Hello all, I am currently trying to use Optaplanner to create a timetable of interviewers, where x students are to be given interviews in a given time slot. The planning variable is Lecturer class, meaning different lecturers can be assigned to different timeslots. Through debugging, I can tell that the planner does optimize (I'm currently using brute force just to be complete for small problems). But the final best solution ALWAYS has the results of the last search the optimizer did, and so the 'best' solution information seems to be lost as soon as the planner does another search. Clearly I'm doing something wrong. Can anybody let me know what the problem is? Thanks a bunch in advance! Karan Timetabler.zip <http://drools.46999.n3.nabble.com/file/n4028442/Timetabler.zip> -- View this message in context: http://drools.46999.n3.nabble.com/Best-Solution-found-but-latest-solution... Sent from the Drools: User forum mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

Hi Ge0ffrey, thanks for your quick reply. Actually I'm quite sure it's my code, I just used brute force to enable completeness on my small problem. I've attached my problem this time just using a first fit heuristic. If you run my code you will see that the solver hits a 0hard/0soft score which is valid, but since it searches for better solutions afterwards it seems to replace the best solution with the last solution it received. I think the TimetableSlot objects' Lecturer planning variable is modified when planner is running and since a new instance isn't made of the objects they are just replaced by the latest value (and I'm guessing wherever the best solution is stored its instances are replaced too since they refer to the same point in memory). I feel I haven't implemented the Solution/Planning Entities and variables correctly so do take a look at those. I think the best way to understand the problem is to just run the main method. It should work out of the box as long as libraries are included! Thanks a lot and awaiting your reply. Karan Timetabler.zip <http://drools.46999.n3.nabble.com/file/n4028475/Timetabler.zip> -- View this message in context: http://drools.46999.n3.nabble.com/Best-Solution-found-but-latest-solution... Sent from the Drools: User forum mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

Hi ge0ffrey, I understand. I forgot mentioning I am using OptaPlanner 6.0.1 Final. Your last email proved useful, I encountered some errors which I tried fixing, finally coming to this error: Exception in thread "main" java.lang.IllegalStateException: HashCode violation: the hashCode of tabu (TimetableSlot@15897e8d) probably changed since it was inserted in the tabu Map or Set. at org.optaplanner.core.impl.localsearch.decider.acceptor.tabu.AbstractTabuAcceptor.adjustTabuList(AbstractTabuAcceptor.java:108) at org.optaplanner.core.impl.localsearch.decider.acceptor.tabu.AbstractTabuAcceptor.stepEnded(AbstractTabuAcceptor.java:98) at org.optaplanner.core.impl.localsearch.decider.LocalSearchDecider.stepEnded(LocalSearchDecider.java:159) at org.optaplanner.core.impl.localsearch.DefaultLocalSearchSolverPhase.stepEnded(DefaultLocalSearchSolverPhase.java:127) at org.optaplanner.core.impl.localsearch.DefaultLocalSearchSolverPhase.solve(DefaultLocalSearchSolverPhase.java:83) at org.optaplanner.core.impl.solver.DefaultSolver.runSolverPhases(DefaultSolver.java:190) at org.optaplanner.core.impl.solver.DefaultSolver.solve(DefaultSolver.java:155) at Main.main(Main.java:29) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:601) at com.intellij.rt.execution.application.AppMain.main(AppMain.java:120) I have run tests to ensure my Timetable (Solution), TimetableSlot (Planning Entity), and Lecturer (Planning Variable) all have working clone(), equals(), and hashCode(). Clone: I always do a deep clone, so Timetable's TimetableSlots and TimetableSlot's Lecturer will be re-instantiated. Same for the other classes. Equals: I always do deep Equals, so even if all objects and sub-objects are different it will equal true if all the base objects are equal. hashCode: I ensured that a cloned object using my cloned methods will output the SAME hashCode. My @PlanningEntityCollectionProperty returns non-cloned (so same instances) collection of TimetableSlots. Hopefully this description of my code can help you spot what I'm doing wrong. Thank you for your responses so far! Karan -- View this message in context: http://drools.46999.n3.nabble.com/Best-Solution-found-but-latest-solution... Sent from the Drools: User forum mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

Oh, I actually decided to implement my own cloning methods because without them FULL_ASSERT gave me the following error: 14:46:41.601 [main] INFO o.o.core.impl.solver.DefaultSolver - Solving started: time spend (7), score (-1hard/0soft), new best score (-1hard/0soft), random seed (0). 14:46:41.620 [main] INFO o.o.c.i.c.DefaultConstructionHeuristicSolverPhase - Phase (0) constructionHeuristic ended: step total (0), time spend (28), best score (-1hard/0soft). 14:46:41.798 [main] DEBUG o.o.c.i.l.DefaultLocalSearchSolverPhase - Step index (0), time spend (207), score (-1hard/0soft), best score (-1hard/0soft), accepted/selected move count (1000/1000) for picked step (TimetableSlot@44c099ef => Thomas). 14:46:41.947 [main] DEBUG o.o.c.i.l.DefaultLocalSearchSolverPhase - Step index (1), time spend (356), score (-1hard/0soft), best score (-1hard/0soft), accepted/selected move count (1000/1889) for picked step (TimetableSlot@5289178b => Thomas). 14:46:42.092 [main] DEBUG o.o.c.i.l.DefaultLocalSearchSolverPhase - Step index (2), time spend (501), score (0hard/0soft), new best score (0hard/0soft), accepted/selected move count (1000/2869) for picked step (TimetableSlot@c7b2b3e => Thomas). Exception in thread "main" java.lang.IllegalStateException: Score corruption: the solution's score (0hard/0soft) is not the uncorruptedScore (-1hard/0soft). at org.optaplanner.core.impl.score.director.AbstractScoreDirectorFactory.assertScoreFromScratch(AbstractScoreDirectorFactory.java:75) at org.optaplanner.core.impl.solver.scope.DefaultSolverScope.assertScoreFromScratch(DefaultSolverScope.java:113) at org.optaplanner.core.impl.bestsolution.BestSolutionRecaller.stepEnded(BestSolutionRecaller.java:90) at org.optaplanner.core.impl.phase.AbstractSolverPhase.stepEnded(AbstractSolverPhase.java:93) at org.optaplanner.core.impl.localsearch.DefaultLocalSearchSolverPhase.stepEnded(DefaultLocalSearchSolverPhase.java:126) at org.optaplanner.core.impl.localsearch.DefaultLocalSearchSolverPhase.solve(DefaultLocalSearchSolverPhase.java:83) at org.optaplanner.core.impl.solver.DefaultSolver.runSolverPhases(DefaultSolver.java:190) at org.optaplanner.core.impl.solver.DefaultSolver.solve(DefaultSolver.java:155) at Main.main(Main.java:29) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:601) at com.intellij.rt.execution.application.AppMain.main(AppMain.java:120) Also, thanks for pointing me to the doc on cloning, I have tried to implement it as follows but now it doesn't seem to be able to update the values and just runs without changing original inputted solution until timelimit: My Solution's clone: @Override public Timetable planningClone() { TimetableSlot[][] clonedTimetableSlotArray = new TimetableSlot[Constants.TIME_SLOTS][Constants.NUMBER_OF_STUDENTS]; for (int i = 0; i < Constants.TIME_SLOTS; i++) { for (int j = 0; j < Constants.NUMBER_OF_STUDENTS; j++) { try { clonedTimetableSlotArray[i][j] = (TimetableSlot) this.timetable[i][j].clone(); } catch (CloneNotSupportedException e) { e.printStackTrace(); } } } Timetable clonedTimetable = new Timetable(clonedTimetableSlotArray); clonedTimetable.setScore(this.getScore()); return clonedTimetable; } My Solution's Entity Collection: @PlanningEntityCollectionProperty public Collection<TimetableSlot> getListOfTimetableSlots() { Collection<TimetableSlot> timetableSlotCollection = new ArrayList<TimetableSlot>(); for (int i = 0; i < Constants.TIME_SLOTS; i++) { for (int j = 0; j < Constants.NUMBER_OF_STUDENTS; j++) { try { timetableSlotCollection.add((TimetableSlot) timetable[i][j].clone()); } catch (CloneNotSupportedException e) { e.printStackTrace(); } } } return timetableSlotCollection; } My Planning entity's clone: @Override protected Object clone() throws CloneNotSupportedException { return new TimetableSlot(this.timeSlot, this.student, this.getLecturer()); } this.getLecturer() ensures that the Lecturer planning variable IS NOT cloned. Both timeSlot and student are int. I think I have met the requirements set out when doing custom cloning as specified in the documentation. I am still stumped as to what the problem is, I hope I've provided more insight. Thanks, Karan -- View this message in context: http://drools.46999.n3.nabble.com/Best-Solution-found-but-latest-solution... Sent from the Drools: User forum mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users