wns::markovchain::MarkovContinuousTime< T > Class Template Reference

Class to represent a Markov Continuous Time Process. More...

#include <markovchain/MarkovContinuousTime.hpp>

Inheritance diagram for wns::markovchain::MarkovContinuousTime< T >:

[legend]

Collaboration diagram for wns::markovchain::MarkovContinuousTime< T >:

[legend]

List of all members.

Public Member Functions
double	calculateNextState (int chainNumber)
	Calculates the next state and the next transition time for the chain chainNumber.
virtual void	calculateStateProbabilities ()
	determine the probability of each state.
	MarkovContinuousTime (int _numberOfStates, int _numberOfChains, std::vector< T > states, TransitionMatrixType matrix, std::vector< int > startStates, simTimeType startTime, simTimeType stopTime)
	Creates a complete MarkovContinuousTime (full specification given).
	MarkovContinuousTime (int _numberOfChains)
	Creates an empty MarkovContinuousTime. the numberOfChains is required as an argument.
void	prepareMatrixOfDistributions ()
	: fills in the matrix of distributions with generators for random values
void	setStartTime (simTimeType _startTime)
	setStartTime
void	setStopTime (simTimeType _stopTime)
	setStopTime
void	setTransitionDistributionSpec (std::string distSpec)
	set a Python compatible specification string for the random number distributions used for flexible transition statistics.
void	startEvents ()
	: first prepares the matrix of distributions, then initializates all the structures needed. Sets the first Timeout
virtual void	stateChangeNotification (const int &chainNumber)
	: stateChangeNotification
std::string	stringify (double x) const
	stringify: converts a double to a string
	~MarkovContinuousTime ()
	destructor MarkovcontinuousTime
Protected Types
typedef wns::container::Matrix < wns::distribution::Distribution *, 2 >	MatrixDistType
	type for the matrix of distributions.
Protected Member Functions
virtual void	onTimeout (const int &chainNumber)
	onTimeout: changes to the new state, sets the next timeout
virtual void	stopMarkovProcess ()
	stop the process (for all chains). No more state changes will happen.
Protected Attributes
MatrixDistType	matrixDistribution
	matrix of distributions
simTimeType	startTime
	start time
simTimeType	stopTime
	stop time
std::string	transDistSpec
	: flexible specification of the transition distribution
double	transitionScale
	transition scale: value to tune up (>1) or down (<1) the rate of state changes
Friends
class	MarkovContinuousTimeTest

---

Detailed Description

template<class T>
class wns::markovchain::MarkovContinuousTime< T >

Author:

Sara Gutierrez de Mesa (sgm@comnets.rwth-aachen.de)

Rainer Schoenen (rs@comnets.rwth-aachen.de)

Definition at line 54 of file MarkovContinuousTime.hpp.

---

Member Typedef Documentation

template<class T>

typedef wns::container::Matrix<wns::distribution::Distribution*, 2> wns::markovchain::MarkovContinuousTime< T >::MatrixDistType [protected]

Definition at line 380 of file MarkovContinuousTime.hpp.

---

Constructor & Destructor Documentation

template<class T>

wns::markovchain::MarkovContinuousTime< T >::MarkovContinuousTime

(

int

_numberOfChains

)

[inline]

Definition at line 63 of file MarkovContinuousTime.hpp.

template<class T>

wns::markovchain::MarkovContinuousTime< T >::MarkovContinuousTime	(	int	_numberOfStates,
		int	_numberOfChains,
		std::vector< T >	states,
		TransitionMatrixType	matrix,
		std::vector< int >	startStates,
		simTimeType	startTime,
		simTimeType	stopTime
	)			[inline]

Definition at line 78 of file MarkovContinuousTime.hpp.

template<class T>

wns::markovchain::MarkovContinuousTime< T >::~MarkovContinuousTime

(

)

[inline]

Definition at line 99 of file MarkovContinuousTime.hpp.

---

Member Function Documentation

template<class T>

double wns::markovchain::MarkovContinuousTime< T >::calculateNextState

(

int

chainNumber

)

[inline]

Definition at line 138 of file MarkovContinuousTime.hpp.

template<class T>

virtual void wns::markovchain::MarkovContinuousTime< T >::calculateStateProbabilities

(

)

[inline, virtual]

Continuous Time Markov Chain: In theory, these equations are solved for vectorN: 1) vectorN * matrixT = 0 (you guess right, matrixT has one linear dependent row) 2) vectorN * vector1 = 1 This is performed by doing a singular value decomposition (SVD). Note that the result is only valid if the transition distribution is NegExp (Markov property). In some other cases and if p[state] is equal for all states, the results are ok, too. Especially the meanRate is Ok then.

Reimplemented from wns::markovchain::MarkovBase< T >.

Reimplemented in wns::markovchain::MarkovContinuousTimeTraffic.

Definition at line 324 of file MarkovContinuousTime.hpp.

template<class T>

virtual void wns::markovchain::MarkovContinuousTime< T >::onTimeout

(

const int &

chainNumber

)

[inline, protected, virtual]

Implements wns::events::MultipleTimeout< int >.

Definition at line 412 of file MarkovContinuousTime.hpp.

template<class T>

void wns::markovchain::MarkovContinuousTime< T >::prepareMatrixOfDistributions

(

)

[inline]

Definition at line 207 of file MarkovContinuousTime.hpp.

template<class T>

void wns::markovchain::MarkovContinuousTime< T >::setStartTime

(

simTimeType

_startTime

)

[inline]

Definition at line 108 of file MarkovContinuousTime.hpp.

template<class T>

void wns::markovchain::MarkovContinuousTime< T >::setStopTime

(

simTimeType

_stopTime

)

[inline]

Definition at line 117 of file MarkovContinuousTime.hpp.

template<class T>

void wns::markovchain::MarkovContinuousTime< T >::setTransitionDistributionSpec

(

std::string

distSpec

)

[inline]

Input: transitionDistributionSpec containing "X" as a placeholder e.g. "Pareto(1.2,X,0.0,1e100)" "Uniform(1.5*X,0.5*X)" "Norm(X,X*X)"

Definition at line 194 of file MarkovContinuousTime.hpp.

template<class T>

void wns::markovchain::MarkovContinuousTime< T >::startEvents

(

)

[inline]

Definition at line 255 of file MarkovContinuousTime.hpp.

template<class T>

virtual void wns::markovchain::MarkovContinuousTime< T >::stateChangeNotification

(

const int &

chainNumber

)

[inline, virtual]

Reimplemented in constanze::GeneratorMMPP.

Definition at line 301 of file MarkovContinuousTime.hpp.

template<class T>

virtual void wns::markovchain::MarkovContinuousTime< T >::stopMarkovProcess

(

)

[inline, protected, virtual]

Definition at line 438 of file MarkovContinuousTime.hpp.

template<class T>

std::string wns::markovchain::MarkovContinuousTime< T >::stringify

(

double

x

)

const [inline]

Definition at line 126 of file MarkovContinuousTime.hpp.

---

Friends And Related Function Documentation

template<class T>

friend class MarkovContinuousTimeTest [friend]

Definition at line 444 of file MarkovContinuousTime.hpp.

---

Member Data Documentation

template<class T>

MatrixDistType wns::markovchain::MarkovContinuousTime< T >::matrixDistribution [protected]

Definition at line 385 of file MarkovContinuousTime.hpp.

template<class T>

simTimeType wns::markovchain::MarkovContinuousTime< T >::startTime [protected]

Definition at line 396 of file MarkovContinuousTime.hpp.

template<class T>

simTimeType wns::markovchain::MarkovContinuousTime< T >::stopTime [protected]

Definition at line 401 of file MarkovContinuousTime.hpp.

template<class T>

std::string wns::markovchain::MarkovContinuousTime< T >::transDistSpec [protected]

Definition at line 406 of file MarkovContinuousTime.hpp.

template<class T>

double wns::markovchain::MarkovContinuousTime< T >::transitionScale [protected]

Definition at line 391 of file MarkovContinuousTime.hpp.

---

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

• framework/library/src/markovchain/MarkovContinuousTime.hpp