then using the optimal basis from the last problem solved to be the initial 
basis of the next problem may reduce execution time.  Using the dual simplex, 
you could do this if only the RHS changes each time.

- If you are using GMPL (the AMPL-like language that is part of GLPK) and then 
glpsol.exe, then write a CPLEX-like .lp file directly and call glpsol.exe 
passing that .lp file, and skipping .mod file.

-----Original Message-----
From: Help-glpk <help-glpk-bounces+marc.meketon=oliverwyman.com@gnu.org> On 
Behalf Of Andrew Makhorin
Sent: Sunday, June 27, 2021 5:40 AM
To: help-glpk@gnu.org
Subject: [Fwd: help for glpk]

-------- Forwarded Message --------
From: yegen06@sina.com
Reply-to: yegen06@sina.com
To: mao <mao@gnu.org>
Subject: help for glpk
Date: Sun, 27 Jun 2021 11:57:36 +0800

> Hi Andrew，
>       How are you? Sorry for bother you. I have some questions about glpk,I
> will be very appreciated if I can get some help from you.
>       I want solve a large scale LP problem,about 20 variables and 600
> constrains like a <= Ax <= b , and I tried glpk to solve it . It
> spends about 1ms to solve my prolem in my PC,very fast.
> But,it still does not satisfy my need, I wonder if there is optimizer
> possibility for the performence. I hope the problem can be solve
> within 100~200us.
>         The constrian max is a sparse matrix,like form:
>       X  X  X
>       X  X  X
>       ... ...  ...
>       X  X  X
>           X  X  X
>           X  X  X
>           ... ...  ...
>           X  X  X
>               X  X  X
>               X  X  X
>                ... ...  ...
>                X  X  X
>                      ... ...  ...
>
>       I wonder if GLPK will do some optimize for this kind of problem?
> look forward to your reply. thanks a lot.


________________________________
This e-mail and any attachments may be confidential or legally privileged. If 
you received this message in error or are not the intended recipient, you 
should destroy the e-mail message and any attachments or copies, and you are 
prohibited from retaining, distributing, disclosing or using any information 
contained herein. Please inform us of the erroneous delivery by return e-mail. 
Thank you for your cooperation.