uk.ac.soton.simulation.jsit.core

Class DistributionCategorical<C extends java.lang.Enum<C>>

java.lang.Object

uk.ac.soton.simulation.jsit.core.AbstractStochasticItem

uk.ac.soton.simulation.jsit.core.Distribution

uk.ac.soton.simulation.jsit.core.DistributionDiscrete

uk.ac.soton.simulation.jsit.core.DistributionCategorical<C>

All Implemented Interfaces:

java.io.Serializable, ContinuousSampler

Direct Known Subclasses:

DistBernoulli, DistCustomCategorical, DistUniformDiscrete

public abstract class DistributionCategorical<C extends java.lang.Enum<C>>
extends DistributionDiscrete
implements java.io.Serializable

JSIT abstract representation of all categorical distributions. The categories can either be a Java enum and/or numbered outcomes 1-K which can be mapped to a sequence of integer ranges [a,b] (where there should be K entries across all the ranges). The nested Range class represents those ranges.

Since:

0.1

Author:

Stuart Rossiter

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DistributionCategorical.Range Nested class representing an integer (inclusive) range [min, max].

Field Summary

Fields inherited from class uk.ac.soton.simulation.jsit.core.DistributionDiscrete

CUMULATIVE_PROB_TOLERANCE

Method Summary

Modifier and Type	Method and Description
int	addRange(int min, int max) Add a range to be used as part of mapping sampled outcomes to an integer.
void	clearRanges() Clear all ranges, which will revert any integer sampling to return a 'raw' 1-K value.
int	getK() Get the distribution's K value; i.e., the number of discrete alternatives (labellable 1-K)
boolean	isLocked() Determine if the distribution is locked.
void	lock() Lock the distribution (which will cause an error on sampling) whilst changes are made to it (e.g., ranges are added to map to numeric outcomes).
C	sampleCategory() Sample a Java enum category value from the distribution.
abstract java.lang.String	toString() Require concrete subclasses to implement toString.
void	unlock() Unlock the distribution (allowing it to be sampled) after changes to it are completed.
boolean	usesEnumCategory() Whether this distribution returns an enum category on sampling or not.
boolean	usesMappedRanges() Whether this distribution uses mapped ranges or not.

Methods inherited from class uk.ac.soton.simulation.jsit.core.DistributionDiscrete

sampleDouble, sampleInt

Methods inherited from class uk.ac.soton.simulation.jsit.core.Distribution

createRegisteredCopy, createUnregisteredCopy

Methods inherited from class uk.ac.soton.simulation.jsit.core.AbstractStochasticItem

getSampler, registerAccessInfo, registerSampler

Methods inherited from class java.lang.Object

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

Method Detail

lock

public void lock()

Lock the distribution (which will cause an error on sampling) whilst changes are made to it (e.g., ranges are added to map to numeric outcomes).

Since:

0.2

unlock

public void unlock()

Unlock the distribution (allowing it to be sampled) after changes to it are completed. Any checks as to the consistency of the distribution will be (re)made at this point.

Since:

0.2

isLocked

public boolean isLocked()

Determine if the distribution is locked.

Returns:

A boolean true/false value.

Since:

0.2

toString

public abstract java.lang.String toString()

Require concrete subclasses to implement toString.

Specified by:

toString in class DistributionDiscrete

Since:

0.1

usesEnumCategory

public boolean usesEnumCategory()

Whether this distribution returns an enum category on sampling or not. (If not, it returns an integer.)

Returns:

A boolean true/false.

Since:

0.1

usesMappedRanges

public boolean usesMappedRanges()

Whether this distribution uses mapped ranges or not. (If not, it returns a 1-K integer when sampling an integer.)

Returns:

A boolean true/false.

Since:

0.2

sampleCategory

public C sampleCategory()

Sample a Java enum category value from the distribution. Will error if the distribution is locked.

Returns:

The enum value sampled.

Since:

0.1

addRange

public int addRange(int min,
                    int max)

Add a range to be used as part of mapping sampled outcomes to an integer. Ranges are applied in the order added; e.g., adding [1,2] and [5,6] ranges means that the 4 possible outcomes map to 1,2,5 and 6 respectively. The distribution should be locked prior to adding the first such range, and unlocked after the last is added (at which point it will be checked that the sequence of ranges includes K elements).

Parameters:

min - The range minimum.

max - The range maximum.

Returns:

The number of entries across all ranges added so far. (This should equal K when the last range is added.)

Since:

0.1

clearRanges

public void clearRanges()

Clear all ranges, which will revert any integer sampling to return a 'raw' 1-K value.

Since:

0.1

getK

public int getK()

Get the distribution's K value; i.e., the number of discrete alternatives (labellable 1-K)

Returns:

The K value.

Since:

0.1