Jens Saak

Efficient handling of complex shift parameters in the low-rank Cholesky factor ADI method

The solution of large-scale Lyapunov equations is a crucial problem for several fields of modern applied mathematics. The low-rank Cholesky factor version of the alternating directions implicit method (LRCF-ADI) is one iterative algorithm that computes approximate low-rank factors of the solution. In order to achieve fast convergence it requires adequate shift parameters, which can be complex if the matrices defining the Lyapunov equation are unsymmetric. This will require complex arithmetic computations as well as storage of complex data and thus, increase the overall complexity and memory requirements of the method. In this article we propose a novel reformulation of LRCF-ADI which generates real low-rank factors by carefully exploiting the dependencies of the iterates with respect to pairs of complex conjugate shift parameters. It significantly reduces the amount of complex arithmetic calculations and requirements for complex storage. It is hence often superior in terms of efficiency compared to other real formulations.

Self-Generating and Efficient Shift Parameters in ADI Methods for Large Lyapunov and Sylvester Equations

Low-rank versions of the alternating direction implicit (ADI) iteration are popular and well estab- lished methods for the numerical solution of large-scale Sylvester and Lyapunov equations. Probably the biggest disadvantage of these methods is their dependence on a set of shift parameters that are crucial for fast convergence. Here we firstly review existing shift generation strategies that compute a number of shifts before the actual itera- tion. These approaches come with several disadvantages such as, e.g., expensive numerical computations and the difficulty to obtain necessary spectral information or data n eeded to initially setup their generation. Secondly, we propose two novel shift selection strategies with the motivation to resolve these issues at least partially. Both ap- proaches generate shifts automatically in the course of the ADI iterations. Extensive numerical tests show that one of these new approaches, based on a Galerkin projection onto the space spanned by the current ADI data, is superior to other approaches in the majority of cases both in terms of convergence speed and required execution time.

A Semi-Discretized Heat Transfer Model for Optimal Cooling of Steel Profiles

Several generalized state-space models arising from a semi-discretization of a controlled heat transfer process for optimal cooling of steel profiles are presented. The model orders differ due to different levels of refinement applied to the computational mesh.

Efficient balancing-based MOR for large-scale second-order systems

Large-scale structure dynamics models arise in all areas where vibrational analysis is performed, ranging from control of machine tools to microsystems simulation. To reduce computational and resource demands and be able to compute solutions and controls in acceptable, that is, applicable, time frames, model order reduction (MOR) is applied. Classically modal truncation is used for this task. The reduced-order models (ROMs) generated are often relatively large and often need manual modification by the addition of certain technically motivated modes. That means they are at least partially heuristic and cannot be generated fully automatic. Here, we will consider the application of fully automatic balancing-based MOR techniques. Our main focus will be on presenting a way to efficiently compute the ROM exploiting the sparsity and second-order structure of the finite element method (FEM) semi-discretization, following a reduction technique originally presented in [V. Chahlaoui, K.A. Gallivan, A. Vandendorpe, and P. Van Dooren, Model reduction of second-order system, in Dimension Reduction of Large-Scale Systems, P. Benner, V. Mehrmann, and D. Sorensen, eds., Lecture Notes in Computer Science and Engineering, Vol. 45, Springer Verlag, Berlin, 2005, pp. 149–172], [Y. Chahlaoui, D. Lemonnier, A. Vandendorpe, and P. Van Dooren, Second-order balanced truncation, Linear Algebra Appl. 415 (2006), pp. 373–384], [T. Reis and T. Stykel, Balanced truncation model reduction of second-order systems, Math. Comput. Model. Dyn. Syst. 14 (2008), pp. 391–406] and [J. Fehr, P. Eberhard, and M. Lehner, Improving the Reduction Process in Flexible Multibody Dynamics by the Use of 2nd Order Position Gramian Matrices, Proceedings ENOC, St. Petersburg, Russia, 2008]. Large-scale sparse solvers for the underlying matrix equations solved in the balancing process are adapted to the second-order structure of the equations and the suitability of our approach is demonstrated for two practical examples.

An Improved Numerical Method for Balanced Truncation for Symmetric Second-Order Systems

We consider balanced truncation model order reduction for symmetric second-order systems. The occurring large-scale generalized and structured Lyapunov equations are solved with a specially adapted low-rank alternating directions implicit (ADI) type method. Stopping criteria for this iteration are investigated, and a new result concerning the Lyapunov residual within the low-rank ADI method is established. We also propose a goal-oriented stopping criterion which tries to incorporate the balanced truncation approach already during the ADI iteration. The model reduction approach using the ADI method with different stopping criteria is evaluated on several test systems.

Riccati-based Boundary Feedback Stabilization of Incompressible Navier--Stokes Flows

In this article a boundary feedback stabilization approach for incompressible Navier--Stokes flows is studied. One of the main difficulties encountered is the fact that after space discretization by a mixed finite element method (because of the solenoidal condition) one ends up with a differential algebraic system of index 2. The remedy here is to use a discrete realization of the Leray projection used by Raymond [J.-P. Raymond, SIAM J. Control Optim., 45 (2006), pp. 790--828] to analyze and stabilize the continuous problem. Using the discrete projection, a linear quadratic regulator (LQR) approach can be applied to stabilize the (discrete) linearized flow field with respect to small perturbations from a stationary trajectory. We provide a novel argument that the discrete Leray projector is nothing else but the numerical projection method proposed by Heinkenschloss and colleagues in [M. Heinkenschloss, D. C. Sorensen, and K. Sun, SIAM J. Sci. Comput., 30 (2008), pp. 1038--1063]. The nested iteration resul...

Numerical Solution of Optimal Control Problems for Parabolic Systems

We discuss the numerical solution of optimal control problems for instationary convection-diffusion and diffusion-reaction equations. Instead of viewing this problem as a large-scale unconstrained optimization problem after complete discretization of the corresponding optimality system, we formulate the problem as abstract linear-quadratic regulator (LQR) problem. Using recently developed efficient solvers for large-scale algebraic Riccati equations, we show how to numerically solve the optimal control problem at a cost proportional to solving the corresponding forward problem. We discuss two different optimization goals: one can be seen as stabilization of the plant model, the second one is of tracking type, i.e., a given (optimal) solution trajectory is to be attained. The efficiency of our approach is demonstrated for a model problem related to an optimal cooling process. Moreover, we discuss how the LQR approach can be applied to nonlinear problems.

Parallel order reduction via balanced truncation for optimal cooling of steel profiles

We employ two efficient parallel approaches to reduce a model arising from a semi-discretization of a controlled heat transfer process for optimal cooling of a steel profile. Both algorithms are based on balanced truncation but differ in the numerical method that is used to solve two dual generalized Lyapunov equations, which is the major computational task. Experimental results on a cluster of Intel Xeon processors compare the efficacy of the parallel model reduction algorithms.

Model Order Reduction for Systems with Moving Loads

Abstract In this contribution we present two approaches allowing to find a reduced order approximant of a full order model featuring a moving load term. First, we apply the Balanced Truncation (BT) method to a switched linear system (SLS) using the special structure given in the spatially discretized model. The second approach treats the variability as a continuous parameter dependence and uses the iterative rational Krylov algorithm (IRKA) to compute a parameter preserving reduced order model.

Efficient Solution of Large-Scale Saddle Point Systems Arising in Riccati-Based Boundary Feedback Stabilization of Incompressible Stokes Flow

We investigate numerical methods for solving large-scale saddle point systems which arise during the feedback control of flow problems. We focus on the instationary Stokes equations that describe instationary, incompressible flows for moderate viscosities. After a mixed finite element discretization we get a differential-algebraic system of differential index two [J. Weickert, Navier-Stokes Equations as a Differential-Algebraic System, Preprint SFB393/96-08, Department of Mathematics, Chemnitz University of Technology, Chemnitz, Germany, 1996]. To reduce this index, we follow the analytic ideas of [J.-P. Raymond, SIAM J. Control Optim., 45 (2006), pp. 790--828] coupled with the projection idea of [M. Heinkenschloss, D. C. Sorensen, and K. Sun, SIAM J. Sci. Comput., 30 (2008), pp. 1038--1063]. Avoiding this explicit projection leads to solving a series of large-scale saddle point systems. In this paper we construct iterative methods to solve such saddle point systems by deriving efficient preconditioners b...