基于灰色动态多属性群决策方法的农业旱灾应急管理

Emergency Management of Agricultural Drought Disaster Based on Grey Dynamic Multi-attribute Group Decision Method

为科学解析农业旱灾形成机理、合理指导防旱减灾工作,针对农业旱灾应急管理中决策信息灰性、决策方案不确定性的动态多属性群决策问题,引入三参数区间灰数进行表征,提出了基于4维矩阵和后悔理论的三参数区间灰数动态多属性群决策模型。首先,针对决策者的风险偏好心理,利用4维矩阵表述决策属性值,提出了基于后悔理论的后悔欣喜函数;其次,构建相应优化模型求解群决策权重、属性权重和阶段权重,根据综合欣喜后悔值对方案排序择优;最后,应用于河南省农业旱灾应急管理中,并采用灰靶群决策方法进行对比分析。结果表明:在河南省农业旱灾应急管理决策中该方法涉及的参数较少,且不需要给定决策者的参考点和权重,评价结果更具客观性。该方法为解决灰信息下的应急管理提供了理论参考,同时为决策部门优选应急管理方案提供理论支撑。

In order to scientifically analyze the formation mechanism of agricultural drought and reasonably guide the work of drought prevention and mitigation, aiming at the dynamic multi-attribute group decision-making problem of decision-making information grey and decision-making scheme uncertainty in agricultural drought emergency management, the three-parameter interval grey number was introduced to characterize it, and a three-parameter interval grey number dynamic multi-attribute group decision-making method based on4-dimensional matrix and regret theory was proposed. Firstly, according to the risk preference psychology of decision makers, a 4-dimensional matrix was used to express the decision attribute value, and a regret euphoria function based on regret theory was proposed;secondly, a corresponding optimization model was constructed to solve the group decision weight, attribute weight and stage weight. Finally, it was applied to the emergency management of agricultural drought in Henan Province, and the grey target group decision-making method was used for comparative analysis. The results showed that the method in this paper involved fewer parameters in the decision-making of agricultural drought emergency management in Henan Province, it did not require the reference point and weight of the decision-maker, and the evaluation results were more objective. This method provides a theoretical reference for solving emergency management under grey information, and provides theoretical support for decision-making departments to optimize emergency management schemes.