Class PbilFrequencyModel

java.lang.Object
com.knezevic.edaf.v3.models.discrete.PbilFrequencyModel
All Implemented Interfaces:
Model<BitString>

public final class PbilFrequencyModel extends Object implements Model<BitString>
Population-Based Incremental Learning (PBIL) probability-vector model.

After each elite selection step, bit probabilities are updated by an exponential moving average:

   p_i(t+1) = (1 - η) p_i(t) + η * mean_elite(x_i)
 
where η is learningRate.

The model clamps each p_i to [1e-6, 1-1e-6] to preserve exploration and avoid numerical degeneracy in long runs.