J.S.C. Prentice
Abstract:We determine the Rolle function in Lagrange polynomial approximation using a suitable differential equation. We then propose a device for improving the Lagrange approximation by exploiting our knowledge of the Rolle function.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to accurately determine the Rolle function in Lagrange polynomial interpolation by solving an appropriate initial - value problem, and use this knowledge to improve the approximation quality of Lagrange interpolation.
Specifically, the paper focuses on the following two aspects:
1. **Determining the Rolle function**:
- The Lagrange interpolation polynomial \( P_n(x) \) interpolates the given data points \( \{f(x_0), f(x_1), \ldots, f(x_n)\} \) at the nodes \( \{x_0, x_1, \ldots, x_n\} \), and its point error is:
\[
\Delta(x; P_n) = f(x) - P_n(x) = \frac{f^{(n + 1)}(\xi(x))}{(n + 1)!}\prod_{k = 0}^n(x - x_k)
\]
where \( \xi(x) \) is the Rolle function. The paper proposes to determine \( \xi(x) \) by solving a differential equation, so as to calculate the interpolation error more accurately.
2. **Improving Lagrange approximation**:
- Once the Rolle function \( \xi(x) \) is determined, it can be used to approximate \( f^{(n + 1)}(\xi(x)) \), and then improve the approximation quality of the Lagrange interpolation polynomial \( P_n(x) \). The specific method is to construct a new polynomial \( P_\xi(x) \) to approximate \( f^{(n + 1)}(\xi(x)) \), and then combine \( P_n(x) \) and \( P_\xi(x) \) to form a higher - order polynomial approximation.
### Main contributions of the paper
- A method for determining the Rolle function by solving a differential equation is proposed.
- Using the knowledge of the Rolle function, a new method for improving the approximation of Lagrange interpolation is proposed, which significantly improves the approximation accuracy.
### Example application
The paper demonstrates the effectiveness of this method through specific examples. For example, for the target function \( f(x)=e^x\sin x \), experiments are carried out using two and three interpolation nodes respectively, verifying that determining the Rolle function can significantly reduce the interpolation error.
### Summary
The core problem of this paper is to improve the approximation quality of Lagrange interpolation by accurately determining the Rolle function, and shows the effectiveness and potential of this method in practical applications.