TSoSRA: A task-oriented resilience assessment framework for system-of-systems

Abstract

System-of-systems (SoS) are inevitably subject to disturbances during task execution, and resilience is a key factor for SoS damage and recovery. System-of-systems resilience (SoSR) assessment holds significant value in optimizing SoS architecture and completing tasks. However, identifying the most resilient SoS among different ones for the same task is a basic but challenging question. This paper proposes a task-oriented SoSR assessment framework (TSoSRA) to identify the most resilient SoS. First, we use a heterogeneous information network and meta-path to describe a SoS network architecture composed of different subsystems and the SoS performance model is developed, based on node and link quality. Based on this model, a system-of-system resilience metric (SoSRM) is proposed to describe the ability of the SoS to recover quickly dealing with disruption. Then, we extend further this metric to consider the impact of mitigation measures that exist in practical scenarios. Finally, we demonstrate the feasibility and effectiveness of TSoSRA through a case study, offering valuable insights to enhance network security, improve SoS reliability and design a more resilient SoS.