广义证据理论中的基本概率指派生成方法

摘要

鉴于基本概率指派生成对证据理论的研究具有重要的意义,文中针对辨识框架不完整情况提出了一种强约束广义基本概率指派赋值方法,新方法可以生成空集不为零的广义基本概率指派赋值,该数值的大小反映了系统是开放世界的可能性,针对辨识框架完整情况提出了一种弱约束基本概率指派赋值方法,该方法在样本与表示模型之间不相交时,也可以给出样本与模型相似性度量的数值,根据所提出的策略生成弱约束的基本概率指派.算例表明了所提出方法的有效性.%Determination of the basic probability assignment (BPA) is the first step to use evidence theory.How to determine BPA is still an open issue.Two methods, namely strong constrain method and weak constrain method, are proposed to deal with open world and close world respectively.In the framework of the proposed generalized evidence theory, a strong constrain method for determination of generalized basic probability assignment (GBPA) is developed.Compared with the existing methods, the new method can generate GBPA without the constrain that the empty set should be set zero.In addition, the value of the empty illustrates the possibility degree that the frame of discernment is not complete.A weak constrain method is also presented to generate BPA in the close world.The weak constrain can determine the BPA even the sample does not cross with the representation model.Some numerical examples are used to illustrate the efficiency of the proposed methods.