Package de.kumpe.hadooptimizer.impl

Some basic und useful implementations of the evolution cycle's interfaces.

See:
          Description

Interface Summary
ReportingHalterWrapper.Reporter<I> The ReportingHalterWrapper.Reporter is used by the ReportingHalterWrapper to report about the passed evaluation-results every cycle.
 

Class Summary
ComputingTimeLimitHalter<I> A Halter which halts after a given amount computation time.
CountingHalter<I> A Halter which halts after the specified number of cycles.
DuplicatingRecombiner<I> A Recombiner which creates the given number of offspring by randomly selecting and duplicating a parents.
EsEvaluatorWrapper Wrapp's a given Evaluator<double[]> to evaluate the double[]-individual nested in an EsIndividual.
EsHalterWrapper A Wrapper that transforms an Halter<double[]> into an Halter<EsIndividual>.
EsMutatorWrapper Wrapp's a given EsWrappableMutator to mutate the double[] -individual nested in an EsIndividual.
EsPopulationReaderWrapper A Wrapper that transforms an PopulationReader<double[]> into an PopulationReader<EsIndividual>.
EsPopulationWriterWrapper A Wrapper that transforms an PopulationWriter<double[]> into an PopulationWriter<EsIndividual>.
EvaluationLimitHalter<I> A Halter halting the evolution process when the evaluation has gone below a given level.
GaussianMutator Mutates a double array by adding gaussian random numbers.
GaussionMutatorBase Base class for Mutators based on gaussian random number.
IdentityMutator<I> The IdentityMutator effectively does nothing, it just passes the input-individual through.
LoggingReporter<I> A ReportingHalterWrapper.Reporter implementation which prints summaries of the evaluation-results.
MemoryPopulationReader<I> A PopulationReader which returns an in memory collection of individuals.
MemoryPopulationWriter<I> A PopulationWriter which stores the given population in memory.
NeedsRandomWrapperBase<T>  
NeverHalter<I> This Halter's NeverHalter.halt(Collection) will always return false, thus never halt.
NullPopulationWriter<I> A PopulationWriter which discards the given population.
PolynomialEvaluator A Evaluator for polynomials.
ReportingHalterWrapper<I> The ReportingHalterWrapper passed the EvaluationResults to each registered ReportingHalterWrapper.Reporter and then delegates the halt-decision the the specified Halter.
TextFileReporter<I> A ReportingHalterWrapper.Reporter implementation which writes the evaluation-results to a simple text-file of the following format: A comment in the first line describing the file's format and one line per EvaluationResult in with the following colums: the cycle number the result number the evaluation the individual formatted by LoggingReporter.formatIndividual(StringBuilder, Object) With an double[] individual the result may look like this: #cycle result evaluation individual 0 0 20000.0 [100.0, 100.0] 1 0 19260.53116542464 [98.73071889720983, 97.38596735781995] 1 1 19468.191322019986 [99.22362530562448, 98.05102529016519] 1 2 19532.212300408937 [97.62563047159728, 99.91503437205381] 1 3 19567.90766494946 [99.72890299878424, 98.03323251603692] 2 0 18794.398942170505 [96.3683774664037, 97.34133324282375] 2 1 18913.606724906247 [98.47298455693839, 95.88758715386727] 2 2 19009.274361645486 [97.58557565219022, 97.25069301017946] 2 3 19068.25006183995 [96.79725135084752, 98.36164874253139]
TimeBasedHalter<I> A Halter implementation halting a the specified Date.
 

Package de.kumpe.hadooptimizer.impl Description

Some basic und useful implementations of the evolution cycle's interfaces.

Author:
Christian Kumpe


Copyright © 2011 Christian Kumpe. All Rights Reserved.