Research On The Network Manufacturing Resources Oriented Multilevel Intelligent Retrieval System

Network manufacturing resources (NMR) search is an important component of network manufacturing system and an area-oriented manufacturing resources retrieval method. The aim of NMR retrieval is to help enterprises find manufacturing resources fast and accurately through computer information retrieval (IR) system. Retrieving and utilizing these NMR efficiently is part of realizing open manufacturing resources model. This paper narrates basis model and characters of NMR and analyzes design principles of user interface, information collecting subsystem, information feedback subsystem and information push-pull subsystem in detail. An ASP-based NMR database is established to classify, describe and store manufacturing resources according to time, quality, cost, service, agility, advance and historical record-based seven guideline. Based on these researches we adopt user model, machine learning, webpage distinguish, information filter, data mining and man-machine interaction etc. preference mining is a key technique in information push-pull subsystem. This paper first comes up with a method of applying user preference-based association rules mining technique to NMR retrieval system. The method of GA-based user preference mining is proposed and user's preference is set as the consequent of the association rule. A case-based reasoning (CBR) method is proposed to search dynamic league efficiently, and a model of partner choice subsystem is established. The way of establishing case and its evaluation value databases is discussed. A method of calculating the similarity degree based on grey-relational theory and fuzzy-set theory is put forward. Thus, similar cases can be found out accurately, and efficiency and accuracy of choice partner enterprises are promoted. Otherwise, information extracting, information preprocessing technique, inquiry interface, and information filter technique based on naive bayes is put forward. NMR multilevel intelligent retrieval system has greatly improved NMR ferret's efficiency and accuracy.