Abundant research efforts were devoted to modeling and optimizing mission systems operating in random shock environments. The existing models assumed that different components are exposed to independent shock processes. However, in many real-world applications (e.g., virtual machines in cloud computing, drones deployed in the same mission), multiple operating components may be simultaneously impacted by common shock processes during the mission, causing them to deteriorate and even crash. This paper contributes by modeling a multi-attempt mission system with heterogeneous components characterized by different performance, shock resistance and cost. Each component may start performing the mission at different times and a common shock process can negatively affect all operating components. Different activation schedules of system components may lead to dramatically different mission success probabilities and expected component losses, contributing to the expected mission losses (EML). We formulate and solve a new optimal component activation schedule (CAS) problem to minimize the EML. The proposed model is demonstrated through a case study of an aerial vehicles delivery mission system. Influences of several key model parameters (shock rate, mission failure penalty, allowed mission time, and component performance) on the mission performance metrics and the optimal CAS solutions are also investigated using the case study.
Read full abstract