The Application of ACA and SVD Algorithms in the Matrix Low-Rank Approximation

Since the adaptive cross approximation (ACA) and the singular value decomposition (SVD) are unsuited for directly decomposing the large-scale matrices, this paper investigates the accelerated algorithm and the accelerated hybrid algorithm for low-rank decomposition of matrices with the scale and elastic wave problems, respectively. The results show that the accelerated algorithm can significantly reduce the low-rank decomposition time for the coefficient matrix of scale wave, and the accelerated hybrid algorithm can effectively compress the tensor form matrix of elastic wave.