Glossary
- compartments
- The possible dynamical states in a compartmented model of disease.
- compartmented model of disease
- A disease model that represents the progression of a disease as
a set of discrete compartments with transitions possible
between them. Transitions typically occur with some base
probability, which might be fixed or might vary across the
course of the simulation. See Hethcote for a survey,
- continuous time
- A simulation mode in which events occur at unique times
represented by real numbers. No two events ever happen
simultaneously, but they can be separated by an arbitrarily
small interval. Continuous-time simulations can be made
statistically exact and run faster for situations in which there
are long periods where no events occur.
- discrete time
- A simulation mode in which time progresses in single integer
timesteps. During each timestep a collection of events can
occur. Discrete-time simulations can be easier to code and
understand.
- dynamical state
- The state of a node or edge at some point in the
simulation. These typically reflect the compartments
of the simulation, but may be more complex and
comprise a vector of information.
- event
- A simulation event that changes the state of the underlying
network or simulation. Events can occur in continuous
time or discrete time.
- event function
- A function called when an event fires to perform the
action required. Event functions take four arguments: the
Dynamics object running the simulation, the
current simulation time, the network, and the element at which
the event occurs (which will be selected by the chosen
process dynamics). Elements are typically either nodes
or edges, depending in the locus at which the event
occurs.
- locus
- A “place” at which dynamics can occur, that is to say, where
nodes can change compartments and any other tasks can happen.
Each event is associated with a particular locus: the
locus contains the set of nodes or edges to which the event may
be applied, while the event defines chat happens.
- posted event
- An event posted for a definite future time. The
process dynamics will execute the posted events at the
appropriate time
- process dynamics
- The simulation approach used, which selects how and when each
event fires. Process dynamics execute events in time
order from two possible sources: a random distribution that
chooses an event based on their relative probability or rate; and
any posted event that has been scheduled.
- stochastic process
- A process whose exact progression is determined by random
variables drawn from particular probability distributions.
- stochastic dynamics
- Also known as Gillespie dynamics, this process dynamics operates
in continuous time with one event occurring at each time
point.
- synchronous dynamics
- A process dynamics using discrete time, where a
simulation passes through a sequence of discrete timesteps which
may include several (or no) events happening.