Abstract Location Based Services are increasingly becoming popular due to increased usage of mobile devices by citizens seeking information on points-of-interest, travel routes, traffic conditions etc. We consider practical variants of the nearest neighbor problem on road networks, wherein the goal is to find the nearest point-of-interest from a query location. Here the notion of proximity is determined by the ease of reaching the point of interest via public transport. Using graphs modeling the road network and transport connectivity, efficient algorithms are presented. Keywords: Nearest Neighbor Search, Computational Geometry, Road Networks, Shortest Path Problem, Location Based Services Transport Optimized Nearest Neighbor Query for Location Based Services Debajyoti Ghosh * , Prosenjit Gupta ** * NIIT University, Neemrana, Rajasthan, India. Email: debajyoti.ghosh@niituniversity.in ** NIIT University, Neemrana, Rajasthan, India. Email: prosenjit_gupta@acm.org Article can be accessed online at http://www.publishingindia.com as Computational Geometry, Pattern Recognition, Interpolation, Probability Theory, Data Mining, Machine Learning and Geographical Information Systems. The basic problem is defned as follows: Given a set of point S and a query point q, the goal is to fnd a point in S nearest to q. In Computational Geometry [1], this is usually studied as a repetitive mode query problem wherein the goal is to preprocess S into a data structure so that given q, we can fnd p ε S nearest to q effciently. Knuth [2] christened this as the “Post Offce Problem” in reference to the age-old problem of locating the nearest post-offce to a query user. In the 2-dimensional Euclidean plane, the solution to the post-offce problem is the Voronoi diagram along with an accompanying planar point location data structure [3]. In this paper, we consider a variant of the nearest neighbor problem which is relevant in the context of Location Based Services and smart cities. We consider a road network, serviced by public transport (e.g. bus, train, metro, auto-rickshaw etc.). There are several points-of- interest (POI’s) spread on the road network. There are also bus stops (or train stations) located at various points on the road network. We are given a set of bus routes serving the road network, each route being an ordered sequence of bus stops. A mobile user, currently at a query location q on this road network wishes to travel to the nearest POI. The user is interested in using public transport (say buses) to reach his/her destination. Given that bus transfers are often inconvenient, lead to delays or increased cost of travel, the user wishes to minimize the number of transfers. However since number of paths to reach a POI using the minimum number of bus transfers may often not be unique, the user may wish to reach the POI using that route that is the shortest among all routes using the minimum number of transfers. We assume that the user will walk to the nearest bus stop from q and also Introducton Location Based Services (LBS) [4][54][55] are increasingly becoming popular due to increased usage of mobile devices by citizens seeking information on points-of-interest, travel routes, traffc conditions etc. LBS aim at providing customized information keeping location as an important factor. Location-based Services include emergency services, navigation and information services, advertising services and tracking services to name a few. Due to rapid urbanization and smart city projects in various countries including India, interest in LBS has further increased. According to market research, the global LBS market is poised to grow signifcantly in the next few years. This has created several challenges leading to research in new technologies, architectures and algorithms [4]. Nearest Neighbor Search [41][43][48][52] and its variants have been well studied in areas as diverse