农业气象灾害等级优化预测仿真研究

Simulation Study on Optimization of Agricultural Meteorological Disaster Grade

对农业气象灾害等级进行准确预测,可对灾害进行有效防御,提高农业生产量。进行灾害等级预测过程应建立在利用农业气象状态观测矩阵,计算农业气象灾害预测加权系数的基础上,而传统方法通过获取灾害关键属性的聚类中心和隶属度矩阵,组建农业气象灾害等级判断准则完成预测,但不能准确组建气象状态观测矩阵,无法精确计算灾害预测加权系数,导致预测误差大的问题。提出基于改进混沌理论的农业气象灾害等级预测优化方法。先融合于混沌理论对农业气象灾害发生等级时间序列数据进行相空间重构,计算气象数据混沌相空间的嵌入维数和嵌入延迟,以卡尔曼滤波方法为基础组建农业气象状态观测矩阵,计算出农业气象灾害预测加权系数,以计算的结果为依据组建农业气象灾害等级预测模型,利用上述模型完成对农业气象灾害等级预测,弥补了传统方法无法预测隐含的灾害发生规律的弊端。仿真结果表明,所提方法可以有效地预测可能即将发生的农业气象灾害及其等级,具有较高的预测精度和良好的适用性。

This research focuses on the optimization method for grade prediction of agricultural meteorological disaster based on modified chaos theory. We carried out phase space reconstruction for time series data of occurrence grade of agricultural meteorological disaster integrated with chaos theory, then calculated the embedded dimension and delay of meteorological data chaotic phase space and built measurement matrix of meteorology state based on Kalman filtering method. The weighting coefficient of disaster prediction was worked out and prediction model of disaster grade was built up according the calculation results. Thus, the grade prediction was completed to made up the disadvantage that traditional method cannot predict connotative regularity of disaster outbreak. Simulation results show that the method mentioned above can predict the impending agricultural meteorological disaster and its grade. It has higher prediction precision and better applicability.