| 校金友,曹衍闯,文立华.快速小波边界元的矩阵后压缩方法[J].计算力学学报,2010,27(6):983~988 |
|
| 
码上扫一扫！ |
| 快速小波边界元的矩阵后压缩方法 |
| A-posteriori compression scheme for wavelet Galerkin BEM |
| 投稿时间：2008-11-14 修订日期：2009-11-24 |
| DOI：10.7511/jslx20106005 |
| 中文关键词: 小波边界元 稀疏矩阵 复杂度 矩阵压缩 |
| 英文关键词:wavelet Galerkin BEM sparse matrix complexity matrix compression |
| 基金项目:西北工业大学博士论文创新基金(CX200601);国家自然科学基金(10674109)资助项目. |
|
| 摘要点击次数: 1959 |
| 全文下载次数: 1574 |
| 中文摘要: |
| 介绍了一种基于传统边界元单元划分的小波 Galerkin 边界元法,该方法具有几乎线性(即 ON,N 为自由度)的求解复杂度。在准消失矩小波的框架下介绍了非标准型系数矩阵的压缩问题,提出了一种后压缩算法以降低小波边界元法的内存消耗。求解 Stokes 方程的算例表明,后压缩算法在保证结果收敛特性的情况下可以将系数矩阵的内存占用量降低 5 倍以上。 |
| 英文摘要: |
| Wavelet Galerkin boundary element method is introduced. We put our emphasis on the matrix compression using the quasi-vanishing moment wavelets. The fact that there are still a large number of entries with small values in the non-standard form matrix after the a-priori compression motivates us to study the a-posteriori compression, in order to further reduce the memory requirement in storing the compressed matrix. An algorithm for the a-posteriori compression is proposed. Numerical example concerning Stokes flow problem clearly shown that, with the a-posteriori compression, the memory requirement for non-standard form can be reduced to a factor larger than 5 while persevering the convergence rate of the underlying Galerkin scheme. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
