Cotton Science 2005,17(4):201-206

 

Design and Implementation of a Knowledge Model-based Decision Support System for Cotton Management
ZHANG Huai-zhi£¬ZHU Yan£¬CAO Wei-xing*
(Hi-tech Key Laboratory of Information Agriculture, Jiangsu Province, Nanjing Agricultural University, Nanjing 210095£¬China)ª¤ª¤

 

Abstract: By summarizing and inducing research achievements and knowledge accumulation in cotton cultivation theory and technology, a knowledge-model-based decision support system for cotton management was developed using system method and structural approach. The system is composed of database, knowledge model, expert consultation and computerª²user interface, according to the target yield and anticipate fiber quality of farmers. It can be used for designing pre-sowing cultivation plan, which include variety type, sowing date, population density, fertilization strategy and water management, suitable cultural plan and growth indices£¬regulating growth dynamics£¬and providing specialist knowledge assistance and system maintenance£¬with wide applicability and good explanation. Using climate and topsoil property data of Nanjing, Jiangsu Province, this study designed pre-sowing cultivation plan, suitable dynamic growth indices for Sumian 9, which was planted widely in Nanjing. Comparing this design results with the practice and expert's knowledge, it indicated that on the basis of local farming conditions, the system generates growth indices for a proper target yield, such as height, leaf area index, the number of fruit branches, square and boll, dry matter accumulation and nutrition accumulation, according with the high yield population quality dynamic growth indices.In addition, the density and sowing rate the system designed were respectively 4.2 ¡Á104plant¡¤hm-2, 31.2 kg¡¤hm-2, which was in the diversification extension of farmers practice. At present, the actual amount of needed fertilizer and water of Sumian 9 aren't clear, but that the system calculated 203.9 kg¡¤hm-2 N, 103.3 kg¡¤hm-2 P, 86.5 kg¡¤hm-2 K and 9112.5 mª¬3¡¤hm-2 water were reasonable based on the understanding of fertilizer and water requirement for other varieties. As for development stages of cotton sowed on April 16 with no plastic film covering, the date of emergence, squaring, flowering and boll opening were April 28, June 6, July 5 and August 20. Those dates were believable according to the farmers experiment and expert knowledge.
Key words: cotton cultivation; decision support system; knowledge model; expert system
            [Full Text,746KB]