摘要
D-S证据理论在处理高冲突证据的融合时有可能得出不符合实际的融合结果。为此引入信息熵概念来确定证据的确定性指数,并对证据可信度和虚假度进行修正,从而构造证据的客观权重,藉此描述证据的冲突性和不确定性;同时以综合权重最大值对应的指标为基准构造折扣因子,实现对证据源的修正。针对南水北调中线工程某高填方渠段实例研究,基于信息熵的证据折扣因子法能较好地进行冲突证据融合,可有效避免融合悖论现象。
An unrealistic fusion result may be produced when fusing high-conflict evidences by using the classical D-S Evidence Theory. For this reason, to describe the conflict and uncertainty of evidence, the certainty indexes of evidence are determined based on information entropy, and the credibility and falsity of evidence are modified by these certainty indexes to construct the objective weight of evidence. Meanwhile, the index corresponding to the maximum of comprehensive weight is used as a benchmark to construct a discount factor to actualize the optimization of evidence source. A case study on a high-filled section of channel in Middle Route of South-to-North Water Diversion Project shows that conflict evidence can be fused well and the paradox of fusion can be avoided effectively by adopting the evidence discount factor method based on information entropy.
引文
[1]程德虎,苏霞.南水北调中线干线工程技术进展与需求[J].中国水利,2018(10).
[2]何金平,等.高拱坝多效应量改进型D-S证据理论融合模型[J].武汉大学学报(信息科学版),2012,37(12).
[3]Shafer G.A Mathematical Theory of Evidence[M].Princeton:Princeton University Press,1976.
[4]Yager R R.On the DempsterShafer framework and New Combination Rules[J].IEEE Trans.on System,1989,41(2).
[5]毕文豪,等.基于新的证据冲突衡量的加权证据融合方法[J].控制与决策,2016,31(1).
[6]李文立,郭凯红.D-S证据理论合成规则及冲突问题[J].系统工程理论与实践,2010,30(8).
[7]敖圣峰,何金平.基于分类修正的水闸健康诊断证据融合模型研究[J]水电与新能源,2018,32(11).