Integration of Sequential Quadratic Programming and Domain Decomposition Methods for Nonlinear Optimal Control Problems
We discuss the integration of a sequential quadratic programming (SQP) method with an optimization-level domain decomposition (DD) preconditioner for the solution of the quadratic optimization subproblems. The DD method is an extension of the well-known Neumann-Neumann method to the optimization context and is based on a decomposition of the first order system of optimality conditions. The SQP method uses a trust-region globalization and requires the solution of quadratic subproblems that are known to be convex, hence solving the first order system of optimality conditions associated with these subproblems is equivalent to solving these subproblems. In addition, our SQP method allows the inexact solution of these subproblems and adjusts the level of exactness with which these subproblems are solved based on the progress of the SQP method. The overall method is applied to a boundary control problem governed by a semilinear elliptic equation.
KeywordsOptimal Control Problem Domain Decomposition Sequential Quadratic Programming Domain Decomposition Method Order Optimality Condition
Unable to display preview. Download preview PDF.
- 4.G. Biros and O. Ghattas. Parallel Lagrange–Newton–Krylov–Schur methods for PDE–constrained optimization. part I: The Krylov–Schur solver. SIAM J. Sci. Comput., 27(2):587–713, 2005.Google Scholar
- 11.M. D. Gunzburger, L. S. Hou, and T. P. Svobotny. Finite element approximations of an optimal control problem associated with the scalar Ginzburg-Landau equation. Comput. Math. Appl., pages 123–131, 1991.Google Scholar
- 12.M. Heinkenschloss and H. Nguyen. Domain decomposition preconditioners for linear–quadratic elliptic optimal control problems. Technical Report TR04–20, Dept. of Computational and Applied Mathematics, Rice U., 2004.Google Scholar
- 14.M. Heinkenschloss and D. Ridzal. Solution of a class of quadratic programs arising in nonlinear programming using inexact linear system solves. Technical report, Dept. of Computational and Applied Math., Rice U., 2006.Google Scholar
- 17.X. Li and J Yong. Optimal Control Theory for Infinite Dimensional Systems. Systems & Control: Foundations & Applications. Birkhäuser-Verlag, Boston, Basel, Berlin, 1995.Google Scholar
- 19.E. Prudencio and X.-C. Cai. Parallel multi-level Lagrange-Newton-Krylov-Schwarz algorithms with pollution removing for PDE-constrained optimization. Technical report, U. of Colorado at Boulder, Dept. of Computer Science, Boulder, Colorado, 2006.Google Scholar
- 21.D. Ridzal. Trust Region SQP Methods With Inexact Linear System Solves For Large-Scale Optimization. PhD thesis, Dept. of Computational and Applied Mathematics, Rice U., Houston, TX, 2006.Google Scholar
- 23.A. Toselli and O. Widlund. Domain Decomposition Methods - Algorithms and Theory. Computational Math., Vol. 34. Springer–Verlag, Berlin, 2004.Google Scholar
- 24.F. Tröltzsch. Optimale Steuerung partieller Differentialgleichungen. Grundlagen, Optimalitätsbedingungen und ihre Anwendungen. Vieweg Verlag, Braunschweig, 2005.Google Scholar