A study of the expression approach to expert knowledge based frame network structure

This paper presents an expression approach to expert knowledge based frame network structure model, and discusses the building principle and achievable program of the model With the frame network structure of knowledge,the author describes the entity units in geo environment,and links the expression of knowledge at different levels with the pointer so as to make it the frame network from knowledge to semantics The data structure can be used to realize an integrated description for geographic entities and their interrelations, therefore the knowledge base of the system can be operated simply and easily with the model, to guarantee the realization of the inference mechanism. Taking the building of Design and Decision Expert System of Rational Land Use (DDES) in Urumqi river basin as an example, the author expounds the expression method of frame network structure for expert knowledge, and discusses the formative structure of knowledge base in expert system, as well as the organization mode of expert knowledge in knowledge base Furthermore, the author probes design principle of inference engine in expert system, and seeks organizational means of inference rules Moreover, the author elaborates control strategies of inference engine and inference algorithms Using the above model, the author makes evaluation and decision further at the geo expert level for planning rational land use in the test region, and the user can choose the practical suggestions for comprehensive management and regional planning of the watershed by consulting the system Finally, the author discusses and appraises the experimental result Also, the expression approach to expert knowledge based on the data structure provides flexible means for the expression of geographic decision knowledge, and the control strategies of inference engine realized by the method are also applicable to the inference algorithms of a large number of data for geographic spatial information