For many applications of location-based services, such as friend finding, buddy tracking,information sharing and cooperative caching in ad hoc networks, it is often important to be able to identify whether the positions of a given set of moving objects are within close proximity. To compute these kinds of proximity relations among large populations of moving objects, continuously available location position information of these objects must be correlated against each other to identify whether a given set of objects are in the specified proximity relation. In this dissertation, we state this problem, referring to it as the location constraint matching problem, both in the Euclidean space and the road network space. In the Euclidean space, we present an adaptive solution to this problem for various environments. We also study the position uncertainty associated with the constraint matching. For the road network space, where the object can only move along the edges of the road network, we propose an efficient algorithm based on graph partitioning, which dramatically restricts the search space and enhances performance. Our approaches reduce the constraint processing time by 80% for Euclidean space and by 90% for road network space respectively. The logical combination of individual constraints with conjunction, disjunction and negation results in more expressive constraint expressions than are possible based on single constraints. We model constraint expressions with Binary Decision Diagrams (BDD). Furthermore, we exploit the shared execution of constraint combinations based on the BDD modeling. All the algorithms for various aspects of the constraint processing are integrated in the research prototype L-ToPSS (Location-based Toronto Publish/Subscribe System). Through experimental study and the development of an analytical model, we show that the proposed solution scales to large numbers of constraints and large numbers of moving objects.

For many applications of location-based services, such as friend finding, buddy tracking,information sharing and cooperative caching in ad hoc networks, it is often important to be able to identify whether the positions of a given set of moving objects are within close proximity. To compute these kinds of proximity relations among large populations of moving objects, continuously available location position information of these objects must be correlated against each other to identify whether a given set of objects are in the specified proximity relation. In this dissertation, we state this problem, referring to it as the location constraint matching problem, both in the Euclidean space and the road network space. In the Euclidean space, we present an adaptive solution to this problem for various environments. We also study the position uncertainty associated with the constraint matching. For the road network space, where the object can only move along the edges of the road network, we propose an efficient algorithm based on graph partitioning, which dramatically restricts the search space and enhances performance. Our approaches reduce the constraint processing time by 80% for Euclidean space and by 90% for road network space respectively. The logical combination of individual constraints with conjunction, disjunction and negation results in more expressive constraint expressions than are possible based on single constraints. We model constraint expressions with Binary Decision Diagrams (BDD). Furthermore, we exploit the shared execution of constraint combinations based on the BDD modeling. All the algorithms for various aspects of the constraint processing are integrated in the research prototype L-ToPSS (Location-based Toronto Publish/Subscribe System). Through experimental study and the development of an analytical model, we show that the proposed solution scales to large numbers of constraints and large numbers of moving objects.