Thursday, July 16

MS51
Symbolic-Numeric Algorithms for Polynomials (Part I of II)

10:30 AM-12:30 PM
Room: Sidney Smith 1073

This minisymposium is intended to present a timely update to the emerging understanding of problems involving multivariate polynomials with inexactly- known coefficients. Such polynomial problems arise, for example, when physical measurements or numerical computations are used to specify a polynomial system. In particular, applications to computer-aided design are important already, and one expects a very wide array of applications to become important in the future.

One of the goals of holding this minisymposium is to stimulate interest in the numerical analysis community. While there is widespread activity in the computer algebra community, with an upcoming volume of the Journal of Symbolic Computation devoted to the topic, there are as yet only a handful of numerical analysts involved. Since the discovery that some important multivariate polynomial problems can and should be phrased instead as eigenvalue problems for nearly-commuting families of very large sparse matrices, it has become clear that there is a significant role for numerical analysis to play.

See Part II, MS66.

10:30 When Are Two Numeric Polynomials Relatively Prime?

George Labahn, University of Waterloo, Canada

11:00 Multivariable Elimination Using Dixon Resultants

Deepak Kapur, State University of New York, Albany; and Tushar Saxena, General Electric Corporate Research and Development

11:30 On the Factorization of Approximate Multivariate Polynomials

Andre Galligo, Université de Nice, France; and Stephen M. Watt, Organizer

12:00 Solving Algebraic Systems Using Multi-Polynomial Resultants: Algorithms, Implementations, and Applications

Dinesh Manocha and Shankar Krishnan, University of North Carolina, Chapel Hill