Research On Product Knowledge Management System

A product knowledge management system (PKMS) is proposed for management and reuse of product knowledge.The key technologies of representation, acquisition, retrieval, supply, publishing and reuse of product knowledge are studied. The Architecture of PKMS is discussed. A software prototype of PKMS is designed and developed. And cases are studied in an enterprise.PKMS is centered on knowledge management, aiming to provide product development for an integrated knowledge environment.Product knowledge representation model is expressed with hierarchical model including three layers: product knowledge object, task and inference. In product knowledge object layer, product knowledge is represented by Semantic Object Semantic Networks (SOSNs). In task layer, the goals of product development are decided. In inference layer, inference methods are defined by using product knowledge objects. SOSNs are based on graph and combine object-oriented techniques. In SOSNs, nodes represent product semantic objects and arcs represent the relationships among nodes. SOSNs can better express product structure, semantic relations, configuration management and constrains. SOSNs make it easy to manage all kinds of knowledge during product development process and help understand, share and transmit knowledge.Knowledge acquisition is the key to product agile development. Knowledge acquisition methods are studied. According to the characteristics of tasks and problems, knowledge acquisition methods and techniques on associated knowledge of parts, task decomposition knowledge, configuration rules, cases and design rationale are addressed.(i) DSM is built to capture the association knowledge among parts during product configuration and variant design.(ii) DSM is decomposed to analyze the relashionships among tasks. Thus task decomposition knowledge is captured during product agile development.(iii) Concept learning algorithm is used to obtain configuration rules.(iv) OLAP, knowledge reduction techniques and NNR are applied to retrieve product cases. In addition, RBF network is used to learn cases.(v) QOC method is extended to capture design rationale during CBR.An intelligent retrieval method based on semantics is presented. In the method, Ontologies are used to retrieve and infer knowledge in the knowledge semantic layer. In addition, push technologies are applied to supply knowledge. Machine learning, knowledge discovering and data mining help to push knowledge. Knowledge publishment can show the most new knowledge to users in time.Knowledge flow management based on Workflow metamodel is designed to...