[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-glpk] Strange Question

From: cgi2
Subject: Re: [Help-glpk] Strange Question
Date: Mon, 08 Mar 2004 10:24:29 -0800

I did not have this problem (GLPK crashing) on the windows dll for version 4.0. 
I did have the problem on Linux.
When I upgraded to 4.4, I started getting this problem on both Windows and 
linux. Since there is a standard makefile
for Linux, I think this rules out the makefile as the culprit.

On Wed, 3 Mar 2004 17:30:00 -0700
"Welson Sun" <address@hidden> wrote:
>Thanks Michael,
>Although I am not quite understand what you said, at least I know that do
>not use MIP routine to retrieve answer if you are using simplex algorithm.
>Regarding the second question, it turns out that it is the problem with the
>JNI.dll. I switched to another Glpk JNI dll from a german website, and it
>runs perfectly! The real problem with the previous dll is that the memory is
>really toasted after you run the simplex algorithm so that you can not
>allocate new objects any more................
>So, still, I think it is necessary that there needs a make file for
>generating JNI dll under Windows, so that we don't need to try out various
>DLLs from various people.
>-----Original Message-----
>From: Michael Hennebry [mailto:address@hidden 
>Sent: Wednesday, March 03, 2004 8:45 AM
>To: Welson Sun
>Cc: address@hidden
>Subject: Re: [Help-glpk] Strange Question
>On Tue, 2 Mar 2004, Welson Sun wrote:
>> Hi all, I have met a strange question with Glpk, can you help me out?
>> I am using Glpk in Cygwin under WindowsXP. Since currently there is no 
>> Windows makefile for Glpk JNI, I used the precompiled version from 
>> address@hidden .
>> My LP problem is quite simple:
>> Minimize C1X1 + C2X2 + ..... + CnXn
>> So that
>>     Xi - Xj <= Wij ( 1 <= i,j <= n )
>> Where all X are integer and all Wij are non-negtive integer.
>> The first strange problem is that I can use simplex() method to solve 
>> this, while I cannot use integer() method to solve this, the later will
>> GlpkMsg: lpx_integer: optimal solution of LP relaxation required
>> I have set the problem to be MIP type, so how can this be possible?
>If you RTFM, you will discover that lpx_integer may only follow lpx_simplex.
>If, as seems likely, the constraint matrix is totally unimodular, you only
>need lpx_simplex anyway.
>If you only use lpx_simplex, don't try to fetch an MIP solution.
>The other, I can't help you with.
>Mike   address@hidden
>"Two roads divurged in a woods, and I took the road less travelled by...
>and they CANCELLED MY FRIKKIN' SHOW. I totally shoulda took that road
>with all those people on it. Damn."                           --  Joss
>Help-glpk mailing list

Don't E-Mail, ZipMail!

reply via email to

[Prev in Thread] Current Thread [Next in Thread]