Victor Valls, Panagiotis Promponas, et al.
IEEE Communications Magazine
We show that the block principal pivot algorithm (BPPA) for the linear complementarity problem (LCP) solves the problem for a special class of matrices in at most n block principal pivot steps. We provide cycling examples for the BPPA in which the matrix is positive definite or symmetric positive definite. For LCP of order three, we prove that strict column (row) diagonal dominance is a sufficient condition to avoid cycling. © 1997 Elsevier Science B.V.
Victor Valls, Panagiotis Promponas, et al.
IEEE Communications Magazine
Rafae Bhatti, Elisa Bertino, et al.
Communications of the ACM
Arun Viswanathan, Nancy Feldman, et al.
IEEE Communications Magazine
Michael C. McCord, Violetta Cavalli-Sforza
ACL 2007