Selective Gram-Schmidt orthonormalization for conic cutting surface algorithms (Q2482686): Difference between revisions

The selective orthonormalization procedure for polyhedral relaxations of the convex feasibility problem of \textit{J.~E.~Mitchell} and \textit{S.~Ramaswamy} [J. Optim. Theory Appl. 125, No. 2, 431--451 (2005; Zbl 1071.90039)] is generalized to conic relaxations. An algorithm and its theoretical analysis, including the convergence, are given. Also the specialization of the algorithm to the cases of semidefinite programming and second order cone programming is present.