Sven Leyffer, Argonne National Laboratory
Nonlinearly constrained optimization problems arise in a broad rage of applications, including optimal experimental design, the control and operation of the power-grid, and the analysis of experimental campaigns. We present a motivating example, and discuss the basic building block of iterative solvers for nonlinearly constrained optimization problems. We show that these building blocks can be presented as a double loop framework that allows us to express a broad range of state-of-the-art nonlinear optimization solvers within a common framework. We have implemented this framework in Uno, a modern, lightweight and extensible C++ solver that unifies the workflow of most derivative-based iterative nonlinear optimization solvers. Uno is meant to enable researchers to experiment with novel optimization strategies while leveraging established subproblem solvers and interfaces to modeling languages. We close by showing some extensions and open questions.