Reasoning Efficiency Research of Expert System for Biomass Soft-Sensor Modeling in Fermentation Process  

作　　者：安莉 王建林 

机构地区：[1]Department of Information Science and Technology,Beijing University of Chemical Technology

出　　处：《Journal of Donghua University(English Edition)》2011年第1期40-44,共5页东华大学学报（英文版）

基　　金：National Natural Science Foundation of China (No. 20476007,No. 20676013)

摘　　要：An expert system for biomass soft-sensor hybrid modeling in fermentation process was decribed in this paper.A production rules representation based on database was presented.The definitions of production rules for biomass soft-sensor hybrid modeling knowledge were proposed.A knowledge base with layered structure was introduced.A breadth-first reasoning approach based on match degree(BFMD) was developed.The definition and calculation method of match degree were illustrated.Compared with the depth-first reasoning approach based on exhaustive method(DFEM),the BFMD needs fewer introduced variables.This expert system could reduce the reasoning steps effectively,and advance reasoning efficiency.Tests shows that reasoning efficiency of the expert system using BFMD in the knowledge base with layered structure is improved 12.9% averagely,compared with using DFEM in the knowledge base with ranking structure.An expert system for biomass soft-sensor hybrid modeling in fermentation process was decribed in this paper.A production rules representation based on database was presented.The definitions of production rules for biomass soft-sensor hybrid modeling knowledge were proposed.A knowledge base with layered structure was introduced.A breadth-first reasoning approach based on match degree（BFMD） was developed.The definition and calculation method of match degree were illustrated.Compared with the depth-first reasoning approach based on exhaustive method（DFEM）,the BFMD needs fewer introduced variables.This expert system could reduce the reasoning steps effectively,and advance reasoning efficiency.Tests shows that reasoning efficiency of the expert system using BFMD in the knowledge base with layered structure is improved 12.9% averagely,compared with using DFEM in the knowledge base with ranking structure.

关 键 词：expert system production rules layered structure match degree 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]