ITR: Collaborative: Innovative Software for Large-Scale Nonlinear Optimization (linked to NSF#0082065)

University Of California-San Diego, La Jolla CA

Investigators

Abstract

Optimization is the process of finding the smallest (or largest) value that a function can return. The demand for advanced optimization software tools is increasing sharply as the importance of optimization methodology in engineering, basic science, and finance is becoming more widely recognized. The dramatic increase in computing power, and the improvements in supporting technologies such as modeling languages and automatic differentiation, are fueling demand for optimization codes. These codes must solve more and more computationally challenging problems, including integer programming and nonlinear optimization problems. This project will develop new algorithms and advanced software for large-scale nonlinear, nonconvex constrained optimization, with special emphasis on interior-point methods. Specific goals include (1) fundamental research on interior-point methods that have appealing convergence properties and robust behavior; (2) development of modern software that is well suited to implementation on advanced computing platforms and interoperable with relevant high-performance software; (3) integration of interfaces to modeling languages and automatic differentiation tools; and (4) free distribution of the software to the manufacturing, engineering, and scientific community through the NEOS web site.