Adaptive Block-Based Krylov Subspace Basis Functions for Solving Bistatic Scattering Problems.

This study aims to improve the efficiency of constructing basis functions for solving the electromagnetic scattering problem of objects using the method of moments combined with compressive sensing and Krylov subspace. To this end, a region decomposition method based on a clustering algorithm is proposed to accelerate the construction process of Krylov subspace basis functions. First, the midpoints of the common edges of triangular pairs are used to form a clustered dataset. Then, the initial clustering center is set, and the processes of clustering center updating and regional decomposition of the constructed dataset are completed iteratively. Finally, each subdomain is expanded according to the average distance from data points to the clustering center to ensure the continuity of currents. The numerical computation results show that the proposed method can achieve significant time efficiency. [ABSTRACT FROM AUTHOR]