Abstract:Spectral methods provide highly accurate numerical solutions for partial differential equations, exhibiting exponential convergence with the number of spectral nodes. Traditionally, in addressing time-dependent nonlinear problems, attention has been on low-order finite difference schemes for time discretization and spectral element schemes for spatial variables. However, our recent developments have resulted in the application of spectral methods to both space and time variables, preserving spectral convergence in both domains. Leveraging Tensor Train techniques, our approach tackles the curse of dimensionality inherent in space-time methods. Here, we extend this methodology to the nonlinear time-dependent convection-diffusion equation. Our discretization scheme exhibits a low-rank structure, facilitating translation to tensor-train (TT) format. Nevertheless, controlling the TT-rank across Newton's iterations, needed to deal with the nonlinearity, poses a challenge, leading us to devise the "Step Truncation TT-Newton" method. We demonstrate the exponential convergence of our methods through various benchmark examples. Importantly, our scheme offers significantly reduced memory requirement compared to the full-grid scheme.

What problem does this paper attempt to address?

### Problems Addressed by the Paper The paper aims to address the time-dependent problem of Nonlinear Convection-Diffusion (NCD) equations. Specifically, the authors propose a Tensor Network (TN)-based method to overcome the curse of dimensionality encountered in traditional time-stepping methods and space-time discretization methods for high-dimensional problems. #### Main Contributions 1. **High-Precision Solution**: Utilizes Spectral Methods for numerical solutions, which have exponential convergence properties. 2. **Joint Handling of Time and Space Variables**: Incorporates both time and space variables into the spectral discretization framework using the Tensor Train (TT) format, maintaining spectral convergence. 3. **Low-Rank Structure Control**: Introduces a new "Step Truncation TT-Newton Method" to control the growth of tensor ranks during Newton iterations, thereby maintaining a low-rank structure. #### Experimental Validation 1. **Benchmark Testing**: Demonstrates the exponential convergence of the method through multiple benchmark experiments and significantly reduces memory requirements compared to full-grid methods. 2. **Numerical Efficiency**: Validates the effectiveness and computational efficiency of the method through three experiments: - Compares the performance of the full-grid Newton method, TT-Newton method with fixed truncation accuracy, and the Step Truncation TT-Newton method. - Uses a Manufactured Solution to verify the method's performance in 3D nonlinear convection-diffusion equations. - Solves the 3D viscous Burger's equation, further showcasing the method's advantages. In summary, the paper proposes an efficient and accurate Tensor Network method for solving the time-dependent problem of nonlinear convection-diffusion equations, particularly suitable for high-dimensional problems.

Tensor Network Space-Time Spectral Collocation Method for Solving the Nonlinear Convection Diffusion Equation

Tensor Network Space-Time Spectral Collocation Method for Time Dependent Convection-Diffusion-Reaction Equations

Tensor Network Space-Time Spectral Collocation Method for Time-Dependent Convection–Diffusion–Reaction Equations

Space-Time Spectral Element Tensor Network Approach for Time Dependent Convection Diffusion Reaction Equation with Variable Coefficients

Tensor Networks for Solving the Time-Independent Boltzmann Neutron Transport Equation

Tensor Networks for Solving Realistic Time-independent Boltzmann Neutron Transport Equation

Solving differential Riccati equations: A nonlinear space-time method using tensor trains

A spectral collocation method for nonlocal diffusion equations

Tensor Train Decomposition for Solving Large-Scale Linear Equations.

Polynomial Spectral Collocation Method for Space Fractional Advection-Diffusion Equation

A Novel Alternating-Direction Implicit Spectral Galerkin Method for a Multi-Term Time-Space Fractional Diffusion Equation in Three Dimensions.

Novel Numerical Methods for Solving the Time-Space Fractional Diffusion Equation in Two Dimensions.

Tensor neural networks for high-dimensional Fokker-Planck equations

Hermite Spectral Methods with A Time-Dependent Scaling for Parabolic Equations in Unbounded Domains

A Spectral Galerkin Method For Nonlinear Delay Convection-Diffusion-Reaction Equations

Tensor-Train WENO Scheme for Compressible Flows

Tensor Neural Controlled Differential Equations

Network of Tensor Time Series

TT-M FE Method for a 2D Nonlinear Time Distributed-Order and Space Fractional Diffusion Equation.

Cross Interpolation for Solving High-Dimensional Dynamical Systems on Low-Rank Tucker and Tensor Train Manifolds

pETNNs: Partial Evolutionary Tensor Neural Networks for Solving Time-dependent Partial Differential Equations