Eigenvalue condition numbers: zero-structured versus traditional. (Q2570035)

The authors study the relationships between the classical theory of conditioning and the theory of the zero-structured conditioning and they derive formulae and relationships for the absolute and relative zero-structured conditioning based on the Frobenius norm, by considering matrix perturbation. Then an algorithm is given in order to compare the zero-structured individual condition numbers of a set of simple eigenvalues with the traditional ones. Numerical cases, such as Lesp matrix, Bessel matrix, Wilkinson matrix and Frank matrix (Hessenberg matrix), are presented in order to demonstrate the applicability of the algorithm providing interesting information about the eigenvalue sensitivity, when the perturbations in the matrix have an arbitrary assigned zero-structure. Further investigation about patterned matrices, i.e. Toeplitz and Hankel matrices, will be presented in the near future.