edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom Class Reference

Inheritance diagram for edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom:

[legend]

List of all members.

Public Member Functions
	MaxWalkSATRoom (WeightedClausalKB kb, PossibleWorld state, WorldVariables vars, Database evidence) throws Exception
void	printBestState ()
void	setState ()
Public Attributes
boolean	alternativeMethod = true
Protected Member Functions
void	SAMove ()
void	walkSATMove ()

---

Detailed Description

This class ia an extension of the MAPMaxWalkSAT-class in which some methods are fitted to our special facility management problem.

Author:

wernickr

Definition at line 15 of file MaxWalkSATRoom.java.

---

Constructor & Destructor Documentation

edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom::MaxWalkSATRoom	(	WeightedClausalKB	kb,
		PossibleWorld	state,
		WorldVariables	vars,
		Database	evidence
	)			throws Exception [inline]

Constructor to instantiate a MAPMaxWalkSATRoom.

Parameters:

	kb	a knowledge base of weighted clauses, which can be instantiated by an Markov Random Field (MRF)
	state	a possible world containing all variables of the Markov Logic Network (MLN)
	vars	world variables of the MLN which is needed by the class
	evidence	the database of the MRF -> given contraints and true/false values of the variables
	sw	an externally instantiated and started Instance of type stopwatch

Exceptions:

java.lang.Exception

Definition at line 31 of file MaxWalkSATRoom.java.

---

Member Function Documentation

void edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom::printBestState

(

)

[inline]

Prints all solutions and necessary informations into a file set at the constructor.

Definition at line 65 of file MaxWalkSATRoom.java.

void edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom::SAMove

(

)

[inline, protected]

Randomly chooses and flips a possible groundatom

Reimplemented from edu::tum::cs::logic::sat::weighted::MaxWalkSAT.

Definition at line 82 of file MaxWalkSATRoom.java.

void edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom::setState

(

)

[inline]

Sets the initial state of the possible world so that the algorithm starts the search at the same positions the staffmembers are sitting before. (nachher = vorher)

Reimplemented from edu::tum::cs::logic::sat::weighted::MaxWalkSAT.

Definition at line 39 of file MaxWalkSATRoom.java.

void edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom::walkSATMove

(

)

[inline, protected]

Jumps to randomly chosen position in the list of unsatisfied constraints. Tries incremental to find an actually unsatisfied constraint which enhances the state of the "world". If no one is found the algorithm executes a SAMove (random flip) with a possibility of 100% (optional value possible, see lines 240 - 243)

Reimplemented from edu::tum::cs::logic::sat::weighted::MaxWalkSAT.

Definition at line 94 of file MaxWalkSATRoom.java.

---

Member Data Documentation

boolean edu::tum::cs::logic::sat::weighted::MaxWalkSATRoom::alternativeMethod = true

whether to use an alternative method that does not actually use the p parameter (Ralf's method)

Definition at line 20 of file MaxWalkSATRoom.java.

---

The documentation for this class was generated from the following file:

• MaxWalkSATRoom.java