Nearest Group Queries with Non-Spatial Constraints

A k nearest group (kNG) search finds k closest groups of objects from multiple data sources, where each group consists of one object from each data source. In this paper, we extend the kNG query to the kNG query with non-spatial constraints, which finds k groups of objects that are closest to q with each group containing one object o from each data source, where o satisfies the additional non-spatial constraints. We present two strategies, namely the aR-tree search in a Depth First Search (DFS) fashion and that in a Best First Search (BFS) fashion. We compare the performances of DFS and BFS strategies and show that the BFS strategy archives better performance than the DFS strategy for answering kNGwNS queries.