Yurkin Maxim A., Maltsev Valeri P., Hoekstra Alfons G.

31 Mar 2007 physics.optics, physics.comp-ph arxiv.org/abs/0704.0035

Abstract. We propose an extrapolation technique that allows accuracy improvement of the discrete dipole approximation computations. The performance of this technique was studied empirically based on extensive simulations for 5 test cases using many different discretizations. The quality of the extrapolation improves with refining discretization reaching extraordinary performance especially for cubically shaped particles. A two order of magnitude decrease of error was demonstrated. We also propose estimates of the extrapolation error, which were proven to be reliable. Finally we propose a simple method to directly separate shape and discretization errors and illustrated this for one test case.

There are no reviews yet.