The optimization design of structural shapes is fundamentally a problem of solving functional extremum. Traditional variational methods often encounter challenges, such as limited functional types and oscillation in the solution process when solving high-dimensional functional extreme value problems. In this paper, a functional extremum numerical solution method based on physics-informed deep learning (PIDL) is proposed by using the high-dimensional nonlinear mapping ability of deep learning model. The method first embeds the physical information (control equations, initial conditions and boundary conditions, etc.) of the shape optimization problem as regularization terms into the deep learning model, and a loss function based on the objective functional extremum is constructed. Then, a random gradient descent algorithm is used to train the deep learning model, further realizing the solution of functional extremum and optimization design of structural shape. The proposed method is verified through numerical examples of optimizing the shape of surfaces and arch axes, and a comparative analysis is conducted with the computational results obtained from genetic algorithms. The results demonstrate that the method has high prediction accuracy and efficiency for the target task of small samples. As a non-parametric modeling technology, the method is of great significance for solving engineering problems characterized by high data acquisition costs and data collection challenges.
LEVINEW .Optimal control theory:an introduction[J].IEEE Transactions on Automatic Control,1972,17(3): 423.
[2]
MAHDYA M S, YOUSSEFE S M .Numerical solution technique for solving isoperimetric variational problems[J].International Journal of Modern Physics C,2021, 32(1): 2150002.
[3]
RAZZAGHIM. Fourier series direct method for variational problems[J].International Journal of Control,1988,48(3): 887-895.
[4]
ZHOUC C, LIUY .The pade approximant based network for variational problems[J]. arXiv: 2004.00711, 2020.
[5]
GORNOVA Y, ZARODNYUKT S, ANIKINA S,et al .Extension technology and extrema selections in a stochastic multistart algorithm for optimal control problems[J].Journal of Global Optimization,2020,76(3):533-543.
[6]
VOGTT, STREKALOVSKIYE, CREMERSD,et al .Lifting methods for manifold-valued variational problems[M]//Handbook of Variational Methods for Nonlinear Geometric Data.Cham:Springer International Publishing,2020:95-119.
[7]
PEDREGALP .A direct algorithm for constrained variational problems in several dimensions[J].Computers & Mathematics with Applications,2018,75(1):105-121.
[8]
YEJ X,REY D, KADAKIAN,et al .Systematic variational method for statistical nonlinear state and parameter estimation[J].Physical Review.E,Statistical,Nonlinear,and Soft Matter Physics,2015,92(5):052901.
[9]
MERAN S, ELLIOTTL, INGHAMD B .Numerical solution of a boundary detection problem using genetic algorithms[J].Engineering Analysis with Boundary Elements,2004,28(4):405-411.
[10]
HOSSEINIS E, KARIMIO, ASEMANBAKHSH MALI .Experimental investigation and multi-objective optimization of savonius wind turbine based on modified non-dominated sorting genetic algorithm-II[J].Wind Engineering,2024,48(3):446-467.
[11]
KORDAV Y, BEREZOVSKYS V, MOLEVA S, et al .Solving variational problems via evolutionary algorithm[J].International Journal of Modern Physics C,2013,24(3):1350009.
[12]
HORNIKK, STINCHCOMBEM, WHITEH .Multilayer feedforward networks are universal approximators[J].Neural Networks,1989,2(5):359-366.
[13]
WEINANE, YUB. The deep ritz method: a deep learning-based numerical algorithm for solving variational problems[J]. Communications in Mathematics and Statistics, 2018, 6(1):1-12.
[14]
ARAZJ Y, CRIADOJ C, SPANNOWSKYM .Elvet: a neural network-based differential equation and variational problem solver[J]. arXiv: 2103.14575,2021.
[15]
ABADIM, BARHAMP, CHENJ M, et al. TensorFlow:a system for large-scale machine learning[C]//Proceedings of the 12th USENIX Symposium on Operating Systems Design and Implementation(OSDI’16). Savannah, GA: USENIX Association, 2016:265-283.
[16]
QIUY T, ARUNACHALAP K, LINDERC .SenseNet:a physics-informed deep learning model for shape sensing[J]. Journal of Engineering Mechanics,2023,149(3):04023002.
[17]
LIUJ Q, CHENR Q, LOUJ H,et al .Deep-learning-based aerodynamic shape optimization of rotor airfoils to suppress dynamic stall[J].Aerospace Science and Technology,2023,133:108089.
[18]
RAISSIM, PERDIKARISP, KARNIADAKISG E. Physics-informed neural networks:a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J].Journal of Computational Physics,2019, 378: 686-707.
[19]
RAMABATHIRANA A, RAMACHANDRANP .SPINN:sparse,physics-based,and partially interpretable neural networks for PDEs[J].Journal of Computational Physics,2021,445:110600.
[20]
LEUNGW T, LING, ZHANGZ C. NH-PINN: neural homogenization-based physics-informed neural network for multiscale problems[J]. Journal of Computational Physics, 2022,470: 111539.
[21]
RAISSIM, YAZDANIA, KARNIADAKISG E .Hidden fluid mechanics:learning velocity and pressure fields from flow visualizations[J].Science, 2020, 367(6481): 1026-1030.
TANGH S, HEZ P, LIAOY Y, et al .Forward and inverse problems of thin plate mechanics based on physics-informed deep transfer learning[J].Engineering Mechanics, 2023, 40(8):1-10.(in Chinese)
[26]
BAYDINA G, PEARLMUTTERB A, RADULA A,et al .Automatic differentiation in machine learning: a survey[J]. The Journal of Machine Learning Research, 2018, 18(1): 5595-5637.
[27]
UDDINM J, MOHEUDDINM M, KOWSHERM .A new study of trapezoidal,Simpson’s 1/3 and Simpson’s 3/8 rules of numerical integral problems[J]. Applied Mathematics and Sciences an International Journal (MathSJ),2019,6(4):1-13.
[28]
KINGMAD P, BAJ L. Adam: a method for stochastic optimization[C]//ICLR 2015-Conference Track Proceedings,2015.
[29]
HAJIS H, ABDULAZEEZA M .Comparison of optimization techniques based on gradient descent algorithm: a review[J].PalArch’s Journal of Archaeology of Egypt/Egyptology, 2021, 18(4):2715-2743.
[30]
SAMANIEGOE, ANITESCUC, GOSWAMIS,et al .An energy approach to the solution of partial differential equations in computational mechanics via machine learning:concepts,implementation and applications[J].Computer Methods in Applied Mechanics and Engineering,2020,362:112790.