冲突证据推理是不确定推理的一个重要研究课题。利用修正数据源对冲突证据进行有效组合的方法有很多, 但是修正证据源使用的证据之间的距离公式计算量较为复杂, 算法的流程也同样较为复杂, 算法对时序证据的组合没有能够进行很好的研究。针对这些问题, 提出一种冲突证据推理的快速算法。首先, 提出一种新的证据度量公式, 并对证据度量公式的有效性、计算量大小进行分析; 然后,给出冲突证据快速算法的一般流程, 并对流程中参数进行分析、确定; 最后,对时序证据的组合规则给出一般的算法步骤, 并通过算例分析, 验证了算法的有效性。

Conflict evidence reasoning is an important research subject of uncertain reasoning.There are many ways to effectively combine conflict evidences by using the revised data source.However, the formula describing the distance between the evidences used to modify the evidence source is complicated, and the flow of the algorithm is also complicated.The algorithm does not probe into the combination rule of the time-sequence evidences.In response to these problems, a fast algorithm for conflict evidence reasoning is proposed.Firstly, a new evidence measurement formula is proposed, and the validity and calculation amount of the evidence measurement formula are analyzed.Then, the general flow of the new algorithm is given, and the parameters in the flow are analyzed and determined.Finally, The algorithmic steps of the combination rule of the time-sequence evidences are given and the effectiveness of the algorithm is verified by example analysis.