Plot experiments of rice and winter wheat respectively at the Beijing West Suburb Farm and the Dongbeiwang Township from 1997 to 1999 were made for questions of reasonable structure and reasonable dynamic during crop yield formation, control and decision-making during crop production. Based on integrate information collection, Culture Complete Collected Information Complex Concert Control Theory(CCCICCC) and Artificial Neural Networks were used to explore, the results were as follows,1.B-P Network used for dynamic description of crop development. After net structure was selected and training method was improved, net model was characteristic of high matching precision, high speed of convergence and powerful functionality. It could realize dynamic description of single-output by single-input and multi-output by multi-input. Prediction of crop development dynamic could be achieved based on this.2. Discussion on methods to understand traits of crop development dynamic. Function length, average of structural action between state variable and time of some development state at some time, was provided as integrate space-time value of development dynamic, based on that, multi-dimension space structure harmonization principle was developed and multi-dimension space-time structure harmonization principle was provided.3. Crop multi-dimension Complete Collected Information Complex Concert Control Theory and B-P NN were combined and Complete Information Harmonization NN (CIHNN) was used to understand, resolve question of reasonable dynamic during crop development. Winter wheat as an example, harmonization structure between yield and its structure and development dynamic and development dynamic of winter wheat could be harmonized and optimized according to target value.4. CIHNN used for decision of crop production project. Winter wheat and rice as examples, their yield target, development dynamic and integrate farming measurement was decided and rice decision-making project was validated that proved veracity of CIHNN... |