Accession Number : ADA292776

Title :   Quasi-Newton Updating for Large Scale Nonlinear Systems.

Descriptive Note : Final rept. 1 Jul 91-30 Jun 94,

Corporate Author : UTAH STATE UNIV LOGAN

Personal Author(s) : Walker, H. F.

PDF Url : ADA292776

Report Date : 1994

Pagination or Media Count : 10

Abstract : Research was directed toward developing robust and efficient iterative methods for large-scale systems of nonlinear algebraic equations, especially discretizations of continuous problems such as partial differential and integral equations. The specific goal was to develop effective Newton iterative methods (implementations of Newton's method using iterative linear algebra methods) and improved iterative linear solvers for use in them. A theoretically sound abstract framework was formulated. Significant progress was made toward a well-developed software implementation. Useful advances were made in the development and understanding of iterative linear solvers. (AN)

Descriptors :   *NONLINEAR SYSTEMS, MATHEMATICAL MODELS, ALGORITHMS, SOFTWARE ENGINEERING, OPTIMIZATION, LINEAR PROGRAMMING, ACCURACY, APPROXIMATION(MATHEMATICS), PARTIAL DIFFERENTIAL EQUATIONS, CONVERGENCE, INTEGRAL EQUATIONS, ITERATIONS, NONLINEAR ALGEBRAIC EQUATIONS.

Subject Categories : Numerical Mathematics
      Operations Research

Distribution Statement : APPROVED FOR PUBLIC RELEASE