A fast algorithm to compute the wave-scattering solution of a large strip (Q686614): Difference between revisions

In many applications the scattering of waves by large objects and structures is an important problem. Computer solutions of large structures consume, however, exorbitant amounts of computer time and computer memory. The reduction in computational complexity can be achieved by the application of translation matrices which are used to change the coordinates of the wave functions describing the scattered field. In this paper the use of a plane-wave basis for the expansion of the scattered field by a finite, nonuniform array of strips is described. This is equivalent to the use of plane waves as a basis for the representation of the translation group, which reduces the translation matrices to diagonal matrices which are easy to compute. It leads to an algorithm with a computational complexity of \(N\log^ 2 N\) for one incident angle and \(N^ 2\log N\) complexity for \(N\) incident angles, with memory requirements of \(N\log N\) and \(N^ 2\) respectively, for the above two cases. This approach is much faster than any other method to solve this scattering problem. The algorithm described in this paper has two advantages: (i) it can solve for the scattering solution from any array of coplanar strips regardless of the nonuniformity of the array, (ii) it has a reduced computational complexity compared to previous methods of solving this problem. This work can be generalized to the case of an array of strips on a dielectric slab.