ALGORITHM OF INTERIOR POINTS FOR NONLINEAR-PROGRAMMING PROBLEM

In this paper, we develop and analyze a new approach, inspired by inte rior points and potential function methods, to solve a nonlinear optim ization problem under inequality constraints. The potential function w e consider involves a parameter p > 0. We study the in fluence of the value of p on asymptotic properties of an algorithmic framework. Actua lly, we show the surprizing result that to obtain a good asymptotic be havior, the only choice is p = q + 1 where q is the number of active c onstraints at the solution. With this choice, quadratic convergence is obtained.