介绍了一种新型的多维数据空间放置算法——SMDPA.该算法使用数据超方体的先验被访问概率以及访问之间的相似度放置超方体.即使在数据超方体的被访问频率不满足均匀分布的情况下，该算法可以有效地放置超方体.模拟结果证明，该算法较传统算法有更良好的性能.

This paper describes a new multidimensional data-space placement algorithm, SMDPA. The algorithm can place the data hyper-cubes efficiently even they have different accessed frequencies. The algorithm uses the data hyper-cubes' prior accessed frequency and their similarities to get the accessed relation between them. The simulation results make clear that the SMDPA algorithm has better performance than traditional algorithms.

This project is supported by the National Natural Science Foundation of China under Grant No.69673046 (国家自然科学基金).