一种基于区间规则的条件证据网络推理决策方法  

A conditional evidential network reasoning and decision method based on interval rules

作　　者：郭强[1] 关欣[2] 潘丽娜[3] 刘俊[1] 潘新龙[1] 

机构地区：[1]海军航空工程学院信息融合技术研究所,山东烟台264001 [2]海军航空工程学院电子信息工程系,山东烟台264001 [3]海军航空工程学院基础部,山东烟台264001

出　　处：《控制与决策》2016年第3期394-402,共9页Control and Decision

基　　金：国家自然科学基金重点项目(61032001);国家自然科学基金项目(61102166;61471379);教育部新世纪优秀人才支持计划项目(NCET-11-0872)

摘　　要：针对证据网络推理方法无法对区间规则进行表示和推理的问题,提出一种基于区间规则的条件证据网络推理决策方法.该方法针对模糊规则的条件概率或信度为不确定区间的情况,可同时表达不确定性和模糊性;并将区间不确定规则转化为区间条件信度函数作为证据网络的结点参数,通过条件推理和证据融合得到条件证据网络中各结点幂集空间中焦元的随机分布作为决策依据.最后,通过空中目标态势评估实例,验证了所提出方法的有效性.To solve the problem that the interval rules can not be expressed and reasoned based on evidential networks, a conditional evidential network reasoning and decision method based on interval rules is proposed. The method extends fuzzy rules to interval uncertain rules, which can express the uncertainty and fuzziness simultaneously. Then, the interval uncertain rules are translated to interval conditional belief functions as the parameters of the nodes in evidential networks. The decision conclusion can be drawn by the random distribution of the nodes＇ power sets, which is obtained by conditional inference and evidential fusion. Finally, the effectiveness of the proposed method is illustrated through the air target situation assessment example.

关 键 词：区间不确定规则 证据网络 信息融合 条件推理 态势评估 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置] TP181[自动化与计算机技术—控制科学与工程]