Comments on the accuracy of the "Ludwig" integration algorithm

The accuracy of the so-called Ludwig algorithm for solution of integrals arising from electromagnetic radiation problems, first reported by the author in 1968 (ibid., vol.AP-16, p.767-9, Nov. 1968), is shown to be better than results reported by others for the case of the pattern of a uniformly illuminated aperture. It is argued that the accuracy is very similar to results obtained using a Romberg-Simpson technique, and computer speed is superior if more than 4-5 sidelobes are computed. A comparison is also made with very recent results, and it is shown that the Ludwig algorithm's accuracy is very competitive with the best results obtained with alternative techniques. The Ludwig algorithm does not exhibit the catastrophic failure shown by the alternative techniques and has the advantage of accepting data at arbitrary spacings.<
>