THE EFFECTIVE SPECTRA OF A MATRIX

Prof. Anne Greenbaum
Department of Mathematics
University of Washington
Monday , March 19th, 103 Schaeffer Hall, 10:30 am - 11:20 am

ABSTRACT:

Let A be a given n by n matrix. In order to estimate ||f(A)|| for various functions f, it is helpful is one can associate A with some set in the complex plane and relate ||f(A)|| to the size of f on this set. For normal matrices, an appropriate set is the spectrum of A, but for nonnormal matrices it is less clear what (if any) set(s) in the complex plane should be associated with A. For a given positive integer k, we identify the largest set Omega in the complex plane satisfying ||p(A)|| >= max_{z in Omega} |p(z)| for all polynomials p of degree k or less. We refer to this as the effective spectrum of A for degree k, since, under the action of kth degree polynomials, A behaves like a normal matrix with eigenvalues throughout this set (or, at least, ||p(A)|| is greater than or equal to ||p(B)|| for any such normal matrix B). For k=1, this set is just the field of values of A, and for k >= m, where m is the degree of the minimal polynomial of A, it is the spectrum of A. For 1 < k < m, these sets are intermediate between the field of values and the spectrum and sometimes closely resemble pseudospectra.

 

  Prof. Anne Greenbaum is currently a Professor in the Math Dept. at the University of Washington. She received her PhD from UC Berkeley in 1981, worked at Lawrence Livermore National Lab from 1974 to 1986, and was at the Courant Institute from 1986-1997. She is ther author of a book entitled "Iterative Methods for Solving Linear Systems," published by SIAM.