From Trial Division to Smooth Prime Indicators: A Framework Based on Fourier Series

Sebastian Fuchs · 2025

Read From Trial Division to Smooth Prime Indicators: A Framework Based on Fourier Series by Sebastian Fuchs (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-18933.txt.

Introduction

Sebastian Fuchs builds a smooth function that flags prime numbers, presented under the head title Fejer-Kernel Prime Indicators. The function comes from a classic kernel identity applied to a sine based code for trial division. For whole numbers from two upward, the value is zero exactly for odd primes.

For non whole inputs above one the value stays positive. The function is smooth in pieces, and its second derivative jumps only at square numbers, with sizes given in exact form. Smoothed versions link in the limit to divisor counts and divisor sums. Tests suggest extra zeros near odd primes for large settings. The author makes no claims about the prime number theorem or related deep results.

Catalog record

CREATORSebastian Fuchs
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-18933

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

From Trial Division to Smooth Prime Indicators: A Framework Based on Fourier Series Sebastian Fuchs Abstract . This work introduces a flexible framework for constructing smooth ( C ∞ C^{\infty} ) analogues of classical arithmetic functions, based on the smoothing of a trigonometric representation of trial division. A foundational function 𝒫 : ℝ → ℝ \mathcal{P}\colon\mathbb{R}\to\mathbb{R} of class C 1 C^{1} is first presented, whose zeros for x > 2 x>2 correspond precisely to the odd primes. The fact that this function is of class C 1 C^{1} but not smoother motivates its generalization into a versatile framework for constructing fully smooth ( C ∞ C^{\infty} ) analogues. This framework is then applied to construct two novel functions: a smooth analogue of the divisor-counting function, 𝒫 τ \mathcal{P}_{\tau} , and a smooth analogue of the sum-of-divisors function, 𝒫 σ \mathcal{P}_{\sigma} . Both functions are proven to be of class C ∞ C^{\infty} . It is shown that these new constructions possess a complete prime-zero property for all integers x ≥ 2 x\geq 2 . The robustness of this property for real numbers is analyzed for each function. It is demonstrated that 𝒫 τ \mathcal{P}_{\tau} provides a robust prime indicator for all real numbers, while the properties of 𝒫 σ \mathcal{P}_{\sigma} in this regard lead to an open question concerning the existence of non-integer zeros. The construction methodology, properties, and potential of these functions are discussed in detail. 2020 Mathematics Subject Classification Primary 11A41; Secondary 11L03, 42A10 1. Introduction Prime numbers occupy a central position in number theory, and explicit functions that identify primes have long been a subject of mathematical inquiry [ 2 ] . Seminal early results include the non-constructive formula of Mills, which relies on the existence of a special constant [ 7 ] , and explicit but computationally infeasible formulae, such as the well-known construction by Willans based on Wilson’s theorem [ 12 ] , which is representative of a class of purely arithmetic expressions for primes. The construction of such functions continues to be an active area of research. Modern approaches are diverse, ranging from purely arithmetic constructions [ 6 ] to various forms of analytic indicators [ 3 , 4 , 10 ] . A recent focus has been the development of smooth, differentiable functions that approximate the characteristic function of the primes, for instance through integral kernel methods [ 9 ] . This note contributes to this line of inquiry by introducing a new type of analytic prime indicator derived from first principles of Fourier analysis. The approach presented here is distinct from several other modern methods. Unlike indicators based on integral kernel methods [ 9 ] or purely arithmetic constructions [ 6 ] , the present work uses the Fejér kernel—a classical tool for ensuring the convergence of trigonometric series [ 13 ] —to regularise a trigonometric representation of trial division. This approach offers a direct and constructive path to an initial C^{1} function whose analytical properties can be precisely controlled. The main contribution of this paper is to show that this initial idea can be extended into a general framework capable of generating smooth ( C^{\infty} ) analogues of various classical arithmetic functions that possess a complete prime-indicating property for all integers x\geq 2 . This paper is organized as follows. Section 2 introduces the foundational function \mathcal{P}(x) and its construction. Sections 3, 4, and 5 detail its analytic properties, its prime-zero property for odd primes, and its quantitative behavior, respectively. These sections establish the groundwork and motivation for generalization. Section 6 introduces the general framework, including a highly refined smooth cutoff function. Sections 7 and 8 apply this framework to construct a smooth divisor-counting function ( \mathcal{P}_{\tau} ) and a smooth sum-of-divisors function ( \mathcal{P}_{\sigma} ), analyzing their properties and their complete prime-zero property in detail. Section 9 discusses an application of the foundational function to the prime-counting function \pi(x) , and Section 10 provides concluding remarks and discusses future research directions. Throughout, \lceil t\rceil denotes the ceiling of a real number t . Numerical verifications of the key properties and the plots presented in this paper is maintained in a public repository. 1 1 1 The repository is available at: https://github.com/SebastianFoxxx/analytic-prime-indicator 2. Analytic construction of \mathcal{P} 2.1. Motivation Trial division classifies an integer x as composite once a divisor i\leq\sqrt{x} is found. This principle can be translated into an analytical form by constructing a quotient Q(x,i) that probes the divisibility of x by i . The numerator, \sin^{2}(\pi x) , serves as a detector for integers, as it vanishes precisely for x\in\mathbb{Z} . The denominator, \sin^{2}(\pi x/i) , can be interpreted as a measure of a "relative divisibility remainder": it approaches zero when x/i is near an integer (i.e., when i is "almost" a divisor of x ) and is large otherwise. The resulting quotient, Q(x,i)=\frac{\sin^{2}(\pi x)}{\sin^{2}(\pi x/i)}, thus relates the integer nature of x to its divisibility by i . For an integer x , this expression conveniently vanishes if i is not a divisor. However, this form becomes indeterminate ( 0/0 ) at the most crucial points: where i is a true divisor of x . The goal is therefore to regularise this expression to resolve the indeterminacy while preserving its underlying prime-indicating property. The Fejér identity, a classical tool for ensuring the convergence of Fourier series, provides a direct way to resolve such trigonometric indeterminacies. 2.2. Regularisation Invoking the Fejér identity \frac{\sin^{2}(ny)}{\sin^{2}y}=n+2\sum_{k=1}^{n-1}(n-k)\cos(2ky) with n=i and y=\pi x/i removes the indeterminacy and yields a finite cosine sum that extends Q to real x . 2.3. Definition The regularised function is constructed by summing the Fejér-kernel-based terms over all potential divisors up to \sqrt{x} . For clarity, let F(x,i) denote the Fejér kernel term: (1) F(x,i)=i+2\sum_{k=1}^{i-1}(i-k)\cos\left(\frac{2\pi kx}{i}\right). Definition 2.1 . For x>1 the function \mathcal{P}(x) is defined as (2) \mathcal{P}(x)=\frac{1}{x}\sum_{i=2}^{\lceil\sqrt{x}\rceil}F(x,i), and \mathcal{P}(x)=0 is set for x\leq 1 . For brevity, let N(x)=\lceil\sqrt{x}\rceil denote the upper summation limit. 3. Smoothness properties of \mathcal{P} 3.1. Global C^{1} -smoothness Property 3.1 . The function \mathcal{P} is continuous on \mathbb{R} and of class C^{1} for x>0 . Proof. On any open interval ((m-1)^{2},m^{2}) for m\geq 2 , the summation limit \lceil\sqrt{x}\rceil is fixed at m , making \mathcal{P} a finite sum of C^{\infty} functions and thus C^{\infty} on these intervals. The critical points are the integer squares x=m^{2} , where the limit jumps from m to m+1 . First, continuity at x=m^{2} is established. As x\to(m^{2})^{-} , the function is \mathcal{P}(x)=\frac{1}{x}\sum_{i=2}^{m}F(x,i) . As x\to(m^{2})^{+} , an additional term \frac{1}{x}F(x,m+1) appears. Continuity requires this new term to be zero at the transition point. At x=m^{2} , the argument of the cosine in F(m^{2},m+1) becomes \frac{2\pi km^{2}}{m+1}=2\pi k(m-1)+\frac{2\pi k}{m+1} . Thus, F(m^{2},m+1)=(m+1)+2\sum_{k=1}^{m}(m+1-k)\cos\left(\frac{2\pi k}{m+1}\right). This is the Fejér kernel K_{m}(t) evaluated at t=2\pi/(m+1) , which equals \frac{\sin^{2}((m+1)t/2)}{\sin^{2}(t/2)}=\frac{\sin^{2}(\pi)}{\sin^{2}(\pi/(m+1))}=0 for m\geq 1 . The added term \frac{1}{m^{2}}F(m^{2},m+1) is zero, proving continuity. Next, C^{1} -smoothness is established. The derivative \mathcal{P}^{\prime}(x) exists on the open intervals. For \mathcal{P} to be of class C^{1} at x=m^{2} , equality of the left-sided and right-sided limits of \mathcal{P}^{\prime}(x) is required. This is equivalent to showing that the derivative of the newly added term, \frac{d}{dx}\left[\frac{1}{x}F(x,m+1)\right] , vanishes at x=m^{2} . By the quotient rule,

Read the complete text