On some efficient interior point methods for nonlinear convex programming (Q1174840)

Two interior-point algorithms for general linearly constrained convex programming are presented: a potential reduction algorithm and a path- following algorithm, respectively. Both of them require the approximate solution of a nonlinear system of equations at each step and they are of \(O(\sqrt{n} | \ln\varepsilon|)\) complexity. The fine technical procedures and the high degree of generality, which are characteristic to this paper, offer large possibilities for different particularizations. Thus, the authors show that for the quadratic programming case, these approaches work quite well and generalize some other methods previously developed by other authors.