Expert System Technology in Aircraft Fleet Management – The Development of SAREX

Abstract

Aircraft check planning and flight assignment are closely interrelated. We describe an expert system supporting both tasks. Search and Constraint techniques from the realm of Artificial Intelligence are paired with Scheduling methods based on certain mathematical models. The objective of Check Planning is to optimise the Maintenance Check Yield across the whole fleet. The problem is how to accommodate the frequently changing user requirements in an expert system environment. Those requirements can be expressed as constraints regarding e.g. fleet availability, hangar space, work load, or time restrictions. In contrast to classical constraint technology, where these problems are solved by instantiating variables and checking consistencies, we make use of a direct link of constraints into scheduling algorithms. As a main advantage, the basic scheduling algorithm could be left unaltered while new constraints were permanently added during the development process. For most of the constraints, SAREX offers dedicated interfaces to edit these planning rules in an easily readable format. This paper gives an overview of the applied technology and of the history of its implementation. For more technical details the reader is referred to [3].