Making Mission Abort Decisions for Systems Operating in Random Environment

In many practical cases, it is important to keep a system operating in hostile or dangerous environment from being destroyed even at a cost of not completing its mission. Therefore, to enhance survivability of many real-world critical systems (e.g., aircrafts and human space flight systems), mission abort decisions are often made in practice. Specifically, the mission objectives of these systems can be aborted in cases where a certain malfunction condition is met or some obstacles/hazards occur. Then a rescue or recovery procedure is initiated to enhance the system survivability. The duration and complexity of the rescue procedure can depend on the stage when the mission is aborted. When a system is experiencing both internal failures and external shocks, each shock increases its degradation. Therefore, the number of experienced shocks can be considered as an abort decision parameter. However, the abort decision can depend not only on the number of experienced shocks but also on the fraction of the mission completed by the time of the shock occurrence. To develop the abort decision rules, one has to consider two interdependent metrics: mission success probability and system survival probability. We demonstrate the tradeoff between these metrics that should be balanced by the proper choice of the mission abort policy and present the constrained abort policy optimization problems. Then we consider the case when the uncompleted mission task and the destruction of the system are associated with monetary losses and find the mission abort policy minimizing the total expected monetary losses. Illustrative examples of a mission performed by an unmanned aerial vehicle are presented.