Research On Uncertainty Temporal Information In Temporal Databases

Time is a very important factor while researching on the changing world. We always explain all kinds of realities and data by time, because the potential relationships among events and entities are often implicated in the temporal information. Up to now, most researches on temporal information focus on certainty temporal information. However, it is hard to determine the precise time of events in most practical situations. It is necessary to build an uncertainty temporal representation model, which can be used to describe uncertainty temporal information of events generally and conveniently.This dissertation was about uncertainty temporal representation model in theory. It focused on uncertainty temporal information management. Combined with the theory of possibility, a possibility theory-based uncertainty temporal representation model was presented. Based on this model, the representation and reasoning of uncertain relations between temporal information was presented. The contents of this dissertation included: Firstly, based on the analysis of temporal information management, a unified possibility theory-based temporal representation model was presented. The property and capability for temporal reasoning of this model was also analyzed; Secondly, the uncertain relations of temporal information were discussed, which can be represented with the possibility values denoting the probability of the basic relations defined by Allen. The reasoning with uncertain relations between the temporal information was represented. It can be used to solve transitivity problem of uncertain; Lastly, a Temporal Element Probabilistic Relation (TEPR) network was defined, which can represent the uncertain relations between the temporal information. Based on the TEPR network, the path consistency algorithm was proposed, witch can be used to solve the constraint satisfaction problem, as well as to test a TEPR network for consistency.