C# Class social_learning.QLearningAgent

A Q-Learning teacher. The teacher learns via the Q-Learning Temporal Difference Algorithm.

Inheritance: NeuralAgent

Public Properties

Property	Type	Description
_prevState	double[]

Method	Description
QLearningAgent ( int id, int speciesId, IBlackBox brain, bool agentsNavigate, bool agentsHide, int numOrientationActions, int numVelocityActions, World world ) : System	Creates a new Q-Learning teacher.
activateNetwork ( double sensors ) : ISignalArray	Called at every step in the world. Given the sensor input, returns the change in orientation and velocity in the range [0,1].

Method	Description
greedyValue ( double sensors ) : double
selectEpsilonGreedy ( double sensors ) : double[]
selectGreedy ( double sensors ) : double[]
selectRandom ( double sensors ) : double[]
updateValueFunction ( double sensors ) : void
world_PlantEaten ( object sender, IAgent eater, Plant eaten ) : void

Creates a new Q-Learning teacher.

public QLearningAgent ( int id, int speciesId, IBlackBox brain, bool agentsNavigate, bool agentsHide, int numOrientationActions, int numVelocityActions, World world ) : System
id	int	The unique ID of this teacher.
speciesId	int
brain	IBlackBox	The neural network value function for this teacher. It should have (2 + # of sensors) input nodes and 1 output node.
agentsNavigate	bool
agentsHide	bool
numOrientationActions	int	The number of buckets to discretize the orientation action spacer into.
numVelocityActions	int	The number of buckets to discretize the velocity action spacer into.
world	World	The world this teacher will be evaluated in.
return	System

Called at every step in the world. Given the sensor input, returns the change in orientation and velocity in the range [0,1].

public activateNetwork ( double sensors ) : ISignalArray
sensors	double
return	ISignalArray

public double[] _prevState
return	double[]