Carleton Coffrin is a staff scientist in Los Alamos National Laboratory’s Advanced Network Science Initiative. His research interests focus on how optimization methods can be used to solve applications in infrastructure networks. His background spans many forms of optimization including mathematical programing, constraint programming, and local search. Recently Carleton has been exploring the potential of novel computing architectures such as, quantum computers, neuromorphic processors and memristors to solve optimization applications.
This work discusses some of the requirements for deploying non-convex nonlinear optimization methods to solve large-scale problems in practice. AC Optimal Power Flow is proposed as a proxy-application for testing the viability of nonlinear optimization frameworks for solving such problems. The current performance of several Julia frameworks for nonlinear optimization is evaluated using a standard benchmark library for AC Optimal Power Flow.