南京大学学报（自然科学版）2016，Vol.52Issue(4)：672-681,10.DOI:10.13232/j.cnki.jnju.2016.04.012
基于置信规则库推理的多属性双边匹配决策方法
Belief rule base inference methodology for two-sided matching decision with multi-attribute
摘要
Abstract
This thesis presents a tentative study on a new two-sided matching approach,which is proposed to solve the two-sided matching problem with uncertain information and multiple attributes.The multi-attributes matching decision making(MAMDM)problem is one of the most important key points in the two-sided matching study,which has evoked great attention for the scholars in recent years.A belief rule-base inference methodology using the evidence reasoning approach(RIMER)has been introduced in this thesis to solve the problem of MAMDM.At the beginning of this thesis,the authors explain the reason why they choose to use belief degree.The current research on the problem of MAMDM is mainly restricted to the study of a kind of two-sided matching,whose evaluation information is linguistic values or interval values.But there exists a lack of study in belief degree as evaluation value. As belief degree can be used to deal with different kinds of uncertain and incomplete information,using it as evaluation value may trigger a new breakthrough in the study of MAMDM.Through the analysis of simulation ex-periments datas and the application of RIMER,belief degrees evaluation information is converted into different levels of confidence information.Then a 0-1 programming model is built by making use of different levels of confidence information to obtain a final matching scheme.It is also pointed out in the thesis that an output error may be caused when BRB(belief rule-base)input is higher than threshold value.To solve this problem,the authors propose that the input value can be incorporated into the uncertainty by the adoption of cutting method.If cutting method is not suitable,linear mapping method can be applied to reduce the influence of the results.The case study analysis shows that it is feasible and effective to adopt the new proposed approach to solve the problem of multi-attributes matching decision making.关键词
置信规则库/多属性双边匹配决策/证据推理/置信度Key words
belief rule base/multi-attribute matching decision-making/evidential reasoning/belief degree分类
信息技术与安全科学引用本文复制引用
方志坚,杨隆浩,傅仰耿,陈建华..基于置信规则库推理的多属性双边匹配决策方法[J].南京大学学报（自然科学版）,2016,52(4):672-681,10.基金项目
国家自然科学基金(70925004,61300026,71501047),福建省自然科学基金(2015J01248),福建省教育厅科技项目(JA13036),福州大学科技发展基金(2014-XQ-26) （70925004,61300026,71501047）
