Mesh-Free Numerical Approximation of the Biharmonic Equation via Optimized Kolmogorov--Arnold Neural Networks

License: See LICENSE file · Updated 2026-09-30
My rating

The paper targets structural mechanics problems involving thin-plate deflection and the biharmonic equation. It develops a mesh-free numerical framework based on Kolmogorov--Arnold physics-informed neural networks, using learnable univariate functions parameterized by radial basis functions on network edges instead of fixed nodal activations in conventional multilayer perceptrons.

Tags: