Efficient Computation of the Skyline Cube

Skyline has been proposed as an important operator for multi-criteria decision making, data mining and visualization, and userpreference queries. In this paper, we consider the problem of efficiently computing a Skycube, which consists of skylines of all possible non-empty subsets of a given set of dimensions. While existing skyline computation algorithms can be immediately extended to computing each skyline query independently, such “shared-nothing” algorithms are inefficient. We develop several computation sharing strategies based on effectively identifying the computation dependencies among multiple related skyline queries. Based on these sharing strategies, two novel algorithms, Bottom-Up and Top-Down algorithms, are proposed to compute Skycube efficiently. Finally, our extensive performance evaluations confirm the effectiveness of the sharing strategies. It is shown that new algorithms significantly outperform the näıve ones.