Repeated integrals of increasing functions

Maxim R. Burke, Maleeha Haris, Madhavendra · 2025

Read Repeated integrals of increasing functions by Maxim R. Burke, Maleeha Haris, Madhavendra (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-2512-02151.txt.

Introduction

Integrating an increasing function once gives its accumulated area; integrating the result again and again builds a tower of smoother curves. Maxim Burke and co-authors characterize in 2025 exactly which triples of values those repeated integrals can take at the endpoint, for increasing functions on the unit interval.

The question grew out of a problem about approximating smooth functions by entire functions in a comonotone way, keeping the same increasing pattern. Knowing the reachable set of first, second, and third integral values draws the boundary of what such functions can do. Results of this kind are the fine print of approximation theory, where they decide which approximation schemes can exist.

Catalog record

CREATORMaxim R. Burke, Maleeha Haris, Madhavendra
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE131 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-02151

Browse the full library · About the Tower of Babel

Full text

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

Repeated integrals of increasing functions Thanks: The first author’s research supported by NSERC. The author thanks Yinhe Peng and the Academy of Mathematics and Systems Science of the Chinese Academy of Sciences for their hospitality in the fall of 2024 when some of this research was carried out. The author also thanks Bill Weiss for a helpful discussion on early results related to this paper. Maxim R. Burke, Maleeha Haris, Madhavendra Abstract . Motivated by a problem on comonotone approximation of C n C^{n} functions by entire functions, for increasing functions f : [ 0 , 1 ] → [ 0 , 1 ] f\colon[0,1]\to[0,1] , we characterize the possible values of ( a , b , c ) (a,b,c) , where a = I ​ ( f ) ​ ( 1 ) a=I(f)(1) , b = I 2 ​ ( f ) ​ ( 1 ) b=I^{2}(f)(1) , c = I 3 ​ ( f ) ​ ( 1 ) c=I^{3}(f)(1) ( I I is the integral operator I ⁡ ( f ) ​ ( x ) = ∫ 0 x f ⁡ ( t ) ​ 𝑑 t I(f)(x)=\int_{0}^{x}f(t)\,dt ), as those which satisfy the conditions 0 ≤ a ≤ 1 0\leq a\leq 1 , a 2 / 2 ≤ b ≤ a / 2 a^{2}/2\leq b\leq a/2 , 2 ​ b 2 ≤ 3 ​ a ​ c 2b^{2}\leq 3ac , a 2 + 4 ​ b 2 + 6 ​ c ≤ 6 ​ a ​ c + 2 ​ a ​ b + 2 ​ b a^{2}+4b^{2}+6c\leq 6ac+2ab+2b , and 0 ≤ c ≤ a / 6 0\leq c\leq a/6 . Our main theorem states that if a , b , c a,b,c are real numbers for which the inequalities are strict, then there is a function f f satisfying a = I ​ ( f ) ​ ( 1 ) a=I(f)(1) , b = I 2 ​ ( f ) ​ ( 1 ) b=I^{2}(f)(1) , c = I 3 ​ ( f ) ​ ( 1 ) c=I^{3}(f)(1) which is C ∞ C^{\infty} with f ⁡ ( 0 ) = 0 f(0)=0 , f ⁡ ( 1 ) = 1 f(1)=1 , D ​ f ​ ( x ) > 0 Df(x)>0 for 0 < x < 1 0<x<1 , and whose derivatives D j ​ f ​ ( 0 ) D^{j}f(0) and D j ​ f ​ ( 1 ) D^{j}f(1) , j ≥ 1 j\geq 1 , are arbitrary as long as they are consistent with the increasing nature of f f . The construction of f f proceeds by starting with a continuous parametrization s ↦ ρ s ∈ C ∞ ​ ( [ 0 , 1 ] ) s\mapsto\rho_{s}\in C^{\infty}([0,1]) defined on an open subset of ℝ 4 \mathbb{R}^{4} , and composing with successive continuous transversals through the open set to fix the values of I j ​ ( ρ s ) ​ ( 1 ) I^{j}(\rho_{s})(1) for j = 0 , 1 , 2 , 3 j=0,1,2,3 . Addressing the aforementioned problem on comonotone approximation, we examine the set V n ⊆ ℝ 2 ​ ( n + 1 ) V_{n}\subseteq\mathbb{R}^{2(n+1)} of possible values D j ​ f ​ ( 0 ) D^{j}f(0) , D j ​ f ​ ( 1 ) D^{j}f(1) , j = 0 , … , n j=0,\dots,n , of the derivatives of a C n C^{n} function at the endpoints when D n ​ f D^{n}f is increasing but not constant. We make a conjecture about the nature of this set and prove our conjecture for n ≤ 3 n\leq 3 as a consequence of the theorem mentioned above. 1991 Mathematics Subject Classification Primary 26A241, 46E10, 26A48; Secondary 54C05, 26D05, 26A51, 41A30. 1. Introduction This paper has its origins in the study of comonotone approximation by entire functions of C^{n} functions having piecewise monotone derivatives. A function f\colon\mathbb{R}\to\mathbb{R} is piecewise monotone if there is a closed discrete set K\subseteq\mathbb{R} such that f is monotone on the components of the complement of K . Two functions f,g are comonotone if there is a closed discrete set K\subseteq\mathbb{R} such that f and g have the same monotonicity on the components of the complement of K . It is known (see [ 2 ] , Theorem A) that if n is a nonnegative integer and for some m>n , f\colon\mathbb{R}\to\mathbb{R} is a C^{m} function such that D^{n+1}f has no flat points (i.e., for each x\in\mathbb{R} , there is an integer i\in[n+1,m] such that D^{i}f(x)\not=0 ), then for any closed discrete set E\subseteq\mathbb{R} , and for any positive continuous function \varepsilon\colon\mathbb{R}\to\mathbb{R} , there is a function g\colon\mathbb{R}\to\mathbb{R} which is the restriction of an entire function, whose derivatives approximate those of f within {\varepsilon} with interpolation on E , in symbols, for all x\in\mathbb{R} we have • |D^{i}g(x)-D^{i}f(x)|<\varepsilon(x) , 0\leq i\leq m ; • D^{i}g(x)=D^{i}f(x) when x\in E , 0\leq i\leq m ; and each D^{k}f is comonotone with D^{k}g , k=0,\dots,n . The following theorem is a partial answer to the question of what can be said if m=n , i.e., f may not have derivatives of order larger than n . In the statement, a platform of a continuous function f\colon\mathbb{R}\to\mathbb{R} is a maximal interval on which f is constant, and a compact platform I=[a,b] is a turning platform if for some {\varepsilon}>0 , the values of f on (a-{\varepsilon},b+{\varepsilon})\setminus I are either all larger than the constant value k of f on I , or all smaller than k . If D^{n}f has a platform unbounded above, then A^{n}_{\max}=A^{n}_{\max}(f) denotes this platform, and A^{n}_{\max}=\emptyset otherwise. Similarly, A^{n}_{\min}=A^{n}_{\min}(f) denotes the platform of D^{n}f which is unbounded below, if there is one, and A^{n}_{\min}=\emptyset otherwise. The theorem makes an assumption (P_{n}) which we discuss shortly. Theorem 1.1 ( [ 3 ] ) . Assume (P_{n}) . Let f\colon\mathbb{R}\to\mathbb{R} be a C^{n} function such that D^{j}f is piecewise monotone for 0\leq j\leq n , and D^{n}f is not constant. Suppose there is a closed discrete set E\subseteq\mathbb{R} , disjoint from A^{n}_{\min}\cup A^{n}_{\max} , having no more than one point on any platform of D^{n}f , and having exactly one point on each turning platform of D^{j}f , 0\leq j\leq n . Let {\varepsilon}\colon\mathbb{R}\to\mathbb{R} be a positive continuous function. Then there is a function g\colon\mathbb{R}\to\mathbb{R} which is the restriction of an entire function and satisfies the following conditions for x\in\mathbb{R} . (1) |D^{j}g(x)-D^{j}f(x)|<{\varepsilon}(x) , 0\leq j\leq n . (2) D^{j}g(x)=D^{j}f(x) when x\in E , 0\leq j\leq n . (3) D^{j}g(x)\not=0 when x\not\in E , 0<j\leq n+1 . On a component I of the complement of E , for 0\leq j\leq n , D^{j}g is monotone by (3) , and D^{j}f is also monotone on I because E has a point on each turning platform of D^{j}f . From (2) it follows that these monotonicities are the same for D^{j}g and D^{j}f if I is bounded (so both endpoints are in E ), and when I is not bounded the same conclusion follows from (2) and (1) if we take {\varepsilon}(x) so that it vanishes at \pm\infty . To state the assumption (P_{n}) , we first define the following set

Read the complete text