Package org.tweetyproject.machinelearning.rl.mdp.algorithms

Class OfflineAlgorithm<S extends State,A extends Action>

java.lang.Object

org.tweetyproject.machinelearning.rl.mdp.algorithms.OfflineAlgorithm<S,A>

Type Parameters:

S - The type of states

A - The type of actions

Direct Known Subclasses:

PolicyIteration, ValueIteration

public abstract class OfflineAlgorithm<S extends State,A extends Action>
extends Object

A general interface for algorithms to determine optimal policies directly from an MDP

Author:

Matthias Thimm

Constructor Summary

Constructors

Constructor	Description
OfflineAlgorithm()	Default Constructor

Method Summary

Modifier and Type	Method	Description
Policy<S,A>	getPolicy(Map<S,Double> utilities, MarkovDecisionProcess<S,A> mdp, double gamma)	Determines the best policy, given the utilities
abstract Policy<S,A>	getPolicy(MarkovDecisionProcess<S,A> mdp, double gamma)	Determines the optimal policy for the given MDP.

Methods inherited from class java.lang.Object

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

Constructor Details

OfflineAlgorithm

public OfflineAlgorithm()

Default Constructor

Method Details

getPolicy

public abstract Policy<S,A> getPolicy(MarkovDecisionProcess<S,A> mdp,
 double gamma)

Determines the optimal policy for the given MDP.

Parameters:

mdp - some MDP

gamma - the used discount factor for utility determination

Returns:

the optimal policy

getPolicy

public Policy<S,A> getPolicy(Map<S,Double> utilities,
 MarkovDecisionProcess<S,A> mdp,
 double gamma)

Determines the best policy, given the utilities

Parameters:

utilities - a mapping of states to utilities

mdp - some MDP

gamma - discount factor

Returns:

the best policy