基于数据分布的随机聚合评价方法与应用  

Random-Aggregation Evaluation Method Based on Data Distribution and Its Applications

作　　者：董乾坤 易平涛[1] 李伟伟[1] 王露 DONG Qiankun;YI Pingtao;LI Weiwei;WANG Lu(School of Business Administration,Northeastern University,Shenyang 110167,China;School of Management,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]东北大学工商管理学院,沈阳110167 [2]沈阳工业大学管理学院,沈阳110870

出　　处：《系统管理学报》2024年第5期1231-1241,共11页Journal of Systems & Management

基　　金：国家自然科学基金资助项目(72171040,72171041,72301062);中央高校教育部基本科研专项资金资助项目(N2406013)。

摘　　要：为了从理论视角对实力差距较大情境下的“以下克上”等特殊评价现象提供合理解释,提出了一种新的基于数据分布的随机聚合评价方法,并从理论求解、结论检验等多个层面对其进行了论证。具体而言,提出了“相对优胜者”理念,并在此基础上考虑评价数据的分布,给出了精确信息的分布式随机转化方法。构建了分布式随机信息的融合框架,该框架具有包容性强、分类计算、独立求解等特征。从整体与局部两个视角给出了分布式随机聚合框架的求解步骤,并对优胜概率出现的边界条件、与影响因素的强度关系进行了数理证明。给出了求解结果的两个检验标准,通过模拟算例与实例应用验证了分布式随机聚合评价方法是有效且实用的,具有理论意义与应用价值。In order to provide a rational explanation for the special evaluation phenomena such as“the weak overcame the strong in the context of power disparity”from a theoretical perspective,a new random-aggregation evaluation method is proposed,which is demonstrated in detail from the solution,test,and other aspects.Specifically,the concept of“relative winner”is proposed,based on which a distributed random transformation method of precise information is put forward by considering the data distribution.A fusion framework for random data is constructed,which has the characteristics of strong inclusiveness,classification calculation,and independent solution.The solution of the fusion framework is discussed from both holistic and local perspectives,and the mathematic proofs of boundary conditions and relationship with influencing factors,concerning the“superiority probability”,are given.Two test criteria for solution results are introduced,which shows that the evaluation method proposed is effective and practical by a simulating arithmetic example and a practical application.

关 键 词：综合评价 数据分布 模拟仿真 相对优胜者理念 随机聚合计算 

分 类 号：C934[经济管理—管理学]