desmoj.core.dist

Class ContDistEmpirical

java.lang.Object

desmoj.core.simulator.NamedObject

desmoj.core.simulator.ModelComponent

desmoj.core.simulator.Reportable

desmoj.core.dist.Distribution

desmoj.core.dist.NumericalDist<java.lang.Double>

desmoj.core.dist.ContDist

desmoj.core.dist.ContDistEmpirical

public class ContDistEmpirical
extends ContDist

Author:

JohannesGoebel

Constructor Summary

Constructors

Constructor and Description
ContDistEmpirical(Model owner, java.lang.String name, boolean showInReport, boolean showInTrace) Constructs an empirical distribution pruducing floating point values.

Method Summary

Modifier and Type	Method and Description
void	addEntry(double value, double frequency) Adds a new entry of an empirical value and its associated cumulative frequency.
Reporter	createDefaultReporter() Creates the default reporter for the RealDistEmpirical distribution.
java.lang.Double	getInverseOfCumulativeProbabilityFunction(double p) Abstract method to map a double p from 0...1 to the distribution's domain by determining the value x that satisfies P(X < x) = p.
boolean	isInitialized() Shows if the RealDistEmpirical distribution already is initialized.
java.lang.Double	sample() Returns the next sample specified by the empirical distribution.

Methods inherited from class desmoj.core.dist.ContDist

sampleTimeSpan, sampleTimeSpan

Methods inherited from class desmoj.core.dist.NumericalDist

sampleObject

Methods inherited from class desmoj.core.dist.Distribution

changeRandomGenerator, getInitialSeed, getNonNegative, getNumSamples, isAntithetic, reset, reset, setAntithetic, setNonNegative, setSeed

Methods inherited from class desmoj.core.simulator.Reportable

getCorrespondingSchedulable, getDescription, getObservations, getReporter, incrementObservations, incrementObservations, reportIsOn, reportOff, reportOn, resetAt, setCorrespondingSchedulable, setDescription, setReporter

Methods inherited from class desmoj.core.simulator.ModelComponent

current, currentEntity, currentEntityAll, currentEvent, currentModel, currentSimProcess, debugIsOn, debugOff, debugOn, getModel, isExperimentCompatible, isModelCompatible, presentTime, sendDebugNote, sendMessage, sendTraceNote, sendWarning, skipTraceNote, skipTraceNote, traceIsOn, traceOff, traceOn

Methods inherited from class desmoj.core.simulator.NamedObject

getName, getQuotedName, toString

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

ContDistEmpirical

public ContDistEmpirical(Model owner,
                         java.lang.String name,
                         boolean showInReport,
                         boolean showInTrace)

Constructs an empirical distribution pruducing floating point values. Empirical distributions have to be initialized manually before use. This is done by calling the addEntry(long, double) method to add values defining the behaviour of the desired distribution.

Parameters:

owner - Model : The distribution's owner

name - java.lang.String : The distribution's name

showInReport - boolean : Flag for producing reports

showInTrace - boolean : Flag for producing trace output

Method Detail

addEntry

public void addEntry(double value,
                     double frequency)

Adds a new entry of an empirical value and its associated cumulative frequency. No two values may be equal, since there can not be two different frequencies for one observed value. Entries with same value/frequency pair as an entry already made before are simply ignored. As soon as there is exactly one value with a cumulative probability of 0 and one value with a cumulative probability of 1, this distribution is considered "initialized", making it possible to obtain samples via the sample() method.

Parameters:

value - double : The empirical value observed

frequency - double : The corresponding cumulative frequency of the empirical value

createDefaultReporter

public Reporter createDefaultReporter()

Creates the default reporter for the RealDistEmpirical distribution.

Overrides:

createDefaultReporter in class Distribution

Returns:

Reporter : The reporter for the RealDistEmpirical distribution

See Also:

ContDistEmpiricalReporter

isInitialized

public boolean isInitialized()

Shows if the RealDistEmpirical distribution already is initialized. Being initialized means that there is exactly one value with a cumulative probability of 0 and one value with a cumulative probability of 1.

Returns:

boolean

sample

public java.lang.Double sample()

Returns the next sample specified by the empirical distribution. Return values are calculated by linear interpolation between the two values surrounding that value given by the random generator.

Overrides:

sample in class NumericalDist<java.lang.Double>

Returns:

Double : The next sample for this empirical distribution or 0 if the distribution has not been properly initialized yet

getInverseOfCumulativeProbabilityFunction

public java.lang.Double getInverseOfCumulativeProbabilityFunction(double p)

Abstract method to map a double p from 0...1 to the distribution's domain by determining the value x that satisfies P(X < x) = p.

Specified by:

getInverseOfCumulativeProbabilityFunction in class NumericalDist<java.lang.Double>

Parameters:

p - double: A value between 0 and 1

Returns:

N : The value x that satisfies P(X < x) = p