com.github.benmanes.caffeine.cache.simulator.policy.sketch.climbing.gradient

Class Stochastic

java.lang.Object

com.github.benmanes.caffeine.cache.simulator.policy.sketch.climbing.AbstractClimber

com.github.benmanes.caffeine.cache.simulator.policy.sketch.climbing.gradient.Stochastic

All Implemented Interfaces:

HillClimber

public final class Stochastic
extends AbstractClimber

Stochastic gradient descent (SGD) optimizer.

w(t+1) = w(t) - alpha * dL/dw(t) where,

• w(t) is the current window size
• alpha is the learning rate (step size)
• dL/dw(t) is the gradient of the curve
• w(t+1) is the new window size configuration

SGC may be enhanced using momentum, either classical or Nesterov's. For details see

• https://towardsdatascience.com/10-gradient-descent-optimisation-algorithms-86989510b5e9
• http://ruder.io/optimizing-gradient-descent/index.html#momentum
• http://cs231n.github.io/neural-networks-3/#sgd

Nested Class Summary

Nested classes/interfaces inherited from interface com.github.benmanes.caffeine.cache.simulator.policy.sketch.climbing.HillClimber

HillClimber.Adaptation, HillClimber.QueueType

Field Summary

Fields inherited from class com.github.benmanes.caffeine.cache.simulator.policy.sketch.climbing.AbstractClimber

hitsInMain, hitsInSample, hitsInWindow, missesInSample, previousHitRate, sampleSize

Constructor Summary

Constructors

Constructor and Description
Stochastic(Config config)

Method Summary

Modifier and Type	Method and Description
protected double	adjust(double hitRate) Returns the amount to adapt by.

Methods inherited from class com.github.benmanes.caffeine.cache.simulator.policy.sketch.climbing.AbstractClimber

adapt, onHit, onMiss, resetSample

Methods inherited from class java.lang.Object

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

Constructor Detail

Stochastic

public Stochastic(Config config)

Method Detail

adjust

protected double adjust(double hitRate)

Description copied from class: AbstractClimber

Returns the amount to adapt by.

Specified by:

adjust in class AbstractClimber