基于高斯隶属函数的模糊定性仿真

Fuzzy qualitative simulation with Gauss membership function

通过分析发现模糊定性仿真系统存在两个缺陷,即模糊子集距离定义和相似原理的不完善。这两个缺陷存在的根本原因是采用了四元数表示的模糊定量空间。用高斯隶属函数代替梯形隶属函数研究模糊定性仿真,并对模糊子集的距离定义以及相似原理作了改进。在FuSim的基础上,给出了到达时间、持续时间的具体表达式并用于时间过滤。仿真实例证明了此算法的有效性。

To overcome drawbacks of conventional fuzzy qualitative simulation, a novel fuzzy qualitative simulation algorithm based on Gauss membership function instead of trapeziform membership function is presented. The definition of distance between fuzzy qualitative spaces is given. The formulas of persistence time and arrival time are given, which are used for temporal filtering technique. The novel algorithm can improve precision and reduce false behaviors. The validity of the algorithm presented is proved by simulation.