Dynamics Reference

GSPNState

This holds the state for the whole system.

Declaration

class afidd::smv::GSPNState<typename Mark, typename TransitionKey, typename UserState>

The Mark is the type of the Marking class. TransitionKey is the transition key for the GSPN. The UserState may be omitted but is whatever state you would like to add to the system state. It will be accessible within transition as the UserState variable and is accessible from the state as GSPNState.user.

Member Functions

double CurrentTime()

Returns the current time in the system.

void SetTime(double time)

Sets the system time.

Members

Mark marking

The type of this is the Mark type passed in.

UserState user

This is the default-instantiated value of the UserState type.

TransitionKey last_transition

The transition key for the last transition to fire.

Stochastic Dynamics

Declaration

class afidd::smv::StochasticDynamics<typename GSPN, typename DynState, typename RandGen>

StochasticDynamics is responsible for advancing the state of the system using the given Propagators. GSPN is the type of the Petri net. DynState is the type of the state of the system. RandGen is the type of the random number generator.

Member Functions

StochasticDynamics(GSPN& gspn, PropagatorVector& pv)

The propagator vector is a type declared in this class. In its constructor, we pass the model for the system and the list of propagators which determine the next state for the system.

Initialize(State* state, RangGen* rng)

There may be internal arrays initialized at this stage once the model and state are known. This must be called before running the StochasticDynamics.

operator()(State& state)

This modifies the state, in place, according to the propagators.

Continuous Propagator

The pure virtual base class for propagators of the state. There are multiple propagators because some propagators are more efficient than others at handling subsets of the transitions enabled in the system.

Declaration

class afidd::smv::ContinuousPropagator<typename TransitionKey, typename RNG>

The arguments are the transition key of the GSPN and the random number generator type.

Pure Virtual Member Functions

bool Include(const TransitionDistribution<RNG>& distribution) const

A propagator can examine a distribution from the core matrix to decide whether it will accept it.

std::tuple<TransitionKey, double> Next(double now, RNG& rng) const = 0

What is the next predicted transition from this propagator.

std::tuple<bool, double> Enabled(const TransitionKey& tkey) const

Is this transition currently enabled?

void Enable(const TransitionKey& tkey, std::unique_ptr<TransitionDistribution<RNG>>& distribution, double when, bool previously_enabled, RNG& rng)

Enable this transition at this time.

void Disable(const TransitionKey& tkey, double when)

Disable this transition at this time.

void Fire(const TransitionKey& tkey, double when, RNG& rng)

This transition has fired.

PropagateCompetingProcesses

This subclass of ContinuousPropagator uses Gillespie’s First Reaction method ([Gillespie:1978]) to choose the next transition. It’s methods are exactly those listed above.

Declaration

class afidd::smv::PropagateCompetingProcesses<typename TransitionKey, typename RNG>

PropagateCompetingProcesses()

The constructor has no arguments.

NonHomogeneousPoissonProcesses

This subclass of ContinuousPropagator uses Anderson’s method ([Anderson:2007]) to choose the next transition. It’s methods are exactly those listed above.

Declaration

class afidd::smv::NonHomogeneousPoissonProcesses<typename TransitionKey, typename RNG>

NonHomogeneousPoissonProcesses()

The constructor has no arguments.