限制分配值 - Choco

Question

Is there a way to limit the number of assignments for a specific value of a variable?

I'm writing a schedule problem using Choco [4.0.1] with courses, teachers and timeslots. The same teacher can teach many courses and I need to define the amount of courses he/she could teach. I was thinking about counting the number of timeslots, like Teacher B can work 12 hours (12 timeslots) , so I could post a constraint like arithm() to ensure that. Any ideas?

[Updated]

Here is some code:

IntVar [] teachers;
IntVar [][] teacherTimeslots;
IntVar [] courses;

For each course there is a teacher and N timeslots, so Teacher A can teach {1,2,3,4} and teacher B {3,4,5} and each course has 4 hours per week. Now imagine that A can work 12 hours. I'd like to limit A to teach only 12 hours (3 courses) while B would teach the remaining courses;

I cannot say how many courses has a teacher but how many he/she could teach, so I'm using

Tuples tuples = new Tuples(true);
tuples.add(1, 1);
...
tuples.add(2, 5);
model.table(teacher, course, tuples).post();

constraint to ensure that he/she is able to teach certain courses.

teacherTimeslots is filled with all possible timeslots and I use model.allDifferent(teacherTimeslots[teacher]).post() to keep each teacher timeslot unique.

My courses are fixed so

course[0] = 1;
...
course[4] = 5;

I thought about getting all timeslots for a teacher but when I'm building my model there is no timeslot selected, so I' get all possibilities.

Answer 1

Well, here is my solution. I've created a Timeslot class to represent a lecture, containing 1 teacher, 1 course, m rooms and n slots (m=n) , so in my model there is a collection of timeslot objects for all courses. I used model.count(...) to check how many timeslots has a teacher and I put zero as lower bound because a teacher can be chosen or not and twelve times as upper bound.

List<IntVar> teachersTimeslotList = new ArrayList<IntVar>();

List<IntVar> teachersList = new ArrayList<IntVar>();

for (int i = 0; i < timeslots.size(); i++) {

    for (int j = 0; j < timeslots.get(i).getTimeslots().size(); j++) {

        IntVar timeslot = timeslots.get(i).getTimeslots().get(j);
        IntVar teacher = timeslots.get(i).getTeacher();

        IntVar sumTeacher = model.intVar("sumTeacher", 0, 100000);

        teachersList.add(teacher);

        model.sum(new IntVar[]{model.intScaleView(timeslot, 1000), teacher}, "=", sumTeacher).post();

        teachersTimeslotList.add(sumTeacher);
    }
}

for (int i = 0; i < teacheresId.length; i++) {
    model.count(teacheresId[i], teachersList.toArray(new IntVar[teachersList.size()]), model.intVar(0, 12)).post();
}

model.allDifferent(teachersTimeslotList.toArray(new IntVar[teachersTimeslotList.size()]), "NEQS").post();

I don't know if this is an optimized solution but at the moment it's working for me and even resolution time was improved by coding this. Thanks!