Abstract:Understanding the distribution of digits in the expansions of perfect powers in different bases is difficult. Rather than consider the asymptotic digit distributions, we consider the base-10 digits of a restricted sequence of powers of two. We apply elementary methods to show that this sequence of powers of two can be constructed to preserve trailing digits while locally maximizing the number of zeroes between nonzero digits. We also provide a heuristic description of the trailing digits of these powers of two.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in the decimal representation, the distribution problem of non - zero digits of powers of 2 (i.e., \(2^n\)). Specifically, the author focuses on how to construct a sequence of powers of 2 such that there are as many zeros as possible between non - zero digits in this sequence, thereby achieving local maximization of the sparsity between non - zero digits.
### Problem Background
Understanding the digit distribution of perfect powers in different bases is a complex mathematical problem. For powers of 2, the digit distribution in decimal is especially difficult to predict. Previous research has mainly focused on asymptotic distributions or attempted to prove the lower bounds of non - zero digit densities. For example, Stewart proposed a logarithmic lower bound \(\frac{\ln(x)}{C+\ln\ln(x)}\) for the number of non - zero digits, and Radcliffe proposed a simpler logarithmic lower bound \(\log_4(x)\).
### Core Problem of the Paper
The author of this paper chose a different approach, focusing on constructing a 10 - adic sequence of powers of 2 through a greedy algorithm to achieve the maximum sparsity between non - zero digits. Specifically, the author's goal is to select an exponent \(p_n\) at each step such that in the decimal representation of \(2^{p_n}\), there are as many zeros as possible between non - zero digits.
### Mathematical Representation
A power of 2 can be uniquely represented as:
\[2^\chi = b_1\cdot10^{d_1}+b_2\cdot10^{d_2}+\cdots\]
where \(d_1 < d_2 <\cdots\) and \(1\leq b_i\leq9\). The goal is to construct an exponent sequence \(p_n\) such that \(d_{n + 1}-d_n\) is as large as possible, that is, the gap between non - zero digits is as large as possible.
### Algorithm Description
The author proposed a greedy algorithm, which is implemented through the following steps:
1. Assume that \(p_n\) has been constructed such that \(2^{p_n}\equiv b_1\cdot10^{d_1}+b_2\cdot10^{d_2}+\cdots + b_n\cdot10^{d_n}\pmod{10^{d_n+1}}\).
2. For each new \(p_{n+1}\), find the largest \(e\) such that \(2^{p_{n+1}}\equiv b_1\cdot10^{d_1}+b_2\cdot10^{d_2}+\cdots + b_n\cdot10^{d_n}\pmod{10^e}\) holds.
3. Keep those \(p_{n+1}\) for which \(e - d_n\) is the largest, and continue iterating.
### Main Conclusions
- Through this method, the author proved that there are at least two consecutive zeros between non - zero digits each time, that is, \(d_{n+1}-d_n\geq2\).
- Further, the author provided a heuristic prediction, indicating that the density of non - zero digits tends to \(\frac{13}{4}\).
### Experimental Results
By calculating a large amount of data, the author verified the effectiveness of his algorithm and showed that the number of zeros between non - zero digits is indeed as expected.
In conclusion, this paper successfully studied the distribution problem of non - zero digits of powers of 2 in decimal representation through a novel greedy algorithm, providing new insights and methods.