Op 07-12-11 21:41, Patrik Dufresne schreef:
Hi all,
I'm trying to create rules to model my problem. So far, I didn't
manage to create rules to make Drools converge to a solution because
of score traps. I don't see any way to avoid it (as I'm not an expert
with Drools). I can express the rule as follow : a senior employee
should work before a less senior employee.
I've implement it as a hard constraint :
rule "hard-Seniority"
when
$employee : Employee()
$assignment : PlanifEventAssignment( $planifEmployee : employee )
What's does
this $planifEmployee do?
not PlanifEventAssignment( employee == $employee )
eval(Helper.compareEmployee($employee, $planifEmployee) < 0)
then
insertLogical(new IntConstraintOccurrence("hard-Seniority",
ConstraintType.NEGATIVE_HARD,
1,
$employee, $assignment));
end
The function Helper.compareEmployee(e1, e2) return -1 if e1 is more
senior then e2 (mostly based on hire date and other boolean fields).
On first sight, in Solution.getProblemFact() I would add these cached
problem facts, for any 2 employees (but just once per combination):
new EmployeeRanking(higherRankedEmployee, lowerRankedEmployee,
numberOfEmployeesInRankBetweenThem)
then you can do
rule "hard-Seniority"
when
EmployeeRanking($higherRankedEmployee, $lowerRankedEmployee,
$numberOfEmployeesInRankBetweenThem)
PlanifEventAssignment(employee == $lowerRankedEmployee)
PlanifEventAssignment(employee == $higherRankedEmployee)
then
insertLogical(new IntConstraintOccurrence("hard-Seniority",
ConstraintType.NEGATIVE_HARD,
$numberOfEmployeesInRankBetweenThem,
$employee, $assignment));
This way, the rule make sure a senior employee is working. But
it's a
score trap since, many moves are required to resolve the constraint. e.g:
PlanifEventAssignment1 = e2
PlanifEventAssignment2 = e3
PlanifEventAssignment3 = e4
PlanifEventAssignment4 = e5
e1 is not working. Multiple move are require to reach the best solution :
PlanifEventAssignment1 = e1
PlanifEventAssignment2 = e2
PlanifEventAssignment3 = e3
PlanifEventAssignment4 = e4
So I'm asking you. What is the best way to make Drools converge ? Do I
need to change my rule, or should I create a BigMove ?
course grained moves will
work mostly, but solving the score trap itself
is a far better long-term solution (and it still allows you to add
course grained moves later).
--
Patrik Dufresne
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
With kind regards,
Geoffrey De Smet