Maintenance strategy selection has been an important link within the management system for airport rigid pavement. Suitable maintenance strategies will guarantee maintenance success. The maintenance personnel investigating pavement distress usually base their maintenance strategy suggestions on expert knowledge and experience. Nevertheless, in practical situations we often see examples of distress taking place following maintenance, which indicates that the maintenance strategies previously selected can not effectively repair the pavement. Other methods and materials have to be adopted for maintenance so as to ensure pavement performance. This paper uses the machine learning theory of the neural network, conducting a questionnaire survey on experts to accumulate knowledge by way of case learning and inferring. This is aimed at enhancing capability of selecting more appropriate strategies for repair of pavement distress, thus to ensure pavement service performance. Furthermore, this paper suggests that feed-back and learning capabilities of maintenance materials be established, allowing the system to learn continuously, ensuring applicability of maintenance materials as well as upgrading maintenance effects.

Included in

Business Commons