A primal-dual infeasible-interior-point algorithm for linear programming (Q1315419)

A so-called `primal-dual infeasible-interior-point algorithm' is proposed for solving linear programming problems. As usual, it takes a new point along the Newton direction at each step; however, the point taken will not be necessarily within the feasible region although within the interior of the region determined by the nonnegative constraints. So one can start from an arbitrary primal-dual pair of points with strictly positive components. A step length rule is used to ensure the global convergence of the algorithm.