betaFeedback

WorksMathematics

A Preconditioned Semismooth Newton Method for Box-Constrained Elliptic Optimal Control Problems with Random Coefficients

Abstract

We consider the optimal control problem of an elliptic PDE with random diffusion under pointwise box constraints on the control. The constraints make the optimality conditions nonsmooth, with an active set that is not known in advance, and the random diffusion couples the state and the adjoint across all stochastic modes. To design a solution algorithm for this problem, we rely on the Fenchel duality viewpoint. Fenchel duality absorbs the box constraint into the objective and yields an exact reduced problem in the elliptic adjoint variable alone. After a stochastic Galerkin–finite element discretization, eliminating the state leaves a strongly monotone, piecewise smooth equation, which we solve directly by a matrix-free semismooth Newton method with the box projection retained exactly and with no outer iteration, penalty or smoothing parameter. Next, we introduce and analyze a current-set preconditioner, which adds the active-set correction of the current Newton step to a mean-diffusion reference operator and is inverted exactly by a rank-1 update. For implementation, we use a cheaper preconditioner that keeps only a fraction of free control values. Numerical experiments show second-order convergence of the state, and Newton and PCG counts that stay bounded under mesh refinement and growth of the stochastic dimension.

Introduction

Optimal control problems governed by partial differential equations (PDEs) are ubiquitous in computational science and engineering, and both a mature theory and a large body of numerical methods are available in the deterministic case [35, 14]. Pointwise control bounds a≤u≤ba \le u \le b change the picture. The first-order conditions are no longer a system of equations but a variational inequality, or equivalently a nonsmooth equation built from the metric projection onto the admissible box [20, 15]. Three consequences follow. First, although the reduced objective is strongly convex, the resulting Karush–Kuhn–Tucker (KKT) system is nondifferentiable, so classical Newton methods do not apply. Second, the set of indices at which the bounds are attained—the active set—is not known in advance and must be discovered as part of the solution process. Third, the standard remedies—penalization, interior-point regularization, smoothing—replace the exact constraint by an approximation governed by an additional parameter, and recover it only in the limit.

In the deterministic problem, the primal–dual active set (PDAS) strategy of Bergounioux, Ito and Kunisch [9] treats the bounds exactly by alternating between a prediction of the active set and the solution of an equality-constrained subproblem. Hintermüller, Ito and Kunisch [18] subsequently showed that PDAS is precisely a semismooth Newton (SSN) method for the projection-based reformulation, which explains its local superlinear and mesh-independent convergence [11, 28, 30, 36, 19].

In most applications the PDE coefficients are uncertain and are modelled by random fields and lead to optimal control problems under uncertainty [27]. The random variable may be discretized by sampling, by stochastic collocation [10, 21, 22, 34], or by stochastic Galerkin (SG) projection onto a global orthogonal polynomial basis [16, 39, 23, 26, 4, 24, 32]. We use SG, which yields a single coupled system: the discrete state operator is a sum of Kronecker products, and state and adjoint interact across all stochastic modes.

Combining pointwise constraints with random coefficients is harder than treating either feature alone. For control constraints, the projection onto the box is available in closed form and need not be regularized, yet stochastic Galerkin discretizations of such problems have been studied mainly from the approximation side: Shen et al. [33] derive a priori error estimates for obstacle-type control constraints, and Winkler and Yücel [38] combine error estimates for a Dirichlet boundary control problem with a primal–dual active set solver and block-diagonal preconditioning. There, the control is deterministic and acts on ∂D\partial D, so the active set stays decoupled from the stochastic tensor structure. Solvers for stochastic control problems are developed in [8, 6, 1, 25, 3, 2, 7]. None of them treats control constraints.

An alternative is the Fenchel duality viewpoint, where the dual absorbs the constraint into the objective. Clason and Kunisch [12] found the duals of certain unconstrained elliptic control problems structurally more favourable than the primal problems, and Wang et al. [37] derived the Fenchel dual of a linear parabolic problem with box constraints on the control, in which the constraints are embedded in an unconstrained, differentiable dual objective, and solved it by conjugate gradients and by semismooth Newton. They also identified the obstacle that reappears here: semismooth Newton applied to the primal problem treats the box and the PDE constraint simultaneously, so the varying active set changes the structure of the Newton equation and the preconditioner has to be adjusted at every step. In the stochastic setting, this obstacle is sharper: control values are indexed by pairs of a spatial node and a quadrature point, so the active-set term of the Jacobian matrix has no Kronecker structure. A preconditioner containing it exactly must be rebuilt at every step, whereas a mean-based Kronecker preconditioner is cheap but blind to the location of the free set.

In this paper, we develop a dual-based SSN framework for optimal control problems governed by elliptic PDEs with random coefficients and pointwise box constraints on the control. We derive the Fenchel dual in Bochner spaces over the parameter domain, which absorbs the control constraints into an unconstrained dual objective in the adjoint variable, and establish the corresponding optimality system and primal–dual recovery relations. After stochastic Galerkin–finite element discretization, we derive the exact finite-dimensional dual and solve the resulting reduced equation directly by a matrix-free semismooth Newton–PCG method, in which the box projection is retained exactly, and no outer iteration, penalty, or smoothing parameter is required. To handle the active set, which changes at every Newton step, we analyze a current-set preconditioner PWP_W that places the current active-set correction beside a mean-based reference operator and applies its inverse exactly through the Sherman–Morrison–Woodbury (SMW) identity [17]. For numerical implementation, we use a cheap approximation PkP_k of the preconditioner PWP_W, which keeps only the fraction of free control values, and its application needs one sparse Cholesky factor of a spatial matrix, which is reused across Newton steps.

The remainder of the paper is organized as follows. Section 2-3 presents the problem and derives the Fenchel dual and the optimality system. Section 4 gives the SG-FE discretization and its exact dual. Section 5 develops the direct SSN–PCG method and the spectral analysis of the current-set preconditioner PWP_W. Section 6 specifies the practical preconditioner PkP_k and reports the numerical experiments. Section 7 draws conclusions and discusses future work.

Problem formulation and well-posedness

The optimal control problem

Let D⊂RdD \subset\mathbb{R}^{d}, d∈{1,2,3}d \in\{1,2,3\}, be a bounded Lipschitz domain with Poincaré constant CPC_{P}, that is, ∥v∥L2(D)≤CP∥∇v∥L2(D)d\lVert v\rVert_{L^{2}(D)} \le C_{P}\lVert\nabla v\rVert_{L^{2}(D)^{d}} for all v∈H01(D)v \in H_{0}^{1}(D), let ξ=(ξ1,…,ξK):(Ω,F,P)→RK\xi= (\xi_{1},\ldots,\xi_{K}):(\Omega,\mathcal{F},\mathbb{P}) \to\mathbb{R}^{K}, K<∞K < \infty, be a random vector with image Γξ:=ξ(Ω)\Gamma_{\xi} := \xi(\Omega), and let (Γξ,B,ρ)(\Gamma_{\xi},\mathcal{B},\rho) be the induced parameter probability space, where B\mathcal{B} is the Borel σ\sigma-algebra on Γξ\Gamma_{\xi} and ρ:=P∘ξ−1\rho:= \mathbb{P} \circ\xi^{-1} is the law of ξ\xi. Let yy denote the state and uu the control, and set

V=H01(D),H=L2(D),V=Lρ2(Γξ;V),H=Lρ2(Γξ;H),V = H_{0}^{1}(D), \qquad H = L^{2}(D), \qquad\mathbb{V} = L_{\rho}^{2}(\Gamma_{\xi};V), \qquad\mathbb{H} = L_{\rho}^{2}(\Gamma_{\xi};H),

where the norms and inner products on V\mathbb{V} and H\mathbb{H} include the expectation with respect to ρ\rho. We equip V\mathbb{V} with the norm ∥v∥V:=∥∇v∥L2(D)d\lVert v\rVert_{\mathbb{V}} := \lVert\nabla v\rVert_{L^{2}(D)^{d}}, so that ∥v∥V2=∫Γξ∥∇v(ξ)∥L2(D)d2 dρ(ξ)\lVert v\rVert_{\mathbb{V}}^{2} = \int_{\Gamma_{\xi}}\lVert\nabla v(\xi)\rVert_{L^{2}(D)^{d}}^{2}\,d\rho(\xi) and ∥v∥H≤CP∥v∥V\lVert v\rVert_{\mathbb{H}} \le C_{P}\lVert v\rVert_{\mathbb{V}} for v∈Vv \in\mathbb{V}. Given a desired state yd∈Hy_{d} \in\mathbb{H}, a regularization parameter γ>0\gamma> 0, finite bounds a<ba < b, and a random diffusion coefficient κ\kappa, the problem studied in this paper is

min⁡y,u12∫Γξ∫D∣y(x,ξ)−yd(x,ξ)∣2 dx dρ(ξ)+γ2∫Γξ∫D∣u(x,ξ)∣2 dx dρ(ξ)subject to−∇⋅(κ(x,ξ)∇y(x,ξ))=u(x,ξ)in D×Γξ,y(x,ξ)=0on ∂D×Γξ,a≤u(x,ξ)≤b(dx⊗dρ)-a.e.(1)\begin{aligned} \min_{y,u}\quad&\frac{1}{2}\int_{\Gamma_{\xi}}\int_{D}\lvert y(x,\xi)-y_{d}(x,\xi)\rvert^{2}\,dx\,d\rho(\xi) +\frac{\gamma}{2}\int_{\Gamma_{\xi}}\int_{D}\lvert u(x,\xi)\rvert^{2}\,dx\,d\rho(\xi)\\ \text{subject to}\quad&-\nabla\cdot\bigl(\kappa(x,\xi)\nabla y(x,\xi)\bigr)=u(x,\xi) \quad\text{in }D\times\Gamma_{\xi},\\ &y(x,\xi)=0 \quad\text{on }\partial D\times\Gamma_{\xi},\\ &a\le u(x,\xi)\le b \quad(dx\otimes d\rho)\text{-a.e.} \tag*{(1)} \end{aligned}

Both integrals are taken with respect to the product of the Lebesgue measure on DD and the probability measure ρ\rho; so, the objective in (1) is exactly

12∥y−yd∥H2+γ2∥u∥H2,\frac{1}{2}\lVert y-y_{d}\rVert_{\mathbb{H}}^{2}+\frac{\gamma}{2}\lVert u\rVert_{\mathbb{H}}^{2},

and we use this compact form from Section 33 onwards.

We note here that the bounds are imposed pointwise in the spatial and the parameter variables, so the admissible set

C={u∈H:a≤u(x,ξ)≤b for (dx⊗dρ)-a.e. (x,ξ)}(2)\mathcal{C}=\{u\in\mathbb{H}:a\le u(x,\xi)\le b\text{ for }(dx\otimes d\rho)\text{-a.e. }(x,\xi)\} \tag*{(2)}

is a subset of the full Bochner space rather than of a deterministic space of controls. Since DD and the parameter space carry finite measure, C\mathcal{C} is nonempty, bounded, closed, and convex in H\mathbb{H}.

The state operator and the reduced problem

Assume that κ\kappa is jointly measurable, symmetric, and uniformly elliptic:

0<κ−≤κ(x,ξ)≤κ+<∞,(3)0<\kappa_{-}\le\kappa(x,\xi)\le\kappa_{+}<\infty, \tag*{(3)}

for all η∈Rd\eta\in\mathbb{R}^{d} and almost every (x,ξ)(x,\xi), with deterministic constants κ−\kappa_{-} and κ+\kappa_{+}. The weak form of the state equation in (1) is governed by the bilinear form

a(y,v)=∫Γξ∫Dκ(x,ξ)∇y(x,ξ)⋅∇v(x,ξ) dx dρ(ξ).a(y,v)=\int_{\Gamma_{\xi}}\int_{D}\kappa(x,\xi)\nabla y(x,\xi)\cdot\nabla v(x,\xi)\,dx\,d\rho(\xi).

Theorem 2.1 (Lax–Milgram [35]). Let WW be a real Hilbert space and let t:W×W→Rt:W\times W\to\mathbb{R} be a bilinear form for which there are constants M≥α>0M\ge\alpha>0 such that

∣t(w,z)∣≤M∥w∥W∥z∥Wandt(w,w)≥α∥w∥W2(w,z∈W).\lvert t(w,z)\rvert\le M\lVert w\rVert_{W}\lVert z\rVert_{W} \quad\text{and}\quad t(w,w)\ge\alpha\lVert w\rVert_{W}^{2} \qquad(w,z\in W).

Then, for every ℓ∈W′\ell\in W', there is a unique z∈Wz \in W such that

t(z,w)=ℓ(w)(w∈W),t(z,w)=\ell(w) \qquad(w\in W),

and ∥z∥W≤α−1∥ℓ∥W′\lVert z\rVert_W \le\alpha^{-1}\lVert\ell\rVert_{W'}.

Proposition 2.2. Under (3), for every u∈Hu \in\mathbb{H} there is a unique y=Su∈Vy=Su \in V such that

a(y,v)=(u,v)H(v∈V).(4)a(y,v)=(u,v)_{\mathbb{H}} \qquad(v\in V). \tag*{(4)}

The solution map

S:H⟶V↪HS:\mathbb{H}\longrightarrow V\hookrightarrow\mathbb{H}

is bounded, linear, positive, injective, and self-adjoint as an operator on H\mathbb{H}.

Proof. Boundedness and coercivity of aa on VV follow from (3). Moreover,

∣(u,v)H∣≤∥u∥H∥v∥H≤CP∥u∥H∥v∥V.\lvert(u,v)_{\mathbb{H}}\rvert\le\lVert u\rVert_{\mathbb{H}}\lVert v\rVert_{\mathbb{H}}\le C_P\lVert u\rVert_{\mathbb{H}}\lVert v\rVert_V.

Theorem 2.1 therefore gives a unique solution and

∥Su∥V≤CPκ−∥u∥H,∥Su∥H≤CP2κ−∥u∥H.\lVert Su\rVert_V\le\frac{C_P}{\kappa_-}\lVert u\rVert_{\mathbb{H}},\qquad \lVert Su\rVert_{\mathbb{H}}\le\frac{C_P^2}{\kappa_-}\lVert u\rVert_{\mathbb{H}}.

References

  1. [1]H. Antil, S. Dolgov, and A. Onwunta, TTRISK: Tensor train decomposition algorithm for risk averse optimization, Numer. Linear Algebra Appl., 30 (2023), p. e2481.arxiv.org/abs/2111.05180
  2. [2]H. Antil, S. Dolgov, and A. Onwunta, Smoothed Moreau–Yosida tensor-train approximation of state-constrained optimization problems under uncertainty, Numer. Linear Algebra Appl., 32 (2025), p. e70028.arxiv.org/abs/2301.08684
  3. [3]H. Antil, S. Dolgov, and A. Onwunta, Smoothed Moreau–Yosida tensor-train approximation of state-constrained optimization problems under uncertainty, Numer. Linear Algebra Appl. (2025), p. 32:e70028, https://doi.org/10.1002/nla.70028.
  4. [4]I. Babuška, R. T. Tempone, and G. E. Zouraris, Galerkin finite element approximations of stochastic elliptic partial differential equations, SIAM J. Numer. Anal., 42 (2004), pp. 800–825.DOI
  5. [5]H. H. Bauschke and P. L. Combettes, Convex Analysis and Monotone Operator Theory in Hilbert Spaces, CMS Books in Mathematics, Springer, Cham, second ed., 2017.
  6. [6]P. Benner, S. Dolgov, A. Onwunta, and M. Stoll, Low-rank solvers for unsteady Stokes–Brinkman optimal control problem with random data, Comput. Methods Appl. Mech. Engrg., 304 (2016), pp. 26–54.DOI
  7. [7]P. Benner, S. Dolgov, A. Onwunta, and M. Stoll, Low-rank solution of an optimal control problem constrained by random Navier–Stokes equations, International Journal on Numerical Methods for Fluids, 92 (2020), pp. 1653–1678.DOI
  8. [8]P. Benner, A. Onwunta, and M. Stoll, Block-diagonal preconditioning for optimal control problems constrained by PDEs with uncertain inputs, SIAM J. Matrix Anal. Appl., 37 (2016), pp. 491–518.DOI
  9. [9]M. Bergounioux, K. Ito, and K. Kunisch, Primal-dual strategy for constrained optimal control problems, SIAM J. Control Optim., 37 (1999), pp. 1176–1194.DOI
  10. [10]P. Chen and A. Quarteroni, Weighted reduced basis method for stochastic optimal control problems with elliptic PDE constraint, SIAM/ASA J. Uncertain. Quantif., 2 (2014), pp. 364–396.DOI
  11. [11]F. H. Clarke, Optimization and Nonsmooth Analysis, Canadian Mathematical Society Series of Monographs and Advanced Texts, John Wiley & Sons, Inc., New York, 1983.
  12. [12]C. Clason and K. Kunisch, A duality-based approach to elliptic control problems in non-reflexive Banach spaces, ESAIM Control Optim. Calc. Var., 17 (2011), pp. 243–266.DOI
  13. [13]I. Ekeland and R. Témam, Convex analysis and variational problems, vol. 28 of Classics in Applied Mathematics, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, english ed., 1999, https://doi.org/10.1137/1.9781611971088, https://doi.org/10.1137/1.9781611971088. Translated from the French.
  14. [14]H. Elman, D. Silvester, and A. Wathen, Finite Elements and Fast Iterative Solvers: with Applications in Incompressible Fluid Dynamics, Oxford University Press, 2014.
  15. [15]F. Facchinei and J.-S. Pang, Finite-Dimensional Variational Inequalities and Complementarity Problems, Springer Series in Operations Research, Springer-Verlag, New York, 2003.
  16. [16]R. G. Ghanem and P. D. Spanos, Stochastic Finite Elements: A Spectral Approach, Springer-Verlag New York, Inc., New York, NY, USA, 1991.DOI
  17. [17]W. W. Hager, Updating the inverse of a matrix, SIAM Rev., 31 (1989), pp. 221–239.DOI
  18. [18]M. Hintermüller, K. Ito, and K. Kunisch, The primal-dual active set strategy as a semismooth Newton method, SIAM J. Optim., 13 (2002), pp. 865–888.DOI
  19. [19]M. Hintermüller and M. Ulbrich, A mesh-independence result for semismooth Newton methods, Math. Program., 101 (2004), pp. 151–184.DOI
  20. [20]K. Ito and K. Kunisch, Lagrange Multiplier Approach to Variational Problems and Appli- cations, vol. 15 of Advances in Design and Control, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2008.
  21. [21]D. P. Kouri, A multilevel stochastic collocation algorithm for optimization of PDEs with uncertain coefficients, SIAM/ASA J. Uncertain. Quantif., 2 (2014), pp. 55–81.DOI
  22. [22]D. P. Kouri, M. Heinkenschloss, D. Ridzal, and B. van Bloemen Waanders, A trust-region algorithm with adaptive stochastic collocation for PDE optimization under uncertainty, SIAM J. Sci. Comput., 35 (2013), pp. A1847–A1879.
  23. [23]O. Le Maître and O. M. Knio, Spectral Methods for Uncertainty Quantification: With Applications to Computational Fluid Dynamics, Scientific Computation, Springer, 2010.
  24. [24]H.-C. Lee and J. Lee, A stochastic Galerkin method for stochastic control problems, Commun. Comput. Phys., 14 (2013), pp. 77–106.DOI
  25. [25]Z. Li, A. Onwunta, and B. Sousedík, Stochastic galerkin method and hierarchical preconditioning for pde-constrained optimization, arXiv preprint arXiv:2512.23804, (2025).arxiv.org/abs/2512.23804
  26. [26]G. J. Lord, C. E. Powell, and T. Shardlow, An Introduction to Computational Stochastic PDEs, Cambridge Texts in Applied Mathematics, Cambridge University Press, 2014.DOI
  27. [27]J. Martínez-Frutos and F. Periago Esparza, Optimal Control of PDEs under Uncertainty: An Introduction with Application to Optimal Shape Design of Structures, Springer Briefs in Mathematics, Springer Cham, 2018.
  28. [28]R. Mifflin, Semismooth and semiconvex functions in constrained optimization, SIAM J. Control Optim., 15 (1977), pp. 959–972.DOI
  29. [29]J. W. Pearson and A. J. Wathen, A new approximation of the Schur complement in preconditioners for PDE-constrained optimization, Numer. Linear Algebra Appl., 19 (2012), pp. 816–829.DOI
  30. [30]L. Qi and J. Sun, A nonsmooth version of Newton’s method, Math. Program., 58 (1993), pp. 353–367.DOI
  31. [31]R. T. Rockafellar, Convex analysis, vol. No. 28 of Princeton Mathematical Series, Princeton University Press, Princeton, NJ, 1970.
  32. [32]E. Rosseel and G. N. Wells, Optimal control with stochastic PDE constraints and uncertain controls, Comput. Methods Appl. Mech. Engrg., 213–216 (2012), pp. 152–167.DOI
  33. [33]W. Shen, T. Sun, B. Gong, and W. Liu, Stochastic Galerkin method for constrained optimal control problem governed by an elliptic integro-differential PDE with stochastic coefficients, Int. J. Numer. Anal. Model., 12 (2015), pp. 593–616.DOI
  34. [34]H. Tiesler, R. M. Kirby, D. Xiu, and T. Preusser, Stochastic collocation for optimal control problems with stochastic PDE constraints, SIAM J. Control Optim., 50 (2012), pp. 2659–2682.DOI
  35. [35]F. Tröltzsch, Optimal Control of Partial Differential Equations: Theory, Methods and Applications, American Mathematical Society, Providence, Rhode Island, 2010.
  36. [36]M. Ulbrich, Semismooth Newton Methods for Variational Inequalities and Constrained Optimization Problems in Function Spaces, vol. 11 of MOS-SIAM Series on Optimization, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2011.
  37. [37]H. Wang, D. Wu, C. Yu, and K. L. Teo, A duality-based approach for linear parabolic optimal control problems, Optimal Control Appl. Methods, 45 (2024), pp. 1140–1165, https://doi.org/10.1002/oca.3094, https://doi.org/10.1002/oca.3094.
  38. [38]M. Winkler and H. Yücel, A stochastic Galerkin method for optimal Dirichlet boundary control problems with uncertain data, 2025, https://arxiv.org/abs/2506.11479.
  39. [39]D. Xiu, Numerical Methods for Stochastic Computations: A Spectral Method Approach, Princeton University Press, 2010.
  40. [40]D. Xiu and G. E. Karniadakis, Modeling uncertainty in flow simulations via generalized polynomial chaos, Journal of computational physics, 187 (2003), pp. 137–167.DOI

Paper details

vlnc:W_3mmvrydp
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 7

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents