Information Technology Reference
In-Depth Information
"
#"
!
!
Fig. 10.
Table a) compares the B
OCHICA
core DSL and the Jadex base transformation with the
ISReal-specific extensions. b) compares the generated code for the SmartFactory case study to
the manually written
Lines of Code
(LOC).
into the ISReal platform, the LSE, sensor interfaces, different service execution seman-
tics, etc.
The ISReal-enabled B
OCHICA
framework has been evaluated in a virtual production
line scenario. For this purpose, a virtual representation of the DFKI SmartFactory
12
was semantically annotated. The SmartFactory is a real existing machine of the DFKI
to evaluate technology for the factory of the future. The B
OCHICA
ISReal extension has
been used to model agents that perform typical workflows such as handling new orders
and fixing problems as they occur. Figure 9 depicts an overview of the modeled ISReal
agents. The current model encompasses basic navigation and object interaction (e.g.
for operating the machine). In order to estimate the effort for the ISReal customization,
Figure 10 depicts two tables. Table a) compares (i) the number of concepts, (ii) the
number of transformation rules, and (iii) the
Lines of Code
(LOC) of the model trans-
formation between B
OCHICA
and the ISReal extension. Table b) compares the gener-
ated code with the manually written code for the SmartFactory scenario. The manual
code mainly implements business logic (e.g. computing the shortest path between two
objects). The ISReal extension required around 10% new concepts and mapping rules.
The result is a modeling environment which addresses the needs of one specific ap-
plication domain. B
OCHICA
is especially suited for large and medium size application
domains and target environments with many end-users (e.g. the ISReal platform) where
customizations pay off. Small applications can be realized with the functionality pro-
vided by the core modeling language and the base transformations (similar to existing
approaches). Probably the biggest obstacle of the approach is the required additional
knowledge of the B
OCHICA
core DSL and MDSD.
6
Related Work
During the recent years, several agent-oriented modeling languages have been pro-
posed. The majority of the modeling languages were created in order to support a
certain agent methodology [1] [9]. One problem of existing methodology-oriented mod-
eling approaches that we see is that they do not clearly distinguish between (i) the agent
platform, (ii) the methodology, and (iii) the modeling language. Two indicators which
support our perception are (i) the development of the modeling languages is not de-
coupled from the methodologies and (ii) none of the languages has an own name (only
the tools have names). However, we think the development of modeling languages is
