A new method of moving asymptotes for large-scale unconstrained optimization (Q2518690)

This article presents a method of moving asymptotes to solve large-scale unconstrained optimization problems of a twice continuously differentiable function. The authors begin with a brief introduction to this general problem and a survey of the relevant literature on moving asymptote methods including several useful theorems. In the third section the authors present their innovative algorithm and prove its convergence. The article concludes with a series of tables presenting the results of the numerical experimentation, ideas for future extensions and a list of relevant references.