径流的模糊模式预报

径流的状态取决于径流成因的状态,在选定预报因子的条件下,每一个径流样本对应着一个因子向量。应用模糊划分聚类分析法,从历史资料中提取数个径流成因模式,它们和各自的径流状态对应,建立起径流状态和成因状态的普遍联系。预报时预报样本的成因状态也构成一个向量,计算它和各成因模式的贴近度。预报样本的径流状态将和贴近度最大的成因模式的径流状态一致,据此作出径流的状态预报。文中给出一个实例,对其他类似问题,该方法也基本适用。

THE FUZZY PATTERN FORECAST OF RUNOFF

The states of rnnoff depend on the states of rnnoff factors. Every sample of run off responds to a vector of factors after forecast factors are chosen. By means of the clustering analysis of fnzzy division several factor patterns can be obtained from historical data which belong to a certain runoff state. Once the factor patterns of runoff states are obtained, a general relationship is established between runoff states and the states of forecast factors. The factor state of a forecast sample can be also showed by a vector. Calculate the nearness degree of the vector to every factor patterns. The forecast runoff will be in accord with the runoff state of the factor pattern which nearness degree is the largest. Based on this, the runoff state forecast can be carried out. The paper gives a practical example. Essentially, the method is available to similar problems.