Large-Scale Optimization

Stanford University, Stanford CA

Investigators

Abstract

The investigators are creating algorithms and software for solving optimization problems of a general yet highly practical nature. Such problems arise in aerospace, biology, engineering, economics, medicine, and many other such fields. They involve thousands of variables and constraints. The intellectual challenge is to design methods that are efficient and reliable on increasingly large problems. A key aspect of the new work is improved handling of the sparse matrices involved. New classes of problems are also targeted, including ones with a mixture of continuous and discrete variables. A large community of researchers already depend critically upon the investigators' optimization software. They include economists exploring international trade agreements and the effect of greenhouse gases. At Boeing and NASA, shape optimization is essential for designing complex new aircraft configurations such as transonic airliners and a quiet supersonic business jet. Future Space Shuttle decisions depend on accurate computation of optimal trajectories for reentry from orbit and for emergency landings during launch. For example, new sharp-nosed vehicles are being compared against conventional blunt-nosed designs. The investigators' work is of national importance through its contribution to the saving of scarce resources and the competitive edge that it provides to US industry and defense.