Originally posted by: SystemAdmin

[shadi said:]

hi,
i have the following array
isTripIncludedInDuty[nbTrips]

i want to assign it like this:
isTripIncludedInDuty[t] = 0     if trip t included in tripSeq[i].
                              = 1    if trip t not included in tripSeq[i].


please help me to write this command in the constraints section in the cp model

i write the following command:
    forall (tr in Trips)
      isTripIncludedInDuty[table][tr][td] == sum (i in tripRng)(tripSeq[i] == tr);

is it right?

thanks
kshieboun shadi
it is not right becuase




#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[alain.chabrier said:]

Hi,

it is not very clear from your post, where all elements come from.
Can you include some more of  the model ?

Alain
#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[shadi said:]

dear Alain,
I found another way I can achieve my goal. i dont need the isTripIncludedInDuty[t], i use the following constraint:


  cNew - sum(j in tripRng) Duals[tripSeq[j]] < 0;<br />

buti get the following error:
CP cannot extract expression: cNew+(sum(j in 1..20) Duals[(tripSeq[(j)])])*(-1)

what i did wrong?


#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[afleischer said:]

Hi,

is Duals defined as an int array or as a float array?
It should be an int array.
Anyway, if this is not the issue, more elements about your model would be welcome

Alex
#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[shadi said:]

Dear alex,
Duals defined as a float array, becuase the dual variables that the cp model receive from the main model is a float and not int.

i changed the declaration to int and i get the following error:
*** ERROR: "Duals": Cannot use type int[range] for float[range]


im using column generation method to minimize the final cost.
i want to write the reduced cost constraint in the cp model to add columns with negative reduced cost only.
so i tried using two ways:

1.  cNew - sum(j in tripRng) Duals[tripSeq[j]] < 0;<br />2.  cNew - sum (j in Trips) Duals[j]*flagTripInColumn[j] < 0;<br />      where
              flagTripInColumn[j]=1    if trip j is covered by the new column
                                      =0    if trip j not covered by the new column    =>

so i assigned it as following (i dont know if it is correct)     
                forall (tr in Trips)
                          flagTripInColumn[table][tr][td] == sum (i in tripRng)(tripSeq[i] == tr);

 
after implementation of the second option, i didnt receive runtime errors but i didnt receive to the optimum solution, i think there is an error in the
reduced cost critirion.


thanks in advenced
Kshieboun Shadi

     



#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[afleischer said:]

Hi,

so maybe in the declaration part you can convert the floats into ints

int DualsI[i in tripRng]=ftoi(round(Duals[i]));

and then your constraint

cNew - sum(j in tripRng) DualsI[tripSeq[j]] < 0; <br />
should work

Alex Fleischer
#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[shadi said:]

thank you Alex,
it works, but i dont know why i still cant reach to the optimum solution using this reduced cost constraint... 






#DecisionOptimization
#OPLusingCPOptimizer