During the design phase of avionics systems, the activity of selecting the physical components of an IMA platform and allocating the avionics system's applications into the chosen components may lead to several options to configure the given IMA architecture. The aeronautical industry already realized that this activity can be assisted by automated methods. Focusing on the communication allocation, this paper presents an approach for allocating application's data flows into physical communication resources of an IMA system able to deal with the main requirements, such as safety and performance, and also able to be extended to other criteria, as for instance maintainability and extensibility requirements. The approach is based on the Boolean satisfiability theory and linear programming. A case study is presented to evaluate its applicability.