5.5 Software Evaluation and Selection
123
5.5.2 Software Selection
In addition to the commercial simulators mentioned in the previous section, there
are many other software solutions that have been developed at research institutes
and universities but are rarely used outside these institutions. AnyLogic and MATLAB are the most widely used software systems in the simulation community for
creating hybrid simulation models. Therefore, and due to the reason, that both
simulators include options to perform simulation-based optimization, these two
simulators are described in more detail in this section. Subsequently, a simulator
is selected for the prototypical implementation of the methodology developed in
sections 5.3 and 5.4.
Many scientists and users claim that AnyLogic is still the only commercial
simulation software developed from the beginning to create hybrid models using
discrete and continuous simulation paradigms [Br+2019, p. 722; Ga+2018, p. 25;
An2019]. AnyLogic allows the use of individual simulation paradigms as well
as any combination of paradigms in one model. The software includes different options to connect discrete and continuous model parts 5 . Through the use of
state charts, the value of continuous variables can be monitored to trigger events
causing state transitions [Bo2013, p. 244; He2008, p. 51]. Furthermore, AnyLogic provides dynamic event objects, which can be used to “schedule a discrete
event based on a condition associated with the value of some model variable that
might be a continuous variable, or based on time” [He2008, p. 52]. If a condition is fulfilled, the event triggers the state change of the model as defined in
the event. AnyLogic provides a third way for the modeler to link discrete and
continuous model parts by treating continuous components as variables instead of
objects. Thus, stock and flow elements of the SD library can be addressed in the
actions coded for DES objects, in dynamic events, as well as in agent-based state
charts. In reverse, attributes of DES objects can be referenced in the mathematical
formulations of the SD variables [He2008, p. 52].
AnyLogic is controlled by an event-discrete engine but includes a numerical solver to perform differential equations to run combined simulations
(Figure 5.23). “This offers a potential solution in providing interoperability and
compatibility between system time-driven and event-driven simulations within a
single environment” [Ga+2018, p. 25].
5 At this point, only a short overview of the linking options of discrete and continuous aspects
in AnyLogic is given. A detailed description can be found in [Bo2013, pp. 241–257].
123
5.5.2 Software Selection
In addition to the commercial simulators mentioned in the previous section, there
are many other software solutions that have been developed at research institutes
and universities but are rarely used outside these institutions. AnyLogic and MATLAB are the most widely used software systems in the simulation community for
creating hybrid simulation models. Therefore, and due to the reason, that both
simulators include options to perform simulation-based optimization, these two
simulators are described in more detail in this section. Subsequently, a simulator
is selected for the prototypical implementation of the methodology developed in
sections 5.3 and 5.4.
Many scientists and users claim that AnyLogic is still the only commercial
simulation software developed from the beginning to create hybrid models using
discrete and continuous simulation paradigms [Br+2019, p. 722; Ga+2018, p. 25;
An2019]. AnyLogic allows the use of individual simulation paradigms as well
as any combination of paradigms in one model. The software includes different options to connect discrete and continuous model parts 5 . Through the use of
state charts, the value of continuous variables can be monitored to trigger events
causing state transitions [Bo2013, p. 244; He2008, p. 51]. Furthermore, AnyLogic provides dynamic event objects, which can be used to “schedule a discrete
event based on a condition associated with the value of some model variable that
might be a continuous variable, or based on time” [He2008, p. 52]. If a condition is fulfilled, the event triggers the state change of the model as defined in
the event. AnyLogic provides a third way for the modeler to link discrete and
continuous model parts by treating continuous components as variables instead of
objects. Thus, stock and flow elements of the SD library can be addressed in the
actions coded for DES objects, in dynamic events, as well as in agent-based state
charts. In reverse, attributes of DES objects can be referenced in the mathematical
formulations of the SD variables [He2008, p. 52].
AnyLogic is controlled by an event-discrete engine but includes a numerical solver to perform differential equations to run combined simulations
(Figure 5.23). “This offers a potential solution in providing interoperability and
compatibility between system time-driven and event-driven simulations within a
single environment” [Ga+2018, p. 25].
5 At this point, only a short overview of the linking options of discrete and continuous aspects
in AnyLogic is given. A detailed description can be found in [Bo2013, pp. 241–257].