Abstract
We present an overview of the Lagrange-Newton-Krylov-Schur (LNKS) method for solution of optimization problems that are governed by systems of partial differential equations. We discuss how to improve LNKS’s work efficiency by carrying out certain computations inexactly, without compromising convergence. LNKS solves the Karush-Kuhn-Tucker optimality conditions using a NewtonKrylov algorithm. Its key component is a preconditioner based on variants of quasiNewton reduced space Sequential Quadratic Programming (QN-RSQP) methods. LNKS combines the fast convergence properties of a Newton method with the ability of preconditioned Krylov methods to solve very large linear systems. Nevertheless, even with good preconditioners, solution of the optimization problem is several times more expensive than solution of the underlying PDE problem. To accelerate LNKS, its computational components are carried out inexactly: premature termination of iterative algorithms, inexact evaluation of gradients and Jacobians, and approximate line searches. We discuss several issues that arise with respect to these inexact computations, and the resulting trade-offs between speed and robustness.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Satish Balay, William D. Gropp, Lois Curfman Mcinnes, and Barry F. Smith. PETSc home page. http://www.mcs.anl.gov, /petsc 1999.
Lorenz T. Biegler, Jorge Nocedal, and Claudia Schmid. A reduced Hessian method for large-scale constrained optimization. SIAM Journal on Optimization, 5:314–347, 1995.
George Biros. Parallel Algorithms for PDE-Constrained Optimization and Application to Optimal Control of Viscous Flows. PhD thesis, Carnegie Mellon University, Pittsburgh, PA, August 2000.
George Biros and Omar Ghattas. Parallel Newton-Krylov algorithms for PDEconstrained optimization. In Proceedings of SC99, The SCxy Conference series, Portland, Oregon, November 1999. ACM/IEEE.
George Biros and Omar Ghattas. Parallel preconditioners for KKT systems arising in optimal control of viscous incompressible flows. In D. E. Keyes, A. Ecer, J. Periaux, and N. Satofuka, editors, Parallel Computational Fluid Dynamics 1999. North-Holland, 1999.
George Biros and Omar Ghattas. Parallel Lagrange-Newton-Krylov-Schur methods for PDE-constrained optimization. Part I: The Krylov-Schur solver. Technical report, Laboratory for Mechanics, Algorithms, and Computing, Carnegie Mellon University, 2000.
George Biros and Omar Ghattas. Parallel Lagrange-Newton-Krylov-Schur methods for PDE-constrained optimization. Part II: The Lagrange Newton solver, and its application to optimal control of steady viscous flows. Technical report, Laboratory for Mechanics, Algorithms, and Computing, Carnegie Mellon University, 2000.
John Dennis E., Jr., Mahmoud El-Alem, and Maria C. Magiel. A global convergence theory for general trust-region-based algorithms for equality constrained optimization. SIAM Journal on Optimization, 7(1):177–207, 1997.
Stanley C. Eisenstat and Homer F. Walker. Globally convergent inexact Newton methods. SIAM Journal on Optimization, 4(2):393–422, 1994.
Stanley C. Eisenstat and Homer F. Walker. Choosing the forcing terms in an inexact Newton method. SIAM Journal on Scientific Computing, 17(1):16–32, 1996.
Roger Fletcher. Practical Methods of Optimization. John Wiley and Sons, second edition, 1987.
Roland W. Freund and Noël M. Nachtigal. An implementation of the QMR method based on coupled two-term recurrences. SIAM Journal of Scientific Computing, 15(2):313–337, March 1994.
Omar Ghattas and Jai-Hyeong Bark. Optimal control of two-and threedimensional incompressible Navier-Stokes flows. Journal of Computational Physics, 136:231–244, 1997.
Max D. Gunzburger. Finite Element for Viscous Incompressible Flows. Academic Press, 1989.
Max D. Gunzburger, editor. Flow Control, volume 68 of IMA Math. Appl. Springer-Verlag, New York, 1995.
Max D. Gunzburger and Roy A. Nicolaides, editors. Incompressible Computational Fluid Dynamics. Cambridge University Press, 1993.
Matthias Heinkenschloss and Luis N. Vicente. Analysis of inexact trust-region SQP algorithms. Technical Report TR99-18, Rice University, Department of Computational and Applied Mathematics, 1999.
C. T. Kelley and David E. Keyes. Convergence analysis of pseudo-transient continuation. SIAM Journal on Numerical Analysis, 35:508–523, 1998.
C.T. Kelley and Ekkehard W. Sachs. Truncated Newton methods for optimization with inaccurate functions and gradients. SIAM Journal on Optimization, 10(1):43–55, 1999.
F. Leibritz and E. W. Sachs. Inexact SQP interior point methods and large scale optimal control problems. SIAM Journal on Control and Optimization, 38(1):272–293, 1999.
Ivan Malčević, Large-scale unstructured mesh shape optimization on parallel computers. Master’ s thesis, Carnegie Mellon University, 1997.
Stephen G. Nash and Ariela Sofer. Linear and Nonlinear Programming. McGraw-Hill, 1996.
Jorge Nocedal and Stephen J. Wright. Numerical Optimization. Springer, 1999.
T. Steihaug. The conjugate gradient method and trust regions in large scale optimization. SIAM Journal on Numerical Analysis, 20:626–637, 1983.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Biros, G., Ghattas, O. (2003). Inexactness Issues in the Lagrange-Newton-Krylov-Schur Method for PDE-constrained Optimization. In: Biegler, L.T., Heinkenschloss, M., Ghattas, O., van Bloemen Waanders, B. (eds) Large-Scale PDE-Constrained Optimization. Lecture Notes in Computational Science and Engineering, vol 30. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-55508-4_6
Download citation
DOI: https://doi.org/10.1007/978-3-642-55508-4_6
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-05045-2
Online ISBN: 978-3-642-55508-4
eBook Packages: Springer Book Archive