基于判断聚合模型对集体判断不一致的化解

发布时间：2018-05-07 23:03

本文选题：社会选择理论 + 集体决策　；参考：《西南大学学报(社会科学版)》2016年01期

【摘要】：阿罗定理证明了将完全理性个体偏好聚合为完全理性集体偏好的不可能性。这个悲观的结论对实际的集体决策过程没有帮助。如何化解聚合产生的集体意见不一致,以保证集体决策公平公正地实现集体最大利益,即实现集体决策的理性化,成为许多学科研究的热点。判断聚合模型将聚合问题放在逻辑框架之下,将集体理性聚合研究提升到一个更高的抽象水平。本文运用判断聚合模型,从分析判断困境入手,对基于前提和结论的判断聚合方法,基于多数判断集的判断聚合方法和基于权重最大的判断聚合方法进行探讨,分析了各种方法在化解集体判断不一致上的优势及局限。这些方法尽管有一定局限,但能够求得一致的集体判断集,并且逻辑形式刻画技术也使这些方法能够运用于计算机和人工智能领域中的推荐系统、多主体决策系统以及搜索引擎技术之中,推动了相关理论和应用技术的发展。
[Abstract]:Arrom theorem proves the impossibility of aggregating the individual preferences of completely rational individuals into the collective preferences of completely rational individuals. This pessimistic conclusion does not help the actual collective decision-making process. How to resolve the collective dissension in order to ensure the collective decision-making to realize the collective maximum interests fairly and fairly, that is, to realize the rationalization of collective decision-making, has become a hot topic in many disciplines. The judgment aggregation model puts the aggregation problem under the logical framework and elevates the research of collective rational aggregation to a higher level of abstraction. In this paper, the judgment aggregation method based on premise and conclusion, the judgment aggregation method based on the majority judgment set and the judgment aggregation method based on the maximum weight are discussed by using the judgment aggregation model, starting with the analysis of the judgment dilemma. This paper analyzes the advantages and limitations of various methods in resolving the inconsistency of collective judgment. In spite of their limitations, these methods are able to obtain a consistent set of collective judgments, and logical representation techniques also enable them to be applied to recommended systems in the field of computers and artificial intelligence. Multi-agent decision system and search engine technology promote the development of relevant theory and application technology.
【作者单位】：西南大学逻辑与智能研究中心;
【基金】：西南大学人文社会科学重大培育项目“中国社会转型下群体理性认知的逻辑分析及实证建模研究”(12XDSKZ003),项目负责人:唐晓嘉
【分类号】：B812

【相似文献】