[jgap-users] Junk gene entries in fittest chromosomes

[Nia W F. <ni...@us...>]

Hi, 

I'm new to JGAP. 

Sometimes the alleles in my StringGene from a call to 
getFittestChromosomes() have junk data in them.  It's usually just 1 out 
of the 5 alleles (the other 4 look valid). 

My population doesn't always have valid data for all 5 alleles in all the 
StringGenes, so I set the values to "" if this is the case.  (I was having 
problems if I set them to null.) 

One weird thing that I noticed is that when I print out the individual 
chromosomes in my evaluate() function where I calculate the fitness 
factor, the chromosomes have junk data in the alleles at this point.  But 
I am pretty sure that the population I use in the Configuration is valid 
and doesn't have any junk data as I tried printing the raw population data 
and it all looked valid.  I'm not sure where the junk data is getting 
generated.  I am explicitly creating a Population for the 
DefaultConfiguration using historical data, creating a new StringGene, new 
Chromosome and calling mypopuation.addChromosome(). 

I don't think I can restrict the individual allele types in the gene when 
they are initialized using the ALPHABET_CHARACTERS_DIGITS, LOWER, SPECIAL, 
UPPER because the allele values are a combination of DIGITS and UPPER. 

Top 3 unique fittest chromosomes : 

1) 0000015R6714        ZF062        ó?°%³Ð(?fǽ?6n??wdö?¥B        340   
20        99  <---- junk data ó?°%³Ð(?fǽ?6n??wdö?¥B 
2) 0000042R6678        GF185        DRIVE ASM , 70GB 3.5 SAS        340   
10        99 
3) 0000097P3000        4327A        MECH ASM  , 70GB TRAY           340   
10        75

Thanks in advance for any help! 

Nia