動作環境

GeForce GTX 1070 (8GB)
ASRock Z170M Pro4S [Intel Z170chipset]
Ubuntu 16.04 LTS desktop amd64
TensorFlow v1.2.1
cuDNN v5.1 for Linux
CUDA v8.0
Python 3.5.2
IPython 6.0.0 -- An enhanced Interactive Python.
gcc (Ubuntu 5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609
GNU bash, version 4.3.48(1)-release (x86_64-pc-linux-gnu)

A presentation given at International Symposium on Electromagnetic Theory (EMTS 2016),Espoo, Finland, 17.08.2016.
https://www.researchgate.net/publication/307477398_Slides_EMTS-2016_iter

p8

asymptotic behavior of N_iter for large scatterers

p9


QMR, QMR2, Bi-CG	complex symmetric
Bi-CGStab, Bi-CGStab(2)	general CG-type method
CGNR, CSYM	solve normalized equations

p26 Estimate of N_iter
for large sizes

N_{iter} \approx \frac{15 ln 10 \rho(Z)}{4 Im(1/(1-\epsilon))}

p28 Estimate of the spectral radius

\rho(Z) \approx 4.5 + 0.14 x

The graph is drawn for x ranging from 10 to 100.

link

number of iterationve becoms large for QMR for specific Re{m} and Im{m} (e.g., m=1.5 + 0.001i)

It gives several contour plots, similar to your ones, and provides a few analytical formula, but they are applicable only to either very small or very large particles. In your "resonant" case of particles comparable to the wavelength it is hard to predict anything, apart from the general trends - #ite inreases with size and refractive index, and almost real or imaginary refractive indices are the toughest.

ADDA + Matplotlib > scatterplot of [the number of iterations] for various values of Re{m} and Im{m}, for different iterative solvers (e.g., bicgstab2, bicg, QMR, etc)