Originally posted by: WeeLoon

Supposed I have interval a -> b -> c in a sequence.

I would like to have a value where a = 1, b = 2, c = 3 and use them in objective function.

I try to use cumulFunction, but cumulFunctionValue can only be extracted after the solution exists.

Which function can I use to achieve this?

Originally posted by: Petr Vilím

Hello,

there is no simple way to access position of an interval variable in a sequence. The main reason is that it does't propagate well and we don't even use it internally.

I would step back first. What kind of objective function do you have in mind? Maybe we can find a way to model it without the actual positions.

Best regards, Petr

Originally posted by: WeeLoon

Hi Petr,

I manage to find a workaround by modifying the objective. Thanks for your help.

Best regards, WeeLoon

Great you found a way.

Otherwise other users could use that small example:

using CP;

int n=5;

range r=1..n;
range r0=0..n;

dvar interval itvs[i in r] size i;

dvar sequence  seq in all(i in r)itvs[i] types all(i in r)i;

dvar int nextType[i in r];
dvar int rank[i in r0];
dvar int rankbis[i in r] in r;
dvar int typeInSequence[r] in r; // compute the type of ith interval in the sequence

subject to
{

// input

noOverlap(seq);
startBeforeStart(itvs[1],itvs[3]);
startBeforeStart(itvs[3],itvs[5]);
startBeforeStart(itvs[5],itvs[2]);
startBeforeStart(itvs[2],itvs[4]);

// and then let s compute typeInSequence

forall(i in 1..n) nextType[i]==(typeOfNext(seq,itvs[i],0));

rank[1]==1;
forall(i in 1..n)  (nextType[i]!=0) => (rank[i]+1==rank[nextType[i]]);

forall(i in r) rankbis[i]==rank[i];

inverse(rankbis,typeInSequence);

 

}

execute
{
writeln("ranks = ",rankbis);
writeln("typeInSequence = ",typeInSequence);

}

which gives

ranks =  [1 4 2 5 3]
typeInSequence =  [1 3 5 2 4]

regards

Alex Fleischer

PS:

Many how to with OPL at https://www.linkedin.com/pulse/how-opl-alex-fleischer/

Many examples from a very good book : https://www.linkedin.com/pulse/model-building-oplcplex-alex-fleischer/

Making optimization simple : https://www.linkedin.com/pulse/making-decision-optimization-simple-alex-fleischer/