宏观到微观模型范式及其应用

Macro-to-Micro Algorithm Model and its Application

宏观到微观模型源于粒计算的思想。该模型的数据结构用O(n)时间建成,并具备高度的并行性,足够的处理器可使之在O(1)时间内建成(n为点集规模)。由于插入、删除、查询等操作都在常数时间内完成,且不会引起树结构不平衡,因此数据结构具有良好的动态性。此外,M2M模型的数据结构及其预处理过程,能够被所有基于M2M模型的算法所共享,从而大大地提高了需要多种算法共同处理的操作效率。实验结果表明,基于该模型的最近邻算法和凸包算法较之对应的传统算法有很大优势。

Macro-to-Micro Model is similar to the idea of Granular Computing. The Macro-to-Micro data structure can be built in O（n） time. It can also be built in 0（1） time by parallel technology. Moreover, dynamic insertion, deletion and query operation can be completed in constant time without breaking the balance of tree. This data structure and preproeessing operation can be shared with Macro-to-Micro algorithms, so that we can greatly improve the efficiency of the multi-operation problem. Experimental results show that the nearest neighbour searching and convex hull algorithm based on Macro-to-Micro model has great advantage than traditional algorithm.