A Research On The Theory And Methods About Forest Resources Statistical Model

The effective prediction and controlling of the forestry resources system is the main basis for us to work out and restructure our forestry policies, to set up projects, to supervise the actual situation of forestry resources; meanwhile, it significantly functions not only in the construction and projecting of the forestry ecological environment, but in the construction of sustainable ecological forest.The forestry resources system is a dynamic and statistics random system with decentralized structure. According to the characteristic of the forestry resources system, in order to combine whole prediction and the part, the centralized control and the decentralized, we select scattered form to indicate the structure of forestry resources system, adopt decentralized statistics random control to predict and make decision for forestry resources system.This thesis adopts a statistics random dynamic system to study the forestry resources system. First, it works out centralized and decentralized solution of the model, and sums up superior gain matrix; second, with the best status of function target, it restructures the model of dispersed statistics random dynamic system to analyze the superior or inferior of the known control vector; last, with the above two steps, it forecasts and estimates the future forestry resources system, thus, works out superior control input vector.Based on the forestry resources dynamic data taken from Jiangsu Province, Shandong Province, and Fujian Province from the year 2001---2005, viewing the above three provinces as regional control units, this thesis respectively forecasts and estimates their total forestry storage and acreage. The statistics show that our previously predicted data and filter data coincide perfectly with the respective actual data of forestry resources. The difference between the total storage and acreage of our predicted data and the actual data is less than 8 %; while the difference between the filter data and the actual is less than 0.05 %. The predicted data and filter data to the future storage and acreage for the mentioned three provinces also coincide with the separate target data. The difference between the predicted data and the target data is less than 3 % in the total storage and acreage. While the difference between the filter data and the target data is less than 0.2 % in the total storage and acreage. To sum up, we may use decentralized statistics random control theory to solve the knotty problem of how to decide superior control input vector in the process of prediction.In the end, the thesis states that the forestry resources model applies to the prediction of forestry ecological benefits.