New Methods for Nonlinear Optimization
Final rept. 15 May 88-14 May 91,
COLORADO UNIV AT BOULDER
Pagination or Media Count:
This research project has investigated topics in unconstrained and constrained optimization, solving systems of nonlinear equations, nonlinear least squares, and parallel optimization. Over the course of this research contract, considerable progress was made in all the areas discussed in the proposal, namely tensor methods for nonlinear equations and optimization, trust regions methods for nonlinearly constrained optimization, orthogonal distance regression, semilocal analysis of quasi-Newton methods for nonlinearly constrained optimization, and parallel unconstrained optimization. In addition, we have worked on several other topics, including a new modified Cholesky factorization, analysis and performance of the symmetric rank one update for unconstrained optimization, secant methods for constrained optimization, the behavior of Broyden class methods for unconstrained optimization, and parallel and sequential methods for global optimization.
- Theoretical Mathematics