Co-optimizing component allocation and activation sequence in heterogeneous 1-out-of-n standby system exposed to shocks

Motivated by real-world applications of cloud computing and unmanned aerial vehicles (UAVs), this paper models a heterogeneous 1-out-of -n standby system with n components that can be allocated to different physical positions exposed to different random shocks. The shocks cause a component to fail with certain probability, depending on the shock rate determined by the component's position and mode (standby or operation). Different components have different shock resistances. The system must perform a specified amount of work to accomplish the mission. The mission success probability (MSP) depends on both position allocation and activation sequence of the system components. We make contributions by formulating and solving the joint optimal allocation and activation sequence (AAS) problem to maximize the MSP. The solution methodology encloses a new numerical algorithm for deriving the MSP of the considered heterogeneous 1-out-of -n standby system under any given AAS, and the application of the genetic algorithm for solving the proposed optimization problem. A detailed case study of a multi-UAV standby system performing a surveillance mission is provided to demonstrate the proposed model and influences of several parameters (component performance rate, component shock resistance parameter, and shock rates) on the MSP and AAS optimization solutions.

Automated summary

Motivated by real-world applications, this paper presents a model for a heterogeneous standby system with n components, which can be allocated to different positions and are subject to random shocks. The system's mission success probability depends on the allocation and activation sequence of components. The paper proposes a joint optimal allocation and activation sequence (AAS) problem and presents a new numerical algorithm and genetic algorithm for solving it. A case study of a multi-UAV standby system is provided to illustrate the proposed model and evaluate its solutions.