Cost Optimization Strategy for Iterative Integration of Multi-critical Functions in IMA and TTEthernet Architecture

Throughout the life cycle of an aircraft, its avionic architecture can be updated or re-configured leading to what we call the iterative integration problem. In particular, new avionics functions of mixed criticality must be added to computation modules. This can cause a modification of the scheduling parameters, inducing a re-certification of modules, whose cost depends on the criticality level of the modified functions. We define a new approach to help the system designer in producing a proper modified scheduling at a minimal cost while satisfying some real-time constraints. The emphasis is put on the iterative process and we focus on Integrated Modular Avionics (IMA) architecture supported by a time-triggered network. First, we give a general formalization of the problem and provide some results on the real-time constraints considered. In particular, we show how they can directly be taken into account in the variable domains. Then, our approach consists in formalizing the problem as a Binary Integer Problem. Thus, using an off-the-shelf solver, an optimal scheduling minimizing the cost can be automatically determined.