Re: [eclipse-clp-users] Two problems with the ic-Solver

From: Andreas Berger <aberger_at_Informatik.tu-Cottbus.DE>
Date: Sun, 28 Feb 2010 14:51:57 +0100
Ok, it was the easiest way to use the third solution, because i have C 
Integers as values. But I see it's not the most efficient way. With the 
domain {0 .. 10^7) it takes a second to calculate the value.
An yes,. i have only a few constraints of this type and so it's ok. 
Maybe I will try to use the "bool_bitand" suggestion of yours, would 
that be much faster?
But then i have to split the Integers into bits first...

Thank's
Andreas

Joachim Schimpf schrieb:
> Andreas Berger wrote:
>   
>> Thank you! The constraint prototype for integer is exactly what I was
>> looking for.
>>     
>
> I was afraid you would say that.  The reason I offered this as the
> last of 3 alternatives is that this is not a constraint that is
> very suited to the way the ic solver represents integer variables.
>
> If you have just a few of these constraints, and the majority of
> constraints are the usual arithmetic or global constraints over
> integer variables, this is probably ok.
>
> But if your problem consists mostly of this kind of constraint,
> you should seriously consider the other possibilities.
>
>
> -- Joachim
>
>
>   
>> Andreas
>>
>>
>> Am 11.02.2010 02:44, schrieb Joachim Schimpf:
>>     
>>> Andreas Berger wrote:
>>> ...
>>>   
>>>       
>>>> I want to express a Constraint of the Form:
>>>>
>>>> "Give me a Number X, which is bitwise connected (C++ '&' ) with a second
>>>> Number A is 0.
>>>>
>>>> Something like:  A /\ X = 0, give me a possible X when A = 9.
>>>>
>>>> I can't find an appropiate predicate for such kind of problem. I tried
>>>> the following construct:
>>>> 9 /\ X #= 0,  but of course it does not work.
>>>>
>>>> Can someone give me a hint to solve this problem please.
>>>>      
>>>>         
>>> First, you should consider whether it is not better to use boolean (0/1)
>>> variables for the individual bits.  Your constraints may then be easier
>>> to express.  For example, if you represent your numbers as lists of
>>> boolean digits, you can define
>>>
>>> bool_bitand(X, Y, Z) :-
>>>     ( foreach(Xi,X), foreach(Yi,Y), foreach(Zi,Z) do
>>>         [Xi,Yi,Zi]::0..1,
>>>         Zi #= (Xi and Yi)
>>>     ).
>>>
>>> and express your example as
>>>
>>> ?- bool_bitand(X, [1, 0, 0, 1], [0, 0, 0, 0]).
>>> X = [0, _368{[0, 1]}, _457{[0, 1]}, 0]
>>> Yes (0.00s cpu)
>>>
>>>
>>> Second, if your bit vectors represent sets, you could also consider
>>> using the ic_set solver's set-variables to represent them.
>>>
>>>
>>> Third, with your original representation, you could prototype such
>>> a constraint as follows:
>>>
>>> :- lib(propia).
>>>
>>> int_bitand(X, Y, Z) :- ( labeling([X,Y,Z]), X/\Y =:= Z ) infers ic.
>>>
>>> which solves your example in this way:
>>>
>>> ?- X :: 0 .. 9, int_bitand(X, 9, 0).
>>> X = X{[0, 2, 4, 6]}
>>> Yes (0.00s cpu)
>>>
>>>
>>> -- Joachim
>>>    
>>>       
>
>
> ------------------------------------------------------------------------------
> SOLARIS 10 is the OS for Data Centers - provides features such as DTrace,
> Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW
> http://p.sf.net/sfu/solaris-dev2dev
> _______________________________________________
> ECLiPSe-CLP-Users mailing list
> ECLiPSe-CLP-Users_at_lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/eclipse-clp-users
>
>
>   
Received on Sun Feb 28 2010 - 14:31:31 CET

This archive was generated by hypermail 2.2.0 : Thu Feb 02 2012 - 02:31:58 CET