[Clp] Fwd: Fwd: C Code of Concorde

Fri Mar 7 10:17:36 EST 2014

Sir,
I'm just repeating here some mail which I sent to Julian Hall,
since I was not able to find this CIp site.
Prof. William Cook from the Univeristy of Waterloo (see mail below)
told me to contact you. We maybe quite similar in age (I'm becoming 70
at the end of April).

My name is Wolfgang M. Hartmann. From 1986 until 2007 I have been working
as a Senior Research Statistician and Principal Developer at SAS Institute
Inc.
in Cary NC. Among many other pieces of software, especially linear algebra
and nonlinear methods in statistics, I did write PROC NLP for the OR
product of SAS.
(I have always been in good terms with Prof. Mike Powell, Prof. Kaj Madsen,
and
Stephen Wright.) For more on CMAT you may have a look at:
www.wcmat.com/cmat
and also at my entry at LinkedIn.

In 2007 I retired from SAS and returned to Germany where my wife was living
since 1999.
For now almost 20 years I'm working on a interactive matrix language which
I call CMAT.
Only a few months ago I started to look into discrete optimization and the
TSP (traveling
salesman problem). The probably best software in this field is Concorde. As
you can see
below, by mail from Prof. William Cook, I've got the source code of
Concorde which has
an interface to the most straight forward LP code QSopt, which is, however,
not
available yet. But Prof. Cook writes that Prof. John Forrest made an
interface between
CIp and Concorde.

I wonder if I could use your code for that?
CMAT is given away with the Gnu license, that means, I'm only requiring
"nonprofit use".
Only for the code of Concorde I had to restrict applications of the PSP
function
to "Academic Research only" which is of course much stronger. Due to some
contributors of other functions of my software CMAT, I cannot make the code
of all CMAT public.

Hoping for an answer,
Wolfgang Hartmann, Heidelberg
Tel.: + 49 6221 616991

---------- Forwarded message ----------
From: Julian Hall <J.A.J.Hall at ed.ac.uk>
Date: Fri, Mar 7, 2014 at 4:03 PM
Subject: Re: Fwd: C Code of Concorde
To: Wolfgang Hartmann <cmat.wolfgang at gmail.com>
Cc: "William J. Cook" <bico at uwaterloo.ca>

Dear Wolfgang

Thanks for this. John Forrest can be contacted via the address on

http://www.fastercoin.com/

He's the only person who can maintain Clp and I hope that he can help you
with the Concorde interface.

With best wishes,

Julian Hall

Sir

 when I went to the website of CIp I found your name. And since I have no
> idea
> how to contact Prof. Forrest I will just bother you with my problem:-)
>
> My name is Wolfgang M. Hartmann. From 1986 until 2007 I have been working
> as a Senior Research Statistician and Principal Developer at SAS Institute
> Inc.
> in Cary NC. Among many other pieces of software, especially linear algebra
> and nonlinear methods in statistics, I did write PROC NLP for the OR
> product of SAS.
> (I have always been in good terms with Prof. Mike Powell, Prof. Kaj Madsen,
> and
> Stephen Wright.) For more on CMAT you may have a look at:
> www.wcmat.com/cmat
> and also at my entry at LinkedIn.
>
> In 2007 I retired from SAS and returned to Germany where my wife was living
> since 1999.
> For now almost 20 years I'm working on a interactive matrix language which
> I call CMAT.
> Only a few months ago I started to look into discrete optimization and the
> TSP (traveling
> salesman problem). The probably best software in this field is Concorde. As
> you can see
> below, by mail from Prof. William Cook, I've got the source code of
> Concorde which has
> an interface to the most straight forward LP code QSopt, which is, however,
> not
> available yet. But Prof. Cook writes that Prof. John Forrest made an
> interface between
> CIp and Concorde.
>
> I wonder if I could use your code for that?
> CMAT is given away with the Gnu license, that means, I'm only requiring
> "nonprofit use".
> Only for the code of Concorde I had to restrict applications of the PSP
> function
> to "Academic Research only" which is of course much stronger. Due to some
> contributors of other functions of my software CMAT, I cannot make the code
> of all CMAT public.
>
> Hoping for an answer,
> Wolfgang Hartmann, Heidelberg
> Tel.: + 49 6221 616991
>
>
>
>
>
>
> ---------- Forwarded message ----------
> From: William J. Cook <bico at uwaterloo.ca>
> Date: Thu, Mar 6, 2014 at 4:14 PM
> Subject: Re: C Code of Concorde
> To: Wolfgang Hartmann <cmat.wolfgang at gmail.com>
>
>
>
>  Dear Wolfgang, I thought you were being a bit optimistic about building
> a concorde interface to your current LP solvers.  Concorde makes use
> of a wide variety of calls for manipulating an LP -- it is used by some
> commercial codes to test their LP libraries.  Commercial codes that have
> interfaces that can support Concorde are CPLEX, Gurobi, and MOSEK.
> One option for your code may be the open source COIN-OR CLP project
>
>  https://projects.coin-or.org/Clp
>
>  I know that John Forrest, the original creator of Clp, created a Concorde
> interface for testing the Clp library.  I do not believe that he made a
> public
> release of the interface, however.
>
>  QSopt is currently only distributed as a pre-buildt library.  I hope to be
> able
> to get the permissions to release QSopt as a source-code library in the
> first
> half of 2015.
>
>  Best regards,
>
>  Bill
>
>
>
>  On 2014-03-06, at 8:23 AM, Wolfgang Hartmann <cmat.wolfgang at gmail.com>
>  wrote:
>
>  Sir,
> looks like that I had my mouth too full when I said that
> the LP code wouldn't be a problem:-)
> I have been fuzzing around with that for some time now.
> Of course, it would be much easier if I would have the QSopt
> LP solver for which you already have an interface.
> May I ask
> - would you have that QSopt solver
> - and if so, what would be the conditions to use it?
>
>  Sorry for my big mouth,
> Wolfgang Hartmann
>
> On Wed, Jan 29, 2014 at 5:33 PM, Wolfgang Hartmann
> <cmat.wolfgang at gmail.com>wrote:
>
>  Sir,
>> please, I don't want to bother.
>> But the term "academic research" use  would not
>> be the same as "nonprofit" use. There are overlaps.
>> And I require for CMAT "nonprofit" from my users,
>> otherwise I would need to be contacted.
>> I could put your exact phrase into my manual and leave it
>> to my users to contact you? naturally, I cannot enforce
>> from my users "academic research" when I ask them until
>> now for "nonprofit" use.
>>
>>  The LP code is no problem,
>>  I have a number of LP codes in CMAT, not only
>> lpsolve which I already mentioned. Steve Wright and E. Ngo gave me the
>> permission to include pcx and I have another beautiful and ver fast LP
>> solver
>> if I can specify a bounded region.
>>
>>  Kind regards, Wolfgang
>>
>>
>>
>>
>> On Wed, Jan 29, 2014 at 5:10 PM, William J. Cook <bico at uwaterloo.ca
>> >wrote:
>>
>>
>>>  Dear Wolfgang, I am happy to permit you to distribute concorde as
>>> part of CMAT, following our license "Permission is granted for
>>> academic research use.  For other uses, contact the authors
>>> for licensing options."   The C code is available on the concorde
>>> Web page
>>>
>>>  http://www.math.uwaterloo.ca/tsp/concorde/downloads/downloads.htm
>>>
>>>  The main exact solver requires also an LP code, so it would not be
>>> easy to include this functionality in CMAT.
>>>
>>>  Best regards,
>>>
>>>  Bill
>>>
>>>
>>>
>>>  On 2014-01-29, at 7:53 AM, Wolfgang Hartmann <cmat.wolfgang at gmail.com>
>>>  wrote:
>>>
>>>
>>> Dear Sir,
>>> my name is Wolfgang Hartmann, living in Heidelberg, Germany.
>>> More about me at our website: www.wcmat.com
>>>
>>>  Starting in 1995, when I was still employed as principal Developer
>>> and Senior Research Statistician at SAS Institute Inc. in Cary NC,
>>> I developed a matrix language similar to Matlab (SAS IML) which
>>> I called CMAT. It is not open source since some people who gave
>>> me their code would not have given me the right to make it public.
>>>
>>>  I only have a very small group of people who use CMAT for very
>>> specific applications and all without making any profit out of it.
>>> I also post on my website that my software comes for free only
>>> for nonprofit applications, but nobody has any use of CMAT for profit:-)
>>>
>>>  Would it be possible for you to grant me access to Concorde
>>> for including it into CMAT? Naturally I would prefer the original C code
>>> (since I always try to stay platform independent) but could probably
>>> also cope with a Windows DLL, as I do with lpsolve.
>>>
>>>  With kind regards,
>>> Wolfgang Hartmann
>>>
>>>  --
>>>  Time is not Onedimensional!
>>> And even if it were - it would still be too difficult for me to
>>> handle.---
>>>  Wolfgang M. Hartmann
>>>
>>>  On YouTube: http://www.youtube.com/watch?v=HE16lQWprwg&feature=youtu.be
>>>  www.wcmat.com/wolfgang
>>>
>>>
>>>
>>>
>>
>
>
>
> --
> Time is not Onedimensional!
> And even if it were - it would still be too difficult for me to handle.---
> Wolfgang M. Hartmann
>
> On YouTube: http://www.youtube.com/watch?v=HE16lQWprwg&feature=youtu.be
>  www.wcmat.com/wolfgang
>
>

Julian Hall (Operational Research MSc Programme Director)
--
Dr. J. A. Julian Hall, Senior Lecturer, School of Mathematics,
University of Edinburgh, JCMB, King's Buildings, EDINBURGH, EH9 3JZ, UK.
Room: 6221   Phone: [+44](131) 650 5075   Email: J.A.J.Hall at ed.ac.uk
Fax: [+44](131) 650 6553   Web: http://www.maths.ed.ac.uk/hall

-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

-- 
Time is not Onedimensional!
And even if it were - it would still be too difficult for me to handle.---
Wolfgang M. Hartmann

On YouTube: http://www.youtube.com/watch?v=HE16lQWprwg&feature=youtu.be
 www.wcmat.com/wolfgang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/clp/attachments/20140307/cdcf9694/attachment.html>