| 摘要: |
| 文章针对现有的定性表示方法和拓扑推理算法存在的问题,提出了一种新的方法.首先,提出了基于概念邻域结构的定性表示方法.然后,给出了不同粒度层次上拓扑关系复合表的计算方法.最后,设计得出分层递阶的拓扑推理算法.文章给出的方法具有较高的认知合理性,所提出的推理算法可根据问题来选择合适的表示和推理层次,在已有推理算法给不出解的情况下,可以给出问题的合理解,对一般定性推理研究有参考价值. |
| 关键词: 拓扑推理,概念邻域结构,复合表,分层递阶,定性空间推理. |
| DOI： |
| 分类号: |
| 基金项目:本文研究得到清华大学智能技术与系统国家重点实验室开放课题基金资助. |
|
| Hierarchical Method for Qualitative Topological Inference |
|
LIAO Shi-zhong,SHI Chun-yi
|
| Abstract: |
| A new method is proposed to solve the problems in existing qualitative represent ation methods and topological inference algorithms. First, a qualitative represe ntation mechanism based on conceptual neighborhood is described. Then, a computa tion procedure for composition tables in different topological relation levels i s put forward. Finally, a hierarchical topological inference algorithm is demons trated. The algorithm can select a suitable representation and inference level f or a given problem, and can work out a reasonable solution when other algorithms fail to give answers. The hierarchical method proposed in this paper is cogniti vely valid, and is helpful for general qualitative reasoning research. |
| Key words: Topological inference, conceptual neighb orhood, composition table, hierarchy, qualitative spatial reasoning. |
