Newton series representation of completely monotone functions

Basic Information

• Paper ID: 2510.11577
• Title: Newton series representation of completely monotone functions
• Authors: Thomas Lamby, Jean-Luc Marichal, Naïm Zenaïdi
• Classification: math.CA (Classical Analysis and ODEs), cs.DM (Discrete Mathematics), math.CO (Combinatorics)
• Publication Date: October 13, 2025
• Paper Link: https://arxiv.org/abs/2510.11577

Abstract

This paper proves that every completely monotone function defined on a right-unbounded open interval admits a Newton series expansion at each point of the interval. This result can be viewed as an analogue of Bernstein's Little Theorem for absolutely monotone functions. As an application, we employ it to study principal indefinite sums, which are constructed through a broad generalization of the Bohr-Mollerup theorem.

Research Background and Motivation

Problem Background

1. Importance of Completely Monotone Functions: Completely monotone functions are an important concept in real analysis, introduced by Bernstein in 1914. A function f is completely monotone on an open interval I if it is infinitely differentiable and satisfies $(-1)^n f^{(n)}(x) \geq 0$ for all $x \in I$ and $n \in \mathbb{N}$.
2. Limitations of Classical Results:
   • Bernstein's Little Theorem establishes that completely monotone functions are real analytic and admit Taylor series expansions
   • Bernstein's Monotone Function Theorem provides Laplace-type integral representations
   • However, a systematic theory of Newton series representations is lacking
3. Special Nature of Newton Series: Newton series expansion is a rather restrictive condition; many analytic functions, such as the classical exponential function $f(x) = e^x$, do not admit Newton series expansions.

Research Motivation

This paper aims to establish a theory of Newton series representations for completely monotone functions, which can be viewed as a discrete analogue of Taylor series representations, providing a new perspective on the theory of function series representations.

Core Contributions

1. Main Theoretical Result: Proves that completely monotone functions defined on right-unbounded open intervals admit Newton series expansions at each point of the interval
2. Generalization Result: Remains valid under weaker assumptions (requiring only that some higher-order derivative is completely monotone)
3. Application Extension: Applies the results to the study of principal indefinite sums, which are constructed through generalizations of the Bohr-Mollerup theorem
4. Theoretical Framework: Establishes deep connections between complete monotonicity and Newton series expansions

Methodology Details

Core Definitions

Completely Monotone Function: A function $f: I \to \mathbb{R}$ is completely monotone on an open interval I if it is infinitely differentiable and satisfies: $(-1)^n f^{(n)}(x) \geq 0 \quad \forall x \in I, n \in \mathbb{N}$

Newton Series Expansion: A function $f: I \to \mathbb{R}$ admits a Newton series expansion at a point $a \in I$ if: $f(x) = \sum_{k=0}^{\infty} \binom{x-a}{k} \Delta^k f(a) \quad (x \in I)$

where $\Delta$ is the forward difference operator and $\binom{x-a}{k}$ denotes the falling factorial.

Main Theorem

Theorem 3.1 (Main Result): Let I be a right-unbounded open interval and $f: I \to \mathbb{R}$ be infinitely differentiable such that $f^{(q)}$ is completely monotone for some $q \in \mathbb{N}$. Then for any $a \in I$, the function f admits a Newton series expansion: $f(x) = \sum_{k=0}^{\infty} \binom{x-a}{k} \Delta^k f(a) \quad (x \in I)$ and the series converges uniformly on compact subsets of I.

Proof Strategy

1. Newton-type Taylor Theorem: Establishes Newton interpolation formulas containing divided difference remainders
2. Higher-order Convexity Theory: Analyzes properties of completely monotone functions using p-convexity theory
3. Remainder Estimation: Controls the asymptotic behavior of remainders through monotonicity properties
4. Uniform Convergence: Establishes uniform convergence on compact sets using Lemma 3.5

Technical Innovations

1. Divided Difference Representation: Expresses Newton series remainders using divided differences $f[x_0, x_1, \ldots, x_n]$
2. Higher-order Convexity Characterization: Reformulates complete monotonicity using higher-order convexity $K^{n-1}_{(-1)^{n-q}}(I)$
3. Asymptotic Analysis: Establishes key asymptotic estimates to control series convergence

Experimental Setup

Theoretical Verification

This paper is primarily theoretical work, with results verified through mathematical proofs. The main verification methods include:

1. Classical Examples Verification:
   • Newton expansion of the reciprocal function $f(x) = 1/x$ on $\mathbb{R}_+$
   • Newton series representation of the logarithmic Gamma function
2. Counterexample Construction: Demonstrates that certain functions do not admit Newton series expansions, such as $f(x) = e^x$

Application Scenarios

Applies the theory to the study of principal indefinite sums, including:

• Generalizations of the Bohr-Mollerup theorem
• Newton representations of harmonic number functions and the Digamma function

Experimental Results

Main Results

Example 3.7 (Reciprocal Function): For $f(x) = 1/x$ on $\mathbb{R}_+$, the Newton series expansion is: $\frac{1}{x} = \frac{1}{a} \sum_{k=0}^{\infty} \frac{(-1)^k \binom{x-a}{k}}{\binom{a+k}{k}} \quad (x > 0)$

Example 4.8 (Logarithmic Gamma Function): $\ln\Gamma(x)$ admits a Newton series expansion: $\ln\Gamma(x) = \sum_{k=1}^{\infty} \binom{x-1}{k} (\Delta^{k-1}_t \ln t)|_{t=1} \quad (x > 0)$

Example 4.9 (Stern Series): Newton representation of the harmonic number function: $H_{x-1} = \sum_{k=1}^{\infty} \binom{x-1}{k} \frac{(-1)^{k-1}}{k} \quad (x > 0)$

Theoretical Findings

1. Domain of Convergence: Newton series converge on the entire right-unbounded interval, not merely in a neighborhood
2. Necessary Conditions: The interval must be right-unbounded, as Newton series require evaluation at points $a+k$
3. Uniform Convergence: Uniform convergence on compact subsets ensures good analytical properties

Related Work

Classical Theory

1. Bernstein Theory:
   • Bernstein's Little Theorem: Real analyticity of completely monotone functions
   • Bernstein's Monotone Function Theorem: Laplace integral representations
2. Taylor Series Theory: Taylor expansions of absolutely monotone functions
3. Newton Interpolation Theory: Classical Newton interpolation formulas and divided difference theory

Modern Developments

1. Higher-order Convexity Theory: Systematic study of p-convex functions
2. Bohr-Mollerup Generalizations: Construction theory of principal indefinite sums
3. Applications of Completely Monotone Functions: Applications in probability theory and number theory

Conclusions and Discussion

Main Conclusions

1. Establishes a complete theory of Newton series representations for completely monotone functions
2. Proves that this representation holds on the entire right-unbounded interval, which is stronger than Taylor series
3. Successfully applies the results to the study of principal indefinite sums

Limitations

1. Interval Restrictions: Requires right-unbounded intervals, limiting the scope of applications
2. Monotonicity Requirements: Requires the strong condition of complete monotonicity
3. Computational Complexity: Computing Newton series may be more complex than Taylor series

Future Directions

The paper proposes important open problems in its conclusion:

1. Characterize the function class $\mathcal{N}(I) = \bigcap_{a \in I} \mathcal{N}_a(I)$ (functions admitting Newton expansions at each point)
2. Develop generalizations of Newton series theory in complex analysis
3. Study relationships between Newton series and other function representations

In-Depth Evaluation

Strengths

1. Theoretical Depth: Establishes profound connections between complete monotonicity and Newton series
2. Methodological Innovation: Cleverly combines higher-order convexity theory and divided difference techniques
3. Result Strength: Newton expansions hold on the entire interval, which is stronger than classical results
4. Application Value: Demonstrates practical utility in the study of principal indefinite sums
5. Clear Exposition: Well-structured paper with detailed proofs

Weaknesses

1. Application Scope: Restrictions to right-unbounded intervals reduce practical application scenarios
2. Computational Practicality: Numerical computation of Newton series may be less efficient than Taylor series
3. Generalization: Further research is needed for generalizations to broader function classes

Impact

1. Theoretical Contribution: Provides a new perspective on function series representation theory
2. Methodological Value: Higher-order convexity methods may be useful in other analytical problems
3. Application Potential: Has potential applications in special function theory and numerical analysis

Applicable Scenarios

1. Special Function Research: Analysis of special functions with monotonicity properties
2. Numerical Computation: May provide better numerical stability in certain cases
3. Theoretical Analysis: Serves as a new tool for studying function properties

References

The paper cites 19 important references, covering:

• Classical works by Bernstein 1,2,3
• Monotone function theory 4,17
• Newton series theory 6,7
• Bohr-Mollerup theorem 5,15,16
• Related work on higher-order convexity theory

Overall Assessment: This is a high-quality theoretical mathematics paper that makes important advances in Newton series representations of completely monotone functions. Although the results are relatively specialized, the paper makes valuable contributions to function analysis theory and demonstrates practical application value in the study of principal indefinite sums. The proof techniques and theoretical framework presented are instructive for related research areas.