An extended physics informed neural network for preliminary analysis of parametric optimal control problems
From MaRDI portal
Publication:6104895
deep learningoptimal control problemparametric partial differential equationsphysics-informed neural network
Navier-Stokes equations (35Q30) Optimality conditions for problems involving partial differential equations (49K20) Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs (65N30) Artificial neural networks and deep learning (68T07) Navier-Stokes equations for incompressible viscous fluids (76D05)
Abstract: In this work we propose an extension of physics informed supervised learning strategies to parametric partial differential equations. Indeed, even if the latter are indisputably useful in many applications, they can be computationally expensive most of all in a real-time and many-query setting. Thus, our main goal is to provide a physics informed learning paradigm to simulate parametrized phenomena in a small amount of time. The physics information will be exploited in many ways, in the loss function (standard physics informed neural networks), as an augmented input (extra feature employment) and as a guideline to build an effective structure for the neural network (physics informed architecture). These three aspects, combined together, will lead to a faster training phase and to a more accurate parametric prediction. The methodology has been tested for several equations and also in an optimal control framework.
Recommendations
- AONN: An Adjoint-Oriented Neural Network Method for All-At-Once Solutions of Parametric Optimal Control Problems
- Optimal control of PDEs using physics-informed neural networks
- A metalearning approach for physics-informed neural networks (PINNs): application to parameterized PDEs
- Control of partial differential equations via physics-informed neural networks
- Optimization with learning-informed differential equation constraints and its applications
Cites work
- \textit{hp}-VPINNs: variational physics-informed neural networks with domain decomposition
- A MULTI-FIDELITY NEURAL NETWORK SURROGATE SAMPLING METHOD FOR UNCERTAINTY QUANTIFICATION
- Adaptive activation functions accelerate convergence in deep and physics-informed neural networks
- All-optical machine learning using diffractive deep neural networks
- Analytical solutions of boundary values problem of 2D and 3D Poisson and biharmonic equations by homotopy decomposition method
- B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy data
- Conservative physics-informed neural networks on discrete domains for conservation laws: applications to forward and inverse problems
- Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
- scientific article; zbMATH DE number 5703572 (Why is no real title available?)
- scientific article; zbMATH DE number 6982909 (Why is no real title available?)
- scientific article; zbMATH DE number 3322278 (Why is no real title available?)
- Lagrange Multiplier Approach to Variational Problems and Applications
- Least-squares finite element methods
- Locally adaptive activation functions with slope recovery for deep and physics-informed neural networks
- Model reduction for parametrized optimal control problems in environmental marine sciences and engineering
- Multi-fidelity regression using artificial neural networks: efficient approximation of parameter-dependent output quantities
- nPINNs: nonlocal physics-informed neural networks for a parametrized nonlocal universal Laplacian operator. Algorithms and applications
- NSFnets (Navier-Stokes flow nets): physics-informed neural networks for the incompressible Navier-Stokes equations
- Numerical approximation of a control problem for advection-diffusion processes
- Numerical approximation of partial differential equations
- Optimal control of Navier-Stokes equations by Oseen approximation
- Optimal Control of the Stationary Navier–Stokes Equations with Mixed Control‐State Constraints
- Optimal flow control for Navier–Stokes equations: drag minimization
- Optimization with PDE Constraints
- Perspectives in Flow Control and Optimization
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- PPINN: parareal physics-informed neural network for time-dependent PDEs
- Reduced basis approximation of parametrized optimal flow control problems for the Stokes equations
- Reduced order methods for parametrized non-linear and time dependent optimal flow control problems, towards applications in biomedical and environmental sciences
- Trends in PDE constrained optimization
Cited in
(17)- A metalearning approach for physics-informed neural networks (PINNs): application to parameterized PDEs
- Relaxation approach for learning neural network regularizers for a class of identification problems
- AONN: An Adjoint-Oriented Neural Network Method for All-At-Once Solutions of Parametric Optimal Control Problems
- Solving high-dimensional parametric engineering problems for inviscid flow around airfoils based on physics-informed neural networks
- An analysis and solution of ill-conditioning in physics-informed neural networks
- Deep mixed residual method for solving PDE-constrained optimization problems
- Regime-aware time weighting for physics-informed neural networks
- Stabilizing and solving unique continuation problems by parameterizing data and learning finite element solution operators
- Solving elliptic optimal control problems via neural networks and optimality system
- Physics informed neural network framework for unsteady discretized reduced order system
- A comparison study of supervised learning techniques for the approximation of high dimensional functions and feedback control
- A PINN approach for the online identification and control of unknown PDEs
- Step-by-step time discrete physics-informed neural networks with application to a sustainability PDE model
- Neural network approaches for parameterized optimal control
- Real-time optimal control of high-dimensional parametrized systems by deep learning-based reduced order models
- A flexible interface learning framework for collaborative physics-informed neural network
- Randomized neural network for solving PDE-constrained optimization problems
This page was built for publication: An extended physics informed neural network for preliminary analysis of parametric optimal control problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6104895)
