Abstract:We aim to fill a gap in the proof of an inequality relating two exponents of uniform Diophantine approximation stated in a paper by Bugeaud. We succeed to verify the inequality in several instances, in particular for small dimension. Moreover, we provide counterexamples to generalizations, which contrasts the case of ordinary approximation where such phenomena do not occur. Our results contribute to the understanding of the discrepancy between small absolute values of a polynomial at a given real number and approximation to the number by algebraic numbers of absolutely bounded degree, a fundamental issue in the famous problem of Wirsing and variants.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to fill a gap in the proof of the inequality between two uniform approximation exponents. Specifically, the author aims to verify the inequality regarding the uniform Diophantine approximation exponent proposed by Bugeaud in his paper, and has successfully verified this inequality in multiple instances, especially in the low - dimensional case.
In addition, the author also provides counter - examples for the generalized cases, which is in contrast to the ordinary approximation case where such a phenomenon does not occur. These results are helpful for understanding the problem of polynomials taking small values at a given point and the difference between approximating real numbers with algebraic numbers of bounded degrees, which is a fundamental problem in the well - known Wirsing problem and its variants.
### Main contributions
1. **Filling the proof gap**: The author corrects and verifies the proof of the inequality of the uniform Diophantine approximation exponent in previous literature.
2. **Providing counter - examples**: For more general cases, the author provides counter - examples, showing the complexity brought by non - separated polynomials.
3. **In - depth study**: By studying the parametric function \(\kappa(n, \xi, X)\) and its extreme values \(\kappa_n(\xi)\), the author reveals the core information of these problems.
### Specific problems and formulas
#### 1. Definitions and notations
- \( w(n, \xi, X)=-\frac{\log |P_X(\xi)|}{\log X} \)
- \( w^*(n, \xi, X)=-\frac{\log (H(\alpha_X)\cdot|\alpha_X - \xi|)}{\log X} \)
- Classical approximation exponents:
\[
w_n(\xi)=\limsup_{X\rightarrow\infty}w(n, \xi, X),\quad w^*_n(\xi)=\limsup_{X\rightarrow\infty}w^*(n, \xi, X)
\]
\[
\hat{w}_n(\xi)=\liminf_{X\rightarrow\infty}w(n, \xi, X),\quad \hat{w}^*_n(\xi)=\liminf_{X\rightarrow\infty}w^*(n, \xi, X)
\]
#### 2. Key inequalities
- Dirichlet's theorem shows that:
\[
w_n(\xi)\geq\hat{w}_n(\xi)\geq n,\quad n\geq1
\]
- Wirsing's conjecture:
\[
w^*_n(\xi)\geq n
\]
#### 3. Parametric functions
- Definition:
\[
\kappa(n, \xi, X):=w(n, \xi, X)-w^*(n, \xi, X)
\]
\[
\kappa_n(\xi):=\liminf_{X\rightarrow\infty}\kappa(n, \xi, X),\quad \kappa_n(\xi):=\limsup_{X\rightarrow\infty}\kappa(n, \xi, X)
\]
#### 4. Key conclusions
- For any algebraic number \(\xi\):
\[
\kappa_n(\xi)=\kappa_n(\xi)=0
\]
- Important inequality:
\[
\kappa_n(\xi)\geq0
\]
### Summary
Through strict mathematical derivations and the construction of counter - examples, the paper fills the proof gaps in the existing literature and provides new perspectives and tools for understanding and solving the Wirsing problem and its variants.