Hi,
if the peak memory is during construction of the engine, it means it
occurs during your problem setup, not during search. Maybe you can
explain a little more what you're doing there - do you use a lot of
intermediate (STL?) data structures, perform symbolic computations,
read in large files o
Lars Kotthoff wrote:
> ./configure --enable-static --disable-dynamic && make
> in a clean source directory builds the static and dynamic libs (by
> which I mean
> the .a and .so files) and links the example binaries dynamically, e.g.
Ah, sorry, it's called --disable-shared, not --disable-dynamic
The Gecode libs themselves cannot be linked statically (e.g. against
libstdc++), they're just archives of object files, after all. One
problem with the current build system is that if you've built dynamic
libs before and they're still in the directory, they will be used for
linking. If yo
Hi Lars,
we currently don't support creating both static and dynamic libraries,
and our configure script has a bug, it doesn't switch off dynamic libs
when you say --enable-static. Please try --enable-static --disable-
dynamic, I think it should work.
Cheers,
Guido
Lars Kotthoff wr
Well, I don't really understand your model. What's an incrementation?
The only way I can imagine how you want your model to work is to just
split the element constraint in two. I.e., have two arrays a0 and a1
of integers (or IntVars), such that is the edge i, and
then use element(a0,b,d0
Guido
Thanks for your quick answers.
The problem is solved with the new package.
You can put the last installer on the download page,
will you change its version number ?
Yours,
Serge
On 3 nov. 09, at 10:30, Guido Tack wrote:
Hi Serge,
apparently Apple has changed the libstdc++ in Snow Le
ks perfectly.
Perhaps there is a problem when using Xcode; I am compiling
without Xcode.
2009/10/30 Guido Tack
serge lemouton wrote:
> I am on MacOs X (10.5.8) and using Xcode.
> When the examples are compiled (with gcc 4.2) and linked against
the
> gecode framework, I get the
Hi Denys,
Denys Duchier wrote:
> A colleague of mine has been having quite a bit of difficulty trying
> to
> install Gecode on her brand new Mac. When compiling from sources for
> 32bits: I added -m32 to CFLAGS and CXXFLAGS, but it turns out that's
> not
> sufficient. I had to tweak the Make
serge lemouton wrote:
> I am on MacOs X (10.5.8) and using Xcode.
> When the examples are compiled (with gcc 4.2) and linked against the
> gecode framework, I get the following error:
>
> dyld: lazy symbol binding failed: Symbol not found:
> __ZSt16__ostream_insertIcSt11char_traitsIcEERSt13basic_os
Ellis Breen wrote:
> 2009/10/29 Guido Tack
> we're currently implementing an interface to the Numberjack Python
> library (http://4c110.ucc.ie/numberjack/home). It's work in
> progress, but we expect enough progress for it to be usable soon.
> The interfac
Hi Ellis,
Ellis Breen wrote:
> I'm currently looking for a free-as-in-beer constraint programming
> library with bindings in both C/C++ and Python. It seems Logilab
> started a Python wrapper for Gecode but never completed it. Are
> there any rival implementations? Would it be non-trivial to
It was fixed in 2.2.0 and we forgot to mention it in the change log,
sorry.
Cheers,
Guido
David Rijsman wrote:
> Hi Guido,
>
> what version has this been fixed, I checked the release notes but
> can not find anything?
>
> thanks,
>
> David
>
>>&g
Hi David,
David Rijsman wrote:
> I have a question on the method "void VarArray::resize(Space&
> home, int m)".
>
> I am on version 2.1.1 and noticed the method has changed somewhere
> between version 2.1.1 and the 3.2.0. I can not find any reason in
> the change list for the change.
>
> I
Hi everyone,
we changed the layout of our web pages to something that hopefully
works better on more browsers and screen resolutions. Additionally,
we will from now on post news about Gecode more regularly, on the
front page as well as our new dedicated news page
(http://www.gecode.org/new
Hi Fred.
There are currently no plans to add that functionality. Actually, the
run method may be in the wrong class, as FlatZincSpace is really just
meant to be a space that you can initialize from FlatZinc. In order
to perform incremental search, just use a DFS search engine as you
wou
nst&, Gecode::IntVarArgs const&, Gecode::PrimArgArray const&,
> bool,
> Gecode::IntConLevel)'
> collect2: ld returned 1 exit status
> make[1]: *** [tools/flatzinc/fz] Error 1
>
>
> This does not occur when removing --disable-scheduling from the
> confi
Thanks, Denys. It was actually a bug in our Makefile that only
occurred on systems that don't have a recent enough flex/bison
installed. It's fixed in the trunk and the 3.2.0 tag now, and we'll
update the source packages accordingly (as it only affects the build
system).
Cheers,
Hi!
The problem is not Propagator::cast, because that method has not been
changed (it was private in the previous release, too). Can you please
let us know what version of Gecode you used previously, and give us an
idea where you call Propagator::cast (e.g. small code sample)?
Cheers,
Dear all,
we are releasing the next version of Gecode.
Gecode 3.2.0
http://www.gecode.org
This release has some important bug fixes (in particular for
global cardinality aka count), the documentation has been
improved (worked around some issues with generation by dox
Fred Spiessens wrote:
I'm building an interface to Gecode from Smalltalk via FlatZinc,
using Gecode Version 3.1.0. (on Mac OSX 10.5 Intel)
FlatZincGecode::runEngine uses "std::cout" to print statistics
information while it uses "out" (the outstream provided as
parameter) to print the results
Chris Mears wrote:
> Hello Gecoders,
>
> I'd like to perform a depth first search with a specific static
> variable
> ordering. I have a matrix of integer variables and I want to label
> the
> top row first, then the left column, then the rest of the matrix in
> standard top-to-bottom, left-t
Hi,
if you just want to encapsulate two IntVars to make modeling easier,
simply implement a wrapper class that forwards the relevant operations
(such as update) to the two component variables. There's nothing
special about IntVars, they're plain C++ objects and you can put them
in whatev
Jan Kelbel wrote:
> I was studying the DivBnd propagator, and in my opinion, there could
> be
> performed some propagation more tight.
>
> It is in method
> DivBnd::propagate(Space& home, const ModEventDelta&).
> In section prop_xpx: starting at line 242 of divmod.hpp file (Gecode
> 3.1.0 and al
Tarek Garoui wrote:
> I created an IntVar like this:
>
> IntVar id = new IntVar(this, "ID", new Integer(0), new Integer(max));
>
> After that, I want to post a constraint consisting of a BExpr where
> I need the value of the IntVar id. But I get the following exception:
> org.gecode.NotAssigned.
Martin Kreißig wrote:
> Alright. Thanks. But that is not the reason in my original code. I
> missed it when writing the example.
>
> How come you get 3 solutions? I get the following output (which are
> the correct solution inserted):
> {{-2,1}, {-1,5}, [2..3], {-4,3}, {-2,2}}
> {-2, -1, 3, -4
oid print( void ) {
> std::cout << v << std::endl;
> }};
>
> int main( void ) {
> int arr[5][2] = {{-2,1},{-1,5},{3,2},{-4,3},{2,-2}};
> CSP* c = new CSP(arr);
> c->print();
> DFS dfs(c);
> delete c;
> CSP* cc;
&
Hi everyone,
I've just upgraded my Mac to Snow Leopard and XCode 3.2 and ran a few
quick checks. It seems everything works well. Here's a couple of
notes:
The installer package from the Gecode download page apparently works
out of the box.
It seems that the installer package for Qt 4.5 h
Martin Kreißig wrote:
>> You can use opt.solutions(0) with DFS, too.
>
> Alright, but how do I pass opt to my CSP object?
You don't. The Options class is for use with the driver
infrastructure. So either you use the driver, and Script::run, or you
don't use options. There's a different opti
Malcolm Ryan wrote:
> I need a search engine for gecode that implements best-first search,
> i.e. search using a priority queue sorted on an heuristic. It doesn't
> look like Gecode current implements this, so I figure I'll have to
> write it myself.
>
> Any recommendations on how I should go abou
Martin Kreißig wrote:
> Hello,
>
> I found out, that GECODe is a very good (if not the best) solution
> to my problem.
> The thing is, that I never used C++ before - everybody has to start
> at a point ;)
>
> So far I got used to it, but now - when finished - I don't get all
> solutions of m
Hi,
this means that libgecodeminimodel.so is not properly linked. Make
sure it's on the path that the linker searches.
Guido
cristian wrote:
> Hi
> I'm working scheduling with cumulatives and others constraints
> using gecode 3.02 ,eclipse and ubuntu.
> I was working correctly until my ubun
amina kemmar wrote:
> Hi,
>
> I am writing a new constraint "NewConstraint", which takes 3 IntVar
> as arguments :
>
> #include
> #include
>
> namespace Gecode {
>
> using namespace Int;
>
> void NewConstraint(Space& home, IntVar x0, IntVar x1, IntVar x3) {
>
> if (home.failed()) retur
Hi Kayler,
you should not use IntArgs for representing your data, only for
posting constraints. So, when you read in your data, just use e.g.
normal integer vectors. Then you post the constraints, and you know
exactly how many you need. For each posted constraint, you construct
the cor
One possible problem could be that you declare a non-template
propagator. In that case, you may have to export all functions that
must be publicly available using the GECODE_INT_EXPORT macro.
Alternatively, you can try running configure with the --disable-gcc-
visibility option in order to
amina kemmar wrote:
thank you for the answer, I changed the version of gecode but now
when i execute the commands :
g++ -I/include -c send-more-money.cpp
g++ -L/lib send-more-money.o \
-lgecodeint -lgecodesearch -lgecodekernel -lgecodesupport
i don't obtain any error, but no executable file g
The version of Gecode you use is too old. You must use the same
version as "Modeling with Gecode", currently 3.1.0.
Guido
amina kemmar wrote:
Hi,
When i try to compile send-more-money.cpp under LINUX mandriva, i
get the following errors :
[am...@localhost modeles]$ g++ -I/usr/local/inc
Dear Alfredo,
I think the problem lies in the initialization of the BoolVarArray.
Even for BoolVars, you have to give an initial domain, so you should
change the code to
m_vars = BoolVarArray(*this,900, 0,1);
The way you wrote it, you just allocated an array of size 900 with
uninitializ
"Modeling with Gecode", the current version on the web, Section 4.2.2.
Guido
Kayler Yao wrote:
Hi Guido,
Sorry, i can't find the index 4.2.2. Did you talk about Gecode
3.1.0? There has no index with X.X.X.
Kayler
2009/8/3 Guido Tack
Kayler Yao wrote:
2. But i still not
Kayler Yao wrote:
2. But i still not understand your reply well about the first
question. whether your means are that distinct constraint is only
used for IntVarArgs, as written in the manual of Gecode ---
" void Gecode::distinct (Space &home, const IntVarArgs &x,
IntConLevel icl=ICL_DEF)"
Kayler Yao wrote:
> 1. the difference bettween IntVarArray and IntVarArgs ，and how to
> use them
>As explained in the manual of Gecode, IntVarArgs is Argument
> array which is just good enough for passing arguments with automatic
> memory management. IntVarArray is Variable array which c
Hi,
there are two papers about views, linked from the publication section
on our web page ("Views and iterators..." and "Perfect derived
propagators"), and a more detailed explanation in my thesis which is
also available there. And you should first look at simpler
propagators than distin
-boun...@gecode.org] On
> Behalf
> Of Kish Shen
> Sent: Tuesday, July 21, 2009 12:54 PM
> Cc: us...@gecode.org
> Subject: Re: [gecode-users] doing multiple minimising searches with
> the same
> space
>
> Guido Tack wrote:
>> Hi Kish,
>>
>> do I un
Hi Kish,
do I understand you correctly, you want to first minimize one cost
variable, and then (independently) minimize another? I.e., when the
first one is minimized, there's potentially still variables left that
are not assigned, and then you branch over them, minimizing the
other? Or
Chris Mears wrote:
> I can check if two IntViews are ground and equal like this:
>
> IntView &x = ...
> IntView &y = ...
> if (x.assigned() && y.assigned() && x.val() == y.val())
>...
>
> Can I do the same thing for sets? There doesn't seem to be an
> equivalent to "val()" in the SetView cl
Hi.
The documentation currently only covers modeling, but not implementing
propagators (that will be added later). You'll have to look at
propagators in the Gecode source code and try to adapt them. However,
I have to admit I can't really see how to turn Euclid's algorithm into
a propag
Hi Gustavo,
sorry, the tutorial documentation was wrong. There is no support for
posting n-ary set operations with arbitrary relations. You have to
decompose by hand using an auxiliary variable. Thanks for pointing
this out!
Cheers,
Guido
Gustavo Gutiérrez wrote:
> Hello,
>
> I
Hi Carlos,
there seem to be several problems in your model. First, you constrain
t[0] instead of t[i]. Then, dom(this,t[i],SRT_EQ, i+1, i+3) states
that t[i] is equal to the set {i+1,...,i+3} - this means that all your
variables are assigned to fixed sets right from the start. I guess
y
Hi Martin,
we removed the pkgconfig stuff because it had very limited use. It
did not work with Windows dlls or Mac OS frameworks. Even worse, as
we didn't really use it ourselves, it often was out of sync with the
actual distribution (e.g. libs missing).
On Linux, all that's necessary to
Malcolm Ryan wrote:
> My apologies, I wasn't clear. I mean detect _within_my_code_. Is there
> some kind of #define that I can check for?
Yes, there's GECODE_HAS_GIST (defined in gecode/support/config.hpp).
Cheers,
Guido
___
Gecode users mail
Dhananjay Thiruvady wrote:
> I wanted to know the limits on variables and constraints in GECODE?
The limits on variable domains are defined in the Int::Limits and
Set::Limits classes (see the reference documentation or the source
code). There is no hard limit on number of variables and constr
Hi everyone,
we had some small build issues with the source packages for Gecode
3.1.0 and Gecode/FlatZinc 1.6. I just fixed these in the released
version and put up new source packages. The changes are completely
source and binary compatible with the previous packages, so we decided
not
Dear all,
we are releasing updates to Gecode and Gecode/FlatZinc.
Gecode 3.1.0
http://www.gecode.org
This release introduces parallel search, features improved memory
management (can double efficiency on MacOS X), and provides a reusable
command line driver upon
Martin Mann wrote:
> Is there a possibility to support user defined generators? e.g. in
> coming versions?
>
> If not available right now, maybe you like the following suggestion:
[...]
A problem with your approach is that a global random number generator
doesn't work in a multi-threaded settin
serge lemouton wrote:
> I am not able to compile the examples included from the Gecode 3.0.2
> Macintosh binary distribution.
> I have installed Gecode with the "Install Gecode 3.0.2.pkg", (after
> having desinstalled the oldest frameworks with the uninstall script)
> created an Xcode project with
Malcolm Ryan wrote:
> Is it possible to find out how many variables are in a particular
> space? Or how many active propagators?
Variables, no. Propagators, yes, using the propagators() method of a
Space. Inactive, i.e. subsumed propagators, are not copied, so
propagators() returns the numb
Malcolm Ryan wrote:
> I am applying two different (custom) branchings to the same constraint
> problem and one of them is reporting a lot more memory use (via
> Statistics) than the other. This is not completely unexpected but the
> difference is big enough that it requires more careful analysis t
Malcolm Ryan wrote:
> Is there any way to get Gist to provide a space as a return value to
> my code?
You can use an Inspector to do that. Here's an example (it's for
Gecode 3.x, I noticed you were still using 2.x but something similar
will work there, too):
class ReturnSpaceInspector : pub
The constraint is even "examined" (i.e. propagated) when variables are
not assigned but only their bounds change. Why don't you just try it
out? Look at the search tree of a simple model with a single
constraint in Gist, and inspect the nodes to see what's going on.
Guido
Jeffrey wrote:
Hi everyone,
it seems that sourcearchive.com just automatically downloads new
Debian packages, unpacks them, and runs doxygen on the result. That's
of course a pretty useless service, but we can't do much about it.
I'll ask the administrator to add a robots.txt to his website, so that
th
Jeffrey wrote:
> I'm optimizing my constraints with Gist at the moment.
I assume you mean that you use Gist in Branch-and-Bound mode, using a
constrain method in your model (or something similar to the
MinimizeExample or MaximizeExample classes).
> I've found that there are lots of pruned tr
Jeffrey wrote:
> I moved from 2.2.0 to 3.0.2.
>
> When I recompile the program, it pops out error saying that
> elementsUnion is not find.
>
> I've included set.hh already.
>
> Is the elementsUnion I used to use in 2.2.0 depreciated in 3.0.2? Is
> there any other function I can use?
The set e
Hi Filip,
the idea was (and still is) to initially allocate all VarArrays from
the space, and only when they are resized (by using add), the array is
copied to the heap. That way, most arrays are still efficiently
allocated in the space (e.g. copying an array always allocates it in
the sp
Dear all,
here's a bugfix release (again):
Gecode 3.0.2
http://www.gecode.org
This is a bug fix release fixing two more embarrassing bugs. However,
this time we redesigned our tests carefully such that they cover all
changes and optimizations done for the transit
Dear Roberto,
you are of course right, you have to use -c. We'll update the
document shortly. Thanks!
Guido
Roberto Castañeda wrote:
> I have just compiled and installed Gecode 3.0.1 in my system (Ubuntu
> 8.04.2, GCC 4.2.4), and followed religiously the steps given in the
> section 2 ("G
Dear all,
we are releasing bugfix versions:
Gecode 3.0.1
http://www.gecode.org
This is a bug fix release fixing an embarassing bug in reified Boolean
linear constraints (plus other minor fixes).
The complete changelog is available at
http://www.gecode.org/c
Nicolas wrote:
> The links to the source codes on
> http://www.gecode.org/benchmark_examples.html are dead. Could you
> please
> correct them?
Thanks for letting us know. They work now.
Cheers,
Guido
___
Gecode users mailing list
us...@geco
Hi everyone.
There's one known issue in the just-released Gecode/FlatZinc 1.4 - it
currently doesn't work with Gist although it lists gist as one of the
mode options. However, there will be another release of Gecode/
FlatZinc in two to three weeks (adapting to upcoming FlatZinc language
ch
Dear all,
we are happy and proud to announce our next major release:
Gecode 3.0.0
http://www.gecode.org
as well as the brand-new tutorial documentation on modeling with Gecode.
This release is a major consolidation release: interfaces have been
cleaned up (consist
Stefano,
thanks for the bug report. I just fixed it in the trunk (empty
IntSets now have min=Int::Limits::max and max=Int::Limits::min).
Cheers,
Guido
Stefano Gualandi wrote:
> Hi,
> there is a strange behavior of one of the SetVarArray constructors,
> in the trunk version.
> In pr
Malcolm Ryan wrote:
> Currently, I create an IntVarArray as:
>
> vars = IntVarArray(space, N);
>
> and then incrementally create variables as:
>
> vars[t] = IntVar(space, domainSet);
>
> My question is: does this actually save memory? I assume that creating
> the array created N 'uninitialised' Int
That was a stupid one, I just hadn't understood that the tab stop
width is measured in pixels, not characters... It's fixed now.
Cheers,
Guido
Guido Tack wrote:
> Thanks for the report, it's now on my agenda and I hope I'll be able
> to fix it for 3.0.
Thanks for the report, it's now on my agenda and I hope I'll be able
to fix it for 3.0.
Cheers,
Guido
Malcolm Ryan wrote:
> I'm using Gist in Gecode 2.2.0 on the Mac (OS X 1.5). I have found
> that tabs don't print in the Gist Console, which means that it is
> difficult to format large
cristian wrote:
> I am a Beginner in Gecode, and the follow error appear when I
> compiling at Eclipse
> The version of gecode is 2.1.1. and the IDE no marked errors in the
> code.
> Thanks for your attention
[...]
> Building target: prueba
> Invoking: GCC C++ Linker
> g++ -O0 -g3 -lgecodesearc
Dear Malcolm,
you're right, the examples are not very helpful as a starting point.
However, the tutorial documentation we're currently writing (and which
is going to be available soon) will provide exactly the kind of
information you're asking for. The new tutorial examples will be self-
Malcolm Ryan wrote:
> Did the experimental Decomposition during search code make it into the
> 2.2.0 release of Gecode? I would like to take advantage of it on my
> current constraint problem. If it is there, is there any description
> of how it works?
Sorry, DDS was planned to be included in 3.0
Mauricio Toro wrote:
> I installed Gecode 2.2.0 universal binary framework
> on a Mac PPC with Mac Os 10.4.11 and gcc 4.0.
> When I try to compile my software, I get the following error:
>
> /usr/bin/ld: /Library/Frameworks/gecode.framework/gecode load
> command 4 unknown cmd field
A quick searc
> Does red triangle means a pruned subtree and a white circle means a branch
> that has not been visited yet?
Yes.
Cheers,
Guido
___
Gecode users mailing list
us...@gecode.org
https://www.gecode.org/mailman/listinfo/gecode-users
Jeffrey wrote:
> What are the colors and symbol shapes mean in the Gist presentation?
Red square = failure, green diamond = solution, blue circle = branching
> We do not have any documentation at this moment about Gist, right?
It will be part of the 3.0.0 documentation.
Cheers,
Guido
Jeffrey wrote:
> As there could be a lot of solutions in an CSP, i'm looking for a
> way to inject randomness during the branching so that the first
> solution everytime I get is random.
>
> I found a post in June 6 saying that a random branching will be
> provided in Gecode 3.0.x.
> Is that
Morten Boysen wrote:
> I am trying to add new propagators to the FlatZincGecode space, and I
> have some question regarding this:
>
> 1) What is the defualt variable order for the FlatZincGecode space?
First branch over the IntVars with VAR_NONE, VAL_MIN, then over the
BoolVars with NONE, VAL_MI
Jérémie Vautard wrote:
I am coding a little stuff that needs to extract some potential
partial solutions from a constraint problem. To do this, I have a
(pointer to a)Gecode::Space S with an array of IntVar, and I do the
following :
- I have initially propared a set of IntVarArgs, each conta
Kish Shen wrote:
> I am not sure what are the trade-offs between the two APIs Guido
> outlined -- is it simply that using reflection will be less flexible?
In a way, it's more flexible. All constraints and variable types
available in Gecode will be automatically available through the
interfa
Patrik Haslum wote:
> Q1: Propagation for the weights constraint. I've posted the
> constraint:
[...]
> So, it seems that the upper bound on the cardinality of deleted has
> been
> propagated to total_cost (it is at most the sum of the 3 largest
> costs),
> but the lower bound has not (total_
Malcolm Ryan wrote:
> As I understand it, you can add multiple Branchings to a Space which
> will each be used in turn until each is satisfied. Is that correct?
> Are they applied in the order that they are added? If a Branching's
> status() returns false, is it permanently removed from the Space,
Hi Kish.
Kish Shen wrote:
> We have been thinking about developing an interface to GeCode for the
> Constraint Logic Programming language ECLiPSe for a while now, and
> finally I can devote some time and resource to this.
>
> The idea is to provide an alternative FD solver for ECLiPSe, in
> addi
Malcolm Ryan wrote:
> What is the procedure for creating a custom BranchingDesc? There don't
> appear to be any examples which do this.
The only requirement is to inherit from BranchingDesc. Then you put
everything in the object that you need for the commit operation.
> What is the role of th
benoitlaur...@neuf.fr wrote:
Hello,
I am working on a scheduling problem. On the instances I have to
deal with, I can reach a number of constraints of 500 000. During
the search process, the memory usage is very high, which eventually
renders my operating system unstable. Is it possible to
Am 17.01.2009 um 15:23 schrieb Jeffrey:
> For constructing an IntSetArgs with instructorOfStudent[], below is
> the way I used.
> It seems a bit stupid to put the instructorOfStudent[] into a
> tmpValue[1] in order to use the constructor.
> Is there are smarter way to do so?
Yes, IntSet(i,i)
Martin Mann wrote:
> I am preparing and updating my code for the next release of Gecode to
> utilize DDS. For that I update everything to work with the current
> trunk.
>
> An important question for me at this point: Is the distinct propagator
> self-decomposing to work properly with DDS or do I'
Jeffrey Mak wrote:
> How to defining constraint to extract Element from an array of
> integer holding repeated values?
> Gecode provide ElementUnion function to extract elements from IntSet
> or VarSet.
> How about an integer array in which elements may hold the same values?
>
> Below is a hyp
Jeffrey wrote:
> Are the following three constraints equivalent?
>
> Objective: To extract the d.val() th element in tmpArray into an
> IntValArray groupStudentVal so that further constraint can be
> written on it.
What's an IntValArray?
> tmpArray is int[]
> groupStudentVal is IntValArray
Gustavo Gutierrez wrote:
I removed all the propagators from my problem. Currently i am only
branching on one variable and i am getting different result from
gist and the other search. As there are no propagators, the shape of
the search tree is easy to guess and should have only 80 solution
Gustavo Gutierrez wrote:
> I am experiencing almost the same problem. There is a difference in
> the number of solutions when i run my problem instance with or
> without gist. The difference with the case reported is that for me
> the number of solutions reported by gist is really large compa
Denys Duchier wrote:
> Guido Tack writes:
>
>> Denys Duchier wrote:
>>> Also:
>>> is there a downside to providing the most general templates?
>>
>> There is actually a downside (although it's not the reason we didn't
>> provide the templ
Denys Duchier wrote:
> Also:
> is there a downside to providing the most general templates?
There is actually a downside (although it's not the reason we didn't
provide the templates). If you use the ModEventDelta during
propagation, you have to access it through all the view types of the
p
Jeffrey wrote:
> I'm a new Gecode user.
> I've tried to build the example in Visual Studio 2008 platform.
>
> I've added the Gecode\lib as the library directory and set linker to
> link all the nine library including GecodeSupport.lib.
>
> It compiles with no error but carries 4 errors when tryi
Maria Garcia de la Banda wrote:
> Guido Tack wrote:
>
>> Do you want to do the extra pruning in the branching? E.g., collect
>> some attributes for particular variables, and then perform the
>> branching based on these attributes? This reminds me of branchings
Maria Garcia de la Banda wrote:
> Total newbie question (from an LPer): assume that, during search, I
> want
> to add some extra prunnings for all search variables that have a
> particular attribute (e.g., they apper in a given constraint, or they
> appear in some data structure I am maintaining,
Denys Duchier wrote:
> I had a couple of problems using some of Gecode macros because they
> cannot all be safely followed by a semicolon. Here is a proposed
> patch
> that uses de "do {...} while (0)" pattern to make these macros safe.
> This is an old trick that is used a great deal in the Li
Denys Duchier wrote:
> It turns out I'll be needing a reified version of set/min. So here
> is a
> 11-patch series that attempts to provide this new constraint (also
> reified set/max). These patches are for Gecode 2.2.0. I am posting
> these here in case you'd like to add something based on t
1 - 100 of 342 matches
Mail list logo
