ITERATION METHOD FOR CONSTRAINED OPTIMIZATION PROBLEMS GOVERNED BY PDE

  • Published : 1998.01.01

Abstract

In this paper we present a new iteration method for solving optimization problems governed by partial differential equations. We generalize the existing methods such as simple gradient methods and pseudo-time methods to get an efficient iteration method. Numerical tests show that the convergence of the new iteration method is much faster than those of the pseudo-time methods especially when the parameter $\sigma$ in the cost functional is small.

Keywords

References

  1. ICASE Report No. 94-52 Multigrid one shot methods for optiman control problems: Infinte Dimensional Control. Arian;S, Ta'asan
  2. Introduction to numerical linear algebra and optimization P. Ciarlet
  3. ICASE Report No. 94-59 Psehdo-time methods for optmal shape design using the Euler equations A. Iollo;G. Kuruvila;S, Ta'asan
  4. Compt. Methods Appl. Mech. Engng. v.118 Analysis, approximation and computation of a coupled solid/fluid temperature control problem Max D. Gunzburger;Hyung, C. Lee
  5. ICASE Report No. 92-2 One-shot method for optiman control of distributed parameter systems Ⅰ: Fine dimensional control S, Ta'asan
  6. ICASE Report No. 95-32 Pseudo-time methods for constrained optimization problems governed by PDE S, Ta'asan
  7. In 30th Aerospace Science Meeting and Exhibit, AIAA 92-005 Aerodynamics Design and optimization in one shot S. Ta'asan;M. Salas;G. Kuruvila