org.jppf.server.scheduler.bundle.rl
Class AbstractRLBundler

java.lang.Object
  org.jppf.server.scheduler.bundle.AbstractBundler
      org.jppf.server.scheduler.bundle.rl.AbstractRLBundler

All Implemented Interfaces:

Bundler

Direct Known Subclasses:

RLBundler

public abstract class AbstractRLBundler
extends AbstractBundler

Bundler based on a reinforcement learning algorithm.

Author:

Laurent Cohen

Field Summary

protected int	action Action to take.
protected int	bundleSize The current bundle size.
protected BundleDataHolder	dataHolder Bounded memory of the past performance updates.
protected int	prevBundleSize The previous bundle size.

Fields inherited from class org.jppf.server.scheduler.bundle.AbstractBundler

bundlerNumber, profile, timestamp

Constructor Summary

AbstractRLBundler(LoadBalancingProfile profile)
Creates a new instance with the specified parameters profile.

Method Summary

void	dispose() Release the resources used by this bundler.
void	feedback(int size, double totalTime) This method computes the bundle size based on the new state of the server.
void	feedback2(int size, double totalTime) This method computes the bundle size based on the new state of the server.
int	getBundleSize() Get the current size of bundle.
BundleDataHolder	getDataHolder() Get the bounded memory of the past performance updates.
void	setBundleSize(int bundleSize) set the current size of bundle.
void	setup() Perform context-independant initializations.

Methods inherited from class org.jppf.server.scheduler.bundle.AbstractBundler

getBundlerNumber, getProfile, getTimestamp, maxSize

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.jppf.server.scheduler.bundle.Bundler

copy

Field Detail

action

protected int action

Action to take.

dataHolder

protected BundleDataHolder dataHolder

Bounded memory of the past performance updates.

bundleSize

protected int bundleSize

The current bundle size.

prevBundleSize

protected int prevBundleSize

The previous bundle size.

Constructor Detail

AbstractRLBundler

public AbstractRLBundler(LoadBalancingProfile profile)

Creates a new instance with the specified parameters profile.

Parameters:

profile - the parameters of the algorithm grouped as a performance analysis profile.

Method Detail

getBundleSize

public int getBundleSize()

Get the current size of bundle.

Returns:

the bundle size as an int value.

See Also:

Bundler.getBundleSize()

setBundleSize

public void setBundleSize(int bundleSize)

set the current size of bundle.

Parameters:

bundleSize - - the bundle size as an int value.

feedback

public void feedback(int size,
                     double totalTime)

This method computes the bundle size based on the new state of the server.

Specified by:

feedback in interface Bundler

Overrides:

feedback in class AbstractBundler

Parameters:

size - the number of tasks executed.

totalTime - the time in milliseconds it took to execute the tasks.

See Also:

AbstractBundler.feedback(int, double)

feedback2

public void feedback2(int size,
                      double totalTime)

This method computes the bundle size based on the new state of the server.

Parameters:

size - the number of tasks executed.

totalTime - the time in milliseconds it took to execute the tasks.

See Also:

AbstractBundler.feedback(int, double)

setup

public void setup()

Perform context-independant initializations.

Specified by:

setup in interface Bundler

Overrides:

setup in class AbstractBundler

See Also:

AbstractBundler.setup()

dispose

public void dispose()

Release the resources used by this bundler.

Specified by:

dispose in interface Bundler

Overrides:

dispose in class AbstractBundler

See Also:

AbstractBundler.dispose()

getDataHolder

public BundleDataHolder getDataHolder()

Get the bounded memory of the past performance updates.

Returns:

a BundleDataHolder instance.