摘要： 针对经典DS证据合成理论，提出一种基于矩阵分析的DS合成算法。该算法采用证据结合的递推方式，计算n组信任函数结合下的融合结果，解决证据冲突。通过矩阵分析避免计算量过大，减轻节点负担。对温室控制采集到的数据进行仿真实验，结果证明该算法的快速性，其计算时间仅为经典DS证据合成理论的1/10。

关键词: 证据理论, 矩阵分析, 融合

Abstract: Aiming at classical DS evidence compositive theory, this paper proposes a DS compositive algorithm based on matrix analysis. It uses reasoning in turn to integrate evidences, computes compositive results in the combinations of n group trust function to sovle the problem of evidence conflict. The algorithm uses analysis matrix to overcome the problem of huge calculation and reduce the burden of the node. Experiment uses the data which collected in greenhouse control as simulation objects, comparative result proves the speediness of the algorithm, the computing time is 1/10 of classical DS evidence compositive theory.

Key words: evidence theory, matrix analysis, fusion

中图分类号: 

• TP391