[gecode-users] Edge-finding evaluation

[Christian Schulte]

Hi Roger,

please find a patch against Gecode 3.4.2 attached: it disables timetabling
propagation for cumulative.

Best
Christian

--
Christian Schulte, www.ict.kth.se/~cschulte/


-----Original Message-----
From: Roger Kameugne [mailto:roger.kameugne at gmail.com] 
Sent: Monday, January 17, 2011 10:33 AM
To: cschulte at kth.se
Cc: users at gecode.org
Subject: Re: [gecode-users] Edge-finding evaluation

 Hi Christian,
 Let me see the patch and may be it can help.

 Best
 Roger

2011/1/14, Roger Kameugne <roger.kameugne at gmail.com>:
> Hi Christian,
> Let me see the patch and may it can help.
>
> Best
> Roger
>
> 2011/1/9, Christian Schulte <cschulte at kth.se>:
>> Hi Roger,
>>
>> What I could do: send you a patch against Gecode 3.4.2 such that all
>> additional features for cumulative are disabled and you only get Vilim's
>> algorithm (including overload-checking). You then would have to rebuild
>> Gecode from the 3.4.2 source after having applied the patch.
>>
>> Would that help?
>>
>> Best
>> Christian
>>
>> --
>> Christian Schulte, web.ict.kth.se/~cschulte/
>>
>>
>>
>> -----Original Message-----
>> From: Roger Kameugne [mailto:roger.kameugne at gmail.com]
>> Sent: Saturday, January 08, 2011 9:45 AM
>> To: cschulte at kth.se
>> Cc: users at gecode.org
>> Subject: Re: [gecode-users] Edge-finding evaluation
>>
>> Hi all,
>> I need help to built Vilim edge-finding propagator with the gecode
>> code source.  Is there someone who can help me? I am still study the
>> philosophy of gecode propagator.
>> Wishing all of you a very happy, prosperous and peaceful NEW YEAR 2011.
>> Roger
>>
>>
>>
>>
>> 2010/12/8, Christian Schulte <cschulte at kth.se>:
>>> Hi Roger,
>>>
>>> No, not right now. What you would have to do is to do a propagator by
>>> yourself: just copy the code from the cumulative propagator and remove
>>> the
>>> part where edge-finding is called (the code is in
>>> gecode/scheduling/cumulative/man-prop.hpp for mandatory tasks).
>>>
>>> Cheers
>>> Christian
>>>
>>> --
>>> Christian Schulte, www.ict.kth.se/~cschulte/
>>>
>>>
>>> -----Original Message-----
>>> From: Roger Kameugne [mailto:roger.kameugne at gmail.com]
>>> Sent: Wednesday, December 08, 2010 10:49 AM
>>> To: cschulte at kth.se
>>> Cc: users at gecode.org
>>> Subject: Re: [gecode-users] Edge-finding evaluation
>>>
>>> Hi all,
>>> Is there any possibility to post the cumulative constraint only only
>>> with time-table and averload checking? In other words, how can I post
>>> the cumulative propagator only for time table and overloard checking?
>>>
>>> Best
>>> Roger
>>>
>>> 2010/11/23, Christian Schulte <cschulte at kth.se>:
>>>> Dear Roger,
>>>>
>>>> That sounds very interesting.
>>>>
>>>> But first a word of warning: the cumulative scheduling propagator in
>>> Gecode
>>>> 3.4.2 is still experimental: we ourselves have not fully evaluated its
>>>> performance and so on (the documentation in MPG marks the scheduling
>>> support
>>>> as still experimental).
>>>>
>>>> As it comes to integration of your algorithm you might want to consult
>>> Part
>>>> P (Programming Propagators) in MPG (Modeling and Programming with
>> Gecode).
>>>> Implementing a new constraint requires the implementation of the
>>>> corresponding propagator as detailed there.
>>>>
>>>> The function stop(t, cout) stops the timer t and prints it on cout (a
>>>> standard output stream). Please consult the files:
>>>> 	gecode/driver/script.hpp
>>>> 	gecode/driver/script.cpp
>>>> In general these pieces of information are found best by searching the
>>>> reference documentation for Gecode.
>>>>
>>>> As an alternative, when you want to print to a textfile you should
>>> redirect
>>>> the output printed in a standard shell to a file, so there is
>>>> definitely
>>> no
>>>> need to hack the script stuff in Gecode.
>>>>
>>>> For setting the time limit, just invoke the -help switch of any of the
>>>> Gecode examples. It tells you how (also consult Chapter 11 in the first
>>> part
>>>> of MPG how to use the commandline driver).
>>>>
>>>> Best
>>>> Christian
>>>>
>>>> --
>>>> Christian Schulte, www.ict.kth.se/~cschulte/
>>>>
>>>> -----Original Message-----
>>>> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On
>> Behalf
>>>> Of Roger Kameugne
>>>> Sent: Tuesday, November 23, 2010 10:31 AM
>>>> To: users at gecode.org
>>>> Subject: [gecode-users] Edge-finding evaluation
>>>>
>>>> Hi all,
>>>> We have developed  a new cumulative edge-finding algorithm and we need
>>>> to evaluate it on RCPSP  with the current best know implement provide
>>>> in gecode 3.4.2  (Petr Vilim). We are new user of gecode and we have
>>>> used the Microsoft Visual Studio 2008.  We have many questions about
>>>> the the cumulative propagator.
>>>>
>>>> First of all, how can we integrate our edge-finding algorithm and used
>>>> the time-table and overload checking?
>>>>
>>>> To print out the runtime in a .txt file, we modify the file script.cpp
>>>> but we didn't know the type of  stop(t, cout)?
>>>>
>>>> The last but not the least question is about the setting of the time
>>>> limit. Where and how can we limit the runtime?
>>>> Thanks in advance.
>>>> Roger
>>>>
>>>> _______________________________________________
>>>> Gecode users mailing list
>>>> users at gecode.org
>>>> https://www.gecode.org/mailman/listinfo/gecode-users
>>>>
>>>>
>>>
>>>
>>
>>
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cumulative-3.4.2.patch
Type: application/octet-stream
Size: 8880 bytes
Desc: not available
URL: <http://www.gecode.org/pipermail/users/attachments/20110117/ec41811e/attachment.obj>