A Study On Knowledge Representation And Knowledge Reasoning Based On Multiple Attributes Knowledge Base Of Hybrid Knowledge

"A study on Knowledge representation and knowledge reasoning based on multiple attributes knowledge base of hybrid knowledge" is a branch of "Management Decision-Oriented Intelligent Systems Study of Isomeric Knowledge representation and knowledge Management" which is sponsored by National Natural Science Foundation of China.With the advent of knowledge economy, knowledge is playing a more and more important role. It is the same case for management decision systems, which can provide more effective decision support only when equipped with rich knowledge and strong knowledge process ability. Thus, it is an important task and new topic to advance the development and utility of knowledge, improve the intelligence level of decision and increase the effectiveness of system function. These efforts can boost the intelligentization of management decision.The intelligence of a system is embodied in its ability of acquiring, representing, processing and utilizing knowledge. The rooted problem of low-level intelligence is a lack of knowledge. The current MIS and some DSS are suffering from it. Even some IDSS only have simple knowledge of fact and rule in their knowledge bases. With the study development of IDSS and knowledge discovery, there emerge some intelligent systems based on NN and GA. They differ from the traditional ES in knowledge acquirement and representation. However, there are still problems of knowledge incompleteness and disagreement. For this reason, more attention is giving to systems with multiple attributes knowledge base. It represents knowledge in different ways, thus improves the representation ability of the system. But on the other hand, the rich of knowledge field which complexes the knowledge structure and the character of decision problem cause the more difficulty on knowledge representation.According to the former research on knowledge base and knowledge representation, knowledge on management decision problem is divided as modelized quantative knowledge, symbol knowledge and case-based sample knowledge. And in regards of management decision, especially when encountered with some complicated problems, model, symbol, case-based and sample knowledge often exist concurrently and support the analysis and solving process. They have different functions and are related and complementary to each other.Under the design principle of object oriented, this paper forms a hybrid knowledge representation combined with kinds of single knowledge representation. It uses BNF to describe the knowledge and represent the knowledge in object oriented method. On the basis of the object-oriented representation of isomeric knowledge or hybrid knowledge and under the control of reason strategy, it constructs a reason system which the meta inference engine runs as core, message sending system is contact and function inference engines work in phase. At last, the representation and reasoning of the isomeric knowledge are explained by the example of solving the management evaluation decision problems.By the object-oriented representation and integrated reason of isomeric knowledge in the papers, the intelligent level of management decision is lifted and the problem solving ability is highly promoted. And it is more important that multiple attributes knowledge base elevates the ability to use knowledge — the integrated reason and application of isomeric knowledge to support management decision. This is not only the objective of research but also the key part of the NNSF which is carrying on.