Package org.cicirello.search.evo
package org.cicirello.search.evo
This package includes classes and interfaces directly related to implementing evolutionary
algorithms. This includes implementations of genetic algorithms and other forms of evolutionary
computation, as well as classes to support such implementation (e.g., classes for maintaining
populations).
-
ClassDescriptionAdaptiveEvolutionaryAlgorithm<T extends Copyable<T>>This class implements an evolutionary algorithm with adaptive control parameters (i.e., crossover rates and mutation rates that evolve during the search).AdaptiveMutationOnlyEvolutionaryAlgorithm<T extends Copyable<T>>This class implements an mutation-only evolutionary algorithm with an adaptive mutation rate that evolves during the search.This class implements a variation of fitness proportional selection that applies a bias function to transform the fitness values.This class implements a variation of Stochastic Universal Sampling (SUS) that we call Biased Stochastic Universal Sampling (Biased SUS), which integrates the use of a bias function with SUS to enable transforming fitness values prior to the stochastic selection decisions.The Best-non-penalized (BNP) is a replacement strategy that promotes proper diversity levels during the evolutionary process.This class implements Boltzmann selection.This class implements Boltzmann selection using Stochastic Universal Sampling (SUS).This class implements exponential rank selection.This class implements exponential rank selection using Stochastic Universal Sampling (SUS).This functional interface is used to provide a bias function to the
BiasedFitnessProportionalSelectionoperator as well as theBiasedStochasticUniversalSamplingoperator.FitnessFunction<T extends Copyable<T>>Fitness function interfaces.FitnessFunction.Double<T extends Copyable<T>>Fitness function interface for double-valued fitnesses.FitnessFunction.Integer<T extends Copyable<T>>Fitness function interface for int-valued fitnesses.This class implements fitness proportional selection, sometimes referred to as weighted roulette wheel, for evolutionary algorithms.FitnessShifter wraps another SelectionOperator, shifting all fitness values by the minimum fitness minus one, such that the least fit population member's transformed fitness is equal to 1, with the wrapped SelectionOperator than performing selection using the transformed fitnesses.GenerationalDisjointOperatorsEvolutionaryAlgorithm<T extends Copyable<T>>This class implements an evolutionary algorithm (EA) with a generational model, where a population of children are formed by applying genetic operators to members of the parent population, and where the children replace the parents in the next generation.GenerationalElitistReplacement keeps the k best-fit distinct members of the current population without application of evolutionary operators, and replaces the other (N-k) members of the population.GenerationalEvolutionaryAlgorithm<T extends Copyable<T>>This class implements an evolutionary algorithm with a generational model, such as is commonly used in genetic algorithms, where a population of children are formed by applying genetic operators to members of the parent population, and where the children replace the parents in the next generation.GenerationalMutationOnlyEvolutionaryAlgorithm<T extends Copyable<T>>This class implements an evolutionary algorithm (EA) with a generational model, such as is commonly used in genetic algorithms, where a population of children are formed by applying mutation to members of the parent population, and where the children replace the parents in the next generation.Generational replacement replaces the entire population each generation.This class is an implementation of a genetic algorithm (GA) with the common bit vector representation of solutions to optimization problems, and the generational model where children replace their parents each generation.InverseCostFitnessFunction<T extends Copyable<T>>This class provides a convenient mechanism for transforming optimization cost values to fitness values.This class implements linear rank selection.This class implements linear rank selection using Stochastic Universal Sampling (SUS).TerminationStrategy that terminates an evolutionary algorithm once the total number of fitness evaluations exceeds a threshold.This class is an implementation of a mutation-only genetic algorithm (GA) with the common bit vector representation of solutions to optimization problems, and the generational model where children replace their parents each generation.NegativeCostFitnessFunction<T extends Copyable<T>>This class provides a convenient mechanism for transforming optimization cost values to fitness values.NegativeIntegerCostFitnessFunction<T extends Copyable<T>>This class provides a convenient mechanism for transforming optimization cost values to fitness values.OnePlusOneEvolutionaryAlgorithm<T extends Copyable<T>>This class implements a (1+1)-EA.This class implements a (1+1)-GA, a special case of a (1+1)-EA, where solutions are represented with a vector of bits.An interface to candidates for the next generation's population, which can be the current members of the population or the children derived from it, depending upon the context.An interface to the candidates for the next generation's population, which can be the current members of the population or the children derived from it, depending upon the context.An interface to the candidates for the next generation's population, which can be the current members of the population or the children derived from it, depending upon the context.An interface to a vector of fitnesses of a population.An interface to a vector of fitnesses, each a double, of a population.An interface to a vector of fitnesses, each an int, of a population.PopulationMetaheuristic<T extends Copyable<T>>This interface defines the required methods for implementations of metaheuristics that maintain a population of candidate solutions such as evolutionary algorithms.This class implements a simple random selection operator that selects members of the population uniformly at random, independent of fitness values.Implement this interface to provide a replacement strategy for use by genetic algorithms and other forms of evolutionary computation.Interface forReplacementStrategyimplementations to specify which members of the parent and child populations serve as replacements into the next generation.Implement this interface to provide a selection operator for use by genetic algorithms and other forms of evolutionary computation.Implements sigma scaling by wrapping your chosen selection operator.This class is an implementation of the simple genetic algorithm (Simple GA) with the common bit vector representation of solutions to optimization problems, and the generational model where children replace their parents each generation.TerminationStrategy that terminates an evolutionary algorithm when a specified number of generations elapses without further improvement to the fitness of the most-fit solution found.This class implements Stochastic Universal Sampling (SUS), a selection operator for evolutionary algorithms.TerminationStrategy that terminates an evolutionary algorithm if a solution with a target fitness value (or better) is found.This interface is used to define a termination strategy.The state of the evolutionary algorithm prior to the start of a generation.TheTerminator enables composition of multiple terminator strategies, terminating an evolutionary algorithm if any of the componentTerminationStrategyinstance's termination criteria are met.This class implements tournament selection for evolutionary algorithms.This class implements truncation selection for evolutionary algorithms.