基于概率语言赋权交叉熵信度分配的应急决策  

Emergency decision-making based on the reliability distribution of weighted cross entropy in probabilistic language

作　　者：白静[1] 陈业华[1] 石蕊[1] 张亚明[1] BAI Jing;CHEN Yehua;SHI Rui;ZHANG Yaming(School of Economic and Management,Yanshan University,Qinhuangdao 066004,China)

机构地区：[1]燕山大学经济管理学院,河北秦皇岛066004

出　　处：《系统工程与电子技术》2021年第2期476-486,共11页Systems Engineering and Electronics

基　　金：国家自然科学基金(71171174,70971115);国家社会科学基金(18BXW118);河北省社会科学基金(HB19GL007)资助课题。

摘　　要：针对多部门协同应急决策带有不确定性和偏好性的多属性决策问题,提出了一种基于概率语言赋权交叉熵信度分配的应急决策方法。首先,采用能充分表达不确定决策信息的概率语言术语集(probabilistic linguistic term set,PLTS)刻画属性指标,并将其作为证据推理算法中的证据信息。同时,从认知科学的角度考虑由部门跨领域及专业偏好等特点造成的证据不确信问题,对证据折扣系数表达式进行改进,构建基于概率语言赋权交叉熵的Mass函数。然后,利用递归算法对多部门多指标形成的证据信息进行合成,得到各方案的综合评估信息。最后,通过算例分析验证了方法的有效性和实用性。For the multi-attribute decision-making problems with uncertainty and preference in multi-department collaborative emergency decision-making,an emergency decision-making method based on the probabilistic language weight cross entropy reliability distribution is proposed.Firstly,the probabilistic linguistic term set(PLTS),which could express the uncertainty decision-making information fully,is used to describe the attribute index.The information is taken as the evidence information of evidence reasoning method.At the same time,from the perspective of cognitive science,the evidence uncertainty caused by the departments with across field and multidimensional preference is taken into account.So the evidence discount coefficient expression is improved,and mass function based on the probability language cross entropy weights is built.Then,the evidence information formed by the multiple departments and indicators is synthesized by the recursive algorithm,and the comprehensive evaluation information of each program is acquired.Finally,an example analysis is given to verify the validity and practicality of the method.

关 键 词：应急决策 多属性决策 概率语言术语集 证据推理法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]