tracks.singlePlayer.tools.ucbOptimizerAgent

Class Agent

java.lang.Object

core.player.Player

core.player.AbstractPlayer

tracks.singlePlayer.tools.ucbOptimizerAgent.Agent

public class Agent
extends AbstractPlayer

Created with IntelliJ IDEA. User: ssamot Date: 14/11/13 Time: 21:45 This is a Java port from Tom Schaul's VGDL - https://github.com/schaul/py-vgdl

Field Summary

Fields

Modifier and Type	Field and Description
static Types.ACTIONS[]	actions
static double	K
static int	NUM_ACTIONS
static double[]	parameters
static int	ROLLOUT_DEPTH
static double	safetyMargin
static UCBEquation	ucb

Constructor Summary

Constructors

Constructor and Description
Agent(StateObservation so, ElapsedCpuTimer elapsedTimer) Public constructor with state observation and time due.

Method Summary

Modifier and Type	Method and Description
Types.ACTIONS	act(StateObservation stateObs, ElapsedCpuTimer elapsedTimer) Picks an action.
void	result(StateObservation stateObservation, ElapsedCpuTimer elapsedCpuTimer) Function called when the game is over.

Methods inherited from class core.player.AbstractPlayer

act

Methods inherited from class core.player.Player

draw, getAllActions, getLastAction, getPlayerID, isHuman, logAction, resultMulti, setPlayerID, setup, teardown

Methods inherited from class java.lang.Object

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

Field Detail

NUM_ACTIONS

public static int NUM_ACTIONS

ROLLOUT_DEPTH

public static int ROLLOUT_DEPTH

K

public static double K

actions

public static Types.ACTIONS[] actions

ucb

public static UCBEquation ucb

parameters

public static double[] parameters

safetyMargin

public static double safetyMargin

Constructor Detail

Agent

public Agent(StateObservation so,
             ElapsedCpuTimer elapsedTimer)

Public constructor with state observation and time due.

Parameters:

so - state observation of the current game.

elapsedTimer - Timer for the controller creation.

Method Detail

act

public Types.ACTIONS act(StateObservation stateObs,
                         ElapsedCpuTimer elapsedTimer)

Picks an action. This function is called every game step to request an action from the player.

Specified by:

act in class Player

Parameters:

stateObs - Observation of the current state.

elapsedTimer - Timer when the action returned is due.

Returns:

An action for the current state

result

public void result(StateObservation stateObservation,
                   ElapsedCpuTimer elapsedCpuTimer)

Function called when the game is over. This method must finish before CompetitionParameters.TEAR_DOWN_TIME, or the agent will be DISQUALIFIED

Overrides:

result in class Player

Parameters:

stateObservation - the game state at the end of the game

elapsedCpuTimer - timer when this method is meant to finish.