詳細資料
04/17(五)邀請李旺謙教授來系演講,請大家踴躍參加
各位老師、同學: Location-Dependent Spatial Queries (LDSQ)s, which search for data objects based on users' locations, are important for many location-based applications. Many existing research works consider Euclidean distance in determining object proximities. In practice, however, these proximities are constrained by underlying spatial networks, e.g., road networks. Therefore, there is a need to take into consideration the network distance in the course of spatial object searches on road networks. In this talk, I will present ROAD, an efficient and scalable framework to support fast evaluation of LDSQs on a road network. Conceptually different from related research studies, ROAD exploits a network space pruning technique that facilitate LDSQs such as range search and nearest neighbor search. To address two costly search operations, namely, the network traversal and object lookup, ROAD consists of two components, Route Overlay and Association Directory, respectively. To achieve better scalability, ROAD organizes a road network as a hierarchy of regional sub-network (Rnets) that abstract both network topology and object presence. This talk will cover (1) the design and implementation of the ROAD framework, (2) efficient search algorithms for various queries, and (3) incremental update techniques for framework maintenance in presence of object and network changes. Extensive experiments have been conducted with real road networks to evaluate ROAD. The experiment result shows the superiority of ROAD over the state-of-the-art approaches. Wang-Chien Lee is currently on sabbatantical leave to the Academic Sinica, Taiwan. He is an Associate Professor of Computer Science and Engineering at Pennsylvania Email address: wlee@cse.psu.edu
相關網址:無 |