DELAUNAY剖分并行算法及其在计算机机群上的实现

Parallel Algorithm for Delaunay Triangulation and Its Implementation on Computer Clusters

基于CORBA技术设计并实现了一种Delaunay剖分的并行算法,该方法可以保证稳定性、高效性、穿行代码可100%复用。Delaunay剖分是并行网格生成的重要方法,该方法在并行网格生成中减少了通信量,并且保证了与串行算法相同的精确度,而且随着计算量的增加,线性加速比增大,并行的优越性表现得越明显。实验结果表明,该算法在机群环境下可以有效地提高Delaunay剖分的计算速度,减少执行时间。

A parallel algorithm for delaunay triangulation was designed and implemented by using CORBA （ Common Object Request Broker Architecture） technology. This algorithm has the characters of stability and high performance, meanwhile the walking through code can be reused at 100 percent. The delaunay triangulation is an important way to generate the simultaneous treatment grid, reduce communication traffic and guarantee accuracy rating in accordance with sequential algorithm. Furthermore, with the accrescence of computational complexity and the augmentation of linear speedup, the superiority of parallel algorithm appeared more obviously. Result indicates that the algorithm can efficiently elevate the calculation speed of delaunay triangulation and reduce the time of operation.