On Completely multiplicative $\pm1$ sequences that omit many consecutive $+1$ values

Yichen You

2024-04-07

Abstract:We investigate the construction of $\pm1$-valued completely multiplicative functions that take the value $+1$ at at most $k$ consecutive integers, which we call length-$k$ functions. We introduce a way to extend the length based on the idea of the "rotation trick" and such an extension can be quantified by the number of modified primes. Under the assumption of Elliott's conjecture, this method allows us to construct length-$k$ functions systematically for $k\geq 4$ which generalizes the work of I. Schur for $k = 2$ and R. Hudson for $k =3$.

Number Theory

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to construct and understand a special class of completely multiplicative functions. These functions take values of ±1 on natural numbers and take the value of +1 on at most consecutive \(k\) integers. Such functions are called functions of length \(k\).

Specifically, the author has studied how to construct such functions so that they take the value of +1 on at most \(k\) consecutive integers. To achieve this goal, the author introduced a method based on the "rotation trick", by modifying the function values at certain prime numbers to extend the length of the function. In addition, the author assumes that Elliott's conjecture holds, and thus systematically constructs functions of length \(k\geq4\).

The following is a detailed description of this problem:

1. **Definition and Background**:
- A completely multiplicative function \(f:\mathbb{N}\to\{+1, -1\}\) satisfies \(f(ab) = f(a)f(b)\) for all \(a, b\in\mathbb{N}\).
- A function of length \(k\) means that there exists some \(n\in\mathbb{N}\) such that \(f(n + 1)=f(n + 2)=\cdots=f(n + k)=+1\), and there is no longer consecutive sequence of +1.

2. **Existing Results**:
- For \(k = 2\), I. Schur has already proved that there are only two possible functions.
- For \(k = 3\), R. Hudson conjectured that there are only thirteen possible functions, and this conjecture was recently proved by O. Klurman et al.

3. **New Contributions**:
- The author attempts to classify functions of length \(k\geq4\) and finds that an infinite number of such functions can be constructed by modifying the function values at certain prime numbers.
- The author proposes a quantification method. By modifying the values of a finite number of prime numbers, the length of the function can be extended to at least \(k\).

4. **Main Results**:
- Theorem 1.2 shows that for any positive integer \(k\), the number of modified prime numbers \(\delta(k)\) can be expressed by the formula:
\[
\delta(k)=\frac{1}{2}k+O(\log k)
\]
- This means that by modifying the values of a finite number of prime numbers, the length of the function can be extended from less than \(k\) to at least \(k\).

5. **Applications and Significance**:
- This research helps to understand the behavior of completely multiplicative functions, especially their value - taking patterns on consecutive integers.
- By introducing the method of modifying prime numbers, the author provides a new way to systematically construct functions of a specific length.

In summary, the main purpose of this paper is to explore and construct completely multiplicative functions of length \(k\), especially in the case of \(k\geq4\), and to extend the length of the function by modifying the values of prime numbers. This not only deepens the understanding of completely multiplicative functions but also provides new tools and methods for further research in related fields.