Package org.spaceroots.mantissa.random

This package provides classes to perform some random draws and statistical anaysis.

See:
Description

Interface Summary

NormalizedRandomGenerator	This interface represent a normalized random generator for scalars.
RandomVectorGenerator	This interface represent a random generator for whole vectors.

Class Summary

CorrelatedRandomVectorGenerator	This class allows to generate random vectors with correlated components.
FourTapRandom	This class implements a powerful pseudo-random number generator studied by Robert M Ziff.
GaussianRandomGenerator	This class is a gaussian normalized random generator for scalars.
MersenneTwister	This class implements a powerful pseudo-random number generator developed by Makoto Matsumoto and Takuji Nishimura during 1996-1997.
ScalarSampleStatistics	This class compute basic statistics on a scalar sample.
UncorrelatedRandomVectorGenerator	This class allows to generate random vectors with uncorrelated components.
UniformRandomGenerator	This class implements a normalized uniform random generator.
VectorialSampleStatistics	This class compute basic statistics on a scalar sample.

Exception Summary

NotPositiveDefiniteMatrixException

This class represents exceptions thrown by the correlated random vector generator.

Package org.spaceroots.mantissa.random Description

This package provides classes to perform some random draws and statistical anaysis.

The aim of this package is to provide the basic components needed to generate random variables (correlated or not in the case of vectorial variables) that could be used in a simulation application and to provide some basic statistical classes in order to analyze the simulation results.

At the lowest level, number generation one at a time for any primitive type is provided by either the FourTapRandom class or the MersenneTwister class which both extend the java.util.Random standard class with much better algorithms. The FourTapRandom algorithm is due to Robert M. Ziff (this class was kindly contributed to mantissa by Bill Maier), whereas the MersenneTwister class is due to Makoto Matsumoto and Takuji Nishimura.

Vectorial generators are build by embedding a normalized scalar generator into the vectorial generator classes CorrelatedRandomVectorGenerator or UncorrelatedRandomVectorGenerator that will be responsible for packaging all numbers into vectors with the specified mean values, standard deviations and correlation coefficients. Since most practical problems make the assumption the probability distribution is a gaussian one, the normalized generator will often be an instance of GaussianRandomGenerator, but uniform distribution are also available using instances of UniformRandomGenerator.

Author:

L. Maisonobe