Optimal termination strategies considering phase redundancy

Abstract

Ensuring the survival of safety-critical systems, such as airplanes and submarines, is of utmost importance, even if it means incomplete task completion. System failures during task execution can have catastrophic consequences. Therefore, prioritizing system survival and temporarily suspending tasks are necessary measures to mitigate the risk of failure. Phase redundancy is a widely employed strategy to enhance the reliability of tasks. By incorporating task redundancy, it becomes possible to successfully accomplish the task even in the event of individual task failures. This study presents a novel condition-based task termination approach that leverages phase redundancy to strike a balance between task success and system survival. Specifically, the proposed approach dynamically terminates tasks based on specific condition information and task progress during each phase. To determine optimal termination thresholds, the study conducts iterative evaluations of task reliability and system survivability. The results demonstrate the effectiveness of the proposed approach in enhancing the probability of system survival by maximizing overall success while prioritizing system safety through active consideration of the system condition and task progress when terminating tasks.