|
Abstract
|
In this article, we present a second-order corrector infeasible interior-point method
based on one-norm large neighborhood for symmetric optimization. We consider the classical
Newton direction as the sum of two other directions associated with the negative and positive
parts of the right-hand side of the centrality equation. In addition to equipping them with different
step lengths, we add a corrector step that is multiplied by the square of the step length in the
expression of the new iterate. The convergence analysis of the algorithm is discussed and it is
proved that the new algorithm has the same complexity as small neighborhood infeasible interiorpoint
algorithms for the Nesterov-Todd (NT) direction, and the xs and sx directions.
|