Comparing models of computation
Proceedings of the 1996 IEEE/ACM international conference on Computer-aided design
Introduction to the art and science of simulation
Proceedings of the 30th conference on Winter simulation
A UML 2.0 profile for SystemC: toward high-level SoC design
Proceedings of the 5th ACM international conference on Embedded software
Seeing errors: model driven simulation trace visualization
MODELS'12 Proceedings of the 15th international conference on Model Driven Engineering Languages and Systems
Proceedings of the 7th Workshop on Models@run.time
| Hi-index | 0.00 |
Model-based engineering is becoming a de facto paradigm for designing complex systems and software. By being executable, models are easier to understand, as well as systems they abstract. UML is the most natural choice for modeling. fUML is an executable subset of UML with precise operational semantics. From causal relations defined in a model, fUML semantics only constructs partial execution orders. When considering fUML in a simulation process, this limitation is possibly an issue for engineers to observe various execution schemes and to have a representative execution of a system model. Usually simulation frameworks control the execution of an application thanks to a dedicated entity. This latter is responsible for the construction of the execution order conforming to the semantics of a specific Model of Computation (MoC). Furthermore it can be used to reflect extra-functional aspects like time. In order to overcome these limitations, this paper proposes principles to use fUML as a simulation environment. We propose to extract the execution control policy from a fUML model and to delegate it to a specific simulation library defining MoCs as fUML models. This library is responsible for controlling the execution and simulating extra-functional aspects. This approach provides the required flexibility and openness needed to support various applications domains. The solution is evaluated on a simple, but representative example.