earticle

영어

This study proposed a quasi-order-based temporal data structure (QOTDS) which differed from conventional, algebraic data management models. Based on this QOTDS, a temporal data index called the temporal quasi-order index (TQOindex) was established. Firstly, the study proposed the concepts of temporal quasi-order (TQO) and linear order partitioning (LOP) of time period sets and discussed the construction algorithm of LOP and the optimum (minimum) properties. On this basis, a temporal data structure was established based on LOP. This structure realized the set-at-a-time data operation-like relational data structure and improved the inquiry efficiency by using multiple threads. Subsequently, in the structural framework of TQO, we discussed the temporal data index (TQOindex) based on quasi-order extensions. This index was effectively applicable to various conventional database platforms depending on the disk (external memory)-based data management and also to big data dynamic index technology relying on the incremental updating mechanism. Finally, a corresponding experimental simulation and comparative evaluation were designed to verify the feasibility and effectiveness of TQOindex. Research and experiments showed that QOTDS were effective at temporal inquiry and management in cases involving the temporal processing and integration mechanisms in new data, such as semantic data, XML data, and moving object data.