Research And Realization Of Expert System For Intelligent Manufacturing Based On Uncertainty And Fuzzy Reasoning

Expert system is defined as the intelligent computer program system which can solve practical problems within a particular field with a large quantity of professional expertise like human experts. At present, China's manufacturing industry is gaining rapid development in intelligence and information technology, and there's a lack in technical personnel with extensive knowledge and experience. In order to solve this problem, developing expert system for intelligent manufacturing industry is imperative, which will strongly promote China's intelligent manufacturing and the whole machinery industry. However, there still exist quite a few problems in the development of intelligent manufacturing expert systems in both theory and practice, such as the lack of deeply study on uncertainty reasoning and other key issues of expert systems. Therefore, this subject aimed to study the theory and methods of developing intelligent manufacturing expert systems, and finally develop the prototype software of intelligent manufacturing expert system, which provided theoretical and methodological basis and software framework for a mature intelligent manufacturing expert system.As for knowledge acquisition and knowledge representation which is the pilot work of developing expert systems, knowledge acquisition method of knowledge list was designed, production rules were used for knowledge representation, SQL Server(DBMS) was used for organization, storage and management of knowledge, and finally the rule base of the expert system was established.Firstly, the model of the evidence and conclusion uncertainty reasoning was established and the related inference engine algorithm was designed. In this issue to resolve the problem of uncertainty in complex knowledge of intelligent manufacturing industry, the reasoning model of uncertainty with threshold and the reasoning model of uncertainty with evidence weighting were established, the synthesis algorithm of the uncertainty of combined evidence and the algorithm of propagating from evidence uncertainty to conclusion credibility was designed, on the basis of CF method. The established model effectively simulated the actual process of expert analyzing and reasoning with uncertain evidence and conclusions, and quantified the uncertainty in a standardized form.Secondly, the model of fuzzy reasoning was established and the related fuzzy inference engine algorithm was designed. In this issue aiming to describe numerical MIMO variables and related knowledge, and to perform numerical reasoning and decisions in manufacturing industry, fuzzy sets and fuzzy theory were applied to the expert system, the model of fuzzy rules and corresponding membership functions were established, fuzzy relationship, synthetic operator, fuzzification of a single point and defuzzification of COG were designed, the fuzzy inference engine algorithm was proposed, and the fuzzy inference system was eventually constructed. The constructed rule-based fuzzy inference system is suitable for the numerical, imprecise reasoning with vague language and uncertainty.Finally, on VC++ platform, the inference engine, rule base, HMI, and the algorithms of uncertainty reasoning and fuzzy reasoning of the expert system were implemented, which was based on the above-mentioned model of the evidence and conclusion uncertainty reasoning and model of fuzzy reasoning. The prototype software of intelligent manufacturing expert system was developed eventually, The developed expert system effectively simulated the actual process of expert analyzing and reasoning with uncertain evidence and conclusions and solving problems. After adding knowledge concerning a specific issue to the rule base, the expert system can be effectively applied to manufacturing industry.