[gecode-users] Questions about element and channel constraint for float variables

[Guido Tack]

Hi,

you could decompose the element constraint by posting implications instead, such as x=i implies y=a[i] for all i.

Regarding the rounding, you're right, your approach should work (and just using a channel directly forces the float var to take an integral value).

Cheers,
Guido

-- 
GUIDO TACK                  
Senior Lecturer

Information Technology
Monash University
Level 6, Room 6.40, Building H, Caulfield Campus
900 Dandenong Road
Caulfield East VIC 3145
Australia

T: +61 3 9903 1214                      
E: guido.tack at monash.edu <mailto:guido.tack at monash.edu>
http://www.csse.monash.edu/~guidot/ <http://www.csse.monash.edu/~guidot/>


> On 21 Nov. 2016, at 6:02 am, Dani G <daniel.garcialorenzo at gmail.com> wrote:
> 
> Hello all, 
> 
> These ares my first question, thanks for this library, I found it very interesting and the tutorial is really great.
> 
> I'm using the last version of Gecode 5.0. In the project I'm working on I need to use float variables (the problem include exponential calculations that need to be precise), although I'm trying to minimize their use as much as possible by converting Floats into Into Ints. Two questions arises:
> 
> - I couldn't find a "element" constraint for FloatVarArray, and it would really help me. I don't know if I could easily reimplemented it using the example of the constraint for Int variables, or there will be some problem? Should I find a different modelling, like using a linear constraint using a 0,1 vector?
> 
> - To minimize the use of Float variables I want to round my Float calculations using two digits and then convert them to Int using a channel constraint. I imagine that I need to take into account the rouding if i want to make it work. This sounds right?
> 
> IntVar time; // Variable used in other constraints
> FloatVar time_float(home, TIME_MIN, TIME_MAX); // conversion to float
> 
> channel(home, time, time_float); // conversion to float for the calculations
> 
> // float calculation, this will obviously wil be any float number
> FloatVar activity_float= expr(home, c1 * exp( c2 * time_float); 
> 
> // Rounded version of activity multiplied by 100 to be converted into Int
> FloatVar activity_float_rounded100(home, ACTIVITY_MIN, ACTIVITY_MAX);
> 
> // Rounding up, and multiplying by 100 to keep the desired precision
> rel(home, activity_float_rounded100   >= 100* activity_float);
> rel(home, activity_float_rounded100-1 <  100* activity_float);
> 
> IntVar activity_int(home, ACTIVITY_MIN, ACTIVITY_MAX); // rounded activity by 100
> channel(home,activity_int, activity_float_rounded100); 
> 
> // now activity_int can be used in other int constraints
> 
> I bellieve that if I apply directly the channel to activity_float, the exponential constraint will fail as channel will force activity to be exactly equal to the Int value, am I right?
> 
> 
> Thank you very much your work and for your thoughts
> 
> Best regards
> 
> Daniel
> 
> 
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gecode.org/pipermail/users/attachments/20161121/fd400972/attachment-0001.html>