农业专家系统生成工具中知识和模型相结合的研究

Study on knowledge and model combination of agriculture expert system tool

专家系统领域中 ,知识规则是由前提和结论组成的。专家系统在进行推理时 ,首先要获取前提 ,有些前提可以直接通过人机交互得到 (称直接前提 ) ,而有些前提需根据模型间接得到 (称间接前提 )。运用面向对象方法学的原则 ,将直接前提用属性表示 ,间接前提用方法表示 ,也将为得出结论所运用的模型用方法表示 ,再将相关的属性和方法封装在一起 ,构成一个类 ,则整个知识规则可抽象成由具有面向对象特点的类组成的知识表示形式 ,它将描述问题空间的属性和方法有机地结合在一起 ;同时 ,在构造推理机时 ,仍然按面向对象的原则匹配对象 ,构成了问题空间和求解空间的一致性。这样的软件具有稳定性好、易维护、易扩充的特点。

In expert system fields, knowledge rules are made up of premises and c onclusions. Before inferring, premises must be acquired, some premises are direc tly acquired by man-computer alternation(called direct premises),but some premis es are acquired indirectly by models(called indirect premises). The authors appl y the principals of object-oriented methodology, direct premises are represented to attrib utes, indirect premises are represented to methods, and models that are needed f or acquiring conclusions are represented to methods, too. Then, interrelated att ributes and methods are encapsulated to a class, the whole knowledge rules are c onstructed of classes that have characteristics of object-oriented. The object-o riented knowledge representation unites knowledge and models; Simultaneously, we construct a inference engine that matches objects by object-oriented principals . The knowledge representation and inference engine keep question space and solv ing space identical. This kind of software has good stability, maintainability a nd expansion ability.