Op 09-01-13 10:32, Michiel Vermandel schreef:
There is a construction heuristic that should have already filled in that planning entity.
The two other planning variables are filled for each planning entity.
I have put a printline in the setter of the troublesome planning variable:

    @PlanningVariable(strengthComparatorClass = TasktypeStrengthComparator.class, nullable = false)
    @ValueRange(type = ValueRangeType.FROM_SOLUTION_PROPERTY, solutionProperty = "tasktypes")
    public TaskType getType() {
        return type;
    }

    public void setType(TaskType type) {
        System.out.println("Type set to " + type);
        this.type = type;
    }
 
And I get the following outprint:

Type set to TL
Type set to null
Type set to TL
Type set to TL
Type set to null
Type set to TL
Type set to null
Type set to TL
Type set to null
Type set to TL
Type set to null
Type set to TL
Type set to TL
Type set to null
Type set to TL
...
Type set to null
Type set to TL
Type set to null
Type set to TL
That's normal. It's setting it and reverting it when doing the moves.
Correlate it with the TRACE log if you want to see why. See manual section "logging".

How many <constructionHeuristics> have you got configured?
One guess is that it's actually being initialized by a construction heuristic,
but at the end of that const heur, the BestSolutionRecaller doesn't put into SolverScope.setBestSolution()...
If you can prove that assumption, it's a bug.

Note that I also tried to set the annotation property nullable to false.
This seems to have no effect.

Thanks,
Michiel
-----------------
http://www.codessentials.com - Your essential software, for free!
Follow us at http://twitter.com/#!/Codessentials
From: Geoffrey De Smet <ge0ffrey.spam@gmail.com>
To: Rules Users List <rules-users@lists.jboss.org>
Sent: Wednesday, January 9, 2013 9:30 AM
Subject: Re: [rules-users] Fw: proble with localSearch


Op 09-01-13 08:21, Michiel Vermandel schreef:
Hi Geoffrey,

Thanks for your advice.
Indeed one of the planning variables is consistently null over all planning entities.
And there was a construction heuristic that should have already filled in that planning entity?
Enable DEBUG logging, so you can see in the construction heuristic log if it does steps which change that variable from null to a non-null value.

If you have 2 <constructionHeuristics> configurations before the <localSearch> configuration (which I don't see how you could with the 5.5 CH behavior):
I just fixed a bug/feature for 6.0 earlier this week to support that.

Should drools planner take care of this or should I impose an extra rule to exclude these selections?
Once again I seem to struggle with nulls being passed in as planning variable...

Thanks,

Michiel
 
-----------------
http://www.codessentials.com - Your essential software, for free!
Follow us at http://twitter.com/#!/Codessentials
From: Geoffrey De Smet <ge0ffrey.spam@gmail.com>
To: Michiel Vermandel <mvermand@yahoo.com>; Rules Users List <rules-users@lists.jboss.org>
Sent: Tuesday, January 8, 2013 4:27 PM
Subject: Re: proble with localSearch


Op 08-01-13 15:24, Michiel Vermandel schreef:
Hi,

I'm trying to implement a basic local search (really new to it).

When I run my project I now get:

Exception in thread "main" java.lang.IllegalStateException: Phase localSearch started with an uninitialized Solution. First initialize the Solution. For example, run a phase constructionHeuristic first.
    at org.drools.planner.core.localsearch.DefaultLocalSearchSolverPhase.phaseStarted(DefaultLocalSearchSolverPhase.java:120)
You have a construction heuristic configured, so all @PlanningVariable's should be different from null after that phase is ended.
So when the local search phase starts, that check should be ok and not throw that exception.

Comment out local search, write the solution to disk an verify that all planning variables are now not null.
Or debug where that exception is throw and inspect phaseScope.getSolverScope().getScoreDirector().getWorkingSolution().

This is my config:

<?xml version="1.0" encoding="UTF-8"?>
<solver>
    <environmentMode>DEBUG</environmentMode>

    <!-- Domain model configuration -->
    <solutionClass>be.axi.planner.app.InspectionSchedule</solutionClass>
    <planningEntityClass>be.axi.planner.domain.Task</planningEntityClass>

    <!-- Score configuration -->
    <scoreDirectorFactory>
        <scoreDefinitionType>HARD_AND_SOFT</scoreDefinitionType>
        <scoreDrl>/inspectionRules.drl</scoreDrl>
    </scoreDirectorFactory>

    <constructionHeuristic>
        <constructionHeuristicType>BEST_FIT_DECREASING</constructionHeuristicType>
        <!-- constructionHeuristicPickEarlyType>FIRST_LAST_STEP_SCORE_EQUAL_OR_IMPROVING</constructionHeuristicPickEarlyType-->
    </constructionHeuristic>

    <localSearch>
        <termination>
            <terminationCompositionStyle>OR</terminationCompositionStyle>
            <maximumSecondsSpend>3600</maximumSecondsSpend>
            <scoreAttained>0hard/0soft</scoreAttained>
        </termination>
        <unionMoveSelector>
            <changeMoveSelector>
                <valueSelector>
                    <planningVariableName>period</planningVariableName>
                </valueSelector>
            </changeMoveSelector>
            <changeMoveSelector>
                <valueSelector>
                    <planningVariableName>spector</planningVariableName>
                </valueSelector>
            </changeMoveSelector>
            <changeMoveSelector>
                <valueSelector>
                    <planningVariableName>type</planningVariableName>
                </valueSelector>
            </changeMoveSelector>
            <swapMoveSelector>
            </swapMoveSelector>
        </unionMoveSelector>

        <acceptor>
            <planningEntityTabuSize>7</planningEntityTabuSize>
        </acceptor>
        <forager>
            <minimalAcceptedSelection>1000</minimalAcceptedSelection>
        </forager>
    </localSearch>

</solver>
 Any thoughts on what I'm doing wrong?

Thanks

Michiel
-----------------
http://www.codessentials.com - Your essential software, for free!
Follow us at http://twitter.com/#!/Codessentials


_______________________________________________
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users








_______________________________________________
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users


_______________________________________________
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users




_______________________________________________
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users