[rules-users] Re: drools 5.0.0-SNAPSHOT Maven repository broken? -> Lingering logically inserted facts problem
Geoffrey De Smet
ge0ffrey.spam at gmail.com
Fri Oct 17 12:59:25 EDT 2008
Hi Tim
>> I am convinced that:
>> 2) Penalty's equals/hashcode uses both "cause" and "weight"?
>> then 3) happens: Keep Penalty(10), add Penalty(5)
>> is an issue in drools core and instead 2) should happen:
>> Remove Penalty(10), add Penalty(5)
>> because Penalty(10) is no longer true
>
> exactly!
> that is the behavior i was expecting. it took me quite some time
> to find out why my program did not find a solution.
I made a testcase and Edson might look into it :)
Watch this issue
https://jira.jboss.org/jira/browse/JBRULES-1804
Once it's resolved, I 'll made the necessary adjustments in the
equals/hashcode of the ConstraintOccurrence implementations.
With kind regards,
Geoffrey De Smet
tim tim schreef:
>
>> I am going to make a testcase to isolate the problem and make it more
>> transparent that it is indeed an issue. And then try to convince Mark and
>> Edson that it's an issue :)
>
> maybe there could be a switch, to tell drools which mode
> is prefered? then they could keep the strange behavior in the engine ;-)
>
>
>> which the HACK indeed that you quothed below:
>> the hack causes the rule to no longer apply (retracting the
>> ConstraintOccurrence), after which it then refires.
>
> argh, you are right! i didnt see that it gets retracted. i still have
> to get used
> to the drools way..
>
>
> best, tim
>
>>> i solved the problem of improving constraints for the same causes that
>>> differ in hashCode by retracting the old fact before inserting the new
>>> one.
>>> but it takes quite some overhead to be able to do so. i could post my
>>> code if anyone is interested.
>>>
>>> unfortunately i could not find in your samples how you handle such a
>>> scenario.
>>>
>>> does the "HACK" part of this rule have something to do with it?
>>>
>>> from curriculumCourseScoreRules.drl
>>>
>>> // MinimumWorkingDays: The lectures of each course must be spread into
>>> a minimum number of days.
>>> // Each day below the minimum counts as 5 points of penalty.
>>> rule "minimumWorkingDays"
>>> when
>>> $course : Course($minWorkingDaySize : minWorkingDaySize);
>>> $dayCount : Number(intValue < $minWorkingDaySize) from accumulate(
>>> $day : Day()
>>> and exists Lecture(course == $course, day == $day),
>>> count($day)
>>> );
>>> // HACK to fix wierd truth maintance behavior in drools
>>> // because making weight part of the equals/hashcode doesn't cut it
>>> not IntConstraintOccurrence(
>>> ruleId == "minimumWorkingDays",
>>> constraintType == ConstraintType.NEGATIVE_SOFT,
>>> causes contains $course,
>>> eval(weight != (($minWorkingDaySize - $dayCount.intValue()) *
>>> 5))
>>> );
>>> // An uninitialized schedule should have no constraints broken
>>> exists Lecture(course == $course);
>>> then
>>> insertLogical(new
>>> IntConstraintOccurrence("minimumWorkingDays",
>>> ConstraintType.NEGATIVE_SOFT,
>>> (($minWorkingDaySize - $dayCount.intValue()) * 5),
>>> $course));
>>> end
>>>
>>> but the hack makes the rule not fire at all, when it had fired before.
>>> even if the error has improved by now.
>>>
>> Feed back is always welcome :)
>>
>> With kind regards,
>> Geoffrey De Smet
>>
>> _______________________________________________
>> rules-users mailing list
>> rules-users at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/rules-users
>>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
More information about the rules-users
mailing list