Spatially dependent node regularity in meshless approximation of partial differential equations

Miha Rot,Mitja Jančič,Gregor Kosec

2024-04-12

Abstract:In this paper, we address a way to reduce the total computational cost of meshless approximation by reducing the required stencil size through spatially varying computational node regularity. Rather than covering the entire domain with scattered nodes, only regions with geometric details are covered with scattered nodes, while the rest of the domain is discretized with regular nodes. A simpler approximation can be used in regions covered by regular nodes, effectively reducing the required stencil size and computational cost compared to the approximation on scattered nodes where a set of polyharmonic splines is added to ensure convergent behaviour.

Numerical Analysis,Fluid Dynamics

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to reduce the total computational cost of the meshless method when approximating partial differential equations. Specifically, the author proposes a new Hybrid Regular - Scattering Node Positioning algorithm (HyNP). By using scattering nodes in geometrically detailed regions and regular nodes in the remaining regions, the required stencil size and computational cost are reduced.

### Summary of Main Problems

1. **Reduce Computational Cost**: Traditional meshless methods require a large stencil size to ensure stability and accuracy, which leads to high computational costs. The method proposed in this paper reduces the stencil size by using regular nodes in areas where high accuracy is not required, thereby reducing the computational complexity.

2. **Handle Geometric Complexity**: For problems with complex geometric shapes, traditional meshless methods usually use scattering nodes throughout the domain. Although this can better adapt to geometric details, it significantly increases the computational burden. The HyNP algorithm proposed in this paper can use as many regular nodes as possible while maintaining the accuracy of geometric description to improve computational efficiency.

3. **Spatial - Dependent Node Regularity**: The author introduces a spatial - dependent node regularity strategy, that is, dynamically adjusting the distribution of nodes according to the geometric characteristics of the region. This method not only improves computational efficiency but also ensures the stability of the numerical solution.

### Specific Solutions

- **HyNP Algorithm**: This algorithm can automatically generate a hybrid regular - scattering node layout, where regular nodes are used for simple geometric regions and scattering nodes are used to handle complex geometric boundaries. The algorithm supports variable node density (h - refinement) and can be applied in domains of different dimensions.

- **Numerical Experiments**: To verify the effectiveness of the proposed method, the author conducted multiple numerical experiments, including nonlinear natural convection problems and contact problems. The experimental results show that the hybrid regular - scattering node layout not only improves computational efficiency but also has an accuracy equivalent to or better than that of the full - scattering node layout.

### Key Formulas

- **Node Spacing Function**:
\[
h: \Omega \subset \mathbb{R}^d \to (0, \infty)
\]
This function defines the distance between nodes and supports local refinement.

- **Separation Distance**:
\[
d_{i,j} = \| p_i - p_{n(i,j)} \|
\]
It represents the distance between the \(i\) - th node and its \(j\) - th nearest neighbor.

- **Maximum Void Diameter**:
\[
s_j = 2 \min_i \| p_i - v_j \|
\]
It represents the maximum hypersphere diameter that can be inscribed between the generated nodes.

Through these methods and formulas, the author successfully shows how to significantly reduce the computational cost of the meshless method while ensuring accuracy.