基于区间矩阵的时态关系的简化方法

用区间代数能表示不确定的时态关系,可以方便地应用于时态推理,表达能力强。将区间代数及区间矩阵中有端点相等的基本关系去掉,只保留现实世界中真正可能发生的基本关系,大大降低了问题的复杂性,然后,对每个原子矩阵建立一个特征集,用特征集的并运算来取代矩阵的复合运算,提高了计算效率。

Simplification of Temporal Relation Based on Interval Matrix

Temporal representation and reasoning are two important research areas in the artificial intelligence.It is not only widely used in the logical basic research but also in the knowledge system.Interval algebra(IA) is a temporal theory,which is independent on the specific applications.Uncertain temporal relations can be represented and reasoned by IA,which possesses strong expressive ability and can represent all the natural temporal relations.By omitting the basic relations of interval algebra and interval matrix where the intervals have at least one same endpoint and keeping those basic relations which is possible to happen in the real world,the expressive ability is still very strong,but the problems become much more simple.By assigning an eigenvalue set to each atomic matrix and using the union of eigenvalue sets to substitute the composition of matrices,the calculation efficiency can be improved.