Real-time Avionics Optimization

We report on the solution of a difficult optimization problem which arises in avionics industry. When constructing the on-board controlling-network of an airplane, the engineers need to solve a computationally highly complex problem. The goal is to assign periodic tasks to the processors on the plane and define a schedule for each processor. Current state-of-the-art approaches to tackle the problem are by far not powerful enough to solve instances of real-world size. With the help of the powerful algorithm engineering paradigm we analyzed the mathematical properties of the scheduling problem and designed sophisticated software based on the structural insights. We were able to design a model that outperformed current state-of-the-art approaches by several orders of magnitude. In particular, we could solve industrial size real-world instances to optimality. Our methods lead, for the first time, to an industrial strength tool to schedule aircraft sized instances.