Study And Application Of Case-based Reasoning Used In The Product Data Management System

With the advent of the era of knowledge economy, product data management (PDM) system based on data management has been far from meeting the needs of enterprises. Thereby there need intelligent knowledge management system urgently which can capture, organize, save, share, reuse and reason knowledge to support intelligent design and product design reuse. Case-based reasoning (CBR) is a good technology that can manage knowledge, achieve intelligent design and design reuse. But now the key technologies in CBR have not been solved satisfactorily. At present these key technologies are on the hot spots. PDM system provides a good platform for the realization of CBR. Therefore, CBR will be applied to PDM system in this thesis, so that it will achieve complementarity between CBR and PDM system. This article focused on the PDM system model that supports CBR, Design Cases Model, Similar design cases retrieval, and other key issues. PDM system that supports CBR is realized. This system can analyze customer needs, be able to carry out effective knowledge management, be able to support intelligent design and product design reuse.First of all, QFD&CBR Design Process Model is putted forward based on Quality Function Deployment (QFD) and CBR. Then the PDM system model that supports CBR is built based on QFD&CBR Design Process Model. This system model realized organic integration between CBR and PDM system.Secondly, Study on the key technologies of the PDM system model that supports CBR, such as Design Case Representation, New Design Case Capture, Similarity Measure and Design Case Retrieval, Design Case Preservation: Object-oriented design case representation integrated with PDM information management model is putted forward. Complex design cases with complex structure can be expressed from different angles well and perfectly through this approach; A method of design case knowledge acquisition based on QFD is proposed. By this method, the designer's knowledge can be switched to computer's internal representation; a hybrid similarity algorithm based on QFD is putted forward. This method not only can calculate the similarity between the design cases accurately and efficiently when design cases has different features and various types of features values, but also consider the relationship between each other's characteristics of design cases and the relationship between characteristics of design cases and customer needs, so that the similarity calculation results can be more reasonable and accurate; The similar design cases retrieval algorithm based on adding clustering algorithm is proposed. The retrieval algorithm has some advantages, such as retrieve fast and efficiently, to be convenient for conservation and management of design cases and so on; Organizational structure model of design cases based on Shared-feature Networks is putted forward. This structure model not only can continue to accumulate knowledge, but also can lay the foundation for achieving design cases automatic classification and retrieving similar design cases.Finally, PDM system that supports CBR be realized by secondary development on Teamcenter Engineering. This system proves correctness, feasibility and practicability of models and technologies which are proposed before.