#1 some bugs in JAGA-Source-v.1.0.b-13.04.04.zip

[Humberto Santos]

I have installed this version and in eclipse I have
bugs in the following packages, all concerning assert
token:

package org.jaga.selection;
package org.jaga.individualRepresentation.booleanFormulas;
package org.jaga.fitnessEvaluation.multiplexer;

some others have the same problem.

When I Run Example2 It gives me the following errors:

...
96) {size=2; repr=00000101010000111101; vals=(25.0,
41.0); fit=-2.5318365560288174}
97) {size=2; repr=00000110110000100100; vals=(18.0,
56.0); fit=-2.8201640821612215}
98) {size=2; repr=00001101000000101110; vals=(39.0,
52.0); fit=-2.391592653589793}
99) {size=2; repr=00011100100001100111; vals=(92.0,
69.0); fit=-1.8082593202564599}java.lang.Error:
Unresolved compilation problem:
Syntax error on token "assert", assert expected

Best result (in generation 0):
{size=2; repr=00011100100000010001; vals=(92.0, 30.0);
fit=-0.07492598692312624}
--------------------------------------------------
at
org.jaga.selection.RouletteWheelSelection.getRelativeFitnessValue(RouletteWheelSelection.java:118)
at
org.jaga.selection.RouletteWheelSelection.calculateCumulativeFitness(RouletteWheelSelection.java:82)
at
org.jaga.selection.RouletteWheelSelection.select(RouletteWheelSelection.java:71)
at
org.jaga.masterAlgorithm.SimpleGA.selectForReproduction(SimpleGA.java:149)
at
org.jaga.masterAlgorithm.SimpleGA.generateNextPopulation(SimpleGA.java:82)
at
org.jaga.masterAlgorithm.SimpleGA.exec(SimpleGA.java:60)
at
org.jaga.masterAlgorithm.ReusableSimpleGA.exec(ReusableSimpleGA.java:50)
at
org.jaga.exampleApplications.Example2.exec(Example2.java:66)
at
org.jaga.exampleApplications.Example2.main(Example2.java:80)
Exception in thread "main"

[Greg Paperin]

Logged In: YES
user_id=1016872

Hi,

thanks very much for bringing this to my attention.

I will look into this over the weekend and post a solution
ASAP.

Best Regards,

Greg.

[Greg Paperin]

Logged In: YES
user_id=1016872

Oh, just thougt of something:

which version of JDK are you using with Eclipse?
JAGA requires JDK 1.4.1 or higher (because it uses the assert
keyword). Also, please check for the options
javac -source 1.4
and
java -ea
or equivalent options in Eclipse.

It could be that you have the assert-keywords disabled.

It is just a thought - I will look into it properly on the
weekend.

Cheers,

Greg.

[Humberto Santos]

Logged In: YES
user_id=657390

I'm using in Eclipse

Compiler compliance level: 1.4

and used J2SDK 1.4.2_04

[Humberto Santos]

Logged In: YES
user_id=657390

I think I found the solution!

In package
org.jaga.individualRepresentation.greycodedNumbers the class
NDecimalsIndividualSimpleFactory I thing the function
getDecimalScale() should be changed to:

public int getDecimalScale() {
//double scal = Math.log(decimalScale) / Math.log(10); //
i.e. log(10, scale)
//return (byte) scal;
return (int) decimalScale;
}

Without this change, the valid() function in the same class
will have problems here:
if (indiv.getDecimalScale() != this.getDecimalScale()
|| indiv.getSize() != this.getIndividualSize()
|| indiv.getPrecision() != this.getPrecision()) {
throw new IllegalArgumentException("The given individual
(" + indiv
+ ") is likely not created by this factory.");
}

and throw off.

Whith that chage my program start to work and also the
Example2.class

What you think?

[Vermaak]

Logged In: YES
user_id=1065315

Yes, I'm another end user and had the same problem,
compiling with JDK1.4 under eclipse. Commented out the
asserts for now.

[Vermaak]

Logged In: YES
user_id=1065315

Excellent! Looking forward to using this library.

I am using Java 1.4.2 as well, so it can't be that.

[Nobody/Anonymous]

Logged In: NO

About that getDecimalScale() problem. Don't you have to
change setDecimalScale(int value) too ?