Retrieving routes of interest over road networks

Abstract

In this paper, we propose route of interest (ROI) query which allows users to specify their interests with query keywords and returns a route such that (i) its distance is less than a distance threshold and (ii) its relevance to the query keywords is maximized. ROI query is particularly helpful for tourists and city explorers. For example, a tourist may wish to find a route from a scenic spot to her hotel to cover many artware shops. It is challenging to efficiently answer ROI query due to its NP-hard complexity. Novelly, we propose an adaptive route sampling framework that adaptively computes a route according to a given response time, and gradually improve the quality of the route with time. Moreover, we design a suite of route sampling techniques under this framework. Experiments on real data suggest that our proposed solution can return high quality routes within a short response time.