C. Sinan Güntürk,Weilin Li
Abstract:Let $\mathscr{C}_\mathbb{Z}([0,1])$ be the metric space of real-valued continuous functions on $[0,1]$ with integer values at $0$ and $1$, equipped with the uniform (supremum) metric $d_\infty$. It is a classical theorem in approximation theory that the ring $\mathbb{Z}[X]$ of polynomials with integer coefficients, when considered as a set of functions on $[0,1]$, is dense in $\mathscr{C}_\mathbb{Z}([0,1])$. In this paper, we offer a strengthening of this result by identifying a substantially small subset $\bigcup_n \mathscr{B}_n$ of $\mathbb{Z}[X]$ which is still dense in $\mathscr{C}_\mathbb{Z}([0,1])$. Here $\mathscr{B}_n$, which we call the ``Bernstein lattice,'' is the lattice generated by the polynomials $$p_{n,k}(x) := \binom{n}{k} x^k(1-x)^{n-k}, ~~k=0,\dots,n.$$ Quantitatively, we show that for any $f \in \mathscr{C}_\mathbb{Z}([0,1])$, $$d_\infty(f, \mathscr{B}_n) \leq \frac{9}{4} \omega_f(n^{-1/3}) + 2 n^{-1/3}, ~~n \geq 1,$$ where $\omega_f$ stands for the modulus of continuity of $f$. We also offer a more general bound which can be optimized to yield better decay of approximation error for specific classes of continuous functions.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to use polynomials with integer coefficients to approximate continuous functions under specific conditions. Specifically, the paper focuses on how to achieve uniform approximation for continuous functions \(f\in C_Z([0,1])\) that take integer values at the endpoints \(0\) and \(1\) on the interval \([0,1]\) through a smaller set of polynomials - the so - called "Bernstein lattice".
### Main contributions of the paper
1. **Strengthening existing results**:
- The authors provide a stronger conclusion than the existing results, that is, to approximate functions in \(C_Z([0,1])\) through a significantly smaller subset \(\bigcup_n B_n\), where \(B_n\) is the lattice generated by Bernstein polynomials.
- Specifically, the Bernstein polynomial \(p_{n,k}(x)=\binom{n}{k}x^k(1 - x)^{n - k}\) is used to construct this lattice.
2. **Quantitative analysis**:
- The authors give a specific error bound, showing that for any \(f\in C_Z([0,1])\), when \(n\geq1\), we have:
\[
d_\infty(f,B_n)\leq\frac{9}{4}\omega_f(n^{-1/3})+2n^{-1/3}
\]
where \(\omega_f\) represents the modulus of continuity of \(f\), defined as:
\[
\omega_f(\delta):=\sup\{|f(x)-f(y)|:x,y\in[0,1],|x - y|\leq\delta\}
\]
3. **More general error bounds**:
- In addition to the above - mentioned specific error bound, the author also provides a more general form, which can be optimized for specific classes of continuous functions, thereby obtaining a better approximation error decay rate.
### Methods and innovation points
- **Bernstein lattice**: By introducing the Bernstein lattice \(B_n\), the authors show how to significantly reduce the number of required polynomials while maintaining the approximation effect.
- **Constructive method**: The paper not only proves the theoretical density result but also provides a specific construction method to achieve approximation by choosing appropriate parameters \(t_n\).
- **Error analysis**: Through detailed error analysis, the authors prove the effectiveness of the proposed approximation method and give specific error bounds.
### Conclusion
This paper presents a new perspective in the theory of polynomial approximation. By introducing the Bernstein lattice, it not only strengthens the existing density theorems but also provides a specific construction method and error analysis. This provides new tools and methods for understanding and applying polynomial approximation with integer coefficients.