An Efficient Parallel-in-Time Method for Optimization with Parabolic PDEs

To solve optimization problems with parabolic PDE constraints, often methods working on the reduced objective functional are used. They are computationally expensive due to the necessity of solving both the state equation and a backward-in-time adjoint equation to evaluate the reduced gradient in each iteration of the optimization method. In this study, we investigate the use of the parallel-in-time method PFASST in the setting of PDE constrained optimization. In order to develop an efficient fully time-parallel algorithm we discuss different options for applying PFASST to adjoint gradient computation, including the possibility of doing PFASST iterations on both the state and adjoint equations simultaneously. We also explore the additional gains in efficiency from reusing information from previous optimization iterations when solving each equation. Numerical results for both a linear and a non-linear reaction-diffusion optimal control problem demonstrate the parallel speedup and efficiency of different approaches.

[1]  Sebastian Götschel,et al.  Parallel-in-Time for Parabolic Optimal Control Problems Using PFASST , 2017 .

[2]  Michael L. Minion,et al.  Efficient Implementation of a Multi-Level Parallel in Time Algorithm , 2014 .

[3]  C. M. Reeves,et al.  Function minimization by conjugate gradients , 1964, Comput. J..

[4]  E. Hairer,et al.  Solving Ordinary Differential Equations II: Stiff and Differential-Algebraic Problems , 2010 .

[5]  Boris Polyak.  The conjugate gradient method in extremal problems , 1969 .

[6]  Michael Minion,et al.  Parallel-In-Time Magnus Integrators , 2019, SIAM J. Sci. Comput..

[7]  Jacob B. Schroder,et al.  A non-intrusive parallel-in-time adjoint solver with the XBraid library , 2017, Comput. Vis. Sci..

[8]  Rolf Krause,et al.  Inexact spectral deferred corrections , 2016 .

[9]  Lars Ruthotto,et al.  Layer-Parallel Training of Deep Residual Neural Networks , 2018, SIAM J. Math. Data Sci..

[10]  Stefan Ulbrich,et al.  Optimization with PDE Constraints , 2008, Mathematical modelling.

[11]  Ya-Xiang Yuan,et al.  A Nonlinear Conjugate Gradient Method with a Strong Global Convergence Property , 1999, SIAM J. Optim..

[12]  Stefan Ulbrich,et al.  Preconditioners Based on “Parareal” Time-Domain Decomposition for Time-Dependent PDE-Constrained Optimization , 2015 .

[13]  Martin Weiser,et al.  Faster SDC convergence on non-equidistant grids by DIRK sweeps , 2015 .

[14]  Eldad Haber,et al.  Stable architectures for deep neural networks , 2017, ArXiv.

[15]  E. Zeidler.  Nonlinear functional analysis and its applications , 1988 .

[16]  M. Heinkenschloss,et al.  A Parallel-inTime Gradient-Type Method for Discrete Time Optimal Control Problems ∗ , 2016 .

[17]  Peter Deuflhard,et al.  Newton Methods for Nonlinear Problems , 2004 .

[18]  E. Polak,et al.  Note sur la convergence de méthodes de directions conjuguées , 1969 .

[19]  Michael L. Minion,et al.  TOWARD AN EFFICIENT PARALLEL IN TIME METHOD FOR PARTIAL DIFFERENTIAL EQUATIONS , 2012 .

[20]  P. Deuflhard.  Newton Methods for Nonlinear Problems: Affine Invariance and Adaptive Algorithms , 2011 .

[21]  Martin Stoll,et al.  Domain decomposition in time for PDE-constrained optimization , 2015, Comput. Phys. Commun..

[22]  M. Minion.  Semi-implicit spectral deferred correction methods for ordinary differential equations , 2003 .

[23]  Michael L. Minion,et al.  A HYBRID PARAREAL SPECTRAL DEFERRED CORRECTIONS METHOD , 2010 .

[24]  A. Bourlioux,et al.  High-order multi-implicit spectral deferred correction methods for problems of reactive flow , 2003 .

[25]  Martin J. Gander,et al.  Schwarz Methods for the Time-Parallel Solution of Parabolic Control Problems , 2016 .

[26]  Matthias Heinkenschloss,et al.  A time-domain decomposition iterative method for the solution of distributed linear quadratic optimal control problems , 2005 .

[27]  M. Carpenter,et al.  Additive Runge-Kutta Schemes for Convection-Diffusion-Reaction Equations , 2003 .

[28]  Sebastian Götschel,et al.  Lossy compression for PDE-constrained optimization: adaptive error control , 2015, Comput. Optim. Appl..

[29]  Rolf Krause,et al.  A multi-level spectral deferred correction method , 2013, BIT Numerical Mathematics.

[30]  Eldad Haber,et al.  Deep Neural Networks Motivated by Partial Differential Equations , 2018, Journal of Mathematical Imaging and Vision.

[31]  Jacob B. Schroder,et al.  A non-intrusive parallel-in-time approach for simultaneous optimization with unsteady PDEs , 2018, Optim. Methods Softw..

[32]  L. Greengard,et al.  Spectral Deferred Correction Methods for Ordinary Differential Equations , 2000 .

[33]  Martin J. Gander,et al.  50 Years of Time Parallel Time Integration , 2015 .

[34]  Liang Zhong,et al.  Efficient estimation of personalized biventricular mechanical function employing gradient‐based optimization , 2018, International journal for numerical methods in biomedical engineering.

[35]  Christian E. Schaerer,et al.  Analysis of Block Parareal Preconditioners for Parabolic Optimal Control Problems , 2010, SIAM J. Sci. Comput..

[36]  Martin J. Gander,et al.  PARAEXP: A Parallel Integrator for Linear Initial-Value Problems , 2013, SIAM J. Sci. Comput..

[37]  Matthias Bolten,et al.  A multigrid perspective on the parallel full approximation scheme in space and time , 2016, Numer. Linear Algebra Appl..

[38]  Stefan Ulbrich,et al.  OPTPDE: A Collection of Problems in PDE-Constrained Optimization , 2014 .

[39]  Fredi Tröltzsch,et al.  On the optimal control of the Schlögl-model , 2013, Comput. Optim. Appl..

[40]  Stephen J. Wright,et al.  Numerical Optimization , 2018, Fundamental Statistical Inference.

[41]  Sebastian Götschel,et al.  Quantitative Defect Reconstruction in Active Thermography for Fiber-Reinforced Composites , 2016 .

[42]  S. Güttel,et al.  A rational deferred correction approach to PDE-constrained optimization , 2016 .