NEWS
xegaPopulation 1.0.0.15
- xegaObservePopulation:
Number of genes with same fitness added as 9th value
in statistics vector. Indicates premature convergence.
xegaPopulation 1.0.0.14
- Added rmpiLapply as parallel execution model.
- Added pid to output for verbose > 1.
- Dropped vector of gene indices with equal fitness from
xegaSummaryPopulation.
xegaPopulation 1.0.0.13
- Added Pipeline method "NoPipe" to lFxegaGaGene in all examples
with an evaluation of a population of genes.
xegaPopulation 1.0.0.12 (2026-02-17)
- asPipelineId() No pipeline configuration.
- asPipelineG() Embed pipeline in gene.
- xegaAsPipelineFactory() Configuration of pipeline configuration.
xegaPopulation 1.0.0.11 (2025-11-26)
- asPipeline() Changed to sequential version.
Reason: 1. Pipeline compilation and rmpi.
xegaPopulation 1.0.0.10 (2025-08-22)
- xegaConfiguration(): Changes problem and grammar parameters
from positional to named arguments.
xegaPopulation 1.0.0.9
- xegaNextPopulation() Moved updates of lF out to main loop in xegaRun.
xegaPopulation 1.0.0.8 (2025-07-30)
- asPipeline() Converts a population into a list of operator pipelines.
- xegaEvalPopulation() Evaluation of operator pipelines added.
- xegaNextPopulation() Elitist support for operator pipelines added.
xegaPopulation 1.0.0.7 (2025-04-16)
- terminateGEQ() Terminate if the (phenotype) value of a solution
is greater or equal than a threshold.
- terminateLEQ() Terminate if the (phenotype) value of a solution
is less or equal than a threshold.
xegaPopulation 1.0.0.6
- Examples for futureLapplyHet, PparLapply, and PparLapplyHet fixed
by Henrik Bengtsson.
xegaPopulation 1.0.0.5
- terminatePAC() Terminate if an epsilon-environment of an
estimated upper PAC bound of the optimal solution
is reached.
xegaPopulation 1.0.0.4
- xegaPopulation.R function xegaBestInPopulation:
- Evaluation of bestGene omitted
and all fields in solution record now derived
from bestGene.
Rationale: xegaRepEval provides aggregated statistics
with any setting of the evaluation function.
All fields derived from fit are now
mean fitnesses which should approximate
the expected fitness of a gene for
a sufficiently high number of repeated evaluations.
- terminate.R contains:
terminatedFalse()
terminatedAbsoluteError()
terminatedRelativeError()
TerminationFactory()
Rationale: Provides a set of standard termination conditions
to avoid their reimplementation in each problem
environment which uses them.
- xegaEvalPopulation.R contains:
xegaEvalPopulation()
xegaRepEvalPopulation()
xegaEvalPopulationFactory()
Rationale: Provides an evaluation function for repeated
evaluations of stochastic functions.
xegaLogEvalsPopulation() Reporting of sigma and obs added.
For assessing stability of mean fitness
of stochastic problems.
xegaPopulation 1.0.0.3
- ApplyFactory:
PparLapplyHet() added. For tasks with a high variance of
execution times.
- ApplyFactory:
futureLapplyHet() added. For tasks with a high variance of
execution times.
xegaPopulation 1.0.0.2
- ApplyFactory:
MClapplyHet() added. For tasks with a high variance of
execution times.
xegaPopulation 1.0.0.1
- CoolingFactory:
Message text corrected.