jmarkov.basic
Class Policy<S extends State,A extends Action>

java.lang.Object
  jmarkov.basic.Policy<S,A>

Type Parameters:

S - The states class

A - Tha Action class

public final class Policy<S extends State,A extends Action>
extends java.lang.Object

Policy is a set of "Decision Rules". It contains the Decision Rule for every stage t. When the problem has infinite horizon, only one decision rule is stored.

Author:

Andres Sarmiento, Germ�n Ria�o

Constructor Summary

Policy(DecisionRule d)
Creates a stationary policy with the given decition rule

Policy(int stages)
Creates a set with the given horizon.

Method Summary

A	getAction(S i, int t) Gets the action to be taken in state i at this stage t
DecisionRule<S,A>	getDecisionRule()
DecisionRule<S,A>	getDecisionRule(int t) Returns the decision rule for statge t
int	getHorizon() Return the time horizon for this Ploicy.
void	print() Prints the policy to the standard output
void	print(java.io.PrintWriter pw) Prints the policy to the given PrintWriter.
void	setDecisionRule(DecisionRule<S,A> pol) Sets a unique decision rule for the policy, for infinite horizon problems.
void	setDecisionRule(DecisionRule<S,A> dr, int t) Sets a decision rule for stage t in the policy
java.lang.String	toString() Gives the sting representation of this Policy

Methods inherited from class java.lang.Object

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

Constructor Detail

Policy

public Policy(int stages)

Creates a set with the given horizon.

Parameters:

stages - The number of stages

Policy

public Policy(DecisionRule d)

Creates a stationary policy with the given decition rule

Parameters:

d - The rule

Method Detail

getHorizon

public int getHorizon()

Return the time horizon for this Ploicy.

Returns:

last stage where actions can be taken

setDecisionRule

public void setDecisionRule(DecisionRule<S,A> dr,
                            int t)

Sets a decision rule for stage t in the policy

Parameters:

dr - decision rule

t - stage

setDecisionRule

public void setDecisionRule(DecisionRule<S,A> pol)

Sets a unique decision rule for the policy, for infinite horizon problems.

Parameters:

pol -

getDecisionRule

public DecisionRule<S,A> getDecisionRule()

Returns:

the unique decision rule, for infinite horizon problems.

getDecisionRule

public DecisionRule<S,A> getDecisionRule(int t)

Returns the decision rule for statge t

Parameters:

t - stage.

Returns:

The decision rule for stage t.

getAction

public A getAction(S i,
                   int t)

Gets the action to be taken in state i at this stage t

Parameters:

i - The state

t - The stage (time) at which action is to be taken.

Returns:

The action.

toString

public java.lang.String toString()

Gives the sting representation of this Policy

Overrides:

toString in class java.lang.Object

print

public void print()

Prints the policy to the standard output

print

public void print(java.io.PrintWriter pw)

Prints the policy to the given PrintWriter.

Parameters:

pw - print writer