A parallelizable augmented Lagrangian method applied to large-scale non-convex-constrained optimization problems (Q2414914)

This article presents an improved solution method for large-scale optimization problems with a convex objective but a non-convex constraint set, using a Lagrangian dual solution method. The authors begin with an introduction to the problem and the notation used in the paper, followed by an overview of the relevant literature on augmented Lagrangian methods, proximal bundle methods and their parallelization as well as their variants. The third and fourth sections present the main result of the article where the new algorithm is described in detail and several properties are proven. The article concludes with a section containing the results of a detailed computational experimentation and suggestions for future work.