欢迎访问《兵工学报》官方网站,今天是 分享到:

兵工学报 ›› 2020, Vol. 41 ›› Issue (6): 1140-1150.doi: 10.3969/j.issn.1000-1093.2020.06.010

• 论文 • 上一篇    下一篇

一种区间距离的冲突证据组合方法

吴斌, 衣晓, 李双明   

  1. (海军航空大学, 山东 烟台 264001)
  • 收稿日期:2019-07-09 修回日期:2019-07-09 上线日期:2020-08-07
  • 通讯作者: 衣晓(1976—),男,教授,博士生导师 E-mail:yxgx_gxyx@163.com
  • 作者简介:吴斌(1992—),男,博士研究生。E-mail: dukenavy@163.com;
    李双明(1986—),男,工程师,博士研究生。E-mail: aminglishuang@126.com
  • 基金资助:
    国防科技卓越青年科学基金项目(2017-JCJQ-ZQ-003);泰山学者工程专项项目(ts201712072)

Conflict Evidence Combination Method Based on Interval Distance

WU Bin, YI Xiao, LI Shuangming   

  1. (Naval Aviation University, Yantai 264001, Shangdong, China)
  • Received:2019-07-09 Revised:2019-07-09 Online:2020-08-07

摘要: 针对含有冲突证据的多条证据组合结果出现反直观的不足,提出基于不确定区间距离的冲突证据组合方法。通过文献[13,15,17]分析发现,由信度及似真函数构成的区间包含事件本身所有的信息,解决了直接由Mass函数构成证据组合过程中信息缺失的问题。根据不确定区间定义了一种区间距离。通过区间距离生成证据间的距离矩阵以及证据间的权重,利用权重对原有证据进行修正,基于Dempster-Shafer证据理论及第6种比例冲突再分配完成证据组合。算例分析结果表明:新方法在冲突度表征变化上与文献[33]的变化趋势一致,证明了该方法的有效性;在多条冲突证据的组合中,新方法的信度值比现有冲突度量方法明显提高,证明了该方法的正确性。

关键词: 冲突证据, 证据理论, 区间距离, 组合方法, 加权

Abstract: A conflict evidence combination method based on uncertain interval distance is proposed for the anti-intuitive problem about combination results of multiple pieces of evidences containing conflict evidence. Through the analysics in Refs. [13, 15, 17], it is found that the interval composed of Bel and Pl function contains all the information of an event, which solves the problem of information missing in the process of evidence combination directly composed of Mass function. The expression of uncertain interval is given, and then an interval distance is defined according to the uncertain interval. The interval distances among the same focal elements of different evidences are calculated, and the interval distances of all focal elements are summed. A distance matrix and weights among evidences are generated through the distance matrix, and the original evidences are corrected according to the weights. The evidence combination is completed with Dempster-Shafer(D-S) and Proportional Conflict Redistribution 6 (PCR 6). The example analysis shows that the proposed method agrees with the methods in Ref. [33] in the representation change of conflict degree, which shows the effectiveness of the proposed method. In the combination of multiple conflict evidences, the Bel of the proposed method is significantly improved compared with the existing methods. Key

Key words: conflictevidence, evidencetheory, intervaldistance, combinationmethod, weighting

中图分类号: