Package org.moeaframework.analysis.collector
package org.moeaframework.analysis.collector
Collectors record information periodically during the execution of an algorithm. The classes contained in this
package implement the various collectors and support utilities, but the preferred method of using collectors is
through the
Instrumenter
. The table below contains the potential data collected by the
instrumenter. The key column shows the string used to retrieve the data from the
Observation
.
Key | Description | Collector |
---|---|---|
NFE |
The number of objective function evaluations. | None (Always Saved) |
Approximation Set |
The approximation set of non-dominated solutions discovered by the algorithm. This collector consumes large quantities of memory. |
ApproximationSetCollector
|
Elapsed Time |
The wall-clock time elapsed since the start of the algorithm. |
ElapsedTimeCollector
|
Number of Restarts |
The number of time continuation or ε-continuation restarts. |
AdaptiveTimeContinuationCollector
|
Number of Improvements |
The number of ε-dominance improvements detected. |
EpsilonProgressCollector
|
Number of Dominating Improvements |
The number of strictly dominating ε-dominance improvements detected. |
EpsilonProgressCollector
|
Population Size |
The size of the population. |
PopulationSizeCollector
|
Archive Size |
The size of the archive. |
PopulationSizeCollector
|
AdditiveEpsilonIndicator |
The value of the additive ε-indicator performance indicator. | IndicatorCollector |
Contribution |
The value of the contribution/coverage performance indicator. | IndicatorCollector |
GenerationalDistance |
The value of the generational distance performance indicator. | IndicatorCollector |
Hypervolume |
The value of the hypervolume performance indicator. | IndicatorCollector |
InvertedGenerationalDistance |
The value of the inverted generational distance performance indicator. | IndicatorCollector |
MaximumParetoFrontError |
The value of the maximum Pareto front error performance indicator. | IndicatorCollector |
Spacing |
The value of the spacing performance indicator. | IndicatorCollector |
Operator Name |
The probability of an operator being selected by adaptive multimethod variation. Use the string name for the
operator used by OperatorFactory , such as "PCX" or "SBX+PM" .
|
AdaptiveMultimethodVariationCollector
|
-
ClassDescriptionCollects the individual operator probabilities from
AdaptiveMultimethodVariation
.Deprecated.Collects the number of restart events resulting fromAdaptiveTimeContinuationExtension
.Collects approximation sets from anAlgorithm
.Identifies objects in an object graph (all public and private fields contained within an object and those it references).Collects information from an object.Collects the elapsed execution time of an algorithm.Collects the number of ε-progress improvements detected in anEpsilonBoxEvolutionaryAlgorithm
.Collects performance indicator values from anAlgorithm
.InstrumentedAlgorithm<T extends Algorithm>Wraps an algorithm to indicate it is instrumented to collect runtime data.Decorates an algorithm to periodically collect information about its runtime behavior.An observation records information about an algorithm at a point in time.The observations collected over the course of a single algorithm run.Collects the population from anEvolutionaryAlgorithm
.Collects the population size and archive size, if available, from anEvolutionaryAlgorithm
.Collects the population from anAbstractPSOAlgorithm
.
AdaptiveTimeContinuationExtensionCollector
instead