Abstract
Traditional skyline queries consider data points with static attributes, e.g., the distance to a beach of a hotel. However, some attributes considered for a skyline query can be dynamic, e.g., the distance to a beach of a moving vehicle. Consider a scenario as follows. A person in a moving vehicle issues a skyline query to find restaurants, taking into account the static attributes of the restaurants as well as the distance to them. This scenario motives us to address a novel skyline query considering both static and dynamic attributes of data points. Given a data set D (e.g., restaurants) with a set of static attributes in a two-dimensional space, a query line segment l (e.g., the route for driving), and a distance constraint of r (for choosing a restaurant), we want to find out the skylines along l considering the static and dynamic attributes of the data points, satisfying the location constraint of r. We propose two methods to solve the problem. In the first method, we find some special location points which partition l into sub-segments and also make the skylines in the adjacent sub-segments different. In the second method, we apply some properties to identify data points which need not be considered for computing skylines. Moreover, to reduce the number of sub-segments, we propose an approximate method to compute skylines and define a similarity function to measure the similarity between the exact and approximate results. A series of experiments are performed to evaluate the exact methods and the results show that the second method is more efficient. We also perform experiments to compare the exact and approximate results and it shows that there is a trade-off between the reduction of the number of sub-segments and the accuracy of the results. © 2014 Springer International Publishing Switzerland.