An inertial minimal-deformation-rate framework for shape optimization

Falai ChenBuyang Li, Jiajie Li, Rong Tang · 2026

Read An inertial minimal-deformation-rate framework for shape optimization by Falai ChenBuyang Li, Jiajie Li, Rong Tang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-22605.txt.

Introduction

Falai Chen, Buyang Li, Jiajie Li, and Rong Tang speed shape design with a flow that has momentum and guards the mesh. Their 2026 preprint tackles PDE-constrained shape optimization, plus hole filling on Willmore-driven surfaces. Two pains rule the field: flat energy zones stall progress, and meshes warp as shapes move.

Their second-order inertial flow pushes through flat zones. A minimal-deformation rule keeps the mesh sound as it goes. Good shapes, reached fast, serve design from parts to implants. The work is a numerics advance with clear uses. It should travel well across fields.

Catalog record

CREATORFalai ChenBuyang Li, Jiajie Li, Rong Tang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE72 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-22605

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

An inertial minimal-deformation-rate framework for shape optimization Thanks: This work is supported in part by National Natural Science Foundation of China under grants (No. 12525111, No. 12494550, No. 12494555 and No. 12401534). Falai ChenBuyang Li, Jiajie Li, Rong Tang Abstract We propose a robust numerical framework for PDE–constrained shape optimization and Willmore-driven surface hole filling. To address two central challenges—slow progress in flat energy landscapes, which can trigger premature stagnation at suboptimal configurations, and mesh deterioration during geometric evolution—we couple a second–order inertial flow with a minimal–deformation–rate (MDR) mesh motion strategy. This coupling accelerates convergence while preserving mesh quality and thus avoids remeshing. To further enhance robustness for non-smooth or non-convex initial geometries, we incorporate surface–diffusion regularization within the Barrett–Garcke–Nürnberg (BGN) framework. Moreover, we extend the inertial MDR methodology to Willmore-type surface hole filling, enabling high-order smooth reconstructions even from incompatible initial data. Numerical experiments demonstrate markedly faster convergence to lower original objective values, together with consistently superior mesh preservation throughout the evolution. 1 Introduction Shape optimization constrained by partial differential equations (PDEs) is a fundamental problem in computational science and engineering, with applications ranging from aerodynamics [ 46 ] to structural design [ 1 , 35 ] . The central topic is to determine a domain \varOmega\subset\mathbb{R}^{d} ( d\in\{2,3\} ) with boundary \Gamma:=\partial\varOmega that minimizes a original objective functional J(\Gamma)=\int_{\varOmega}j(x,u) , subject to a PDE constraint [ 18 , 53 , 36 , 33 , 34 ] . A prototypical example is the Poisson equation–constrained shape design problem: \min_{\varOmega\in\mathcal{O}_{\rm ad}}J(\Gamma)=\int_{\varOmega}\frac{1}{2}|u-u_{d}|^{2}\quad\text{subject to}\quad\left\{\begin{aligned} -\varDelta u+u&=f&&\text{in }\varOmega,\\ u&=0&&\text{on }\Gamma,\end{aligned}\right. (1) where u_{d} denotes prescribed observation data and \mathcal{O}_{\rm ad} is the admissible set of Lipschitz domains. The practical relevance of such problems has stimulated extensive research on numerical methods for evolving geometries [ 13 , 32 , 17 , 56 , 47 , 26 ] as well as rigorous convergence analyses for time-dependent domains [ 44 , 11 , 4 , 23 , 21 , 29 , 50 , 24 , 25 , 40 , 42 , 30 ] . A standard solution strategy relies on shape gradient flows [ 12 , 16 , 20 , 31 ] , in which the descent velocity \tilde{\bm{w}} is defined via the Riesz representation in a Hilbert space \mathbb{H} by \displaystyle b(\tilde{\bm{w}},\bm{v})=-\,\mathrm{d}J(\Gamma;\bm{v})\quad\forall\,\bm{v}\in\mathbb{H}, (2) where b(\cdot,\cdot) denotes the inner product on \mathbb{H} and \mathrm{d}J(\Gamma;\cdot) is the Eulerian derivative. The choice of \mathbb{H} dictates the regularity of the flow; common examples include \mathbb{H}=\textbf{H}^{1}(\varOmega) , \mathbb{H}=\textbf{L}^{2}(\Gamma) , or \mathbb{H}=\textbf{H}^{1}(\Gamma) [ 52 , 51 ] . In two dimensions, conformal-mapping-based flows have also been explored to enhance mesh quality for shape minimization [ 39 ] . Let \varOmega(t):=\tilde{\bm{X}}(t)(\varOmega^{0}) denote the evolving domain induced by the flow map \tilde{\bm{X}}(t):\varOmega^{0}\to\mathbb{R}^{d} . The classical first-order shape gradient flow then reads \displaystyle\partial_{t}\tilde{\bm{X}}(t) \displaystyle=\tilde{\bm{w}}(t)\circ\tilde{\bm{X}}(t) \displaystyle\text{in }\varOmega^{0}, (3a) \displaystyle b(\tilde{\bm{w}}(t),\bm{v}) \displaystyle=-\,\mathrm{d}J(\Gamma(t);\bm{v}) \displaystyle\forall\,\bm{v}\in\mathbb{H}(t). (3b) Despite their simplicity, first-order dynamics ( 3 ) often converge slowly and cannot escape shallow local basins. Since the Riesz isomorphism \mathcal{R}:\mathbb{H}\to\mathbb{H}^{\ast} gives \tilde{\bm{w}}=-\mathcal{R}^{-1}\mathrm{d}J , the velocity vanishes as \mathrm{d}J\to 0 , causing stagnation in flat regions and premature convergence. This motivates inertial effects, which introduce kinetic memory to traverse plateaus and escape shallow minima. Conceptually, second-order (inertial) flows are inspired by accelerated dynamics in convex optimization [ 3 , 22 ] and can be viewed as continuous-time analogues of Nesterov-type acceleration [ 49 , 54 ] . This perspective has recently been extended to infinite-dimensional problems, including nonlinear second-order PDEs for Bose–Einstein condensates [ 15 ] and dissipative inertial dynamics on Sobolev spaces for non-convex energies [ 14 ] . Motivated by these developments, we consider an inertial second-order evolution for PDE-constrained shape optimization of the form \epsilon_{0}\,a\big(\partial_{tt}^{\bullet}\tilde{\bm{X}}(t),\,\partial_{t}^{\bullet}\tilde{\bm{X}}(t),\,\bm{v}\big)+\eta(t)\,b\big(\partial_{t}^{\bullet}\tilde{\bm{X}}(t),\,\bm{v}\big)=-\mathrm{d}J\big(\Gamma(t);\,\bm{v}\big),\qquad\forall\,\bm{v}\in\mathbb{H}(t), (4) where \partial_{tt}^{\bullet}\tilde{\bm{X}}(t) denotes the second material derivative, \epsilon_{0}\geqslant 0 and \eta(t)>0 are inertia and damping parameters, and the first-order gradient flow ( 3 ) is recovered by setting \epsilon_{0}\equiv 0 and \eta(t)\equiv 1 . The term a(\cdot,\cdot,\cdot) accounts for Reynolds-type transport effects induced by the evolving geometry. This formulation admits the mechanical energy \mathcal{H}(t):=J\big(\Gamma(t)\big)+\frac{\epsilon_{0}}{2}\,\|\partial_{t}^{\bullet}\tilde{\bm{X}}(t)\|_{\textbf{L}^{2}(\varOmega(t))}^{2}, (5) which satisfies the energy dissipation property \frac{\mathrm{d}}{\mathrm{d}t}\mathcal{H}(t)\leqslant 0 . While the inertial term allows for temporary non-monotonicity of the original objective functional J(\Gamma(t)) , the damping term \eta(t) ensures energy dissipation and suppresses spurious oscillations. At the discrete level, the resulting scheme is linear at each time step, computationally efficient, and exhibits significantly faster convergence in practice compared to the corresponding first-order flow ( 3 ), as demonstrated in Section 6 . However, inertia alone does not prevent mesh degradation, which can occur for both first- and second-order flows. In the continuous setting, the shape of \varOmega(t) is determined only by the normal boundary velocity; the tangential boundary motion and the extension into the interior are shape-invariant degrees of freedom, yet crucial for mesh-quality control. Standard interior extensions, such as harmonic extension [ 11 , 21 , 42 ] , may lead to severe mesh distortion or even element entanglement under large deformations. This motivates incorporating a dedicated mesh-motion strategy to preserve mesh quality throughout the evolution and avoid remeshing. To address this mesh-quality issue while respecting the prescribed shape evolution, we adopt and extend the Minimal Deformation Rate (MDR) framework [ 5 ] . In this framework, we impose the normal component of the boundary velocity as a hard constraint, namely \bm{w}\cdot\bm{n}|_{\Gamma(t)}=\tilde{\bm{w}}\cdot\bm{n} , where \tilde{\bm{w}} is the velocity prescribed by the inertial flow ( 4 ), with \tilde{\bm{w}}=\partial_{t}^{\bullet}\tilde{\bm{X}}(t) . The bulk velocity field \bm{w} is then obtained by minimizing the instantaneous deformation-rate energy \frac{1}{2}\int_{\varOmega(t)}\bigl|\varepsilon(\bm{w})\bigr|^{2},\quad\text{subject to }\bm{w}\cdot\bm{n}|_{\Gamma(t)}=\tilde{\bm{w}}\cdot\bm{n}, (6) where \displaystyle\varepsilon(\bm{w}):=\tfrac{1}{2}(\nabla\bm{w}+\nabla\bm{w}^{\top}) (7) denotes the symmetric strain-rate tensor. This constrained minimization yields a linear variational problem at each time step, requiring only a single linear solver. By directly minimizing the deformation rate, the proposed method maintains high mesh quality throughout the evolution. It significantly outperforms both first- and second-order flows as well as classical approaches based on standard harmonic extension, offering a robust solution where specialized two-dimensional remedies, such as conformal mappings [ 39 ] , do not readily generalize to three dimensions. Despite their favorable theoretical properties, both classical gradient flows and the proposed inertial MDR framework remain prone to numerical instabilities when applied to non-smooth (e.g., cornered) or non-convex initial geometries. Such geometric irregularities typically induce severe mesh entanglement near corners and reentrant boundaries, often precipitating solver failure. Consequently, much of the existing literature has focused primarily on convex [ 58 , 30 , 37 ] or smooth star-shaped domains [ 2 ] . To address these robustness challenges, we regularize the second-order flow ( 4 ) by incorporating a surface-diffusion term. Surface diffusion, which can be interpreted as the H^{-1} -gradient flow of the area functional [ 48 , 6 , 7 , 43 ] , is particularly suitable in this context, as it effectively smooths geometric singularities while preserving the enclosed volume—an essential property for PDE-constrained shape optimization. We implement this regularization by modifying the normal-velocity constraint as \displaystyle\hat{\bm{w}}\cdot\bm{n}=\bigl(\tilde{\bm{w}}+\alpha\,\varDelta_{\Gamma}H\,\bm{n}\bigr)\cdot\bm{n}\quad\text{on }\Gamma(t), (8) where \tilde{\bm{w}} is the velocity prescribed by the inertial flow ( 4 ), and H is the mean curvature. The regularization parameter is adapted dynamically as \alpha=\|\tilde{\bm{w}}\|_{{\bf L}^{2}(\Gamma(t))} , so that geometric smoothing is dominant in the early stages to resolve singularities and gradually diminishes as the flow approaches equilibrium. To discretize the regularized

Read the complete text