I'm wondering if we broke something in SPEA in some of the recent
changes to ECJ16's population model.

1. Does the same problem occur in ECJ 15?
2. Can you provide a small test example which exhibits this property
in ECJ 16?

Sean

On May 11, 2007, at 8:09 AM, Eva Alfaro wrote:

> Hi all,
>
> My name is Eva Alfaro-Cid and I work in a research institute in
> Valencia
> (Spain).
>
> Lately I have been working on the creation of bankruptcy prediction
> models
> using GP. I am trying to use the multiobjective SPEA2 package for my
> problem but I get an error message that says:
>
> _______________________________________________________
>
> Initializing Generation 0
>
> Truncating the archive
> Exception in thread "EvolutionAgent.0789779.3185923"
> java.lang.NullPointerException
> at ec.select.TournamentSelection.produce
> (TournamentSelection.java:170)
> at ec.gp.koza.CrossoverPipeline.produce
> (CrossoverPipeline.java:260)
> at ec.breed.MultiBreedingPipeline.produce
> (MultiBreedingPipeline.java:130)
> at ec.multiobjective.spea2.SPEA2Breeder.breedPopChunk
> (SPEA2Breeder.java:177)
> at ec.multiobjective.spea2.SPEA2Breeder.breedPopulation
> (SPEA2Breeder.java:108)
> at ec.simple.SimpleEvolutionState.evolve
> (SimpleEvolutionState.java:118)
> at ec.drm.EvolutionAgent.run(EvolutionAgent.java:484)
> at java.lang.Thread.run(Thread.java:595)
> ____________________________________________________________
>
> I guess I have done something wrong while setting the parameters in
> the .params file. (See attachment)
>
> Can anyone help me to figure out what I am doing wrong?
>
> Best regards, Eva
>
> <Y:ecjSambaecappquiebrasMOCopy of quiebras.params>