Linear Algebra for Large Nonlinear Algebra Over the Reals (Short Abstract)

J. Maurice Rojas (MIT)

We point out some of the gaps between computational algebraic geometry and numerical linear algebra, and then state some results bridging these gaps. More precisely, we describe some recent efficient methods for reducing polynomial system solving to large linear algebra. One application of these techniques is a new fast algorithm for counting the number of real roots of a polynomial system. Structured matrices, arising from the use of sparse resultants, figure prominently.


The EXTENDED version, in PostScript format, can be found by clicking here.