Module es.upm.fi.cig.multictbnc

Package es.upm.fi.cig.multictbnc.learning.structure.optimisation.tabusearch

Class CTBNTabuSearchIndividual

java.lang.Object

es.upm.fi.cig.multictbnc.learning.structure.optimisation.hillclimbing.implementation.CTBNHillClimbingIndividual

es.upm.fi.cig.multictbnc.learning.structure.optimisation.tabusearch.CTBNTabuSearchIndividual

All Implemented Interfaces:

HillClimbingImplementation

public class CTBNTabuSearchIndividual
extends CTBNHillClimbingIndividual

Implements the tabu search algorithm for continuous-time Bayesian networks. It finds the parent set for a single node, so it can be used to find the structure of a continuous-time Bayesian network in parallel.

Constructor Summary

Constructors

Constructor	Description
CTBNTabuSearchIndividual(CTBNScoreFunction scoreFunction, int tabuListSize, int maxNumNotImprovements)	Initialises the tabu search algorithm by proving a score function and a tabu list size.

Method Summary

Modifier and Type	Method	Description
protected HillClimbingSolution	findBestNeighbor(CTBN<? extends Node> ctbn, int idxNode, boolean[][] adjacencyMatrix, Map<Long,Double> cache)	Finds the best neighbour for a CTBN node.
protected boolean[][]	findStructureNode(CTBN<? extends Node> ctbn, int idxNode, boolean[][] adjacencyMatrix)	Optimises the function score to find the parent set of a given node.
String	getIdentifier()	Returns a unique identifier for the hill climbing-based algorithm.
Map<String,String>	getParametersAlgorithm()	Returns the parameters that are used by the hill climbing implementation.

Methods inherited from class es.upm.fi.cig.multictbnc.learning.structure.optimisation.hillclimbing.implementation.CTBNHillClimbingIndividual

computeScore, findStructure, findStructure, findStructure, getInfoScoreFunction, increaseNumEdgesTested

Methods inherited from class java.lang.Object

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

Constructor Details

CTBNTabuSearchIndividual

public CTBNTabuSearchIndividual(CTBNScoreFunction scoreFunction,
 int tabuListSize,
 int maxNumNotImprovements)

Initialises the tabu search algorithm by proving a score function and a tabu list size.

Parameters:

scoreFunction - score function

tabuListSize - tabu list size

maxNumNotImprovements - maximum number of iterations to continue without improvements in the score before stopping the search

Method Details

getIdentifier

public String getIdentifier()

Description copied from interface: HillClimbingImplementation

Returns a unique identifier for the hill climbing-based algorithm.

Specified by:

getIdentifier in interface HillClimbingImplementation

Overrides:

getIdentifier in class CTBNHillClimbingIndividual

Returns:

unique identifier for the hill climbing-based algorithm

getParametersAlgorithm

public Map<String,String> getParametersAlgorithm()

Description copied from interface: HillClimbingImplementation

Returns the parameters that are used by the hill climbing implementation.

Specified by:

getParametersAlgorithm in interface HillClimbingImplementation

Overrides:

getParametersAlgorithm in class CTBNHillClimbingIndividual

Returns:

a Map with the parameters used by the algorithm

findBestNeighbor

protected HillClimbingSolution findBestNeighbor(CTBN<? extends Node> ctbn,
 int idxNode,
 boolean[][] adjacencyMatrix,
 Map<Long,Double> cache)

Description copied from class: CTBNHillClimbingIndividual

Finds the best neighbour for a CTBN node.

Overrides:

findBestNeighbor in class CTBNHillClimbingIndividual

Parameters:

ctbn - continuous-time Bayesian network that contains the node

idxNode - node index

adjacencyMatrix - current adjacency matrix

cache - cache used to avoid recomputing scores

Returns:

a HillClimbingSolution with the adjacency matrix and score of the best neighbour

findStructureNode

protected boolean[][] findStructureNode(CTBN<? extends Node> ctbn,
 int idxNode,
 boolean[][] adjacencyMatrix)

Description copied from class: CTBNHillClimbingIndividual

Optimises the function score to find the parent set of a given node.

Overrides:

findStructureNode in class CTBNHillClimbingIndividual

Parameters:

ctbn - continuous-time Bayesian network that contains the node

idxNode - index of the node

adjacencyMatrix - current adjacency matrix

Returns:

adjacency matrix that includes the best parent set of the node that was found