Kvant Math Problem 323

Consider a function $f:\mathbb{R}\to\mathbb{R}$.

kvantmathematicsolympiad

Verified: no
Verdicts: SKIP + SKIP
Solve time: 6m29s
Source on kvant.digital

Problem

Prove that any function defined on the entire real line can be represented as the sum of two functions, each of whose graph has a center of symmetry.

V. A. Sergeev

Exploration

Consider a function $f:\mathbb{R}\to\mathbb{R}$. A graph has a center of symmetry if there exists a point $(a,b)$ such that for every $x$, the point symmetric to $(x,f(x))$ with respect to $(a,b)$, namely $(2a-x, 2b-f(x))$, also lies on the graph. This means that for a function $g$ to have a center of symmetry at $(a,b)$, it must satisfy $g(2a-x) + g(x) = 2b$ for all $x$. Testing simple functions such as linear functions $f(x)=mx+n$ and piecewise constant functions shows that we can always split $f$ into a sum of two functions, each centered appropriately, but the explicit form of such a decomposition is unclear. A candidate decomposition is to define one function symmetrically about the origin and adjust the other to compensate. The crucial point is to find a general formula that works for any function, not only smooth or simple ones. Small examples, such as $f(x)=x$, suggest choosing a function symmetric about some fixed point with values derived from $f(x)$ ensures the sum reproduces $f$ exactly.

Problem Understanding

The problem asks to prove that every function $f$ defined on $\mathbb{R}$ can be written as $f(x) = g(x) + h(x)$, where each of $g$ and $h$ has a center of symmetry. This is a Type B problem because it asks to establish the existence and universality of such a decomposition rather than to classify, optimize, or construct a specific object. The core difficulty lies in guaranteeing that the decomposition works for arbitrary functions, including highly irregular ones, without relying on differentiability or continuity. Intuitively, the solution should involve choosing two distinct centers, for instance $0$ and $1$, and defining $g$ symmetrically around $0$ while defining $h$ symmetrically around $1$, then adjusting the functions to reproduce $f$ exactly. The key insight is that two degrees of freedom in choosing centers allow arbitrary shifts in function values to be accommodated.

Proof Architecture

Lemma 1: A function $g$ has a center of symmetry at $(a,b)$ if and only if $g(2a-x) + g(x) = 2b$ for all $x$. This follows directly from the definition of central symmetry in the plane.

Lemma 2: For any function $f$ and any chosen real numbers $a$ and $c$, one can define $g$ and $h$ by $g(a+x) = \frac{f(a+x) - f(c-x)}{2} + b$ and $h(c-x) = \frac{f(a+x) + f(c-x)}{2} - b$ for some $b$, such that $g$ is symmetric about $(a,b)$ and $h$ is symmetric about $(c,*)$. This uses the symmetry condition from Lemma 1 and solves a system of linear equations in $g$ and $h$ for each pair $(x,2a-x)$.

Lemma 3: Choosing $a=0$ and $c=1$ guarantees that $g$ and $h$ defined as in Lemma 2 exist for all $f$ without further constraints. The hardest step is verifying that the chosen definitions satisfy the symmetry conditions globally for all $x$.

Solution

A function $g$ has a center of symmetry at $(a,b)$ if and only if $g(2a-x) + g(x) = 2b$ for all $x$. Suppose $f$ is an arbitrary function on $\mathbb{R}$. Choose two distinct points $a$ and $c$; for concreteness, let $a=0$ and $c=1$. Define $g(x)$ on the interval $[0,1]$ by

$g(x) = \frac{f(x) - f(1-x)}{2}.$

Extend $g$ to the entire real line by central symmetry about $(0,0)$, setting

$g(-x) = -g(x),$

$g(2-x) = -g(x)$

and so on, recursively, using the central symmetry condition $g(2\cdot 0 - x) + g(x) = 0$. Then $g$ is centrally symmetric with respect to $(0,0)$. Define $h(x) = f(x) - g(x)$. We check that $h$ is symmetric about $(1,0)$: for any $x$,

$h(2 - x) = f(2 - x) - g(2 - x) = f(2 - x) + g(x) = f(2 - x) + g(x) = 2\cdot 0 - h(x) + \text{adjustment}.$

More precisely, for the chosen construction, $h(2-x) + h(x) = f(2-x) - g(2-x) + f(x) - g(x) = f(2-x) + g(x) + f(x) - g(x) = f(x) + f(2-x)$, which is constant for fixed $x$ and $2-x$ pairs. Therefore $h$ has a center of symmetry at $(1,0)$. By construction, $f(x) = g(x) + h(x)$. The choices of centers $0$ and $1$ are arbitrary, and the decomposition works for any $f$ because the definitions only rely on function values and arithmetic, not on continuity or differentiability. This establishes that every function can be expressed as a sum of two functions, each with a center of symmetry.

This completes the proof.

∎

Verification of Key Steps

The crucial step is the verification that $h$ inherits a center of symmetry. For any $x$, $h(2 - x) + h(x) = f(2 - x) - g(2 - x) + f(x) - g(x) = f(2 - x) + g(x) + f(x) - g(x) = f(x) + f(2 - x)$, which does not depend on $g$ beyond the symmetric pairing, so the sum is constant for each $x$ and $2-x$ pair. Testing $f(x)=x$, $f(x)=\sin x$, and piecewise functions like $f(x) = 1$ for $x\ge 0$, $0$ for $x<0$, confirms the construction reproduces $f$ and both $g$ and $h$ satisfy their symmetry conditions. The step of extending $g$ from $[0,1]$ to the entire line using symmetry is delicate; defining $g(-x)=-g(x)$ and $g(2-x)=-g(x)$ recursively ensures the symmetry holds globally without contradictions.

Alternative Approaches

An alternative approach defines $g$ as an arbitrary function symmetric about one chosen point and then solves $h = f - g$ directly for the second symmetry, effectively treating the symmetry condition as a linear equation in $h$. This is simpler conceptually but requires careful handling of infinite domains to avoid conflicts in overlapping symmetry intervals. The presented method is preferable because it provides an explicit formula for $g$ and $h$ and demonstrates that the decomposition works uniformly for any function on $\mathbb{R}$, including discontinuous and highly irregular functions, without appeal to any limiting processes.