タイトル | A hybrid Pade-Galerkin technique for differential equations |
本文(外部サイト) | http://hdl.handle.net/2060/19940008816 |
著者(英) | Andersen, Carl M.; Geer, James F. |
著者所属(英) | Institute for Computer Applications in Science and Engineering |
発行日 | 1993-07-01 |
言語 | eng |
内容記述 | A three-step hybrid analysis technique, which successively uses the regular perturbation expansion method, the Pade expansion method, and then a Galerkin approximation, is presented and applied to some model boundary value problems. In the first step of the method, the regular perturbation method is used to construct an approximation to the solution in the form of a finite power series in a small parameter epsilon associated with the problem. In the second step of the method, the series approximation obtained in step one is used to construct a Pade approximation in the form of a rational function in the parameter epsilon. In the third step, the various powers of epsilon which appear in the Pade approximation are replaced by new (unknown) parameters (delta(sub j)). These new parameters are determined by requiring that the residual formed by substituting the new approximation into the governing differential equation is orthogonal to each of the perturbation coordinate functions used in step one. The technique is applied to model problems involving ordinary or partial differential equations. In general, the technique appears to provide good approximations to the solution even when the perturbation and Pade approximations fail to do so. The method is discussed and topics for future investigations are indicated. |
NASA分類 | NUMERICAL ANALYSIS |
レポートNO | 94N13289 NASA-CR-191494 NAS 1.26:191494 ICASE-93-38 AD-A269127 |
権利 | No Copyright |
|