Integration of learning approaches for maintenance tasks

This paper introduces the concept of a knowledge-based diagnostic tool to be applied in production. A learning module facilitates the knowledge acquisition process. To build up an initial knowledge-base, an explanation-based learning approach generates rules which are based on a functional model of the manufacturing system to be diagnosed. While the system is applied in the production process, the learning module continues to improve the system's diagnostic capabilities using analogical inference and a similarity-based learning method. The model is also employed here to control these processes.