Primal-dual target-following algorithms for linear programming

In this paper, the authors propose a method for linear programming with the property that, starting from an initial non-central point, it generates iterates that simultaneously get closer to optimality and closer to centrality. The iterates follow paths that in the limit are tangential to the central path. Together with the convergence analysis, the authors provide a general framework which enables us to analyze various primal-dual algorithms in the literature in a short and uniform way.