dk.deepthought.sidious.rules
Class ConstrainingRule

java.lang.Object
  dk.deepthought.sidious.rules.Rule
      dk.deepthought.sidious.rules.ConstrainingRule

public class ConstrainingRule
extends Rule

Rule to keep conventionally unconstrained sensors constrained. I.e. adds cost to otherwise "free" adjustments.

Author:

Deepthought

Field Summary

private java.util.Collection<SuperLinkID>

idsToConstrain A list of sensor id's which will be constrained by this.

Constructor Summary

ConstrainingRule(SuperLinkID... sensorIDs)
Constructs a constraining rule that constrains sensors with input IDs.

Method Summary

double	desire(State currentState, State newState, Step step) Method returns the calculated desire associated with the change from currentState to newState.
java.util.Collection<Goal>	getGoals() Method returns a collection of immediate goals.
private double	normalizedDistance(SensorInput currentSensor, SensorInput newSensor) Calculates the normalized distance between two sensors.

Methods inherited from class dk.deepthought.sidious.rules.Rule

getAdjustableSetting, getAdjustableSettingFromParent, getExplanation, getParentID, getSensorValue, setParentID

Methods inherited from class java.lang.Object

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

Field Detail

idsToConstrain

private java.util.Collection<SuperLinkID> idsToConstrain

A list of sensor id's which will be constrained by this.

Constructor Detail

ConstrainingRule

public ConstrainingRule(SuperLinkID... sensorIDs)

Constructs a constraining rule that constrains sensors with input IDs.

Parameters:

sensorIDs - the ids of the sensors to constrain

Method Detail

desire

public double desire(State currentState,
                     State newState,
                     Step step)

Description copied from class: Rule

Method returns the calculated desire associated with the change from currentState to newState.

The calculated desire must evaluate to [0,1], where 0 represents no desire for changing state, and 1 represents maximum desire for change.

Some rules are allowed to evaluate to values (much) larger than 1. This exception is only allowed if the outcome of not respecting the rule is fatal.

Specified by:

desire in class Rule

Parameters:

currentState - the current state

newState - the new state

step - the step

Returns:

the calculated desire

getGoals

public java.util.Collection<Goal> getGoals()

Description copied from class: Rule

Method returns a collection of immediate goals.

If there are no immediate goals, an empty collection is returned.

Specified by:

getGoals in class Rule

Returns:

the goals of this rule

normalizedDistance

private double normalizedDistance(SensorInput currentSensor,
                                  SensorInput newSensor)

Calculates the normalized distance between two sensors.

It takes two sensors and normalizes the distance between them to a value between 0 and 1.

Parameters:

currentSensor - the first sensor

newSensor - the second sensor

Returns:

the normalized value of the distance.