A primal-dual interior point method for nonlinear optimization: Global convergence, convergence rate and numerical performance for large scale problems (Q2722337)

This paper contains a survey of primal-dual interior point algorithms for general nonlinear optimization problems, using Newton-like iteration and the barrier penalty function. In this context, both methods are shown to be convergent in the global sense. An important example is given in order to show that it is possible to combine the superlinear convergence property and the global convergence property. The experiments considered here show that the methods used by the authors are indeed efficient also for small scale Hoch and Schittkowski problems and for large scale problems.NEWLINENEWLINEFor the entire collection see [Zbl 0958.00047].