农业自然灾害系统灰色分析方法的探讨

对灰色关联方法在农业自然灾害系统分析中的应用进行了探讨。认为，农业自然灾害对干农业生产的影响完全有别于生产资料投入的作用。当系统特征行为序列（Ｘｏ）为增长序列，相关因素序列（Ｘｉ）亦为增长序列时，不能就此判定Ｘｏ与Ｘｏ为正相关关系而直接进行建模。当促进因素的作用大于抑制因素的作用时，即使灾害的危害有所增加，农业生产仍可表现为一个增长序列。提出判定关联性的命题应为：设系统特征行为序列为Ｘｏ，相关因素序列为Ｘｉ，则，１°当影响因素的性质一致时（如均为促进或抑制），设Ｘｏ为增长序列，若Ｘ１为增长序列，则Ｘ１与Ｘｏ呈正相关关系；若Ｘ１为衰减序列，则Ｘｉ与Ｘｏ呈负相关关系。２°当系统主行为受性质不一致的因索（包含促进和抑制因素）影响时，若Ｘｉ的增长促使Ｘｏ增长或衰减趋势减弱，Ｘｉ的衰减使Ｘｏ衰减或增长趋势减弱，则Ｘｉ与Ｘｏ呈正相关关系；若Ｘｉ增长促使Ｘｉ衰减或增长趋势减弱、Ｘｉ衰减使Ｘｏ增长或衰减趋势减弱，则Ｘｉ与Ｘｏ为负相关关系。

On the Application of the Grey Relational Method to the Natural Disaster Systematic Analysis of Agriculture

The application of grey relational method to the natural disaster systematic analysis of agriculture was discusse the author''s own view on it has been raised which are opposite sides for agriculture be influenced by the disasters and affecting for the means of producible. When the system action characteristic series (SACS) x.was a growth series and the factors series (FS) was also a growth series, the relationship was not necessarily to be positive between x_o and x_o and thus not directly to construct the grey model. When the effect of promoting factors was stronger than thus not directly factors, though the endangerment of disasters rised to some extent, the agricultural production may remain showing a growth series'' In the present paper, the theorem for diciding rolational character was proposed. Defining there are the SACS (Reference Series) x_o and FS x_i (i=1, 2,.., n). Theorem 1° when the factors character shows no difference (all progress or all decrease), Let x_o is a growth series, thus:if x_i is also the growth series, then the relational character is positive between x_i and x_o; if x_i is the decrease series, then the relational character is negative between x_i and x_o. 2° when the factors character shows difference (include progress and decrease), thus:if the x_i increase impels the x_o increase or abate its trend of declining, and the x_i decrease causes the x_o decrease or weakens its trend of increase, then the relational character is positive between x_i and x_o; if the x_i. increase makes the x_o decrease or weakens its increase trend, and the x_i decrease makes the x_o increase or alleviates its decrease trend, then the relational character is negative between x_i. and x_o.