---
title: Fibonacci Floor Function Overview
url: https://www.emergentmind.com/topics/fibonacci-floor-function
type: topic
---

# Fibonacci Floor Function Overview

The expression **Fibonacci floor function** is used in several adjacent but non-identical senses. In the most specific recent usage, it denotes a unary function on \(\mathbb Z\) that returns the greatest Fibonacci number of even index less than or equal to a given integer, introduced in the model theory of \(\langle \mathbb Z,<,+,\lfloor \varphi x\rfloor,0\rangle\) [2508.02303]. In a broader arithmetic sense, the same phrase also covers explicit floor-function formulas governed by the golden ratio and Fibonacci asymptotics, including a closed formula for the non-Fibonacci numbers [1105.1127], Beatty-type descriptions of Fibonacci-related index sets [1812.02107], generating functions with floor indices [2303.15478], and floor sums attached to Fibonacci and Lucas coefficients in linear Diophantine counting [2604.10294].

## 1. Definition, conventions, and scope

A first point of terminology is that the literature does not fix a single convention for the Fibonacci sequence. The model-theoretic paper on the named **Fibonacci floor function** uses the shifted convention
\[
F_0=1,\qquad F_1=1,\qquad F_{n+2}=F_n+F_{n+1},
\]
whereas the complementary and analytic papers use the standard convention
\[
F_0=0,\qquad F_1=1,\qquad F_{n+2}=F_{n+1}+F_n.
\]
This distinction affects index parity statements and must be tracked explicitly [2508.02303].

In the model-theoretic setting, let
\[
f(x)=\lfloor \varphi x\rfloor,\qquad [\varphi x]=\varphi x-\lfloor \varphi x\rfloor,
\]
with \(\varphi=(1+\sqrt5)/2\). The paper defines functions \(F\) and \(G\) by
\[
F(x)=y \Longleftrightarrow 0<y\le x \ \wedge\ [\varphi y]=\min\{[\varphi w]:0<w\le x\},
\]
\[
G(x)=y \Longleftrightarrow 0<y\le x \ \wedge\ [\varphi y]=\max\{[\varphi w]:0<w\le x\}.
\]
It then states that \(F\) maps each integer \(x\) to the greatest Fibonacci number of even index and less than or equal to \(x\), and remarks that a “more telling notation” for \(F\) would be the symbol \({-}\), called the **Fibonacci floor function** [2508.02303].

The surrounding literature uses the phrase more loosely for any floor-based encoding of Fibonacci structure. The main usages in the cited works are summarized below.

| Context | Object | Role |
|---|---|---|
| Model theory | \({-}(x)=F(x)\) | Largest even-index Fibonacci \(\le x\) |
| Complement formulas | Explicit floor expression | Enumerates non-Fibonacci numbers |
| Floor-indexed sequences | \(F_{\lfloor n/k\rfloor}\) | Block-sampled Fibonacci sequence |
| Beatty indexing | \(\lfloor n\varphi^2\rfloor-1\) | Positions of odd fibbinaries |
| Diophantine counting | Floor sums \(\sum \lfloor\cdot\rfloor\) | Exact formulas for special triplets |

A common misconception is that the Fibonacci floor function is simply the Beatty map \(x\mapsto\lfloor \varphi x\rfloor\). The model-theoretic paper explicitly separates these two objects: \(f(x)=\lfloor \varphi x\rfloor\) is the ambient Beatty function, whereas the Fibonacci floor function \(F(x)={-}(x)\) is a new definable function extracted from the order of the fractional parts \([\varphi n]\) [2508.02303].

## 2. Beatty-theoretic characterization and logical role

The model-theoretic construction is organized around the decimal parts \([\varphi x]\). A key characterization is that, for \(a\in\mathbb N\), \(a\) is a Fibonacci number of even index iff \([\varphi a]\) is the minimum of
\[
\{[\varphi n]:0<n\le a\},
\]
and \(a\) is a Fibonacci number of odd index iff \([\varphi a]\) is the corresponding maximum. This identifies Fibonacci numbers as extremal points of the orbit of the Beatty rotation \(n\mapsto [\varphi n]\), with parity detected by min-versus-max behavior [2508.02303].

The paper also defines a “decimal order”
\[
x<^* y \Longleftrightarrow x\neq y\ \wedge\ f(y-x)=f(y)-f(x),
\]
and notes that this coincides with the order relation \([\varphi x]<[\varphi y]\). From the basic axioms for \(f\), the authors prove that \(<^*\) is a dense linear order without endpoints on \(\mathbb Z\), and they obtain the constructive Kronecker lemma
\[
[\varphi x] < [\varphi(f(y-x)+y)] < [\varphi y]\qquad (x<y).
\]
This embeds the Fibonacci floor function into a first-order framework where extremal decimal parts on intervals are definable [2508.02303].

Within that framework, the Fibonacci floor function interacts tightly with the Beatty function \(f\). If \(\bar f(x)=x+f(x)\), then the first even-index Fibonacci strictly greater than \(x\) is
\[
\bar f(F(x))=F(x)+f(F(x)).
\]
Moreover, \(F\) is a step function in the precise sense that
\[
F(x)<y<\bar f(F(x)) \to F(y)=F(x).
\]
The paper also provides an explicit addition rule for \(F(m+n)\), expressed in terms of \(F(m)\), \(F(n)\), \(f(F(m))\), \(f(F(n))\), and the definable inverse \(f^{-1}\) on the relevant ranges [2508.02303].

These constructions are then incorporated into a recursive theory \(T\) in the language
\[
\mathcal L_{\mathrm{fib}}=\{<,+,f,0,{-}\}.
\]
Its axioms include Presburger arithmetic, the Beatty axioms for \(f\), interval-extrema axioms for decimal parts, the defining axioms for \(F\) and \(G\), and the explicit addition law for \(F\). The central logical result is that \(T\) is model-complete; the standard structure \(\mathbb Z\) is its prime model; and the theory is complete and decidable. In this sense, the Fibonacci floor function is not an isolated combinatorial gadget but a named definable function that stabilizes the model theory of a Presburger expansion by a Beatty sequence [2508.02303].

## 3. Exact floor descriptions of the complement of the Fibonacci sequence

A distinct line of work studies floor expressions that do not return Fibonacci numbers but rather their complement. Let \((u_n)\) be an increasing sequence of integers and let \(\psi:[0,\infty)\to\mathbb R\) be continuous, increasing, and unbounded, subject to
\[
u_n-n<\psi(n)\le u_n-n+1.
\]
A general complement theorem then shows that
\[
\bigl\lfloor k+\psi^{-1}(k)\bigr\rfloor\qquad (k\ge u_0+1)
\]
generates exactly the complement of \(\{u_n:n\in\mathbb N\}\) in \([u_0,\infty)\) [1105.1127].

Applying this theorem to
\[
u_n=F_{n+2}
\]
with the standard Fibonacci convention yields an explicit formula for the non-Fibonacci numbers. If \(\Phi=(1+\sqrt5)/2\), then for each integer \(n\ge2\),
\[
a_n=\left\lfloor n+\log_{\Phi}\!\left(\sqrt5\bigl(\log_{\Phi}(\sqrt5 n)+n\bigr)-5+\frac3n\right)-2\right\rfloor
\]
is exactly the \(n\)-th positive integer that is not a Fibonacci number, in increasing order [1105.1127].

The construction is based on Binet’s formula
\[
F_n=\frac{\Phi^n-(1-\Phi)^n}{\sqrt5},
\]
together with the approximation \(F_n\approx \Phi^n/\sqrt5\). The specific nested logarithmic structure arises from inverting the asymptotic relation between \(n\) and \(F_n\), while the terms \(-5\) and \(3/n\) serve as fine-tuning corrections so that the required one-unit window inequality holds for all \(n\), not only asymptotically [1105.1127].

This formula is analogous in spirit to classical complement formulas such as
\[
\lfloor n+\sqrt n\rfloor
\]
for non-squares and
\[
\left\lfloor n+\sqrt{2n}+\tfrac12\right\rfloor
\]
for non-triangular numbers. The Fibonacci case is more complicated because the underlying growth is exponential with irrational base \(\Phi\), so inverse logarithmic terms necessarily appear [1105.1127].

A second common misconception is therefore that a “Fibonacci floor function” must generate Fibonacci numbers themselves. One major explicit floor formula in the literature does the opposite: it generates the positive integers that are not Fibonacci numbers.

## 4. Floor indices, block sampling, and generating functions

A different usage of floor in Fibonacci analysis appears in the study of **dual floor sequences**
\[
\bigl(a_{\lfloor n/k\rfloor}\bigr)_{n\ge0},
\]
defined for any sequence \((a_n)_{n\ge0}\) and integer \(k\ge1\). If \(F(z)=\sum_{n\ge0} a_n z^n\) is the ordinary generating function, then
\[
\sum_{n=0}^\infty a_{\lfloor n/k\rfloor} z^n
=
\frac{1-z^k}{1-z}\,F(z^k),\qquad |z|<1.
\]
The factor \((1-z^k)/(1-z)=1+z+\cdots+z^{k-1}\) records the block structure induced by the floor: each value \(a_j\) is repeated exactly \(k\) times [2303.15478].

Specializing to the classical Fibonacci and Lucas generating functions,
\[
\sum_{n=0}^\infty F_n z^n=\frac{z}{1-z-z^2},\qquad
\sum_{n=0}^\infty L_n z^n=\frac{2-z}{1-z-z^2},
\]
gives the floor-indexed identities
\[
\sum_{n=0}^{\infty}F_{\lfloor n/k\rfloor}z^n
=
\frac{(1-z^k)z^k}{(1-z)(1-z^k-z^{2k})},
\]
\[
\sum_{n=0}^{\infty}L_{\lfloor n/k\rfloor}z^n
=
\frac{(1-z^k)(2-z^k)}{(1-z)(1-z^k-z^{2k})}.
\]
These formulas are the basic structural description of Fibonacci and Lucas sequences sampled at floor-compressed indices [2303.15478].

The same paper derives explicit closed forms for mixed series of the types
\[
\sum_{n=0}^{\infty}\frac{F_{\lfloor n/k\rfloor}F_{n+m}}{p^{n+1}},\qquad
\sum_{n=0}^{\infty}\frac{F_{\lfloor n/k\rfloor}L_{n+m}}{p^{n+1}},
\]
\[
\sum_{n=0}^{\infty}\frac{L_{\lfloor n/k\rfloor}F_{n+m}}{p^{n+1}},\qquad
\sum_{n=0}^{\infty}\frac{L_{\lfloor n/k\rfloor}L_{n+m}}{p^{n+1}},
\]
as well as alternating block-weighted variants involving \((-1)^{\lfloor n/k\rfloor}\). The method combines the floor-sequence generating-function identity with Binet’s formulas and substitutions \(z=\alpha/p\), \(z=\beta/p\), where \(\alpha\) and \(\beta\) are the roots of \(x^2=x+1\) [2303.15478].

Here the floor operator does not select extremal Fibonacci numbers, as in the model-theoretic setting, nor does it describe a complement. Instead, it creates a controlled blockwise deceleration of the index map \(n\mapsto \lfloor n/k\rfloor\), and the resulting rational factorization of generating functions makes a broad family of Fibonacci and Lucas series explicitly computable.

## 5. Beatty position formulas and odd fibbinary numbers

Another Fibonacci-adjacent floor phenomenon arises from **fibbinary numbers**, the positive integers whose binary expansion contains no consecutive \(1\)s. Using Zeckendorf representation, the paper defines \(\mathrm{fib}(n)\) as the increasing sequence of all fibbinary numbers, and \(\mathrm{odfib}(n)\) as the increasing sequence of odd fibbinary numbers. If the \(n\)-th odd fibbinary is the \(j\)-th fibbinary overall, then
\[
Z(n)=j
\]
is its position in the full fibbinary ordering [1812.02107].

The main theorem gives a Beatty-type closed form:
\[
Z(n)=\left\lfloor n\varphi^2\right\rfloor-1
=
\lfloor n\varphi\rfloor+n-1,
\]
with \(\varphi=(1+\sqrt5)/2\) and \(\varphi^2=\varphi+1\). Thus the positions of odd fibbinary numbers form a shifted Beatty sequence associated with the golden ratio [1812.02107].

The proof is recursive. The paper establishes
\[
\mathrm{odfib}(F_n+k)=2^n+\mathrm{odfib}(k)
\]
and
\[
Z(F_n+k)=F_{n+2}+Z(k)\qquad (1\le k\le F_{n-1}),
\]
then shows that the floor expression \(\lfloor n\varphi^2\rfloor-1\) satisfies the same block recurrences and initial conditions. Fractional-part control is handled via \(\tau=\varphi-1\) and the identity
\[
F_n\tau^n=-(-\tau)^{n-1},
\]
which makes it possible to compare the jumps of \(Z(n)\) with the jumps of the Beatty sequence [1812.02107].

This result does not define a unary Fibonacci floor function on integers, but it is a canonical example of a floor formula whose arithmetic content is governed by Fibonacci growth and Zeckendorf structure. It shows that Beatty sequences associated with \(\varphi\) and \(\varphi^2\) can encode the exact locations of a combinatorially defined Fibonacci-related subsequence.

## 6. Floor sums in Diophantine counting with Fibonacci and Lucas triplets

Floor functions also enter through counting formulas for non-negative solutions of
\[
ax+by+cz=n.
\]
Binner’s general formula expresses
\[
N(a,b,c;n)
\]
as a rational term plus three sums of floor functions:
\[
\sum_{i=1}^{b'_1-1}\left\lfloor \frac{i c'_1}{a}\right\rfloor,\qquad
\sum_{i=1}^{c'_2-1}\left\lfloor \frac{i a'_2}{b}\right\rfloor,\qquad
\sum_{i=1}^{a'_3-1}\left\lfloor \frac{i b'_3}{c}\right\rfloor.
\]
In general, these sums are difficult to evaluate explicitly [2604.10294].

For the special triplets
\[
(a,b,c)=(F_i,F_{i+1},F_{i+2})
\quad\text{and}\quad
(a,b,c)=(L_i,L_{i+1},L_{i+2}),
\]
the paper shows that Cassini-type identities force the modular parameters to collapse to
\[
c'_1=1,\qquad a'_2=1,\qquad b'_3=\text{(largest modulus)}-1.
\]
Consequently, the floor sums simplify drastically. In the Fibonacci case,
\[
\sum_{k=1}^{b'_1-1}\left\lfloor \frac{k}{F_i}\right\rfloor=0,\qquad
\sum_{k=1}^{c'_2-1}\left\lfloor \frac{k}{F_{i+1}}\right\rfloor=0,
\]
and
\[
\sum_{k=1}^{A'_3-1}\left\lfloor \frac{k(F_{i+2}-1)}{F_{i+2}}\right\rfloor
=
\frac{(A'_3-1)(A'_3-2)}{2}.
\]
The Lucas case is formally identical with \(F_{i+2}\) replaced by \(L_{i+2}\) and \(A'_3\) replaced by \(A''_3\) [2604.10294].

This yields exact formulas for
\[
N(F_i,F_{i+1},F_{i+2};n)
\quad\text{and}\quad
N(L_i,L_{i+1},L_{i+2};n),
\]
with the difficult floor-sum component reduced to a triangular number. The significance is methodological: Fibonacci and Lucas recurrences do not merely appear as coefficients, but force a nontrivial floor-sum reciprocity problem into an explicitly solvable regime [2604.10294].

Across these works, the phrase **Fibonacci floor function** therefore has a layered meaning. In the strictest sense it names the definable step function \(x\mapsto\) “largest even-index Fibonacci \(\le x\).” In a wider technical sense it refers to the use of floor operators to encode complementary sets, Beatty-indexed subsequences, block-sampled recurrences, and exact lattice-counting formulas controlled by Fibonacci or Lucas arithmetic.

Source: https://www.emergentmind.com/topics/fibonacci-floor-function