Convenience class to store information associated to a specific state. More...

Inheritance diagram for frodo2.algorithms.duct.SearchNodeChild< V extends Addable< V > >:

Public Member Functions
	SearchNodeChild (int numberOfValues, boolean maximize, boolean ignore_inf)
	Constructor.
AddableReal	storeCost (int valueIndex, AddableReal costSample, AddableReal childBound, AddableReal infeasibleUtility, final boolean maximize)
	Stores the sum of the costs reported by the children for this particular state.
String	toString (double delta)
Public Member Functions inherited from frodo2.algorithms.duct.SearchNode< V >
	SearchNode (int numberOfValues, boolean maximize, boolean ignore_inf)
	Constructor.
void	initSampling (UtilitySolutionSpace< V, AddableReal > space, AddableReal infeasibleUtility, String[] contextVariables, V[] context, V[] domain)
	Finds the first feasible local solution, given the current context.
void	visited (Bound< V > b)
	Method called whenever the state is sampled.
int	solveLocalProblem (UtilitySolutionSpace< V, AddableReal > space, AddableReal infeasibleUtility, String[] contextVariables, V[] context, V[] domain)
	Finds the next feasible local solution.
AddableReal	storeCost (int valueIndex, AddableReal costSample, AddableReal infeasibleUtility, final boolean maximize)
	Stores the sum of the costs reported by the children for this particular state.
double	convergenceBound (int index, double delta)
	With probablity 1- `delta` the error between the real value and the estimated value lies in.
String	toString (double delta)
int	getRandomUnknowLocal (UtilitySolutionSpace< V, AddableReal > space, AddableReal infeasibleUtility, String[] contextVariables, V[] context, V[] domain)
	Method to randomly pick a value that has not been sampled yet.

Public Attributes
double[]	childB
	The index with the max/min value is to be sampled.
boolean	complete
	`true` when for all domain values the optimal cost is known
Public Attributes inherited from frodo2.algorithms.duct.SearchNode< V >
AddableReal[]	localCosts
	For each domain value, the local costs assocated to it.
ArrayList< Integer >	unknowLocalSolutions
	List of local solutions that have not yet been sampled.
int	nbrFeasibleLocalSolutions
	Counts the number of local infeasible solutions.
boolean	feasible
	`true` when the local problem has at least one feasible solution, `false` otherwise
int	numberOfValues
	the size of the domain
double[]	averageUtil
	For each domain value, the average utility received so far.
double[]	bestUtil
	The best util received so far.
double[]	bounds
	For each domain value, the current bound.
int[]	actionFrequencies
	For each posible domain values, the number of times it has been selected.
int	frequencie
	The number of times the corresponding state has been visited.
double	maxValue
	The value woth the currently best utility.
double	maxBound
	The value with the currently highest upper bound.
int	maxValueIndex
	The index of the currently best value.
int	maxBoundIndex
	The index of the currently highest bound.

Additional Inherited Members
Protected Attributes inherited from frodo2.algorithms.duct.SearchNode< V >
final boolean	IGNORE_INF
	when `true`, infeasible values should always be ignored, when `false` the should be sampled once
boolean	random
	`true` when not all domain values have been sampled, and `false` otherwise
int	lastVisited
	Last value visited.
Package Functions inherited from frodo2.algorithms.duct.SearchNode< V >
String	arrayToString (Object[] arr, int highlight)
	Method for printing an array, highlighting the entry at `highlight`.
Package Attributes inherited from frodo2.algorithms.duct.SearchNode< V >
ArrayList< Integer >	samplingOrder
	The sampling order.

Detailed Description

Convenience class to store information associated to a specific state.

Author: Brammert Ottens, 29 aug. 2011

Parameters

<V>	type used for domain values

Constructor & Destructor Documentation

◆ SearchNodeChild()

frodo2.algorithms.duct.SearchNodeChild< V extends Addable< V > >.SearchNodeChild	(	int	numberOfValues,
		boolean	maximize,
		boolean	ignore_inf )

Constructor.

Parameters

numberOfValues	the domain size of the variable
maximize	`true` when maximizing, and `false` otherwise
ignore_inf	`true` when infeasible utilities should be totally ignored, `false` otherwise

References childB, and frodo2.algorithms.duct.SearchNode< V >.numberOfValues.

Member Function Documentation

◆ storeCost()

AddableReal frodo2.algorithms.duct.SearchNodeChild< V extends Addable< V > >.storeCost	(	int	valueIndex,
		AddableReal	costSample,
		AddableReal	childBound,
		AddableReal	infeasibleUtility,
		final boolean	maximize )

Stores the sum of the costs reported by the children for this particular state.

Author: Brammert Ottens, 24 aug. 2011

Parameters

valueIndex	the index of the sampled value
costSample	the sum of the costs reported by the children
childBound	bound reported by the child
infeasibleUtility	utility of infeasible assignments
maximize	`true` when maximizing, and `false` otherwise