Search papers, labs, and topics across Lattice.
This paper introduces a scaled Traveling Wave Physics-Informed Neural Network (TW-PINN) to solve n-dimensional reaction-diffusion equations with variable coefficients. By applying a scaling transformation, the original problem is reduced to a simpler 1D equation, allowing a single trained PINN to generalize across different coefficients and dimensions. Numerical results demonstrate the method's accuracy, flexibility, and improved performance compared to existing wave-PINN approaches.
A single, pre-trained neural network can now solve a wide range of reaction-diffusion equations with varying coefficients and dimensions, thanks to a clever scaling transformation.
We propose an efficient and generalizable physics-informed neural network (PINN) framework for computing traveling wave solutions of $n$-dimensional reaction-diffusion equations with various reaction and diffusion coefficients. By applying a scaling transformation with the traveling wave form, the original problem is reduced to a one-dimensional scaled reaction-diffusion equation with unit reaction and diffusion coefficients. This reduction leads to the proposed framework, termed scaled TW-PINN, in which a single PINN solver trained on the scaled equation is reused for different coefficient choices and spatial dimensions. We also prove a universal approximation property of the proposed PINN solver for traveling wave solutions. Numerical experiments in one and two dimensions, together with a comparison to the existing wave-PINN method, demonstrate the accuracy, flexibility, and superior performance of scaled TW-PINN. Finally, we explore an extension of the framework to the Fisher's equation with general initial conditions.
