基于团树传播的证据网络推理算法  

Reliability inference method for evidential network based on the clique tree propagation algorithm

作　　者：吴根秀[1] 黄涛[1] 赖岳[1] 杨莎莎[1] WU Gen-xiu HUANG Tao LAI Yue YANG Sha-sha(School of Mathematics and Information Science, Jiangxi Normal University, Nanchang 330022, Chin)

机构地区：[1]江西师范大学数学与信息科学学院,南昌330022

出　　处：《控制与决策》2017年第2期305-310,共6页Control and Decision

基　　金：江西省自然科学基金项目(20151BAB207030);江西省教育厅科技项目(GJJ14244)

摘　　要：将团树传播算法应用于证据网络中,解决复杂的多连通知识网络结构下的信度推理问题.将复杂多连通网络构造成一棵团树,并将联合信度作为团节点的参数实现复杂多连通网络结构上的证据网络信度推理.在进行联合信度函数信息融合的过程中,通过引入两种新的交并运算实现对DSm T组合规则的改进,减少不确定性.最后通过实例验证了所提出方法的可行性.For solving the reliability inference problem in the multi-connected knowledge network model, a clique tree propagation algorithm is innovatively applied to the evidential network. Firstly, the multi-connected knowledge network model is clustered as a clique tree, and the joint belief function is regarded as the main parameter of the cluster nodes,therefore, the information each node is obtained, which facilitates the possibility of reliability inference within the multiconnected knowledge network model. In the process of evident fusion of the joint belief function, two new union and intersection methods are introduced to improve the existing DSm T theory, which helps to eliminate the influence of conflicting evident information on other evident variables. Finally, an example is given to illustrate the feasibility of the proposed method.

关 键 词：团树传播算法 证据网络 信度推理 信息融合 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]