Laboratory for Analysis and Architecture of Systems
To manage the complexity of analysis introduced by the coexistence of several different formalisms or levels we will use the concept of component-based systems because it contributes to the interconnection of heterogeneous models (interdisciplinarity and interoperability), reuse and virtual prototyping. This process of simulation and in particular co-simulation will enhance the process of partitioning with the inclusion of joint constraints of "meet in the middle" (reuse, technology, levels of abstraction, ...)
Our activity stands within a general framework of system evaluation by simulation. The objective of system evaluation is to ensure compliance with the different requirements at every stage of the life-cycle. It builds on various means for a better effectiveness and sufficiency of system validation and verification. These means are: search for model intrinsic properties and mapping with system properties, simulation, and virtual (and/or real) prototyping. System simulation and virtual prototyping processes are considered, at different abstraction levels and all life-cycle stages, as design support techniques and often start much earlier than the design cycle. Within the frame of simulation formalisation and the possibility to quantify findings, a simulation product must systematically specify three interlinked elements: an experimental setting, a reference model, and the model of the system to be simulated. Several studies show the need to build complex or heterogeneous systems in composition to get the maximum benefit of all formal results that can be expected from a model or a given formalism.
This method allows propagating certain properties by construction. It is recognized that the composition enables encapsulation of information, a natural heterogeneity and availability at a given level of abstraction of relevant information.
The composition is therefore seen as a process of integration for which the simulation will allow a consistent analysis. As suggested in the issue description we break the simulation into three parts:
The creation of an enabling environment to implement the demands of the real system.
The formalization of the simulation in order to generalize and take into account all the elements proposed by each formalism (observation) and that can contribute to the improved results of the simulation.
The generation of stimuli that aims to address what the present formalisms cannot afford to deal with formally. The objective is to reduce the testing and trying to assess what the tests do not cover. Moreover, this generation may be a way to validate strategies for V V&A.
Our contrbution to virtual prototyping is to propose:
Approaches to build environements of distributed simulation or co-simulation in order to structure the sub-systems for their individual evaluation in a global context. With this approach it is possible, in one hand to keep the original models and in the other hand to create an efficient observation allowing the optimsation of the global covering of systme evaluation
Approaches of partionning (hardware, software or system). This virtual partitionning allows to take into account the reuse and the non functional constraints, and also to evaluate different solutions for the architecture.
The objective is to see the new systems understanding either physical systems or biological system as interacting systems with their own rules that may have some common basis of understanding. Systems engineering tried and succeeded to have a common basis in systems development. The short term objective is willing to study emergent properties in systems development; however the main stream of actual research is to further understand the social and biologic and physical systems and hence the eventuality to set up an institute in this context.