Redundancy allocation in automotive systems using multi-objective optimisation

Redundancy allocation is a widely used method for the reliability improvement of complex embedded systems in the automotive domain. The allocation of redundant components can affect other non-functional quality attributes of the system, which are as important as reliability and very often conflicting with each other, such as cost and response time. Consequently, to find the good system design options, multi-objective optimisation methods can be employed which find a trade-off between these conflicting attributes. While cost and reliability have been already studied in the context of an optimisation problem, the impact of redundancy allocation to the response time has not yet been considered. The approach taken in this paper, considers reliability, cost and response time as optimisation criteria and employs a multi-objective Ant Colony Optimisation algorithm to find the near optimal set of system designs. Our approach has been implemented in the ArcheOpterix framework and illustrated with a case study from the automotive industry