Convergence of the discrete dipole approximation. II. An extrapolation
technique to increase the accuracy
Maxim A. Yurkin
Valeri P. Maltsev
Alfons G. Hoekstra
Maxim A. Yurkin update to 2008-07-29
https://arxiv.org/abs/0704.0035
We propose an extrapolation technique that allows accuracy improvement of thediscrete dipole approximation computations. The performance of this techniquewas studied empirically based on extensive simulations for 5 test cases usingmany different discretizations. The quality of the extrapolation improves withrefining discretization reaching extraordinary performance especially forcubically shaped particles. A two order of magnitude decrease of error wasdemonstrated. We also propose estimates of the extrapolation error, which wereproven to be reliable. Finally we propose a simple method to directly separateshape and discretization errors and illustrated this for one test case.