Aggregate k-nearest neighbors queries in time-dependent road networks

In this paper we present an algorithm for processing aggregate nearest neighbor queries in time-dependent road networks, i.e., given a road network where the travel time over an edge is time-dependent, a set of query points Q, a set of points of interest (POIs) P and an aggregate function (e.g., sum), we find the k POIs that minimize the aggregated travel time from the query points. For instance, considering a city's road network at a given departure time and a group of friends at different locations wishing to meet at a restaurant, the time-dependent aggregate nearest neighbor query, considering the sum function, would return the restaurant that minimizes the sum of all travel times to it. The main contribution of our work is the consideration of the time-dependency of the network, a realistic characteristic of urban road networks, which has not been considered previously when addressing aggregate nearest neighbor queries. Our approach is based on the ANNQPLB algorithm proposed by Htoo et al. and uses Hub Labels, proposed by Abraham et al., to compute optimistic travel times efficiently. In order to compare our proposal we extended the previously proposed ANNQPLB algorithm aimed at non-time dependent aggregate nearest neighbor queries, enabling it to deal with the time-dependency. Our experiments using a real road network have shown our proposed solution to be up to 94% faster than the temporally extended previous solution.