jp.ac.hiroshima_u.sci.math.saito.tinymt
Class TinyMT32

java.lang.Object
  org.apache.commons.math.random.AbstractRandomGenerator
      jp.ac.hiroshima_u.sci.math.saito.tinymt.TinyMT32

All Implemented Interfaces:

org.apache.commons.math.random.RandomGenerator

public final class TinyMT32
extends org.apache.commons.math.random.AbstractRandomGenerator

TinyMT is a pseudo random number generator.

To get an instance, call getDefault(), getDefault(long), getDefault(String) or getDefault(int[]).

This class is not synchronized, One way to use TinyMT in concurrent environment is to use an instance per a thread.

This class supports jump function. User can get an array of pseudo random number generators by calling getDefaultArray(int, long, long), or getDefaultArray(int, String, long).

This class supports discrete characteristic polynomial generators. User can get an array of pseudo random number generators whose characteristic polynomials are discrete by calling getTinyMTArray(int, long).

Author:

M. Saito

See Also:

TinyMT web page

Method Summary

java.lang.String	getCharacteristic() return characteristic polynomial in hexadecimal format.
static TinyMT32	getDefault() Factory method which returns the TinyMT with the first generated parameter of TinyMTDC.
static TinyMT32	getDefault(int[] seeds) get default TinyMT32 with seeding by array.
static TinyMT32	getDefault(long seed) Factory method which returns the TinyMT with the first generated parameter of TinyMTDC.
static TinyMT32	getDefault(java.lang.String seed) Factory method which returns the TinyMT with the first generated parameter of TinyMTDC.
static TinyMT32[]	getDefaultArray(int count, long seed, long jump) make and return an array of TinyMT.
static TinyMT32[]	getDefaultArray(int count, java.lang.String seed, long jump) Make and return an array of TinyMT.
int	getDelta() return Delta of TinyMT.
int	getId() return ID of TinyMT.
TinyMT32[]	getJumpedArray(int count, long jump) Make and return an array of TinyMT.
static TinyMT32[]	getTinyMTArray(int count, long seed) Make and return an array of TinyMT.
int	getWeight() return Hamming weight of characteristic polynomial of TinyMT.
double	nextDouble() returns double r, 0 <= r < 1.0.
float	nextFloat() returns float r, 0 <= r < 1.0.
int	nextInt() returns 32-bit integer.
long	nextLong() returns 64-bit integer.
void	setSeed(int seed) internal set seed function This seeding is compatible with C language implementation.
void	setSeed(int[] seeds) seeding by array of integers.
void	setSeed(long seed) initialize internal state by seed.
void	setSeed(java.lang.String seed) seeding by string, This will be convenient.

Methods inherited from class org.apache.commons.math.random.AbstractRandomGenerator

clear, nextBoolean, nextBytes, nextGaussian, nextInt

Methods inherited from class java.lang.Object

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

Method Detail

nextInt

public int nextInt()

returns 32-bit integer.

Specified by:

nextInt in interface org.apache.commons.math.random.RandomGenerator

Overrides:

nextInt in class org.apache.commons.math.random.AbstractRandomGenerator

Returns:

next int

nextLong

public long nextLong()

returns 64-bit integer.

Specified by:

nextLong in interface org.apache.commons.math.random.RandomGenerator

Overrides:

nextLong in class org.apache.commons.math.random.AbstractRandomGenerator

Returns:

next long

setSeed

public void setSeed(long seed)

initialize internal state by seed.

Specified by:

setSeed in interface org.apache.commons.math.random.RandomGenerator

Specified by:

setSeed in class org.apache.commons.math.random.AbstractRandomGenerator

Parameters:

seed - seed of randomness

setSeed

public void setSeed(java.lang.String seed)

seeding by string, This will be convenient. This function does not compatible from other language implementation because coding of characters are different.

Parameters:

seed - seed of pseudo random numbers

setSeed

public void setSeed(int[] seeds)

seeding by array of integers. This seeding is compatible with other language implementation.

Specified by:

setSeed in interface org.apache.commons.math.random.RandomGenerator

Overrides:

setSeed in class org.apache.commons.math.random.AbstractRandomGenerator

Parameters:

seeds - seeds of pseudo random numbers.

setSeed

public void setSeed(int seed)

internal set seed function This seeding is compatible with C language implementation.

Specified by:

setSeed in interface org.apache.commons.math.random.RandomGenerator

Overrides:

setSeed in class org.apache.commons.math.random.AbstractRandomGenerator

Parameters:

seed - seed of pseudo random numbers

getDefault

public static TinyMT32 getDefault(java.lang.String seed)

Factory method which returns the TinyMT with the first generated parameter of TinyMTDC.

Parameters:

seed - seed of pseudo random numbers.

Returns:

TinyMT with the first parameter.

getDefault

public static TinyMT32 getDefault(long seed)

Factory method which returns the TinyMT with the first generated parameter of TinyMTDC.

Parameters:

seed - seed of pseudo random numbers.

Returns:

TinyMT with the first parameter.

getDefault

public static TinyMT32 getDefault(int[] seeds)

get default TinyMT32 with seeding by array.

Parameters:

seeds - seeds for initialization.

Returns:

random number generator TinyMT32

getDefault

public static TinyMT32 getDefault()

Factory method which returns the TinyMT with the first generated parameter of TinyMTDC. System.nanoTime() and Thread.getId() are used for seed.

Returns:

TinyMT with the first parameter.

getDefaultArray

public static TinyMT32[] getDefaultArray(int count,
                                         long seed,
                                         long jump)

make and return an array of TinyMT. Each element has the same characteristic polynomial with TinyMT gotten by getDefaultMT. Especially, the first element is just same as default TinyMT. The second element has the state of jump * 2⁶⁴ steps after the first element. In other word, the first element will generate the same sequence with the second element, after jump * 2⁶⁴ pseudo random number generation.

Parameters:

count - number of TinyMT to be created.

seed - seed of first element

jump - step is jump * 2⁶⁴

Returns:

array of TinyMT

getDefaultArray

public static TinyMT32[] getDefaultArray(int count,
                                         java.lang.String seed,
                                         long jump)

Make and return an array of TinyMT. Each element has the same characteristic polynomial with TinyMT gotten by getDefaultMT. Especially, the first element is just same as default TinyMT. The second element has the state of jump * 2⁶⁴ steps after the first element. In other word, the first element will generate the same sequence with the second element, after jump * 2⁶⁴ pseudo random number generation. This is equals to TinyMT32.getDefault(seed).getJumpedArray(count, jump);

Parameters:

count - number of TinyMT to be created.

seed - seed of first element

jump - step is jump * 2⁶⁴

Returns:

array of TinyMT

getJumpedArray

public TinyMT32[] getJumpedArray(int count,
                                 long jump)

Make and return an array of TinyMT. Each element of the array has the same characteristic polynomial with this. Especially, the first element is just same as this. The second element has the state of jump * 2⁶⁴ steps after the first element. In other word, the first element will generate the same sequence with the second element, after jump * 2⁶⁴ pseudo random number generation.

Note: Do not call any setSeed methods after jump. Seeding will cancel the effect of jump.

Parameters:

count - number of arrays

jump - jump step

Returns:

jumped array of TinyMT32.

getTinyMTArray

public static TinyMT32[] getTinyMTArray(int count,
                                        long seed)
                                 throws java.io.IOException

Make and return an array of TinyMT. Each element of the array has discrete characteristic polynomial. This function does not return defaultTinyMT. User can call getJumpedArray(int, long) using each element and can get more TinyMTs.

Calling setSeed for each element is O.K and is a nice idea to assure independency.

Parameters:

count - number of TinyMT you want.

seed - seed of each element

Returns:

array of TinyMT, length may be smaller than count

Throws:

java.io.IOException - when can't read resource file.

nextDouble

public double nextDouble()

returns double r, 0 <= r < 1.0.

Specified by:

nextDouble in interface org.apache.commons.math.random.RandomGenerator

Specified by:

nextDouble in class org.apache.commons.math.random.AbstractRandomGenerator

Returns:

next double

nextFloat

public float nextFloat()

returns float r, 0 <= r < 1.0.

Specified by:

nextFloat in interface org.apache.commons.math.random.RandomGenerator

Overrides:

nextFloat in class org.apache.commons.math.random.AbstractRandomGenerator

Returns:

next float

getCharacteristic

public java.lang.String getCharacteristic()

return characteristic polynomial in hexadecimal format.

Returns:

characteristic polynomial

getId

public int getId()

return ID of TinyMT. ID is not unique in TinyMT.

Returns:

ID

getDelta

public int getDelta()

return Delta of TinyMT.

Returns:

Delta

getWeight

public int getWeight()

return Hamming weight of characteristic polynomial of TinyMT.

Returns:

Hamming weight