Due to the complexity of today's embedded systems and time-to-market competition between companies developing embedded systems, system architects have to perform a complex task. To design a system which meets all its quality requirements becomes increasingly difficult because of customer demand for new innovative user functions. Methods and tools are needed to assist the architect during system design. The goal of this paper is to show how metaheuristic optimization approaches can improve the process of designing efficient architectures for a set of given quality attributes. A case study is conducted in which an architecture optimization framework is applied to an existing sub-system in the automotive industry. The case study shows that metaheuristic optimization approaches can find efficient solutions for all quality attributes while fulfilling given constraints. By optimizing multiple quality attributes the framework proposes revolutionary architecture solutions in contrast to human architects, who tend to propose solutions based on previous architectures. Although the case study shows savings in manual effort, it also shows that the proposed architecture solutions should be assessed by the human architect. So, the paper demonstrates how an architecture optimization framework complements the domain knowledge and experience of the architect.