On the convergence of an infeasible primal-dual interior-point method for convex programming

Abstract
We consider the infeasible primal-dual algorithm for smooth convex programming recently introduced by Vial [15]. We show, under mild assumptions, that a “SUMT” or “long-step path following” version of the algorithm is globally convergent. The stepiength on each iteration is based on a merit function which is a modification of the potential function proposed by Vial.