A Satisficing MOMIP Framework for Reliable Real-time Application Scheduling

In a networked embedded system a real-time application is distributed and scheduled over a set of nodes, which are connected by means of heterogeneous networks. The application designer for such a system is often confronted with the challenge of providing better response time and low failure probability for the distributed application. These objectives are conflicting with each other hence some trade-off between these two design objectives is needed to ensure a satisfactory application schedule. More often the designer's satisfaction can be expressed as a set of aspiration bounds on the design objectives, that if attained will be a good schedule. This paper proposes a satisfying multiobjective mixed integer programming (MOMIP) framework for deriving such a schedule