Abstract : The avionic domain has seen a profound evolution by the introduction of Integrated Modular Avionics (IMA). This defines a standardized execution and communication support in order to reduce the complexity of the physical architecture. Nevertheless, due to the sharing of resources, this reduction of complexity is opposed by an increased difficulty in application conception and integration, which necessitates dedicated tools for assisting system designers. This thesis' contributions concern two major resource allocation problems: i) the multiprocessor scheduling of strictly periodic tasks and ii) the routing of messages exchanged between the avionic functions. The first problem was formulated using integer linear programming so as to guarantee a maximum evolution potential for the task execution durations. The inefficiency of this exact approach for large problem instances led us to develop an original heuristic, inspired from Game Theory, and further enhance it with a multi-start algorithm. The routing problem was formulated as an optimization one so as to minimize the maximum link loads. Two methods were proposed for this purpose, the first is exact based on node-link formulations, and the other is a two phase heuristic based on link-path formulations.