Automated selection, sizing, and mapping of Integrated Modular Avionics Modules

A novel automated design method for Distributed Integrated Modular Avionics (DIMA) architectures is presented herein. With the presented multi-objective method software and hardware mapping are carried out in parallel, while the optimal module types, the number of instances, and their resources are determined. Selecting, sizing, and mapping of DIMA modules in parallel takes automated DIMA design one step ahead from a single-aspect optimization to multi-aspect optimization. Two approaches are proposed for calculating the set of Pareto optimal architectures for contradicting design objectives. With optimization experiments on a multi-aircraft-system example the performance, feasibility, and solution quality of the solving techniques are compared. A large optimization potential is revealed. In addition, it is revealed what automatically designed architectures look like compared to a manually created reference design. Benefits, potentials, and limitations of the novel methodology are discussed.