Two methods for the calculation of the degree of an approximate greatest common divisor of two inexact polynomials

The calculation of the degree d of an approximate greatest common divisor of two inexact polynomials f(y) and g(y) reduces to the determination of the rank loss of a resultant matrix, the entries of which are functions of the coefficients of f(y) and g(y). This paper considers this issue by describing two methods to calculate d, such that knowledge of the noise level imposed on the coefficients of f(y) and g(y) is not assumed. One method uses the residual of a linear algebraic equation whose coefficient matrix and right hand side vector are derived from the Sylvester resultant matrix S(f,g), and the other method uses the first principal angle between a line and a hyperplane, the equations of which are calculated from S(f,g). Computational results on inexact polynomials whose exact forms have multiple roots of high degree are shown and very good results are obtained. These results are compared with the rank loss of S(f,g) for the calculation of d, and it is shown that this method yields incorrect results for these examples.

[1]  Erich Kaltofen,et al.  Approximate factorization of multivariate polynomials via differential equations , 2004, ISSAC '04.

[2]  Petre Stoica,et al.  Common factor detection and estimation , 1997, Autom..

[3]  Stephen A. Vavasis,et al.  Solving Polynomials with Small Leading Coefficients , 2005, SIAM J. Matrix Anal. Appl..

[4]  Joab R. Winkler,et al.  The calculation of the degree of an approximate greatest common divisor of two polynomials , 2011, J. Comput. Appl. Math..

[5]  S. U. Pillai,et al.  Blind image deconvolution using a robust 2-D GCD approach , 1997, Proceedings of 1997 IEEE International Symposium on Circuits and Systems. Circuits and Systems in the Information Age ISCAS '97.

[6]  E. W. Ng Symbolic and Algebraic Computation , 1979, Lecture Notes in Computer Science.

[7]  S. Barnett Polynomials and linear control systems , 1983 .

[8]  Ben Liang,et al.  Blind image deconvolution using a robust GCD approach , 1999, IEEE Trans. Image Process..

[9]  I. Emiris,et al.  Certified approximate univariate GCDs , 1997 .

[10]  David S. Watkins,et al.  Fundamentals of matrix computations , 1991 .

[11]  S. Klasa,et al.  Polynomial scaling , 1990 .

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

[13]  Zhonggang Zeng Computing multiple roots of inexact polynomials , 2005, Math. Comput..

[14]  Stephen M. Watt,et al.  The singular value decomposition for polynomial systems , 1995, ISSAC '95.

[15]  Zhonggang Zeng,et al.  The approximate GCD of inexact polynomials , 2004, ISSAC '04.

[16]  S. Smale,et al.  The mathematics of numerical analysis , 1996 .

[17]  Donna Kastle Dunaway A composite algorithm for finding zeros of real polynomials , 1972 .

[18]  Dario Bini,et al.  Structured matrix-based methods for polynomial ∈-gcd: analysis and comparisons , 2007, ISSAC '07.

[19]  Joab R. Winkler,et al.  A non-linear structure preserving matrix method for the low rank approximation of the Sylvester resultant matrix , 2010, J. Comput. Appl. Math..