jmarkov.solvers
Class MtjSolver

java.lang.Object
  jmarkov.solvers.Solver
      jmarkov.solvers.SteadyStateSolver
          jmarkov.solvers.MtjSolver

All Implemented Interfaces:

JMarkovElement

public class MtjSolver
extends SteadyStateSolver

This class uses MTJ to solve Steady State probabilities.

Author:

Germ�n Ria�o. Universidad de los Andes.

Nested Class Summary

static class	MtjSolver.EnumPrecond This is the list of preconditioner offered in MTJ.
static class	MtjSolver.EnumSolver This is the list of solvers provided by MTJ.

Field Summary

Fields inherited from class jmarkov.solvers.Solver

mp

Constructor Summary

MtjSolver(MarkovProcess mp)
Default constructor.

MtjSolver(MarkovProcess mp, MtjSolver.EnumSolver solver)
Construct a solver for the given SimpleMarkovProcess.

MtjSolver(MarkovProcess mp, MtjSolver.EnumSolver solver, boolean tryOthers)
Construct a solver for the given SimpleMarkovProcess.

Method Summary

java.lang.String	description() This method return a complete verbal describtion of this element.
MtjSolver.EnumSolver	getCurrentIterSolver()
MtjSolver.EnumPrecond	getCurrentPreConditioner()
no.uib.cipr.matrix.Matrix	getGenerator() Returns the Generator matrix.
no.uib.cipr.matrix.Matrix	getGenMatrix()
no.uib.cipr.matrix.sparse.IterativeSolver	getIteartiveSolver(no.uib.cipr.matrix.Vector pi)
no.uib.cipr.matrix.sparse.IterativeSolver	getIterativeSolver(no.uib.cipr.matrix.Vector pi0, MtjSolver.EnumSolver solver)
long	getProcessTime()
double[]	getSteadyState() This process should be extended in order to compute the steady State probabilities of the MarkovChain.
boolean	isTryOthers()
java.lang.String	label() The name of this solver.
void	setCurrentIterSolver(MtjSolver.EnumSolver iterSolver) Sets the solver to use.
void	setCurrentPreConditioner(MtjSolver.EnumPrecond preConditioner)
void	setIterSolver(MtjSolver.EnumSolver iterSolver, boolean tryOthers) Sets the solver to use.
void	setTryOthers(boolean tryOthers) Sets whether the solver shall try other solvers when it fails.

Methods inherited from class jmarkov.solvers.Solver

getMP, toString

Methods inherited from class java.lang.Object

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

Methods inherited from interface jmarkov.basic.JMarkovElement

equals

Constructor Detail

MtjSolver

public MtjSolver(MarkovProcess mp)

Default constructor. Uses as default solver BiCGStab first. If it fails, it tries other solvers.

Parameters:

mp -

MtjSolver

public MtjSolver(MarkovProcess mp,
                 MtjSolver.EnumSolver solver)

Construct a solver for the given SimpleMarkovProcess. It uses the given solver enumeration to determine what solver to use.

Parameters:

mp - Tha Markov process solved

solver - The solver to use.

MtjSolver

public MtjSolver(MarkovProcess mp,
                 MtjSolver.EnumSolver solver,
                 boolean tryOthers)

Construct a solver for the given SimpleMarkovProcess. It uses the given solver enumeration to determine what solver to use. If it fails it tries other solvers.

Parameters:

mp - the Markov Process to solve.

solver - the iterative solver form the enumeration IterSolver.

tryOthers - whether a different solver should be tried if the first one fails.

Method Detail

getGenerator

public no.uib.cipr.matrix.Matrix getGenerator()

Returns the Generator matrix.

Returns:

the Generator Matrix G.

getSteadyState

public double[] getSteadyState()
                        throws NotUnichainException

Description copied from class: SteadyStateSolver

This process should be extended in order to compute the steady State probabilities of the MarkovChain. The user can get information of the SimpleMarkovProcess associated with this solver though the methods getRates(), getGenerator, and getRate(State,State)

Specified by:

getSteadyState in class SteadyStateSolver

Returns:

an array with the Steady state probabilities for the given problem.

Throws:

NotUnichainException

See Also:

MarkovProcess.getGenerator(), MarkovProcess.getRates(), MarkovProcess.getRate(State, State)

label

public java.lang.String label()

Description copied from class: Solver

The name of this solver. This should be implemented by the extending classes..

Specified by:

label in interface JMarkovElement

Specified by:

label in class Solver

Returns:

A String label.

See Also:

Solver.label()

getCurrentIterSolver

public MtjSolver.EnumSolver getCurrentIterSolver()

Returns:

Returns the currentIterSolver.

setIterSolver

public void setIterSolver(MtjSolver.EnumSolver iterSolver,
                          boolean tryOthers)

Sets the solver to use.

Parameters:

iterSolver - The currentIterSolver to set.

tryOthers - whether other solvers should be tryed if this fails.

setCurrentIterSolver

public void setCurrentIterSolver(MtjSolver.EnumSolver iterSolver)

Sets the solver to use. It will not try other solvers if this one fails.

Parameters:

iterSolver -

getCurrentPreConditioner

public MtjSolver.EnumPrecond getCurrentPreConditioner()

Returns:

Returns the currentPreConditioner.

setCurrentPreConditioner

public void setCurrentPreConditioner(MtjSolver.EnumPrecond preConditioner)

Parameters:

preConditioner - The currentPreConditioner to set.

isTryOthers

public final boolean isTryOthers()

Returns:

Returns true if the solver shall try other solvers when it fails.

setTryOthers

public final void setTryOthers(boolean tryOthers)

Sets whether the solver shall try other solvers when it fails.

Parameters:

tryOthers - true if the solver shall try other solvers when it fails.

getGenMatrix

public final no.uib.cipr.matrix.Matrix getGenMatrix()

Returns:

Returns the generator Matrix.

getProcessTime

public final long getProcessTime()

Returns:

Returns the Process Time of the last solved problem.

getIteartiveSolver

public no.uib.cipr.matrix.sparse.IterativeSolver getIteartiveSolver(no.uib.cipr.matrix.Vector pi)

Parameters:

pi - The probability vector.

Returns:

Returns the iteartiveSolver.

getIterativeSolver

public no.uib.cipr.matrix.sparse.IterativeSolver getIterativeSolver(no.uib.cipr.matrix.Vector pi0,
                                                                    MtjSolver.EnumSolver solver)

Parameters:

pi0 - Initial guess value.

solver - The solver used.

Returns:

Returns the iterativeSolver.

description

public java.lang.String description()

Description copied from interface: JMarkovElement

This method return a complete verbal describtion of this element. This description may contain multiple text rows.

Returns:

A String describing this element.

See Also:

JMarkovElement.label()