Journal of Computational Physics最新文献

筛选
英文 中文
Stochastic Norton dynamics: An alternative approach for the computation of transport coefficients in dissipative particle dynamics 随机诺顿动力学:耗散粒子动力学中输运系数计算的另一种方法
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-25 DOI: 10.1016/j.jcp.2025.114316
Xinyi Wu, Xiaocheng Shang
{"title":"Stochastic Norton dynamics: An alternative approach for the computation of transport coefficients in dissipative particle dynamics","authors":"Xinyi Wu,&nbsp;Xiaocheng Shang","doi":"10.1016/j.jcp.2025.114316","DOIUrl":"10.1016/j.jcp.2025.114316","url":null,"abstract":"<div><div>We study a novel alternative approach for the computation of transport coefficients at mesoscales. While standard nonequilibrium molecular dynamics (NEMD) approaches fix the forcing and measure the average induced flux in the system driven out of equilibrium, the so-called “stochastic Norton dynamics” instead fixes the value of the flux and measures the average magnitude of the forcing needed to induce it. We extend recent results obtained in Langevin dynamics to consider the generalisation of the stochastic Norton dynamics in the popular dissipative particle dynamics (DPD) at mesoscales, important for a wide range of complex fluids and soft matter applications. We demonstrate that the responses profiles for both the NEMD and stochastic Norton dynamics approaches coincide in both linear and nonlinear regimes, indicating that the stochastic Norton dynamics can indeed act as an alternative approach for the computation of transport coefficients, including the mobility and the shear viscosity, as the NEMD dynamics. In addition, based on the linear response of the DPD system with small perturbations, we derive a closed-form expression for the shear viscosity, and numerically validate its effectiveness with various types of external forces. Moreover, our numerical experiments demonstrate that the stochastic Norton dynamics approach clearly outperforms the NEMD dynamics in controlling the asymptotic variance, a key metric to measure the associated computational costs, particularly in the high friction limit.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114316"},"PeriodicalIF":3.8,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144988009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Cluster dynamics simulations using stochastic algorithms with logarithmic complexity in number of reactions 用反应数具有对数复杂度的随机算法模拟聚类动力学
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-24 DOI: 10.1016/j.jcp.2025.114318
Rohit Vasav , Thomas Jourdan , Gilles Adjanor , Achraf Badahmane , Jérôme Creuze , Manuel Athènes
{"title":"Cluster dynamics simulations using stochastic algorithms with logarithmic complexity in number of reactions","authors":"Rohit Vasav ,&nbsp;Thomas Jourdan ,&nbsp;Gilles Adjanor ,&nbsp;Achraf Badahmane ,&nbsp;Jérôme Creuze ,&nbsp;Manuel Athènes","doi":"10.1016/j.jcp.2025.114318","DOIUrl":"10.1016/j.jcp.2025.114318","url":null,"abstract":"<div><div>We show how to adapt and modify the Stochastic Simulation Algorithm (also sometimes referred to as the Gillespie algorithm or the kinetic Monte Carlo algorithm for rate equations) to efficiently simulate cluster dynamics equations for large systems, bypassing the linear complexity of the SSA by associating an internal time scale and priority queues with binary heaps for each mobile species. The internal time of a binary heap renormalises the physical time thus making it independent of the number of clusters of the mobile species concerned, while the binary heaps allow efficient sorting advances. The resulting algorithm has an algorithmic complexity that is linear with respect to the number of mobile cluster types, but logarithmic with respect to the number of immobile cluster species, thus being extremely effective when the number of mobile species is small, a situation satisfied in most models of cluster dynamics. As a physical application, we simulate the time evolution of defect clusters in a FeCu<span><math><msub><mrow></mrow><mrow><mn>0.1</mn><mspace></mspace><mo>%</mo></mrow></msub></math></span> system under irradiation, by integrating the associated cluster dynamics equations using the stochastic algorithm. The speed-up of the algorithm, compared to the direct method is substantial, about several orders of magnitude, while consuming much less memory than deterministic simulations.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114318"},"PeriodicalIF":3.8,"publicationDate":"2025-08-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144911744","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Quasi-Minnaert resonances in high-contrast acoustic structures and applications to invisibility cloaking 高对比声学结构中的准minnaert共振及其在隐形斗篷中的应用
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-23 DOI: 10.1016/j.jcp.2025.114310
Weisheng Zhou , Huaian Diao , Hongyu Liu
{"title":"Quasi-Minnaert resonances in high-contrast acoustic structures and applications to invisibility cloaking","authors":"Weisheng Zhou ,&nbsp;Huaian Diao ,&nbsp;Hongyu Liu","doi":"10.1016/j.jcp.2025.114310","DOIUrl":"10.1016/j.jcp.2025.114310","url":null,"abstract":"<div><div>This paper investigates a novel quasi-Minnaert resonance phenomenon in acoustic wave propagation through high-contrast medium in both two and three dimensions, occurring in the sub-wavelength regime. These media are characterized by physical properties significantly distinct from those of a homogeneous background. The quasi-Minnaert resonance is defined by two primary features: boundary localization, where the <span><math><msup><mi>L</mi><mn>2</mn></msup></math></span>-norms of the internal total field and the external scattered field exhibit pronounced concentration near the boundary, and surface resonance, marked by highly oscillatory behavior of the fields near the boundary. In contrast to classical Minnaert resonances, which are associated with a discrete spectral spectrum tied to physical parameters, quasi-Minnaert resonances exhibit analogous physical phenomena but with a continuous spectral spectrum. Using layer potential theory and rigorous asymptotic analysis, we demonstrate that the coupling between a high-contrast material structure, particularly with radial geometries, and a carefully designed incident wave is critical for inducing quasi-Minnaert resonances. Extensive numerical experiments, involving radial geometries (e.g., unit disks and spheres) and general-shaped geometries (e.g., hearts, corner, and clovers in <span><math><msup><mi>R</mi><mn>2</mn></msup></math></span>, and spheres in <span><math><msup><mi>R</mi><mn>3</mn></msup></math></span>), validate the occurrence of these resonances. Furthermore, we numerically demonstrate that quasi-Minnaert resonances induce an invisibility cloaking effect in the high-contrast medium. These findings have significant implications for mathematical material science and the development of acoustic cloaking technologies.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114310"},"PeriodicalIF":3.8,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145019860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
An improved machine learning strategy for shear analysis in structural concrete based on the Newton’s method and the solvability region of the steel constitutive model 基于牛顿法和钢本构模型可解区域的结构混凝土剪切分析改进机器学习策略
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-22 DOI: 10.1016/j.jcp.2025.114303
E. Lorente-Ramos , A.M. Hernández-Díaz , J. Pérez-Aracil , S. Salcedo-Sanz
{"title":"An improved machine learning strategy for shear analysis in structural concrete based on the Newton’s method and the solvability region of the steel constitutive model","authors":"E. Lorente-Ramos ,&nbsp;A.M. Hernández-Díaz ,&nbsp;J. Pérez-Aracil ,&nbsp;S. Salcedo-Sanz","doi":"10.1016/j.jcp.2025.114303","DOIUrl":"10.1016/j.jcp.2025.114303","url":null,"abstract":"<div><div>The Compression Field Theories (CFTs) can predict the full non-linear response of reinforced and prestressed concrete beams subjected to shear, considering the equilibrium and compatibility conditions in the cracked web of the beam, and the corresponding stress-strain relationships of the involved materials. These theories are supported by a set of non-linear algebraic governing equations whose numerical solution (which contains, among others, the inclination of the diagonal concrete struts, or crack angle) has been calculated through several strategies in the last years. Between them, machine learning methods arise as the best and most effective procedure for this aim, since, contrary to the traditional Newton-type methods, they do not require taking initial approximations to the numerical solution. Thus, in this work we present a new variant of the application of machine learning to the CFTs, based on Newton’s method as optimizer. Moreover, the training of this new strategy considered the existence of a solvability region for the steel constitutive model, as well as the location of the steel apparent yield strain in such region. In this sense, a classification sub-procedure about the location of such strain is implemented. As result, this new hybrid regression model, based on the Newton’s method and trained with previous classification, not only does not depend on initial approximations, but it predicts significantly better the experimental shear response of cracked reinforced and prestressed concrete beams than those machine learning strategies developed in previous works. In this work we propose an improved machine learning strategy for shear analysis in structural concrete based on the Newton’s method and the solvability region of the steel constitutive model</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114303"},"PeriodicalIF":3.8,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144903267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
A second-order relaxation flux solver for compressible Navier-Stokes equations based on generalized Riemann problem method 基于广义Riemann问题方法的可压缩Navier-Stokes方程二阶松弛通量求解器
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-21 DOI: 10.1016/j.jcp.2025.114314
Tuowei Chen , Zhifang Du
{"title":"A second-order relaxation flux solver for compressible Navier-Stokes equations based on generalized Riemann problem method","authors":"Tuowei Chen ,&nbsp;Zhifang Du","doi":"10.1016/j.jcp.2025.114314","DOIUrl":"10.1016/j.jcp.2025.114314","url":null,"abstract":"<div><div>In the finite volume framework, a Lax-Wendrof type second-order flux solver for the compressible Navier-Stokes equations is proposed by utilizing a hyperbolic relaxation model. The flux solver is developed by applying the generalized Riemann problem (GRP) method to the relaxation model that approximates the compressible Navier-Stokes equations. The GRP-based flux solver includes the effects of source terms in numerical fluxes and treats the stiff source terms implicitly, allowing a CFL condition conventionally used for the Euler equations. The trade-off is to solve local linear systems of algebraic equations. The resulting numerical scheme achieves second-order accuracy in both space and time within a single stage, and the linear systems are solved only once in a time step. The parameters to establish the relaxation model are allowed to be locally determined at each cell interface, improving the adaptability to diverse flow regions. Numerical tests with a wide range of flow problems, from nearly incompressible to supersonic flows with strong shocks, for both inviscid and viscous problems, demonstrate the high resolution of the current second-order scheme.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114314"},"PeriodicalIF":3.8,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144911743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Improving sampling by modifying the effective diffusion 通过改变有效扩散来改善采样
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-21 DOI: 10.1016/j.jcp.2025.114313
T. Lelièvre , R. Santet , G. Stoltz
{"title":"Improving sampling by modifying the effective diffusion","authors":"T. Lelièvre ,&nbsp;R. Santet ,&nbsp;G. Stoltz","doi":"10.1016/j.jcp.2025.114313","DOIUrl":"10.1016/j.jcp.2025.114313","url":null,"abstract":"<div><div>Markov chain Monte Carlo samplers based on discretizations of (overdamped) Langevin dynamics are commonly used in the Bayesian inference and computational statistical physics literature to estimate high-dimensional integrals. One can introduce a non-constant diffusion matrix to precondition these dynamics, and recent works have optimized it in order to improve the rate of convergence to stationarity by overcoming entropic and energy barriers. However, the introduced methodologies to compute these optimal diffusions are generally not suited to high-dimensional settings, as they rely on costly optimization procedures. In this work, we propose to optimize over a class of diffusion matrices, based on one-dimensional collective variables (CVs), to help the dynamics explore the latent space defined by the CV. The form of the diffusion matrix is chosen in order to obtain an efficient effective diffusion in the latent space. We describe how this class of diffusion matrices can be constructed and learned during the simulation. We provide implementations of the Metropolis–Adjusted Langevin Algorithm and Riemann Manifold (Generalized) Hamiltonian Monte Carlo algorithms, and discuss numerical optimizations in the case when the CV depends only on a few degrees of freedom of the system. We illustrate the efficiency gains by computing mean transition durations between two metastable states of a dimer in a solvent.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114313"},"PeriodicalIF":3.8,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144903266","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Provably positivity-preserving constrained transport scheme for 2D and 3D ideal magnetohydrodynamics 二维和三维理想磁流体力学的可证明保正约束输运方案
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-20 DOI: 10.1016/j.jcp.2025.114312
Dongwen Pang , Kailiang Wu
{"title":"Provably positivity-preserving constrained transport scheme for 2D and 3D ideal magnetohydrodynamics","authors":"Dongwen Pang ,&nbsp;Kailiang Wu","doi":"10.1016/j.jcp.2025.114312","DOIUrl":"10.1016/j.jcp.2025.114312","url":null,"abstract":"<div><div>This paper proposes and analyzes a robust second-order positivity-preserving constrained transport (PPCT) scheme for ideal magnetohydrodynamics (MHD) on non-staggered Cartesian meshes. The PPCT scheme provably preserves two crucial physical constraints: a globally discrete divergence-free (DDF) condition on the magnetic field and the positivity of density and pressure. Achieving both properties in a single framework is essential for stability and physical fidelity in MHD simulations, yet this has proven to be a challenging endeavor (existing works, e.g., [S. Ding &amp; K. Wu, <em>SIAM J. Sci. Comput.</em>, 46: A50–A79, 2024], achieve positivity and only a locally DDF property simultaneously). The PPCT method is motivated by a novel splitting technique proposed in [T.A. Dao, M. Nazarov &amp; I. Tomas, <em>J. Comput. Phys.</em>, 508: 113009, 2024], which splits the MHD system into an Euler subsystem with a steady magnetic field and a magnetic subsystem with steady density and internal energy. To achieve structure-preserving properties, the PPCT scheme uses a novel finite volume-finite difference (FV-FD) hybrid approach: a PP finite volume method for the Euler subsystem and a CT finite difference method for the magnetic subsystem. The two are coupled using Strang splitting. The finite volume method is based on a new PP limiter, which is proven to maintain the second-order accuracy of the reconstruction. The PP limiter enforces the positivity of the reconstructed values for density and pressure, as well as an a priori condition for the PP property of the updated cell averages. A rigorous theoretical proof of the PP property is provided using the geometric quasilinearization (GQL) approach [K. Wu &amp; C.-W. Shu, <em>SIAM Review</em>, 65:1031–1073, 2023]. For the magnetic subsystem, we construct an implicit finite difference CT method that conserves energy and preserves a globally DDF constraint on non-staggered Cartesian meshes. The resulting nonlinear algebraic system is solved with an iterative algorithm, reducing the residual error to machine precision within a few iterations. Unique solvability and convergence of this algorithm are theoretically proven under a CFL-like condition. Since the finite difference CT method for the magnetic subsystem is unconditionally energy-stable and preserves steady density and internal energy, the time step for the PP property and stability of the PPCT scheme is restricted only by a mild CFL condition for the scheme of the Euler subsystem. While the primary focus is on the 2D case for clarity, the 3D extension of the proposed PPCT framework is also presented. Several challenging 2D and 3D numerical experiments, including highly magnetized MHD jets with extremely high Mach numbers, validate the accuracy, robustness, and high resolution of the PPCT scheme.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114312"},"PeriodicalIF":3.8,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144906971","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Learning-enhanced variational regularization for electrical impedance tomography via Calderon's method 基于Calderon方法的电阻抗断层扫描的学习增强变分正则化
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-19 DOI: 10.1016/j.jcp.2025.114309
Kai Li , Kwancheol Shin , Zhi Zhou
{"title":"Learning-enhanced variational regularization for electrical impedance tomography via Calderon's method","authors":"Kai Li ,&nbsp;Kwancheol Shin ,&nbsp;Zhi Zhou","doi":"10.1016/j.jcp.2025.114309","DOIUrl":"10.1016/j.jcp.2025.114309","url":null,"abstract":"<div><div>This paper aims to numerically solve the two-dimensional electrical impedance tomography (EIT) with Cauchy data. This inverse problem is highly challenging due to its severe ill-posed nature and strong nonlinearity, which necessitates appropriate regularization strategies. Choosing a regularization approach that effectively incorporates the <em>a priori</em> information of the conductivity distribution (or its contrast) is therefore essential. In this work, we propose a deep learning-based method to capture the <em>a priori</em> information about the shape and location of the unknown contrast using Calderón’s method. The learned <em>a priori</em> information is then used to construct the regularization functional of the variational regularization method for solving the inverse problem. The resulting regularized variational problem for EIT reconstruction is then solved using the Gauss-Newton method. Extensive numerical experiments demonstrate that the proposed inversion algorithm achieves accurate reconstruction results, even in high-contrast cases, and exhibits strong generalization capabilities. Additionally, some stability and convergence analysis of the variational regularization method underscores the importance of incorporating <em>a priori</em> information about the support of the unknown contrast.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114309"},"PeriodicalIF":3.8,"publicationDate":"2025-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144903268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Orthogonal greedy algorithm for linear operator learning with shallow neural network 浅神经网络线性算子学习的正交贪心算法
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-18 DOI: 10.1016/j.jcp.2025.114308
Ye Lin , Jiwei Jia , Young Ju Lee , Ran Zhang
{"title":"Orthogonal greedy algorithm for linear operator learning with shallow neural network","authors":"Ye Lin ,&nbsp;Jiwei Jia ,&nbsp;Young Ju Lee ,&nbsp;Ran Zhang","doi":"10.1016/j.jcp.2025.114308","DOIUrl":"10.1016/j.jcp.2025.114308","url":null,"abstract":"<div><div>Greedy algorithms, particularly the orthogonal greedy algorithm (OGA), have proven effective in training shallow neural networks for fitting functions and solving partial differential equations (PDEs). In this paper, we extend the application of OGA to the task of linear operator learning, which is equivalent to learning the kernel function through integral transforms. First, we develop a novel greedy algorithm for kernel estimation with respect to a new semi-inner product, enabling the approximation of the Green’s function for linear PDEs from data. Second, we introduce OGA-based point-wise kernel estimation to further improve the approximation rate, achieving orders of accuracy improvement across various tasks over baseline models. Additionally, we provide a theoretical analysis on the kernel estimation problem with the semi-inner product, deriving the optimal approximation rates for both algorithms. Our results demonstrate their efficacy and potential for future applications in PDE solving and operator learning.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114308"},"PeriodicalIF":3.8,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144911742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Adversarial physics-informed neural networks with hard constraints for optimal control of PDEs 具有硬约束的对抗物理信息神经网络用于偏微分方程的最优控制
IF 3.8 2区 物理与天体物理
Journal of Computational Physics Pub Date : 2025-08-18 DOI: 10.1016/j.jcp.2025.114307
Yuandong Cao , Chi Chiu So , Yifan Dai , Siu Pang Yung , Jun-Min Wang
{"title":"Adversarial physics-informed neural networks with hard constraints for optimal control of PDEs","authors":"Yuandong Cao ,&nbsp;Chi Chiu So ,&nbsp;Yifan Dai ,&nbsp;Siu Pang Yung ,&nbsp;Jun-Min Wang","doi":"10.1016/j.jcp.2025.114307","DOIUrl":"10.1016/j.jcp.2025.114307","url":null,"abstract":"<div><div>Physics-informed neural networks (PINNs) have emerged as a promising deep learning approach for solving optimal control problems of partial differential equations (PDEs). Balancing the trade-offs between competing loss terms remains a significant challenge when using PINNs to solve optimal control problems of PDEs. The balance between different competing loss terms is crucial for control performance. Generative adversarial networks (GANs) have been proven to significantly improve the accuracy of PINNs in solving PDEs by introducing adversarial training methods to handle the weight relationships of different loss terms. In order to resolve the challenge, we propose the hard-constrained PINN adversarial method (hPINN-Adver), an innovative approach that integrates the PINN framework with GANs. The method aims to “learn the loss function” and dynamically adjust the weight relationships of different loss terms through adversarial training, thereby optimizing the balance between competing loss terms. We conduct detailed and comprehensive experiments to compare hPINN-Adver with soft-constrained PINN line search method (sPINN-Line), hard-constrained PINN line search method (hPINN-Line), hard-constrained PINN penalty method (hPINN-Penalty) and hard-constrained PINN augmented Lagrangian method (hPINN-Augmented) in solving five typical and representative optimal control problems of PDEs. Extensive numerical experiments demonstrate the great potential of hPINN-Adver method in solving optimal control problems of PDEs.</div></div>","PeriodicalId":352,"journal":{"name":"Journal of Computational Physics","volume":"541 ","pages":"Article 114307"},"PeriodicalIF":3.8,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144896571","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
相关产品
×
本文献相关产品
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信