Package org.uncommons.watchmaker.framework

This package provides a framework for evolutionary computation.

See:
Description

Interface Summary

CandidateFactory<T>	Creates new populations of candidates.
EvolutionaryOperator<T>	An evolutionary operator is a function that takes a population of candidates as an argument and returns a new population that is the result of applying a transformation to the original population.
EvolutionEngine<T>	Operations for classes that provide an evolution implementation.
EvolutionObserver<T>	Call-back interface so that programs can monitor the state of a long-running evolutionary algorithm.
FitnessEvaluator<T>	Calculates the fitness score of a given candidate of the appropriate type.
SelectionStrategy<T>	Strategy interface for "natural" selection.
TerminationCondition	Interface for implementing conditions used to terminate evolutionary algorithms.

Class Summary

AbstractEvolutionEngine<T>	Base class for EvolutionEngine implementations.
CachingFitnessEvaluator<T>	A wrapper that provides caching for FitnessEvaluator implementations.
EvaluatedCandidate<T>	Immutable wrapper class for associating a candidate solution with its fitness score.
EvolutionStrategyEngine<T>	General purpose engine for implementing Evolution Strategies.
EvolutionUtils	Utility methods used by different evolution implementations.
FitnessEvaluationWorker	This is the class that actually runs the fitness evaluation tasks created by a EvolutionEngine.
GenerationalEvolutionEngine<T>	This class implements a general-purpose generational evolutionary algorithm.
PopulationData<T>	Immutable data object containing statistics about the state of an evolved population and a reference to the fittest candidate solution in the population.
SteadyStateEvolutionEngine<T>	An implementation of steady-state evolution, which is a type of evolutionary algorithm where a population is changed incrementally, with one individual evolved at a time.

Package org.uncommons.watchmaker.framework Description

This package provides a framework for evolutionary computation. It defines generic interfaces for evolutionary operators, fitness functions and selection strategies. It also provides an all-purpose EvolutionEngine.

Author:

Daniel Dyer