Page 167 - DCAP405_SOFTWARE_ENGINEERING

Software Engineering

Notes continuous fashion over the length of the observation interval. It must, however, be realities
introduced by the computational process. It is simply infeasible for any practical procedure to
actually yield data at every value of time within the continuum of the observation interval.
Thus, from the perspective of the observer, state changes do apparently occur with discrete
‘jumps’ as the solution unfolds over the observation interval. Our presentation in this textbook
may give the erroneous impression that models neatly separate into the two broad categories
that we refer to as DEDS models and CTDS models. This is an oversimplification. There is, in fact
a third category of models that are usually called combined models where the name reflects the
combination of elements from both the discrete and continuous domains. As an illustration
consider the parts in a manufacturing plant that move from one workstation to another on the
way to assembly into a final product. At these workstations, queues form and the service function
provided by the workstation may have random aspects (or may become inoperative at random
points in time). Thus the basic elements of a DEDS model are present. At some workstations the
operation may involve heating the part to a high temperature in a furnace. This heating operation
and the control of it would best fall in the realm of a CTDS model. Hence the overall model that
is needed has components from the two basic domains.
Work on the development of modelling formalisms and tools for handling this third category
of combined models has a long history. The interested reader wishing to explore this topic in
more detail will find relevant discussions in Cellier, Ören and Praehofer, the initial portion of
a simulation experiment.
Modern systems that may be found in various domains like automotive, defense, medical and
communications, integrate continuous and discrete models. In a recent ITRS study covering the
domain of mixed continuous discrete systems, the conclusion is a “shortage of design skills and
productivity arising from lack of training and poor automation with needs for basic design
tools” as one of the most daunting challenges in this domain (ITRS, 2003). One of the main
difficulties in the definition of CAD tools for Continuous/Discrete (C/D) systems is due to the
heterogeneity of concepts manipulated by the discrete and the continuous components. Therefore,
in the case of validation tools, several execution semantics have to be taken in consideration in
order to perform global simulation:
In Discrete Models (DM), the time represents a global notion for the overall system and advances
discretely when passing by time stamps of events, while in Continuous Models (CM), the time
is a global variable involved in data computation and it advances by integration steps that may
be variable.
In discrete models, processes are sensitive to events while in continuous models processes are
executed at each integration step.
Currently, co-simulation is a popular validation technique for heterogeneous systems. This
technique was successfully applied for discrete systems, but very few applied it for C/D systems.
The co-simulation allows joint simulation of heterogeneous components. This requires the
elaboration of a global execution model, where the different components communicate through
a co-simulation bus via simulation interfaces performing adaptation. For C/D systems co-
simulation, the simulation interfaces have to provide efficient synchronization models in order
to cope with the heterogeneous.
This Unit presents CODIS (Continuous/Discrete Systems simulation), a co-simulation framework
for C/D systems validation. This framework assists designers in building global simulation
models. The supported simulators are Simulink for continuous models and OSCI SystemC
simulator for discrete models.

160 LOVELY PROFESSIONAL UNIVERSITY