uk.ac.liv.auction.agent
Class DiscreteLearnerStrategy

java.lang.Object
  uk.ac.liv.auction.agent.AbstractStrategy
      uk.ac.liv.auction.agent.FixedQuantityStrategyImpl
          uk.ac.liv.auction.agent.AdaptiveStrategyImpl
              uk.ac.liv.auction.agent.DiscreteLearnerStrategy

All Implemented Interfaces:

AdaptiveStrategy, AuctionEventListener, java.lang.Cloneable, FixedQuantityStrategy, Parameterizable, Prototypeable, Resetable, java.io.Serializable, Strategy

Direct Known Subclasses:

MDPStrategy, StimuliResponseStrategy

public abstract class DiscreteLearnerStrategy

extends AdaptiveStrategyImpl

implements java.io.Serializable

A class representing a strategy in which we adapt our bids using a discrete learning algorithm.

Parameters

base.markupscale double >= 0

(scaling factor by which to multiply the output from the learner)

Version:

$Revision: 1.17 $

Author:

Steve Phelps

See Also:

Serialized Form

Field Summary

protected double

markupScale A scaling factor used to multiply-up the output from the learning algorithm.

Fields inherited from class uk.ac.liv.auction.agent.AbstractStrategy

agent, auction, currentShout

Constructor Summary

DiscreteLearnerStrategy()

DiscreteLearnerStrategy(AbstractTradingAgent agent)

Method Summary

abstract int	act() Generate an action from the learning algorithm.
void	endOfRound(Auction auction)
double	getMarkupScale()
void	initialise()
abstract void	learn(Auction auction) Perform learning.
boolean	modifyShout(Shout.MutableShout shout) Modify the price and quantity of the given shout according to this strategy.
void	setMarkupScale(double markupScale)
void	setup(ec.util.ParameterDatabase parameters, ec.util.Parameter base) Initialise this object from a parameter database.

Methods inherited from class uk.ac.liv.auction.agent.FixedQuantityStrategyImpl

determineQuantity, getQuantity, setQuantity

Methods inherited from class uk.ac.liv.auction.agent.AbstractStrategy

eventOccurred, getAgent, modifyShout, protoClone, reset, setAgent

Methods inherited from class java.lang.Object

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

Methods inherited from interface uk.ac.liv.auction.agent.AdaptiveStrategy

getLearner, setLearner

Methods inherited from interface uk.ac.liv.auction.agent.Strategy

determineQuantity, modifyShout, setAgent

Methods inherited from interface uk.ac.liv.util.Prototypeable

protoClone

Methods inherited from interface uk.ac.liv.util.Resetable

reset

Methods inherited from interface uk.ac.liv.auction.core.AuctionEventListener

eventOccurred

Field Detail

markupScale

protected double markupScale

A scaling factor used to multiply-up the output from the learning algorithm.

Constructor Detail

DiscreteLearnerStrategy

public DiscreteLearnerStrategy(AbstractTradingAgent agent)

DiscreteLearnerStrategy

public DiscreteLearnerStrategy()

Method Detail

initialise

public void initialise()

Overrides:

initialise in class AbstractStrategy

setup

public void setup(ec.util.ParameterDatabase parameters,
                  ec.util.Parameter base)

Description copied from interface: Parameterizable

Initialise this object from a parameter database.

Specified by:

setup in interface Parameterizable

Overrides:

setup in class FixedQuantityStrategyImpl

endOfRound

public void endOfRound(Auction auction)

Overrides:

endOfRound in class AdaptiveStrategyImpl

modifyShout

public boolean modifyShout(Shout.MutableShout shout)

Description copied from class: AbstractStrategy

Modify the price and quantity of the given shout according to this strategy.

Overrides:

modifyShout in class FixedQuantityStrategyImpl

getMarkupScale

public double getMarkupScale()

setMarkupScale

public void setMarkupScale(double markupScale)

act

public abstract int act()

Generate an action from the learning algorithm.

learn

public abstract void learn(Auction auction)

Perform learning.