Decision support methodology for evolutionary embedded system design

Design decisions are made in an early-design phase of system development. These decisions have a big impact on the resulting system design and realization. Making design decisions in this stage is a complex and risky task, because there are a lot of uncertainties regarding their impact on the system qualities. Moreover, in many cases, trade-offs have to be made when there are conflicting objectives. This paper presents a methodology to support decision making for evolutionary systems, to decide on the most suitable design. The focus is on the embedded part of a system. Information is structured with explicit relations, such that the realization of a design can be traced towards the concerns a stakeholder has for the system. This structure enables architects and designers to understand and reason about the impact of a design with a system-wide perspective. The method consists of a calibration step which imports a current design, a design exploration step in which designs are selected, and a decision step where the most suitable design is chosen. The methodology is demonstrated by exploring and revealing the trade-offs for platooning designs from the automotive domain.