Solution of Linear Generalized Eigenvalue Problems Containing Singular Matrices
Alan Jennings, J.R. Halliday, M. J. COLE · IMA Journal of Applied Mathematics · 1978
A general method is presented for the solution of the linear generalized eigenvalue problem Ax = λBx, where matrices A and B may both be singular but ‖A — λB‖ ≢ 0. The procedure can easily be implemented on a digital computer using a Gaussian elimination algorithm with pivoting together with a QR eigensolution package. Results are given for two problems with sensitive eigenvalues.