Group Nearest Neighbor Queries for Fuzzy Geo-Spatial Objects

Abstract

A geo-spatial object with non-deterministic boundaries and compositions is commonly known as a fuzzy geo-spatial object. The advancement of data capturing devices such as sensors and satellite imaging technologies enable us to identify fuzzy geo-spatial objects from a large and complex image of an area. The nearest neighbor (NN) query processing on fuzzy objects, which finds the nearest fuzzy object to the given query point, has been addressed recently. In this paper, we envision a new set of applications that require finding the nearest fuzzy geo-spatial object for a group of fuzzy geo-spatial query objects. For example, when an oil spill occurs at a sea, the primary concern of an emergency response planner is to find an environmentally sensitive area, e.g., port or harbor, that will be affected the most by the oil spill. To support such applications, in this paper, we propose a new query type, called a fuzzy group nearest neighbor (FGNN) query. Given a set of fuzzy geo-spatial data objects, and a group of fuzzy geo-spatial query objects, an FGNN query returns a fuzzy geo-spatial object that minimizes the aggregate distance to the group. To solve FGNN queries, we develop an efficient technique in this paper. Our extensive experimental study reveals the efficacy and efficiency of our proposed technique.