Research On Expert System Application Technology Based On CLIPS

This thesis is based on the system of key action assessment of aircraft which is one of the prjects in our laboratory. Because both the user interface and the reasoning kernel of this system were developed in C++ language, it has the following problems. Firstly, the presentation of knowledge is difficult and its scalability is poor. Secondly, the management and maintenance of the reasoning facts and rules is inconvenient. Thirdly, it is inefficient and not user-friendly because of the loss of the monitoring of time consuming and the details of the reasoning process.To solve the aforementioned problems, we studies the key technology of the expert system, the expert system tool CLIPS(C Language Integrated Production System), and the expert system application technology based on CLIPS. CLIPS has good openness, scalability and high reasoning efficiency. It is easy to separate the knowledge and the reasoning engine. Knowledge in CLIPS has high scalability and it is easily expressed in the form of templates or production rules. In addition, CLIPS is small and compact and can be applied in real-time applications. To this end, we study the key technology of the expert system based on CLIPS and try to improve our system of key action assessment of aircraft by taking CLIPS as its reasoning engine.Firstly, the overall structure of the assessment expert system is studied and designed, which is divided into evaluation system interface, assess knowledge base module, working memory module, inference engine module, knowledge acquisition module, the interpretation module. Through the research of knowledge base and inference engine of assessment expert system of aircraft critical action, we use CLIPS to develop the evaluation system. Secondly, we build the knowledge base using knowledge representation of "framework + production rules". Then, the assessment knowledge which is collected is wrote to CLIPS knowledge base, completing the construction of knowledge base of the entire system. Thirdly, we use CLIPS as the reasoning core to manage the fact base and knowledge base dynamically when the aircraft is performing mission, and determine the forward chaining inference method and breadth-first search reasoning strategy. According to the tree structure characteristics of the aircraft key action assessment, we design the reasoning process and conflict digestion strategy based on CLIPS, meanwhile apply the fuzzy inference mechanism to the evaluation system to improve the credibility of reasoning. Finally, we study the interpretation mechanism based on CLIPS and apply it to the system to explain the reasoning process. At last, we use VS2010 to build user interface, embed CLIPS into VS2010 and research software interface technology, analyze the limitations of the way of taking text file as the interaction intermediary of CLIPS and VS2010 and propose an improved interaction method, and ultimately achieve the entire functions of assessment expert system.By combinating the aforementioned techniques in our system of key action assessment of aircraft, we can make decision on whether the key actions of the aircraft is normal or not in real time. It is precise, efficient and portable. In the end, this paper gives useful cues for constructing an expert system based on CLIPS efficiently.