基于误差传递和隶属度的动态灰靶多属性决策  

A Dynamic Grey Target Multi-attribute Decision-making Method Based on Error Propagation and Membership Degree

作　　者：陈伟[1] 李金秋 杨早立[2] 

机构地区：[1]哈尔滨工程大学经济管理学院,哈尔滨150001 [2]北京工业大学经济管理学院,北京100124

出　　处：《系统管理学报》2017年第6期1034-1042,共9页Journal of Systems & Management

基　　金：国家社会科学基金资助项目(14BGL007);黑龙江省自然科学基金资助项目(G201209);黑龙江省软科学基金资助项目(GY2015RK0074)

摘　　要：针对当前区间数动态多属性决策方法存在的不足,提出了一种基于误差传递和隶属度的动态灰靶多属性决策方法。该方法从区间数型属性值的误差视角,运用误差传递模型确定了属性权重区间;考虑决策目标在时序上的差异性和波动性,引入时间度的概念,并基于最大熵原理和方差最小的思想,建立了多目标非线性规划模型确定时间权重;依据灰靶理论计算各方案的正负靶心距,并采用相对隶属度法对各时间段的正负靶心距进行集结,得到备选方案的优劣性排序。以突破性技术创新研发联盟伙伴选择为例,验证了该方法的可行性和有效性。To overcome the deficiencies in existing methods of dynamic multi-attribute decision-making with interval numbers, we propose a novel decision-making method based on error propagation and membership degree. From the perspective of error of attribute values in the form of interval numbers, an error propagation model is applied to determine the intervals of attribute weights. Considering the heterogeneity and volatility of decision-making objectives in the timing sequence, we introduce the concept of time degree, and then establish a multi-objective nonlinear programming model to determine the tim weights based on the principle of maximum entropy and minimum variance. Moreover, we calculate th positive and negative off-target distances based on the grey target theory, and aggregate the off-targe distances in all of the periods by using relative membership degree, to obtain the ranking of al alternatives. Finally, an example of partner selection is given to illustrate the feasibility and effectivenes of the proposed method.

关 键 词：区间数 误差传递 靶心距 隶属度 动态决策 

分 类 号：C934[经济管理—管理学]