Diagonalizing the Simple GA Mixing Matrix
Gary J. Koehler
Department of Decision and Information Sciences
College of Business Administration
University of Florida
ABSTRACT
The simple GA mixing matrix, M, is diagonalized by a congruence
transformation involving a lower-triangular matrix, L. L is sparse.
Several applications using L are given. One application gives the rank
of M under various choices of GA parameter values. A second
application shows how L might prove useful in studying the GA Fixed
Point Problem. In this problem L is used to change a quadratic
equation having l(l+1)/2 terms to one having no cross terms and only
2l squared terms.