Developing and Understanding Methods for Large-Scale Nonlinear Optimization
Abstract
During the course of this grant, we have made further progress on our optimization approach for protein structure prediction. This includes development of a preconditioned limited-memory quasi-Newton method, which should have broader applicability. We have also developed and analyzed a method for constrained optimization that uses a linear programming subproblem as an active set predictor. In our work on tensor methods we have developed a tensor method for large-scale nonlinear systems of equations using a Krylov iteration to compute the step.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 24, 2006
- Accession Number
- ADA454804
Entities
People
- Richard H. Byrd
- Robert B. Schnabel
Organizations
- University of Colorado Boulder