Abstract
•Components that can reduce their tasks during the mission are considered.•Joint task reduction and abortion policy is introduced.•Algorithm for estimating the expected mission losses is developed.•Optimal policy minimizing expected mission losses is found.•The interplay between task reduction and task aborting policies is analyzed.
Existing aborting models typically assume “all or nothing”, meaning a system component either completes its required operation, contributing fully to the mission requirement, or contributes nothing due to failure or mission abort. Motivated by practical applications like scientific computing and payload delivery, which allow task reduction in the event of deterioration, this paper pioneers the joint modeling and optimization of task reduction and abort policies (TRAP) to minimize expected mission losses (EML). Both policies depend on the number of shocks occurring during the task execution and on the operation time elapsed from the beginning of the mission. Different components may have varying TRAPs. A universal generating function-based approach is proposed for assessing the EML of the considered multi-component system. The genetic algorithm is further implemented to determine the optimal TRAP that minimizes the EML. A delivery mission performed by multiple unmanned aerial vehicles is analyzed under homogeneous and heterogeneous scenarios to demonstrate the proposed model. The case study also investigates influences of key parameters, including mission demand, per unit score deficiency penalty, shock rates, shock resistance parameters, and component loss cost, on the EML and the optimal solutions. The interplay between task reduction and task aborting policies is also illustrated.