信念变化的多—核收缩算子研究

【作者】 苏静；

【导师】 肖奚安； 朱朝晖；

【作者基本信息】 南京航空航天大学 ， 计算机软件与理论， 2004， 硕士

【摘要】 信念修正理论是目前人工智能的一个重要的研究方向，很多专家学者对此进行了广泛而深入的研究，并且根据不同的应用领域的不同需要，提出了许多信念修正的方法，其中最具有代表性的就是AGM理论。本文在AGM理论和单核收缩的基础上，讨论了被收缩的对象是一般语句集合的信念收缩理论。具体而言，主要完成以下工作： (1)给出包-核收缩算子的构造及其公理假设系统并证明了相应的表示定理。 (2)建立了饱和的包-核收缩算子和平滑的包-核收缩算子的表示定理并证明了两者的等价性，然后讨论了后者和基于部分交收缩的包收缩算子之间的关系。 (3)给出选择-核收缩算子的构造并且给出相应的公理假设系统及表示定理，然后讨论了子核收缩算子以及相应的表示定理。更多还原

【Abstract】 The theory of belief revision is an important branch in the research of artificial intelligence. Many researchers have discussed about it, and put forward lots of methods according to different requirements in different applicable areas, the most representative of which is the AGM theory. On the basis of the AGM theory and kernel contraction, this thesis mainly discusses the theory of belief contraction in which new information is a general set of sentences. The main work can be divided into the following parts:(1) The thesis gives the construction of package-kernel contraction and its postulates and proof of corresponding representation theorem.(2) It gives representation theorems of saturated package-kernel contraction and smooth package-kernel contraction, and proves that the two contraction operations coincide with each other. It discusses the relation between smooth package-kernel contraction and package contraction basing on partial meet contraction.(3) It constructs the choice-kernel contraction and gives corresponding postulates and representation theorem, and discusses sub kernel contraction and its representation theorem.更多还原