Comments on "Application of FFT and the conjugate gradient method for the solution of electromagnetic radiation from electrically large and small conducting bodies"

In the original paper, (ibid., vol.AP-34, no.5, p.635-40, 1986) T.K. Sarkar et al. describe how the fast Fourier transform (FFT) and the conjugate gradient method (CGM) can be used to efficiently make use of the convolutional form of the electric field integral equation for straight wire antennas. The commentators agree that the CGM converges more rapidly than the previously used method of steepest descent and the spectral iteration technique. They also point out that experience with the dyadic Green's function encountered in the two-dimensional transverse electric (TE) and three-dimensional equations for dielectric bodies is not so encouraging. It is found that for these cases the CGM convergence can be quite slow, degrading the efficiency of the algorithm. In addition, the solution obtained using the pulse-basis point-matching method to discretize the integral equation are very inaccurate for the two-dimensional TE and three-dimensional cases. A reply to these comments from the authors of the original paper is included