A new self-dual embedding method for convex programming (Q557368)

The aim of this paper is to present a self-dual embedding method for solving conic optimization problems of type \[ \roman{(CCP)}\quad \min\{\bar c^T \bar x \mid \bar A \bar x= \bar b, \bar x \in K\}, \] obtained by reformulating an original standard convex programming problem \[ \roman{(CP)}\quad \min\{c^Tx \mid Ax=b, f_1(x)\leq 0,\dots,f_r(x) \leq 0\}. \] Under appropriate assumptions, the self-concordant barrier function theory of Nesterov and Nemirovski is used to ensure polynomial-time complexity of the involved path-following procedure.