The Construction Of Wheat Sharp Eyespot Warning And Decision-making Expert System For Henan Province

Wheat Sharp eyespot Warning and Decision-making Expert System (WSWDES) for Henan Province is intelligently built based on the knowledge, experience and research results of the plant protection experts in Henan Province. It adopts oriented-object programming technology, Visual Basic, using Access for the management of data and SQL(Structure Query Language) for the query and editing of data, utilizing SPSS for statistics and coupling multimedia techniques & expert system techniques. This paper expounds its forecasting arithmetic, the representation of knowledge, the development of knowledge base and database and the realization of inference engine, etc, minutely. On the function of WSWDES, the intelligent decision-making system integrates the diagnosis of wheat diseases, forecasting, loss estimation and control decision. It has the function of reasoning and query and can give users the reference to terminal decision. The database and knowledge base are expandable.WSWDES adopts 3 methods to forecast wheat sharp eyespot, the qualitative forecasting of disease degree on steady stage, the quantitative forecasting of the rate of ill plants before winter and the disease index on steady stage. The former one is realized by fuzzy list and the latter two by step-wise regression. The three methods are integrated and complementary each other. They almost cover the whole growth period of wheat and can provide the users with the foundation of control.The representation of knowledge of WSWDES adopts complex method: frame, rule skeleton plus rule body, procreant rule and data dictionary. It includes the frame of wheat disease, the frame of pathogen, the frame of the characteristic of wheat sharp eyespot, the frame of the resistant trait of wheat variety to wheat sharp eyespot, the frame of forecasting, the frame of loss estimate and the frame of control decision-making. The forecasting rules and mathematics models in the frames of forecasting and loss estimate are described by two layers,namely, rule skelecton and rule body. The knowledge base of diagnosis adopts procreant rule. Data dictionary is used to translate Chinese to codes.65The comprehensive knowledge base of WSWDES is made up of rule base, frame base and multimedia base. Rule base is made up of diagnosis knowledge base, forecast knowledge base, loss estimate knowledge base, control decision-makirig knowledge base. The multimedia base contains ten items, the conditions of occurrence, the cycle characteristic of wheat sharp eyespot, etc. It interact users with photographs, characters and voice. The knowledge base is described more minutely as following.(1) Diagnosis knowledge base contains the diagnosis of 29 common diseases of wheat, using the position of disease and symptom as the basis of diagnosis, attaching characters and photographs. Users can get the results by choosing from combo box and can learn about the occurrence principle and control measure of each disease.(2) Forecasting knowledge-base contains 625 qualitative forecasting rules, 5 quantitative models for forecasting the rate of ill plants before winter and 5 quantitative models for forecasting the disease index on steady stage, aiming at 5 areas, namely, the east of Henan province, the west of Henan province, the south of Henan province, the north of Henan province and the middle area. It provides the foundation of control decision.(3)Loss estimate knowledge-base contains 5 models, namely, the relationship between the loss of seed weight and the disease grade on steady stage, the relationship between the loss of seed quantity per ear and the disease grade on steady stage, the relationship between the loss of ear weight and the disease grade on steady stage, the relationship between the loss of yield and the disease index on steady stage, the relationship between the loss of yield and the rate of ill plants on steady stage.(4)Control decision-making knowledge base provides users with the threshold of control and the measure of control.The database of WSWDES includes the database o...