Guiding Genetic Algorithms using importance measures for reliable design of embedded systems

Reliability importance measures (IMs) support analysts in understanding the contributions of components to the reliability of the system under investigation. This understanding can be of use to improve the reliability of a system and at the same time, restrict the cost penalty by upgrading only the highly important components to more reliable ones. This paper studies how IMs can enhance the design of embedded systems, more specifically to guide the optimization process. The observations are later employed to modify a well-known Genetic Algorithm (GA) to create new offsprings using the IMs of the components of their parents. The experimental results prove the efficiency of the proposed algorithm which not only seeks for more reliable designs, but also reckons with other design objectives-in this paper resource cost and power consumption-concurrently to ensure that they are not degraded through the optimization process.