Study On Knowledge Model-based Expert System For Soybean Management

The total production of soybean hesitated and the imports increased year by year in recent years. Traditional farming modes and low technological content of soybean production lead to the low yield and poor quality and effect the soybean industry development. transforming production and management modes and accelerating the transformation and extension of soybean scientific research results have became the main direction of the soybean revival industry in China. Basing on the former soybean researching data of many years and the existing research results and using the constructing theory and technology of expert system and knowledge models, this study established a soybean production and management expert system based on WEB. This expert systemwas able to provide expert decisions and advisory service in the prenatal, delivery andpostnatal of soybean production by using information technology and made traditionalproduction modes develop to half-pricisie-quantity modes.1. The knowledge model was used in the wheat, cotton, rice, canola and corn before it was applied again in the soybean production and management expert system. The knowledge model for pre-sowing planting plan included crop rotation, cultivars selection, yield target, determination of planting density, irrigation and fertilization.In order to determine the rotation crops and cultivation area, and construct a reasonable rotation mode, so as to solve the continuous and alternate cropping problem of soybean and develop the sustainable agriculture, the ecological and economical benefits were considered adequately in the reasonable rotation model. The model regarded three years as a rotation range, and considered yield levels, input costs, market prices, the effect of preceding crop on ensuing crop and so on. The model used linear programmingprinciple, and made Y=sum from i=1 to 3 A_ij(j= 1,2,...,m) which was equal to the economic benefitof three years in unit yield maximal.Cultivars-selection model wasZXD = (?),when the confidence limitZXD=1-multiply from i=1 to N((|X_iy-X_ix|/(X_ix) was higher than 80%, the system will select the cultivars outand provide the selection of drought-resistant cultivars. The planting density model of soybean ascertained 341,400 strains per hectare as the basic planting density through collecting data and analyzing models for past years. This model was amended with following factors such as the plant-type, soil fertility, fertilizer, cultivation and irrigation, etc. The amended result was indicated by the equationMD=JCMD×ZX×NSL×ZP×GS.Knowledge models such as ecological adaptive judgment of soybean, yield targets identification, sowing date determination, irrigation and fertilization were conducted in this study. And the relevant parameters were also determined.2. The model for the dynamics leaf area index, colony photosynthesis, dry matter accumulation, nitrogen, phosphorus and potassium uptaking of the ground part and their mutual relationships was established. All those target physiological indicators during every stage were determined with the objective yield and made as the basis of managements.WhenTK=((Xy-Xs)/(Xy))≥10% or TK=((Xy-Xy)/(Xy))≤-10%, appropriate cultivationmeasures should be applied in order to make the crop production develop target.3. An expert advisor system (EAC) for the whole producing process of soybean including morphological knowledge, physiological knowledge, reasonable rotation, pre-sowing preparation, sowing, seedling management, branching management, fluorescence management, drum graining management, the control of disease and pest and harvesting and storage timely was established. Expert advisory service by using the way of forward reasoning and reverse reasoning could be offered to users.4. The soybean production and management expert system based on WEB was designed on the platform of Windows 2003 Sever andⅡS6.0, and was based on the data-base Sql Server2000 of Microsoft. The asp.net mode was used in its design. This expert system used the convenient internet as a link and used the information technology to integrate the existing research. It could provide technology information service about soybean production to users.