edu.unm.cs.cs351.tdrl.f07.p2
Class XRandom

java.lang.Object
  java.util.Random
      edu.unm.cs.cs351.tdrl.f07.p2.XRandom

All Implemented Interfaces:

Serializable

public class XRandom
extends Random

A small extension of java.util.Random, providing a couple of handy methods Random is missing.

Version:

1.1

Author:

ackley

See Also:

Serialized Form

Field Summary

static long

serialVersionUID

Constructor Summary

XRandom()
Create an XRandom initialized with a randomly-chosen seed, so results will tend to differ from run to run even if nothing else changes.

XRandom(long seed)
Create an XRandom initialized with the specified seed, so results will remain identical from run to run -- so long as the same seed is used, and all other random objects and methods (such as Math.random()) are scrupulously avoided.

Method Summary

int	drawFromCDF(double[] cdf) Draws an index from an arbitrary (discrete) CDF.
int	drawFromPMF(double[] pmf) Draws a value from a discrete probability mass function (PMF) and returns the corresponding outcome.
<T> T	drawFromPMF(Map<T,Double> pmf) Draws a value from a discrete probability mass function (PMF) and returns the corresponding outcome.
double	nextExponential(double mean) Draw a sample from an exponentially distributed random variable with the given mean value.
int	nextGeometric(double p) Draw a sample from a geometric distribution with parameter p.
boolean	nextProbability(double probability) Return true at random the specified fraction of the time.

Methods inherited from class java.util.Random

next, nextBoolean, nextBytes, nextDouble, nextFloat, nextGaussian, nextInt, nextInt, nextLong, setSeed

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

serialVersionUID

public static final long serialVersionUID

See Also:

Constant Field Values

Constructor Detail

XRandom

public XRandom()

Create an XRandom initialized with a randomly-chosen seed, so results will tend to differ from run to run even if nothing else changes.

XRandom

public XRandom(long seed)

Create an XRandom initialized with the specified seed, so results will remain identical from run to run -- so long as the same seed is used, and all other random objects and methods (such as Math.random()) are scrupulously avoided.

Parameters:

seed - the seed to initialize the generator with.

Method Detail

nextExponential

public double nextExponential(double mean)

Draw a sample from an exponentially distributed random variable with the given mean value.

Parameters:

mean - the specified average value of the produced numbers

Returns:

a sample drawn from exponentially distributed random variable with average value 'mean'

Throws:

IllegalArgumentException - if the mean is equal to zero

nextGeometric

public int nextGeometric(double p)

Draw a sample from a geometric distribution with parameter p. The mean of the resulting geometric distribution will be 1/p.

Parameters:

p - Parameter of the geometric distribution.

Returns:

Sample drawn from geometric distribution with parameter p.

Throws:

IllegalArgumentException - if p is not a probability (i.e., if p<0.0 || p>1.0).

nextProbability

public boolean nextProbability(double probability)

Return true at random the specified fraction of the time. Setting probability to 0.5 makes this method act the same as nextBoolean(). Setting it to 0.0 makes this method always return false, setting it to 1.0 makes it always return true.

Parameters:

probability - the chance of returning true

Returns:

true with the given probability, false with 1-probability

Throws:

IllegalArgumentException - if probability is less than 0.0 or greater than 1.0

drawFromCDF

public int drawFromCDF(double[] cdf)

Draws an index from an arbitrary (discrete) CDF. This selects one entry out of an array containing cumulative probabilities, and returns the index of the selected entry.

The contents of the array are assumed to be CDF values, sorted in increasing order. That is, it is required that:

 cdf[j]>=cdf[i]  for all j>i
 cdf[cdf.length-1]==1.0
 

If violations of these conditions are detected, an IllegalStateException is generated.

Parameters:

cdf - Array containing conditional distribution function data

Returns:

index of selected outcome

Throws:

IllegalStateException - if cdf appears to be non-normalized.

drawFromPMF

public <T> T drawFromPMF(Map<T,Double> pmf)

Draws a value from a discrete probability mass function (PMF) and returns the corresponding outcome. The PMF is required to be represented as a Map from the outcome space onto probabilities. It is required that:

• All probability values in the map are >=0.0
• The sum of all probabilities in the map is 1.0.

If a violation of these conditions is detected, this throws a IllegalStateException.

Type Parameters:

T - Type of the outcome space

Parameters:

pmf - Map from outcome space (T) onto probabilities.

Returns:

Element drawn from the outcome space according to the pmf.

Throws:

IllegalStateException - if the pmf is not properly normalized.

drawFromPMF

public int drawFromPMF(double[] pmf)

Draws a value from a discrete probability mass function (PMF) and returns the corresponding outcome. The PMF is required to be represented as an array of probabilities, and the returned value is the index of the selected element. It is required that:

• All probability values in the array are >=0.0
• The sum of all probabilities in the array is 1.0.

If a violation of these conditions is detected, this throws a IllegalStateException.

Parameters:

pmf - Array of probability values for each index

Returns:

Index drawn according to the pmf.

Throws:

IllegalStateException - if the pmf is not properly normalized.