Extrema of Elementary Symmetric Polynomials of the Eigenvalues of the Matrix P*KP+L

摘要

The report derives extrema of elementary symmetric polynomials of a matrix P*KP+L, where K and L may be considered normalized noise and signal covariance matrices, respectively. The matrix P is a right-orthonormal, rectangular matrix transformation that may be applied to a normalized sampled input signal. The polynomials are generalizations of specific error and information content criteria in analog communications and may serve as mixed criteria. (Author)