Design And Implementation Of Reasoning Machine In Expert System For Embedded Meter Design

The ultimate aim of artificial intelligence is to simulate the human thought and conduct reasoning process. As one application-oriented branch of artificial intelligence, expert system is the incorporate of the utility of artificial intelligence. Expert system is basically one computer program which can apply knowledge and consequence in one specific area in settling the problems with the human expert level. It has the ability to work with expert level of knowledge, experience and capability.Ever since 1960s when the first generation of expert system was developed in the USA, it has been widely applied in various fields of medicine, mathematics, education, engineering, geography, biochemistry, fault-judgment and military and so on. However, expert system was rarely used in electronic design automation, especially for embedded system hardware knowledge type. With the prevailing application of electronic products, developing the expert system that can aid the electronic design has become an urgent affair. This essay aims to research on the reasoning machine of "embedded meter design expert system"."Embedded meter design expert system" establishes the intelligent meter design expert system knowledge database, rule database and case database on the basis of expert knowledge accumulation and data mining technology, and apply artificial intelligence meter to aid professional intelligent meter designers with intelligent and constructive help on the design. Distinct from the traditional process, it has simulated the expert thinking process and used such knowledge to get the results, therefore its structure can be concluded as "expert system =knowledge + consequence" with knowledge database as foundation and consequence machine as its core. This essay focuses on the design and realization of this reasoning machine.This paper studies on the consequence method combined with the one based on cases and the one based on rules. The existing data on cases of intelligent meter have very important reference value for designers since under most circumstances the design have strong similarity with the existing successful cases. However, the case-based consequence method provides little support on the innovation design, thus this system uses rule-based and linguistic network consequence methods to support innovation in design.For the purpose of settling the vast-amount data consequence, and the usability in time, we bring the multi-sorting index technology and SPA preference technology tothe case-based consequence process on the basis of principle, design feature and key technology of embedded intelligent meter design. This is also done with the successful application of oriented customization on traditional genetic algorithm on reasoning machine.With this system, even designers with basic design knowledge can give a mature combined blue-print and suggested precept on relative software, hardware and parts on the basis of demand in design and current circumstances. In this way, the designers' ability to acquire knowledge is raised, the cycle of the intelligent meter design is shorten and time and platform for innovation on the basis of lifting design quality, stability and universal adaptation is provided. The application of expert system in embedded system intelligent meter developing platform can greatly improve the competitive strength of enterprises and therefore has genuine urgency in the promoting of industry development.