The abstract representation of a continuous phenomenon in a simulation model requires that all events be presented in discrete terms.
With some confusion in the notation, discrete-event computer simulations can be cast either in discrete time or in continuous time.
With discrete time, time is broken into regular (equi-spaced) time slices (∆t) and the simulator calculates the variation of state variables for all the elements of the simulated model between one point in time and the next. Nothing is known about the order of the events that happen within each time period: discrete events (marriage, job loss, etc.) could have happened at any moment in∆t while inherently continuous events (aging, wealth accumulation, etc.) are best thought to progress linearly between one point in time and the next.
By contrast, simulations cast in continuous time are characterized by irregular timeframes that are punctuated by the occurrence of discrete events. Between consecutive events, no change in the system is assumed to occur; thus the simulation can directly jump in time from one event to the next. Inherently continuous events must be discretized.
The event list order the events and the simulation is performed by extracting the event that is closest in time and submitting it to the model’s agents, which change their state according to the signal (corresponding to the event) they have received. In the case of continuous simulations, the order of the processes that are applied must be exogenously assumed (and the assumption must be coherent with the specification of the model used for estimating the coefficients governing each process). The events may be generated and scheduled also while running the simulation and not only in the initial planning phase.