Optimal task aborting and sequencing in time constrained multi-task multi-attempt missions

Mission abort may reduce the risk of system losses in many critical applications. Diverse aborting policies have been designed for single-attempt and multi-attempt missions performing a single task. Very little work addressed the aborting policy for multi-task multi-attempt missions with the assumption of unlimited mission time. In practice, the time for accomplishing a mission is often restrictive, which further limits the number of attempts for the task execution and thus serves as a crucial parameter for mission success. This paper makes contributions by modeling and optimizing the aborting policy for a system that must complete multiple distinct tasks within a pre-specified mission time. Each task may be attempted multiple times until its successful completion or termination due to the system failure or mission time expiration. The attempt aborting policy (AAP) is defined by two parameters (the number of shocks experienced, the system operation time threshold) and may vary from task to task. The AAP for each task as well as the execution sequence of multiple tasks may greatly impact the mission performance. Therefore, we formulate and solve an optimization problem that determines the task-dependent AAPs and task execution sequence to minimize the expected mission losses (EML). The methodology includes a new numerical procedure suggested for assessing the EML and the genetic algorithm implemented for solving the proposed EML minimization problem. We demonstrate the proposed model using an unmanned aerial vehicle that performs a five-task multi-attempt surveillance mission.