论文信息 - Symmetric centrosymmetric matrix-vector multiplication

Symmetric centrosymmetric matrix-vector multiplication

Abstract We present a method for the multiplication of an arbitrary vector by a symmetric centrosymmetric matrix, requiring 5 4 n 2 + O (n) floating-point operations, rather than the 2n 2 operations needed in the case of an arbitrary matrix. Combining this method with Trench's algorithm for Toeplitz matrix inversion yields a method for solving Toeplitz systems with the same complexity as Levinson's algorithm.

Aaron Melman

[1] W. F. Trench. An Algorithm for the Inversion of Finite Toeplitz Matrices , 1964 .

[2] James R. Bunch,et al. Stability of Methods for Solving Toeplitz Systems of Equations , 1985 .

[3] J. L. Hock,et al. An exact recursion for the composite nearest‐neighbor degeneracy for a 2×N lattice space , 1984 .

[4] W. Gragg,et al. The generalized Schur algorithm for the superfast solution of Toeplitz systems , 1987 .

[5] James Durbin,et al. The fitting of time series models , 1960 .

[6] Philippe Delsarte,et al. The split Levinson algorithm , 1986, IEEE Trans. Acoust. Speech Signal Process..

[7] N. Levinson. The Wiener (Root Mean Square) Error Criterion in Filter Design and Prediction , 1946 .

[8] Gene H. Golub,et al. Matrix computations , 1983 .