org.apache.hadoop.tools.rumen

Class CDFPiecewiseLinearRandomGenerator

java.lang.Object

org.apache.hadoop.tools.rumen.CDFRandomGenerator

org.apache.hadoop.tools.rumen.CDFPiecewiseLinearRandomGenerator

public class CDFPiecewiseLinearRandomGenerator
extends CDFRandomGenerator

Constructor Summary

Constructors

Constructor and Description
CDFPiecewiseLinearRandomGenerator(LoggedDiscreteCDF cdf)
CDFPiecewiseLinearRandomGenerator(LoggedDiscreteCDF cdf, long seed)

Method Summary

Modifier and Type	Method and Description
long	valueAt(double probability) TODO This code assumes that the empirical minimum resp.

Methods inherited from class org.apache.hadoop.tools.rumen.CDFRandomGenerator

floorIndex, getDatumAt, getRankingAt, initializeTables, randomValue

Methods inherited from class java.lang.Object

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

Constructor Detail

CDFPiecewiseLinearRandomGenerator

public CDFPiecewiseLinearRandomGenerator(LoggedDiscreteCDF cdf)

Parameters:

cdf - builds a CDFRandomValue engine around this LoggedDiscreteCDF, with a defaultly seeded RNG

CDFPiecewiseLinearRandomGenerator

public CDFPiecewiseLinearRandomGenerator(LoggedDiscreteCDF cdf,
                                         long seed)

Parameters:

cdf - builds a CDFRandomValue engine around this LoggedDiscreteCDF, with an explicitly seeded RNG

seed - the random number generator seed

Method Detail

valueAt

public long valueAt(double probability)

TODO This code assumes that the empirical minimum resp. maximum is the epistomological minimum resp. maximum. This is probably okay for the minimum, because that likely represents a task where everything went well, but for the maximum we may want to develop a way of extrapolating past the maximum.

Specified by:

valueAt in class CDFRandomGenerator