Abstract: This paper presents a general math expression of step by step combination results for evidence combination problem, and studies the theoretical rationality of step by step combination method and the convergence of the combination results. The step by step combination algorithm mixing Dempster evidence combination formula and the weighted average method is proposed for combination of severe conflicting evidence. The corresponding combination formula is selection and used according to the preset threshold of conflict. The simulation results show that the proposed algorithm has simpler calculation process and better convergence effectiveness than representative alternative rules.

Key words: D-S evidence combination, step by step combination algorithm, convergence

摘要： 针对证据分步合成问题，给出了多证据分步合成结果的一般表达式，对分步合成方法理论上的合理性以及合成结果的收敛性进行了研究。针对高度冲突证据的分步合成问题，提出一种将Dempster合成公式与加权平均法混合使用的分步合成算法;根据预设的冲突阈值，选用相应的合成公式。对比仿真结果表明该算法比国内外一些代表性改进方法合成计算过程更为简捷，合成结果的收敛效果更好。

关键词: D-S证据合成, 分步合成算法, 收敛性