IG-Tree: an efficient spatial keyword index for planning best path queries on road networks

Abstract

Due to the popularity of Spatial Databases, many search engine providers have started to expand their text searching capability to include geographical information. Because of this reason, many new queries on spatial objects affiliated with textual information, known as the Spatial Keyword Queries, have taken significant research interest in the past years. Unfortunately, most of existing works on Spatial Keyword Queries only focus on objects retrieval. There is barely any work on route planning queries, even though route planning is often needed in our daily life. In this research, we propose the Best Path Query, which we find the best optimum route from two different spatial locations that visits or avoids the objects that are specified by the textual data given by the user. We show that Best Path Query is an NP-Hard problem. We propose an efficient indexing technique, namely IG-Tree, and three different algorithms with different trade-offs to process the Best Path Queries on Road Networks. Our extensive experimental study demonstrates the efficiency and accuracy of our proposed approach.