A primal projective interior point method for linear programming

The authors present a new projective interior point for linear programming with unknown optimal value. This algorithm requires only that an interior feasible point be provided. It generates a strictly decreasing sequence of objective values and within polynomial time, either determines an optimal solution, or proves that the problem is unbounded. The authors also analyze the asymptotic convergence rate of the present method and discuss its relationship to other polynomial time projective interior point methods and the affine scaling method.