I am trying to use Optaplanner to do some sort of Course Scheduling but, instead of have isolated courses, I have a group of them that have to occur one after other. I am not sure if it is clear, let me give an example: Group1(c1,c2,c3,c4),Group2(c1,c2,c3,c4,c5,c6), Group3(c5,c6,c7,c8). My problem is: In order to choose the best period, I create the combination of all possible periods to each course and I wrote the rules to assert a lot of conditions. For example, if course2Group1 is not just after course1Group1 I break a hard constraint. OptaPlanner is solving the problem, but it takes 12 hours to achieve the best solution. The big problem is the amount of data, there are a lot of periods that are not used because it should pick only one of the possibilities to each Course and verify if the periods choosen for courses of the same group are satisfying the conditions.
Do you have any tip about the way that I created my Domain model? Any tip about performance? I am using some selector filters already...
