Abstract:The Local Randomized Neural Networks with Discontinuous Galerkin (LRNN-DG) methods, introduced in [42], were originally designed for solving linear partial differential equations. In this paper, we extend the LRNN-DG methods to solve nonlinear PDEs, specifically the Korteweg-de Vries (KdV) equation and the Burgers equation, utilizing a space-time approach. Additionally, we introduce adaptive domain decomposition and a characteristic direction approach to enhance the efficiency of the proposed methods. Numerical experiments demonstrate that the proposed methods achieve high accuracy with fewer degrees of freedom, additionally, adaptive domain decomposition and a characteristic direction approach significantly improve computational efficiency.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: to extend the Local Randomized Neural Networks with Discontinuous Galerkin (LRNN - DG) method from linear partial differential equations to nonlinear partial differential equations (PDEs), specifically the Korteweg - de Vries (KdV) equation and the Burgers equation. The authors introduce the space - time method, adaptive domain decomposition, and the characteristic direction method to improve the efficiency and accuracy of solving these nonlinear equations. ### Specific background of the problem 1. **KdV equation**: - The KdV equation is used to study wave behavior on shallow water surfaces and has a nonlinear term and a third - order derivative term. - Its mathematical form is: \[ u_t+u_{xxx}+uu_x = 0, \quad (t, x)\in I\times\Omega \] - The boundary conditions and initial conditions are as follows: \[ u(t, x_0)=g_0, \quad \frac{\partial u}{\partial x}(t, x_N)=g_1, \quad \frac{\partial^2 u}{\partial x^2}(t, x_N)=g_2, \quad u(0, x)=u_0 \] 2. **Burgers equation**: - The Burgers equation was originally proposed by Burgers for turbulence modeling and is widely used in the fields of physics and engineering. - Its mathematical form is: \[ u_t+u(\nabla u\cdot I)-\epsilon\Delta u = 0, \quad (t, x)\in I\times\Omega \] - The boundary conditions and initial conditions are as follows: \[ u(t_0, x)=u_0, \quad u(t, x)=g \quad \text{on} \quad I\times\partial\Omega \] ### Research objectives - **Extension of the LRNN - DG method**: Extend the LRNN - DG method from linear PDEs to nonlinear PDEs (such as the KdV and Burgers equations). - **Improve computational efficiency**: By introducing adaptive domain decomposition and the characteristic direction method, reduce the number of degrees of freedom and significantly improve computational efficiency. - **Verify the effectiveness of the method**: Verify the high - precision and low - degree - of - freedom requirements of the proposed method through numerical experiments. ### Solutions 1. **Space - time method**: Combine the time - space discretization method to reduce error accumulation and improve the solution accuracy. 2. **Adaptive domain decomposition**: Dynamically adjust the mesh division according to the error distribution of the numerical solution and concentrate resources in areas with larger errors. 3. **Characteristic direction method**: Utilize known characteristic directions (such as wave propagation directions) to pre - divide sub - regions and enhance the approximation ability of the local network. ### Conclusion The paper demonstrates the efficiency and accuracy of the LRNN - DG method in solving the nonlinear KdV and Burgers equations and verifies its superior performance through numerical experiments.

Local Randomized Neural Networks with Discontinuous Galerkin Methods for KdV-type and Burgers Equations

Local Randomized Neural Networks with Discontinuous Galerkin Methods for Partial Differential Equations

An Ultra-weak Local Discontinuous Galerkin Method with Generalized Numerical Fluxes for the KdV–Burgers–Kuramoto Equation

Local Randomized Neural Networks with Hybridized Discontinuous Petrov-Galerkin Methods for Stokes-Darcy Flows

Analysis of a local discontinuous Galerkin scheme for fractional Korteweg-de Vries equation

Randomized neural network with Petrov–Galerkin methods for solving linear and nonlinear partial differential equations

A Local Discontinuous Galerkin Method for Nonlinear Parabolic SPDEs

GNRK: Graph Neural Runge-Kutta method for solving partial differential equations

An Ultra-Weak Discontinuous Galerkin Method with Implicit–Explicit Time-Marching for Generalized Stochastic KdV Equations

Numerical analysis of variable-order fractional KdV-Burgers-Kuramoto equation

Local discontinuous Galerkin method for nonlinear BSPDEs of Neumann boundary conditions with deep backward dynamic programming time-marching

A Discontinuous Galerkin Method Based On A Bgk Scheme For The Navier-Stokes Equations On Arbitrary Grids

Multi-Grade Deep Learning for Partial Differential Equations with Applications to the Burgers Equation

Enhancing Low-Order Discontinuous Galerkin Methods with Neural Ordinary Differential Equations for Compressible Navier--Stokes Equations

Runge–Kutta Discontinuous Galerkin Methods for Convection-Dominated Problems

A BGK-Based Discontinuous Galerkin Method for the Navier-Stokes Equations on Arbitrary Grids

Numerical analysis of a class of penalty discontinuous Galerkin methods for nonlocal diffusion problems

The Runge--Kutta discontinuous Galerkin method with stage-dependent polynomial spaces for hyperbolic conservation laws

Local Extreme Learning Machines and Domain Decomposition for Solving Linear and Nonlinear Partial Differential Equations

Semi-discrete and fully discrete HDG methods for Burgers' equation