Research On The Case-Based Reasoning Assistant Decision Support System Of Airworthiness Directives Issue

At present, the issue of Airworthiness Directive (AD) has been paid more and more attention to by the airworthy management departments of countries in the world. It is considered the essential condition of assurance of aircraft safety and airworthiness to excute and accomplish the ADs. Applying the technology relating to case-based resoning decision support system to airworthy information management field can improve the past artificial experiential decision-making condition, has extremely vital significance for the realization of the scientific decision-making.At first, after comparing the AD issue procedure between FAA (Federal Aviation Administration) and CAAC(General Administration of Civil Aviation of China), the article deeply studies the current status of Chinese ADs issue work, analyzes and summarizes its insufficiencies, and indicates the existing problem and difficulties in ADs information management.By means of comunicating and discussing with relating staff of our Airworthy Management Department, the function and demand of the system mentioned above are clear.Secondly, the article compares advantages and disadvantages between the two kinds of reasoning methord: rule-based reasoning and case-based reasoning after studying the basic method and principle of case-based reasoning, and decides to apply the method of case-based reasoning to the system.After that, the article deeply studies the key techniques of Case-Based Reasoning Decision Support System of AD Issue based on detailed data analyzing in the NTSB aviation accident database applied in this paper. Including construction of AD case base (which contains the expression, index strategy of case base and the case storage), retrieve, reuse, revise and retain of AD case. According to the reality of ADs issue, detailedly introduce the method of establishment of case base and take an example. And establish the model of case retrieve combining the technology of classified indexing with K-Nearest Neighbors method. After working out a reasonable arithmetic, the paper takes an example to validate the work. The example for the model shows that the results are quite good.Finally develops the software of Decision Support System of AD Issue on the base of researches above. In this part, the article introduces software demand, function, structure, database design and interface design.