Using multiple-precision arithmetic to prevent low-frequency breakdowns in the diagonalization of the green's function

2014-08-28
Multiple-precision arithmetic (MPA) is used to prevent low-frequency breakdowns in the diagonalization of the Green's function that is required to implement the multilevel fast multipole algorithm (MLFMA). The breakdown problem is considered at a numerical level, where rounding errors are reduced by increasing the precision as much as required. Using MPA seems to provide a direct solution to low-frequency breakdowns of the standard diagonalization, which may lead to straightforward implementations of broadband MLFMA.

Suggestions

Low-frequency multilevel fast multipole algorithm using an approximate diagonalization of the Green's function
Ergül, Özgür Salih (2014-08-23)
We present an approximate diagonalization of the Green's function to implement a stable multilevel fast multipole algorithm (MLFMA) for low-frequency problems. The diagonalization is based on scaled spherical functions, leading to stable computations of translation operators at all distances and for all frequencies. Similar to the conventional diagonalization, shift operators are expressed in terms of complex exponentials, while radiated and incoming fields are expanded in terms of scaled plane waves. Even ...
Stabilization of the Fast Multipole Method for Low Frequencies Using Multiple-Precision Arithmetic
Karaosmanoglu, Bariscan; Ergül, Özgür Salih (2014-08-23)
We stabilize a conventional implementation of the fast multipole method (FMM) for low frequencies using multiple-precision arithmetic (MPA). We show that using MPA is a direct remedy for low-frequency breakdowns of the standard diagonalization, which is prone to numerical errors at short distances with respect to wavelength. By increasing the precision, rounding errors are suppressed until a desired level of accuracy is obtained with plane-wave expansions. As opposed to other approaches in the literature, u...
Rigorous Solutions of Large-Scale Scattering Problems Discretized with Hundreds of Millions of Unknowns
Guerel, L.; Ergül, Özgür Salih (2009-09-18)
We present fast and accurate solutions of large-scale scattering problems using a parallel implementation of the multilevel fast multipole algorithm (MLFMA). By employing a hierarchical partitioning strategy, MLFMA can be parallelized efficiently on distributed-memory architectures. This way, it becomes possible to solve very large problems discretized with hundreds of millions of unknowns. Effectiveness of the developed simulation environment is demonstrated on various scattering problems involving canonic...
Optimal Control of Diffusion Convection Reaction Equations Using Upwind Symmetric Interior Penalty Galerkin SIPG Method
Karasözen, Bülent; Yücel, Hamdullah (2012-05-01)
We discuss the numerical solution of linear quadratic optimal control problem with distributed and Robin boundary controls governed by diffusion convection reaction equations. The discretization is based on the upwind symmetric interior penalty Galerkin (SIPG) methods which lead to the same discrete scheme for the optimize-then-discretize and the discretize-then-optimize.
Benchmark Solutions of Large Problems for Evaluating Accuracy and Efficiency of Electromagnetics Solvers
Gurel, Levent; Ergül, Özgür Salih (2011-07-08)
We present a set of benchmark problems involving conducting spheres and their solutions using a parallel implementation of the multilevel fast multipole algorithm (MLFMA). Accuracy of the implementation is tested by comparing the computational results with analytical Mie-series solutions. Reference solutions are made available on an interactive website to evaluate and compare the accuracy and efficiency of fast solvers. We also demonstrate the capabilities of our solver on real-life problems involving compl...
Citation Formats
Ö. S. Ergül, “Using multiple-precision arithmetic to prevent low-frequency breakdowns in the diagonalization of the green’s function,” 2014, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/55826.