A Combined Skyline Algorithm Based on Quickhull and BNL

Skyline computation has been widely used in many research areas such as database visualization, data mining and multi-criteria decision. In this paper, we focus on a simple and effective skyline algorithm, namely, Block Nested Loop (BNL) since it has two advantages, i.e. simplicity and universality. In order to improve the performance of BNL for processing massive data, we develop a combined skyline algorithm based on Quickhull and BNL. Firstly, all supported non-dominated points are computed with Quickhull algorithm. Then the remaining unsupported non-dominated points are computed using the original BNL algorithm. We test the algorithm on three different types of randomly produced data sets and compare it with the original BNL algorithm. Experimental result shows that our proposed algorithm performs better than the original BNL algorithm in all the test data sets.