Preconditioning Techniques in Optimization and Applications

10:45 AM-12:45 PM

*Room: Atlanta 1 and 2*

Large-scale nonlinear optimization problems require special numerical routines for efficient solution. Usually, methods like Sequential Quadratic Programming are used to solve the nonlinear optimization problem. However, at each iteration a large-scale quadratic optimization problem occurs. Its solution is obtained by an iterative solver like CG or GMRES. The efficiency of these methods and hence of the overall method is heavily dependent on the use of proper preconditioners. The speakers will present several methods and an important application to the numerical solution in groundwater modeling.

**Organizers: Astrid Battermann and Ekkehard W. Sachs**

*Universität Trier, Germany*

**10:45-11:10 Simultaneous Multigrid Methods for Large Optimization Problems**

- Volker Schulz, Universität Stuttgart, Germany

**11:15-11:40 Uzawa Type Algorithms for Nonsymmetric Saddle Point Problems**

- Apostol Vassilev, Schlumberger Technology Corporation

**11:45-12:10 Domain Decomposition Preconditioners for Nonlinear Equations in Hydrology**

*Lea Jenkins*and C. T. Kelley, North Carolina State University

**12:15-12:40 Efficient Preconditioners for Quadratic Control Problems**

- Astrid Battermann, Organizer

*MMD, 12/21/98*