ec.coevolve

Interface GroupedProblemForm

All Known Implementing Classes:

CoevolutionaryECSuite, CompetitiveMaxOnes, GEProblem, MasterProblem

public interface GroupedProblemForm

GroupedProblemForm.java

GroupedProblemForm is an interface which defines methods for Problems to implement simple coevolutionary evaluation. In particular, the evaluate method receives as parameters a set of individuals that need to be evaluated. An additional vector-parameter (updateFitness) marks which individual fitnesses need to be updated during the evaluation process.

Version:

1.0

Author:

Sean Luke & Liviu Panait

Method Summary

Methods

Modifier and Type	Method and Description
void	evaluate(EvolutionState state, Individual[] ind, boolean[] updateFitness, boolean countVictoriesOnly, int[] subpops, int threadnum) Evaluates the individuals found in ind together.
void	postprocessPopulation(EvolutionState state, Population pop, boolean[] assessFitness, boolean countVictoriesOnly) Finish processing the population (such as fitness information) after evaluation.
void	preprocessPopulation(EvolutionState state, Population pop, boolean[] prepareForFitnessAssessment, boolean countVictoriesOnly) Set up the population pop (such as fitness information) prior to evaluation.

Method Detail

preprocessPopulation

void preprocessPopulation(EvolutionState state,
                        Population pop,
                        boolean[] prepareForFitnessAssessment,
                        boolean countVictoriesOnly)

Set up the population pop (such as fitness information) prior to evaluation. Although this method is not static, you should not use it to write to any instance variables in the GroupedProblem instance; this is because it's possible that the instance used is in fact the prototype, and you will have no guarantees that your instance variables will remain valid during the evaluate(...) process. Do not assume that pop will be the same as state.pop -- it may not. state is only provided to give you access to EvolutionState features. Typically you'd use this method to set the Fitness values of all Individuals to 0.

countVictoriesOnly will be set if Individuals' fitness is to be based on whether they're the winner of a test, instead of based on the specifics of the scores in the tests. This really only happens for Single-Elimination Tournament one-population competitive coevolution.

prepareForFitnessAssessment will indicate which subpopulations will have their fitness values updated this time around, during postprocessPopulation. It may not be the same as updateFitness[] in evaluate(...).

If you are basing fitness on trials, this method should create the initial trials if the prepareForFitnessAssessment[...] is true for that subpopulation.

postprocessPopulation

void postprocessPopulation(EvolutionState state,
                         Population pop,
                         boolean[] assessFitness,
                         boolean countVictoriesOnly)

Finish processing the population (such as fitness information) after evaluation. Although this method is not static, you should not use it to write to any instance variables in the GroupedProblem instance; this is because it's possible that the instance used is in fact the prototype, and you will have no guarantees that your instance variables will remain valid during the evaluate(...) process. Do not assume that pop will be the same as state.pop -- it may not. state is only provided to give you access to EvolutionState features.

countVictoriesOnly will be set if Individuals' fitness is to be based on whether they're the winner of a test, instead of based on the specifics of the scores in the tests. This really only happens for Single-Elimination Tournament one-population competitive coevolution. If this is set, probably would leave the Fitnesses as they are here (they've been set and incremented in evaluate(...)), but if it's not set, you may want to set the Fitnesses to the maximum or average or the various trials performed.

assessFitness will indicate which subpopulations should have their final fitness values assessed. You should not clear the trials of individuals for which assessFitness[] is false. Instead allow trials to accumulate and ultimately update the fitnesses later when the flag is set. assessFitness[] may not be the same as updateFitness[] in evaluate(...).

evaluate

void evaluate(EvolutionState state,
            Individual[] ind,
            boolean[] updateFitness,
            boolean countVictoriesOnly,
            int[] subpops,
            int threadnum)

Evaluates the individuals found in ind together. If updateFitness[i] is true, then you should use this evaluation to update the fitness of the individual in ind[i]. Individuals which are updated should have their fitnesses modified so that immediately after evaluation (and prior to postprocessPopulation(...) being called) individuals' fitnesses can be checked to see which is better than which. Do not assume that the individuals in ind will actually be in state.pop (they may not -- this method may be called at the end of a run to determine the best individual of the run in some kind of contest).

countVictoriesOnly will be set if Individuals' fitness is to be based on whether they're the winner of a test, instead of based on the specifics of the scores in the tests. This really only happens for Single-Elimination Tournament one-population competitive coevolution. If this is set, you should increment the Fitness of the winner each time. If it's not set, you should update Fitness as you see fit, then set the final Fitness in preprocessPopulation.