A Large-Update Primal-Dual Interior-Point Method for Second-Order Cone Programming

摘要

A large-update primal-dual interior-point algorithm is presented for solving second order cone programming. At each iteration, the iterate is always following the usual wide neighborhood N_∞(r), but not necessary staying within it. However, it must stay within a wider neighborhood N(t,β). We show that the method has O{yrrL) iteration complexity bound which is the best bound of wide neighborhood algorithm for second-order cone programming.