On Waldron Interpolation on a Simplex in $\mathbb{R}^d$

Len Bos,Sione Ma'u,Shayne Waldron

2023-06-14

Abstract:We introduce explicit families of good interpolation points for interpolation on a triangle in $\mathbb{R}^2$ that may be used for either polynomial interpolation or a certain rational interpolation for which we give explicit formulas.

Numerical Analysis

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: to find a set of good interpolation points on a multi - dimensional simplex, which can be used for polynomial interpolation or specific rational function interpolation. Specifically, the author proposes the Waldron interpolation method, which improves the traditional barycentric coordinate interpolation method by introducing an appropriate weight function.

### Core of the problem

1. **Selection of interpolation points**:
- The selection of interpolation points on a two - dimensional triangle is a classic problem, but how to extend it to higher dimensions and ensure the stability and accuracy of interpolation is a challenge.
- The author proposes a new set of interpolation points - Waldron points, which are based on an admissible weight function \( w(x) \) and can provide better interpolation performance in some cases.

2. **Stability of interpolation**:
- The stability of interpolation is usually measured by the Lebesgue constant. The smaller the Lebesgue constant, the better the stability of interpolation.
- The author compares the Lebesgue constants of Waldron points with those of other known sets of interpolation points (such as Simplex points and concentric triangle points) to evaluate their performance.

3. **Geometric interpretation and application**:
- The paper also explores the geometric meaning of these interpolation points, especially their mapping on the sphere and their distribution in high - dimensional space.
- By mapping Waldron points onto the sphere, the author shows the uniform distribution characteristics of these points on the sphere and analyzes the distances between them.

### Summary of formulas

- **Definition of Waldron points**:
For a simplex \( S_d \) of degree \( n \), Waldron points are defined as:
\[
W_n := \left\{ x_\alpha = \sum_{j = 1}^{d + 1} \omega_j V_j:|\alpha| = n \right\}
\]
where
\[
\omega_j := w\left(\frac{\alpha_j}{n}\right)+\frac{1-\sum_{i = 1}^{d + 1} w\left(\frac{\alpha_i}{n}\right)}{d + 1}, \quad 1\leq j\leq d + 1
\]

- **Admissible weight function**:
The weight function \( w(x) \) must satisfy the following conditions:
- \( w(x) \) is monotonically increasing.
- \( w(0)=0 \) and \( w(1) = 1 \).
- For any non - negative \( \theta_i \) satisfying \( \sum_{i = 1}^{d + 1} \theta_i = 1 \), we have \( \sum_{i = 1}^{d + 1} w(\theta_i)\leq1 \).

- **Lebesgue constant**:
The Lebesgue constant \( \Lambda_n \) is defined as:
\[
\Lambda_n := \max_{x\in S_d}\sum_{i = 1}^N|\ell_i(x)|
\]
where \( \ell_i(x) \) is the Lagrange basis function.

### Conclusion

This paper provides a new interpolation point selection strategy by introducing the Waldron interpolation method, which is especially suitable for interpolation problems on high - dimensional simplexes. Through numerical experiments, the author shows the superiority of Waldron points in terms of the Lebesgue constant and provides a detailed theoretical analysis and geometric interpretation.