On optimal condition based task termination policy for phased task systems

Abstract

Phased task systems commonly perform different tasks in sequence and have been widely applied in various engineering fields. Failures of phased mission systems may result in irretrievable economic losses and significant damages. It is crucial to take effective measures to enhance the survivability of phased task systems. Existing literature mainly focuses on optimizing preventive maintenance policy upon the completion of each phase to mitigate the failure risk. In addition to carrying out preventive maintenance, task termination in case of high failure risk is another intuitive way to improve system survivability. We investigate the condition based task termination and preventive maintenance policies for phased task systems with observable internal degradation. Dynamic preventive maintenance and task termination decisions are considered based on the degradation level as well as the number of completed phases. The problem is formulated as a Markov decision process to minimize the expected costs of task failure and system failure. Some structural properties are derived and several heuristic policies are compared. Case studies are presented to illustrate the proposed policies.