Information Technology Reference
In-Depth Information
other hand, if after winning the contract, it is discovered in later design phases that
the selected components are actually not reusable, their development from scratch
may result in a project loss that can challenge the SME's economic viability. Thus,
a careful investigation has to take place quickly.
To summarize the SME and control system related challenges:
•
Control system development is a customer- and project-oriented business.
Although all engines are somehow similar on an abstract level, they differ in
detail, mainly due to hardware issues [
7]
. This precludes long-term planning of
product families due to the individuality of the developed solutions.
•
There is a high frequency of innovations. The knowledge changes and grows
quite fast. With each new development project, new engine components, sensors,
actuators, and construction styles may arise.
•
The offer must be developed within a short time frame dictated by the customer.
The SME thus must fast and reliably identify reusable components from earlier
projects without being able to build on a full product line approach.
•
Control systems are realized in software but base on physical features of
the controlled system. Its development therefore requires the interdisciplinary
investigation of combined software and control requirements and solutions.
•
The benefits of model-based approaches during later development stages in both
disciplines especially in regard to quality assurance, reliability, and safety [
37]
need to be taken advantage of.
3 Domain Model Based RE for Control Systems
In this section, our agent and goal-oriented approach to domain model based
requirements engineering is introduced in response to the challenges above. The
usage of the approach is exemplified for the domain of control systems.
3.1 Agent- and Goal-Oriented Requirements Engineering
The foundation of our approach is a suitable common modeling notation that tack-
les the interdisciplinary issues. Goal-oriented approaches have proven to be suitable
here. “Goals have long been recognized to be the essential components involved in
the requirements engineering (RE) process” [
43]
. Accordingly, they are also suitable
as a common ground for different disciplines. A second suitable concept is that of an
“agent” [
46]
. Disregarding any major fundamental distinctions at the detailed level,
any stakeholder, environment, legacy, or system-to-be-developed component (irre-
spective whether physically, hardware, or software motivated) can be represented
by an agent and thus be related to other components. Again, this simple concept
is amenable for any discipline and suited to prepare a common understanding of a
problem.
Search WWH ::
Custom Search