Theory and Modern Applications

# On a functional equation involving iterates and powers

## Abstract

We present a complete list of all continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of the equation ${f}^{2}\left(x\right)=\gamma {\left[f\left(x\right)\right]}^{\alpha }{x}^{\beta }$, where α, β and $\gamma >0$ are given real numbers.

MSC:39B22, 39B12, 26A18.

## 1 Introduction

In this note we give a complete list of all continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of the equation

${f}^{2}\left(x\right)=\gamma {\left[f\left(x\right)\right]}^{\alpha }{x}^{\beta },$
(1.1)

where α, β and $\gamma >0$ are given real numbers; here and throughout, ${f}^{2}$ denotes the second iterate of f. The motivation for writing this note was two problems concerning continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of some special cases of equation (1.1) (see [, Problem 11, p.312] and [, Problem 5, p.22]) as well as conference reports and papers on both problems (see [3, 4] and ). Let us mention that the problem from booklet  is wrongly solved in this booklet. To see that the problem from  concerns really equation (1.1) observe that from Remark 1.1 below it follows that in the case where $\beta \ne 0$ equation (1.1) can be rewritten in the form

$f\left(x\right){\left[{f}^{-1}\left(x\right)\right]}^{-\beta }=\gamma {x}^{\alpha }.$

Remark 1.1 Assume $\beta \ne 0$. Then every continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1) is strictly monotone and maps $\left(0,+\mathrm{\infty }\right)$ onto $\left(0,+\mathrm{\infty }\right)$.

Proof Fix $x,y\in \left(0,+\mathrm{\infty }\right)$ and assume that $f\left(x\right)=f\left(y\right)$. Then by (1.1) we get

$\gamma {\left[f\left(x\right)\right]}^{\alpha }{x}^{\beta }={f}^{2}\left(x\right)={f}^{2}\left(y\right)=\gamma {\left[f\left(y\right)\right]}^{\alpha }{y}^{\beta }=\gamma {\left[f\left(x\right)\right]}^{\alpha }{y}^{\beta },$

and since $\beta \ne 0$, we obtain $x=y$. Thus f is injective. This jointly with continuity implies strict monotonicity.

Now suppose that, contrary to our claim, ${lim}_{x\to 0}f\left(x\right)\in \left(0,+\mathrm{\infty }\right)$. Then, by the continuity of f and (1.1), we obtain

$f\left(\underset{x\to 0}{lim}f\left(x\right)\right)=\underset{x\to 0}{lim}{f}^{2}\left(x\right)=\underset{x\to 0}{lim}\gamma {\left[f\left(x\right)\right]}^{\alpha }{x}^{\beta }\in \left\{0,+\mathrm{\infty }\right\},$

a contradiction. Thus ${lim}_{x\to 0}f\left(x\right)\in \left\{0,+\mathrm{\infty }\right\}$. In the same manner we can prove that ${lim}_{x\to +\mathrm{\infty }}f\left(x\right)\in \left\{0,+\mathrm{\infty }\right\}$. □

## 2 Main results

To give a complete list of all continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1), we will split our consideration into the following three cases: $\beta =0$, $\alpha =0\ne \beta$ and $\alpha \ne 0\ne \beta$. It turns out that the description of all continuous solutions of equation (1.1) in the first case is quite easy; whereas in the third case it is much more complicated than in the second one.

### 2.1 The case $\beta =0$

If $\beta =0$, then equation (1.1) reduces to the equation

${f}^{2}\left(x\right)=\gamma {\left[f\left(x\right)\right]}^{\alpha }.$
(2.1)

Equation (2.1) was examined in ; cf. also [12, 13] and the references therein.

We begin with a (rather obvious and simple) characterization of general solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.1).

Proposition 2.1 A function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ satisfies (2.1) if and only if

$f\left(x\right)=\gamma {x}^{\alpha }$
(2.2)

for all $x\in f\left(\left(0,+\mathrm{\infty }\right)\right)$.

Proof () If a function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ satisfies (1.1), then for every $x=f\left(z\right)\in f\left(\left(0,+\mathrm{\infty }\right)\right)$ we have $f\left(x\right)={f}^{2}\left(z\right)=\gamma {\left[f\left(z\right)\right]}^{\alpha }=\gamma {x}^{\alpha }$.

() Fix a function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ satisfying (2.2) for all $x\in f\left(\left(0,+\mathrm{\infty }\right)\right)$. Then, for every $x\in \left(0,+\mathrm{\infty }\right)$, we have $f\left(x\right)\in f\left(\left(0,+\mathrm{\infty }\right)\right)$. Now, putting $f\left(x\right)$ in place of x in (2.2), we obtain (2.1). □

From Proposition 2.1 we obtain the following description of all continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.1).

Corollary 2.2 Let $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ be a continuous solution of equation (2.1). Then either f has form (2.2) for all $x\in \left(0,+\mathrm{\infty }\right)$ or there exists a proper subinterval I (open or closed or closed on one side; possible infinite or degenerated to a single point) of the half-line $\left(0,+\mathrm{\infty }\right)$ satisfying

(2.3)

such that f has form (2.2) for all $x\in I$,

(2.4)

and

$f\left(\left(0,+\mathrm{\infty }\right)\setminus I\right)\subset I.$
(2.5)

Moreover:

1. (i)

If $\alpha <-1$, then $I=\left\{{\gamma }^{\frac{1}{1-\alpha }}\right\}$;

2. (ii)

If $\alpha =-1$, then $clI=\left[A,\frac{\gamma }{A}\right]$ with arbitrary $A\in \left(0,\sqrt{\gamma }\right]$;

3. (iii)

If $\alpha \in \left(-1,0\right)$, then $clI=\left[A,B\right]$ with arbitrary $A\in \left(0,{\gamma }^{\frac{1}{1-\alpha }}\right]$ and $B\in \left[{\gamma }^{\frac{1}{1-\alpha }},+\mathrm{\infty }\right)$;

4. (iv)

If $\alpha \in \left[0,1\right)$, then either $clI=\left[0,B\right]$ or $clI=\left[A,B\right]$ or $clI=\left[A,+\mathrm{\infty }\right)$ with arbitrary $A\in \left(0,{\gamma }^{\frac{1}{1-\alpha }}\right]$ and $B\in \left[{\gamma }^{\frac{1}{1-\alpha }},+\mathrm{\infty }\right)$;

5. (v)

If $\alpha =1<\gamma$, then $clI=\left[A,+\mathrm{\infty }\right)$ with arbitrary $A\in \left(0,+\mathrm{\infty }\right)$;

6. (vi)

If $\alpha =1=\gamma$, then no restriction on I;

7. (vii)

If $\alpha =1>\gamma$, then $clI=\left(0,B\right]$ with arbitrary $B\in \left(0,+\mathrm{\infty }\right)$;

8. (viii)

If $\alpha \in \left(1,+\mathrm{\infty }\right)$, then either $I=\left\{{\gamma }^{\frac{1}{1-\alpha }}\right\}$ or $clI=\left[0,B\right]$ or $clI=\left[A,+\mathrm{\infty }\right)$ with arbitrary $A\in \left[{\gamma }^{\frac{1}{1-\alpha }},+\mathrm{\infty }\right)$ and $B\in \left(0,{\gamma }^{\frac{1}{1-\alpha }}\right]$.

Proof Put $I=f\left(\left(0,+\mathrm{\infty }\right)\right)$.

If $I=\left(0,+\mathrm{\infty }\right)$, then (2.2) holds for all $x\in \left(0,+\mathrm{\infty }\right)$ by Proposition 2.1. Therefore, to the end of the proof, we assume that $I\ne \left(0,+\mathrm{\infty }\right)$.

Since f is continuous, it follows that I is an interval. By Proposition 2.1 we see that (2.2) holds for all $x\in I$. Thus (2.3) holds. Condition (2.4) follows from the continuity of f and condition (2.5) is a consequence of the definition of I. This completes the proof of the main part of the result.

To prove the moreover part put $A=infI$ and $B=supI$. Since $I\ne \left(0,+\mathrm{\infty }\right)$, it follows that $0 or $B<+\mathrm{\infty }$.

Assume first that $\alpha <0$.

We will show that $0 and $B<+\mathrm{\infty }$.

If $B=+\mathrm{\infty }$, then (2.3) implies $0={lim}_{x\to B}\gamma {x}^{\alpha }\in clI$, which contradicts $I\ne \left(0,+\mathrm{\infty }\right)$. Similarly, if $A=0$, then (2.3) implies $+\mathrm{\infty }={lim}_{x\to {A}^{+}}\gamma {x}^{\alpha }\in clI$, which contradicts $I\ne \left(0,+\mathrm{\infty }\right)$.

Applying condition (2.3), we get $0 and $\gamma {A}^{\alpha }\le B<+\mathrm{\infty }$. Hence $\gamma {A}^{\alpha }\le B\le {\gamma }^{-\frac{1}{\alpha }}{A}^{\frac{1}{\alpha }}$.

1. (i)

If $\alpha <-1$, then ${\gamma }^{\frac{1}{1-\alpha }}\le A\le B\le {\gamma }^{\frac{1}{1-\alpha }}$.

2. (ii)

If $\alpha =-1$, then $\frac{\gamma }{A}\le B\le \frac{\gamma }{A}$.

3. (iii)

If $\alpha \in \left(-1,0\right)$, then $A\le {\gamma }^{\frac{1}{1-\alpha }}\le B$.

Assume now that $\alpha \ge 0$.

Then (2.3) yields $A\le \gamma {A}^{\alpha }$ provided that $A>0$ and $\gamma {B}^{\alpha }\le B$ provided that $B<+\mathrm{\infty }$.

1. (iv)

If $\alpha \in \left[0,1\right)$, then $A\le {\gamma }^{\frac{1}{1-\alpha }}$ if $A>0$ and ${\gamma }^{\frac{1}{1-\alpha }}\le B$ if $B<+\mathrm{\infty }$.

2. (v)

If $\alpha =1<\gamma$, then $B=+\mathrm{\infty }$ and no restriction on A.

3. (vi)

If $\alpha =1=\gamma$, then no restriction on A and B.

4. (vii)

If $\alpha =1>\gamma$, then $A=0$ and no restriction on B.

5. (viii)

If $\alpha >1$, then $A\ge {\gamma }^{\frac{1}{1-\alpha }}$ if $A>0$ and $B\le {\gamma }^{\frac{1}{1-\alpha }}$ if $B<+\mathrm{\infty }$.

□

As a consequence of Corollary 2.2, we have the following direct construction of all continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.1) (cf. [, Theorem 15.15]).

Corollary 2.3 Let I be a suitably chosen interval from Corollary  2.2 for a given $\alpha \in \mathbb{R}$, and let ${f}_{0}:I\to I$ be a function given by ${f}_{0}\left(x\right)=\gamma {x}^{\alpha }$. Then every extension of ${f}_{0}$ to a continuous function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ satisfying (2.5) is a solution of equation (2.1).

### 2.2 The case $\alpha =0\ne \beta$

If $\alpha =0\ne \beta$, then equation (1.1) reduces to the equation

${f}^{2}\left(x\right)=\gamma {x}^{\beta }.$
(2.6)

Charles Babbage was probably the first who looked for solutions of equation (2.6) in the case where $\beta =\gamma =1$ (see ). For the case where $\gamma =-\beta =1$, see ; cf. also . Equation (2.6) is a particular case of the equation of iterative roots (see [12, 13, 1921]). According to known results on the equation of iterative roots, we can formulate a theorem on continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.6).

Denote by $I\left(x,y\right)$ the closed interval $\left[min\left\{x,y\right\},max\left\{x,y\right\}\right]$ with $x,y\in \left(0,+\mathrm{\infty }\right)$.

Theorem 2.4

1. (i)

Assume $\beta <0$. Then equation (2.6) has no continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$.

2. (ii)

Assume $\beta =\gamma =1$.

(ii1) Then the formula $f\left(x\right)=x$ defines the unique continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.6).

(ii2) If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous and decreasing solution of equation (2.6), then there exists ${x}_{0}\in \left(0,+\mathrm{\infty }\right)$ such that $f\left({x}_{0}\right)={x}_{0}$ and f maps $\left(0,{x}_{0}\right]$ bijectively onto $\left[{x}_{0},+\mathrm{\infty }\right)$. Conversely, if ${x}_{0}\in \left(0,+\mathrm{\infty }\right)$, then every decreasing bijection ${f}_{0}:\left(0,{x}_{0}\right]\to \left[{x}_{0},+\mathrm{\infty }\right)$ such that $f\left({x}_{0}\right)={x}_{0}$ can be uniquely extended to a continuous and decreasing solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.6).

1. (iii)

Assume $\beta =1\ne \gamma$.

(iii1) Let ${x}_{0}\in \left(0,+\mathrm{\infty }\right)$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous and increasing solution of equation (2.6), then f maps $I\left({x}_{0},f\left({x}_{0}\right)\right)$ bijectively onto $I\left(f\left({x}_{0}\right),\gamma {x}_{0}\right)$. Conversely, if ${x}_{1}\in IntI\left({x}_{0},\gamma {x}_{0}\right)$, then every increasing bijection ${f}_{0}:I\left({x}_{0},{x}_{1}\right)\to I\left({x}_{1},\gamma {x}_{0}\right)$ can be uniquely extended to a continuous and increasing solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.6).

(iii2) Equation (2.6) has no continuous and decreasing solution from $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$.

1. (iv)

Assume $0<\beta \ne 1$.

(iv1) Let ${x}_{0}\in \left(0,{\gamma }^{\frac{1}{1-\beta }}\right)$ and let ${y}_{0}\in \left({\gamma }^{\frac{1}{1-\beta }},+\mathrm{\infty }\right)$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous and increasing solution of equation (2.6), then f maps $I\left({x}_{0},f\left({x}_{0}\right)\right)\cup I\left({y}_{0},f\left({y}_{0}\right)\right)$ bijectively onto $I\left(f\left({x}_{0}\right),\gamma {x}_{0}^{\beta }\right)\cup I\left(f\left({y}_{0}\right),\gamma {y}_{0}^{\beta }\right)$. Conversely, if ${x}_{1}\in IntI\left({x}_{0},\gamma {x}_{0}^{\beta }\right)$ and ${y}_{1}\in IntI\left({y}_{0},\gamma {y}_{0}^{\beta }\right)$, then every increasing bijection ${f}_{0}:I\left({x}_{0},{x}_{1}\right)\cup I\left({y}_{0},{y}_{1}\right)\to I\left({x}_{1},\gamma {x}_{0}^{\beta }\right)\cup I\left({y}_{1},\gamma {y}_{0}^{\beta }\right)$ can be uniquely extended to a continuous and increasing solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.6).

(iv2) If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous and decreasing solution of equation (2.6), then f maps $\left(0,{\gamma }^{\frac{1}{1-\beta }}\right]$ bijectively onto $\left[{\gamma }^{\frac{1}{1-\beta }},+\mathrm{\infty }\right)$ and $\gamma {\left[f\left(x\right)\right]}^{\beta }=f\left(\gamma {x}^{\beta }\right)$ for all $x\in \left(0,+\mathrm{\infty }\right)$. Conversely, every decreasing bijection ${f}_{0}:\left(0,{\gamma }^{\frac{1}{1-\beta }}\right]\to \left[{\gamma }^{\frac{1}{1-\beta }},+\mathrm{\infty }\right)$ such that $\gamma {\left[{f}_{0}\left(x\right)\right]}^{\beta }={f}_{0}\left(\gamma {x}^{\beta }\right)$ for all $x\in \left(0,{\gamma }^{\frac{1}{1-\beta }}\right]$ can be uniquely extended to a continuous and decreasing solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (2.6).

Proof All the assertions can be derived from [, Chapter XV]) as it has been noticed earlier. However, most of the assertions have evident proofs, so we present them for the convenience of the reader.

1. (i)

Suppose that, contrary to our claim, equation (2.6) has a continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$. Then the second iterate ${f}^{2}$ of f is strictly increasing. Now by (2.6) we conclude that $\beta >0$, a contradiction.

(ii1) It is clear that the identity function on $\left(0,+\mathrm{\infty }\right)$ satisfies (2.6). Suppose the contrary, and let $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ be another increasing solution of equation (2.6). Then there exists $x\in \left(0,+\mathrm{\infty }\right)$ such that $f\left(x\right)\ne x$. By the monotonicity of f, we conclude that ${f}^{2}\left(x\right)\ne x$, which contradicts (2.6).

(ii2) The first assertion is clear. To prove the second one, fix a decreasing bijection ${f}_{0}:\left(0,{x}_{0}\right]\to \left[{x}_{0},+\mathrm{\infty }\right)$ such that $f\left({x}_{0}\right)={x}_{0}$ and extend it to a function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ putting $f\left(x\right)={f}_{0}^{-1}\left(x\right)$ for all $x\in \left({x}_{0},+\mathrm{\infty }\right)$. It is easy to see that f is a decreasing bijection satisfying (2.6).

(iii1) The first assertion is evident. The second one can be deduced from [, Lemma 15.6].

(iii2) Suppose, to derive a contradiction, that $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous and decreasing solution of equation (2.6). Then there exists $x\in \left(0,+\mathrm{\infty }\right)$ such that $f\left(x\right)=x$. Hence ${f}^{2}\left(x\right)=x\ne \gamma x$, a contradiction.

(iv1) The first assertion is easy to verify. The second one can be inferred from [, Theorem 15.7].

(iv2) Let f be a continuous and decreasing solution of equation (2.6). Then there exists ${x}_{0}\in \left(0,+\mathrm{\infty }\right)$ such that $f\left({x}_{0}\right)={x}_{0}$. Hence by (2.6) we get ${x}_{0}={f}^{2}\left({x}_{0}\right)=\gamma {x}_{0}^{\beta }$, and thus ${x}_{0}={\gamma }^{\frac{1}{1-\beta }}$. Consequently, f maps $\left(0,{\gamma }^{\frac{1}{1-\beta }}\right]$ bijectively onto $\left[{\gamma }^{\frac{1}{1-\beta }},+\mathrm{\infty }\right)$. Moreover, (2.6) yields $f\left(\gamma {x}^{\beta }\right)={f}^{3}\left(x\right)=\gamma {\left[f\left(x\right)\right]}^{\beta }$ for all $x\in \left(0,+\mathrm{\infty }\right)$. To prove the second part of the assertion, fix a decreasing bijection ${f}_{0}:\left(0,{\gamma }^{\frac{1}{1-\beta }}\right]\to \left[{\gamma }^{\frac{1}{1-\beta }},+\mathrm{\infty }\right)$ such that $\gamma {\left[{f}_{0}\left(x\right)\right]}^{\beta }={f}_{0}\left(\gamma {x}^{\beta }\right)$ for all $x\in \left(0,{\gamma }^{\frac{1}{1-\beta }}\right]$ and extend it to a function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ putting $f\left(x\right)={f}_{0}^{-1}\left(\gamma {x}^{\beta }\right)$ for all $x\in \left({\gamma }^{\frac{1}{1-\beta }},+\mathrm{\infty }\right)$. It is easy to calculate that f is a decreasing bijection satisfying (2.6). □

### 2.3 The case $\alpha \ne 0\ne \beta$

In this case an explicit description of all continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1) is much more involved than in both the previous cases. We begin with an observation which allows us to rewrite equation (1.1) in an equivalent form.

Lemma 2.5 If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a solution of equation (1.1), then the function $g:\mathbb{R}\to \mathbb{R}$ given by $g\left(x\right)=logf\left({e}^{x}\right)$ satisfies

${g}^{2}\left(x\right)=log\gamma +\alpha g\left(x\right)+\beta x.$
(2.7)

Conversely, if $g:\mathbb{R}\to \mathbb{R}$ is a solution of equation (2.7), then the function $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ given by $f\left(x\right)={e}^{g\left(logx\right)}$ satisfies (1.1).

Equation (2.7) is a special case of the polynomial-like iterative inhomogeneous equation

$\sum _{n=0}^{N}{a}_{n}{f}^{n}\left(x\right)=b,$
(2.8)

where all ${a}_{n}$’s and b are real numbers and f is an unknown self-mapping; here ${f}^{n}$ denotes the n th iterate of f. For the theory of equation (2.8) and its generalizations, we refer the readers to . The problem of finding all continuous solutions of equation (2.8) seems to be very difficult. It is completely solved in  for $N=2$, but it is still open even in the case where $N=3$ (see ). It turns out that the nature of continuous solutions of equation (2.8) depends on the behavior of complex roots ${r}_{1},\dots ,{r}_{N}$ of its characteristic equation ${\sum }_{n=0}^{N}{\alpha }_{n}{r}^{n}=0$. This characteristic equation is obtained by putting $f\left(x\right)=rx$ into (2.8) with $b=0$; in this way we can determine all linear solutions of the homogeneous counterpart of equation (2.8), and then all affine solutions of equation (2.8). Therefore, to formulate our result on continuous solutions $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1), denote by ${r}_{1}$ and ${r}_{2}$ the complex roots of the equation

${r}^{2}-\alpha r-\beta =0.$

By our assumption, we have ${r}_{1}+{r}_{2}=\alpha \ne 0$ and ${r}_{1}{r}_{2}=-\beta \ne 0$.

Combining Lemma 2.5 with the results from , we get the following theorem.

Theorem 2.6

1. (i)

Assume ${r}_{1},{r}_{2}\in \mathbb{C}\setminus \mathbb{R}$. Then equation (1.1) has no continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$.

2. (ii)

Assume ${r}_{1},{r}_{2}\in \mathbb{R}\setminus \left\{1\right\}$. Let $\xi ={\gamma }^{\frac{1}{\left({r}_{1}-1\right)\left({r}_{2}-1\right)}}$ and ${f}_{1},{f}_{2}:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ be defined by

${f}_{1}\left(x\right)={\gamma }^{-\frac{1}{{r}_{2}-1}}{x}^{{r}_{1}},\phantom{\rule{2em}{0ex}}{f}_{2}\left(x\right)={\gamma }^{-\frac{1}{{r}_{1}-1}}{x}^{{r}_{2}}.$

(ii1) Assume either $1<{r}_{1}<{r}_{2}$ or $0<{r}_{1}<{r}_{2}<1$. Let $0<{x}_{0}<\xi <{y}_{0}$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous solution of equation (1.1), then $f\left(\xi \right)=\xi$, ${f}_{0}=f{|}_{I\left({x}_{0},f\left({x}_{0}\right)\right)\cup I\left({y}_{0},f\left({y}_{0}\right)\right)}$ is continuous,

${f}_{0}\left({x}_{0}\right)\in I\left({f}_{1}\left({x}_{0}\right),{f}_{2}\left({x}_{0}\right)\right),\phantom{\rule{2em}{0ex}}{f}_{0}^{2}\left({x}_{0}\right)=\gamma {\left[{f}_{0}\left({x}_{0}\right)\right]}^{\alpha }{x}_{0}^{\beta },$
(2.9)
${f}_{0}\left({y}_{0}\right)\in I\left({f}_{1}\left({y}_{0}\right),{f}_{2}\left({y}_{0}\right)\right),\phantom{\rule{2em}{0ex}}{f}_{0}^{2}\left({y}_{0}\right)=\gamma {\left[{f}_{0}\left({y}_{0}\right)\right]}^{\alpha }{y}_{0}^{\beta }$
(2.10)

and

${\left(\frac{y}{x}\right)}^{{r}_{1}}\le \frac{{f}_{0}\left(y\right)}{{f}_{0}\left(x\right)}\le {\left(\frac{y}{x}\right)}^{{r}_{2}}$
(2.11)

for all $x,y\in I\left({x}_{0},f\left({x}_{0}\right)\right)\cup I\left({y}_{0},f\left({y}_{0}\right)\right)$. Conversely, every continuous function ${f}_{0}:I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)\cup I\left({y}_{0},{f}_{0}\left({y}_{0}\right)\right)\to \mathbb{R}$ such that (2.9) and (2.10) are satisfied and (2.11) holds for all $x,y\in I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)\cup I\left({y}_{0},{f}_{0}\left({y}_{0}\right)\right)$ can be uniquely extended to a continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(ii2) Assume $0<{r}_{1}<1<{r}_{2}$. Let ${x}_{0}>0$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous solution of equation (1.1), then either $f{|}_{\left(0,\xi \right]}\in \left\{{f}_{1}{|}_{\left(0,\xi \right]},{f}_{2}{|}_{\left(0,\xi \right]}\right\}$ and $f{|}_{\left(\xi ,+\mathrm{\infty }\right)}\in \left\{{f}_{1}{|}_{\left(\xi ,+\mathrm{\infty }\right)},{f}_{2}{|}_{\left(\xi ,+\mathrm{\infty }\right)}\right\}$ or ${f}_{0}=f{|}_{I\left({x}_{0},f\left({x}_{0}\right)\right)}$ is continuous,

${f}_{0}\left({x}_{0}\right)\notin I\left({f}_{1}\left({x}_{0}\right),{f}_{2}\left({x}_{0}\right)\right),\phantom{\rule{2em}{0ex}}{f}_{0}^{2}\left({x}_{0}\right)=\gamma {\left[{f}_{0}\left({x}_{0}\right)\right]}^{\alpha }{x}_{0}^{\beta }$
(2.12)

and (2.11) holds for all $x,y\in I\left({x}_{0},f\left({x}_{0}\right)\right)$. Conversely, every continuous function ${f}_{0}:I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)\to \mathbb{R}$ such that (2.12) is satisfied and (2.11) holds for all $x,y\in I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)$ can be uniquely extended to a continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(ii3) Assume either ${r}_{1}<{r}_{2}<-1$ or $-1<{r}_{1}<{r}_{2}<0$. Let ${x}_{0}>\xi$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous solution of equation (1.1), then $f\left(\xi \right)=\xi$ and ${f}_{0}=f{|}_{I\left({x}_{0},{f}^{2}\left({x}_{0}\right)\right)}$ is continuous,

${f}_{0}\left({x}_{0}\right)\in I\left({f}_{1}\left({x}_{0}\right),{f}_{2}\left({x}_{0}\right)\right),\phantom{\rule{2em}{0ex}}{f}_{0}^{3}\left({x}_{0}\right)=\gamma {\left[{f}_{0}^{2}\left({x}_{0}\right)\right]}^{\alpha }{x}_{0}^{\beta },$
(2.13)

and (2.11) holds for all $x,y\in I\left({x}_{0},{f}^{2}\left({x}_{0}\right)\right)$. Conversely, every continuous function ${f}_{0}:I\left({x}_{0},{f}_{0}^{2}\left({x}_{0}\right)\right)\to \mathbb{R}$ such that (2.13) is satisfied and (2.11) holds for all $x,y\in I\left({x}_{0},{f}_{0}^{2}\left({x}_{0}\right)\right)$ can be uniquely extended to a continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(ii4) Assume either ${r}_{1}<{r}_{2}=-1$ or $-1={r}_{2}<{r}_{1}<0$. Then every continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1) is of the form

with some $a\in \left[1,+\mathrm{\infty }\right]$.

(ii5) Assume either ${r}_{1}<0<{r}_{2}$ or ${r}_{1}<-1<{r}_{2}<0$. Then ${f}_{1}$ and ${f}_{2}$ are the only continuous solutions from $\left(0,+\mathrm{\infty }\right)$ to $\left(0,+\mathrm{\infty }\right)$ of equation (1.1).

1. (iii)

Assume ${r}_{2}=1\ne {r}_{1}$. Let ${f}_{3}:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ be defined by

${f}_{3}\left(x\right)={\gamma }^{-\frac{1}{{r}_{1}-1}}x.$

(iii1) Assume ${r}_{1}>1\ne \gamma$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous solution of equation (1.1), then ${f}_{0}=f{|}_{I\left({x}_{0},f\left({x}_{0}\right)\right)}$ is continuous,

${f}_{0}\left({x}_{0}\right)\le {f}_{3}\left({x}_{0}\right),\phantom{\rule{2em}{0ex}}{f}_{0}^{2}\left({x}_{0}\right)=\gamma {\left[{f}_{0}\left({x}_{0}\right)\right]}^{\alpha }{x}_{0}^{\beta }$
(2.14)

and

$\frac{y}{x}\le \frac{{f}_{0}\left(y\right)}{{f}_{0}\left(x\right)}\le {\left(\frac{y}{x}\right)}^{{r}_{1}}$
(2.15)

for all $x,y\in I\left({x}_{0},f\left({x}_{0}\right)\right)$. Conversely, every continuous function ${f}_{0}:I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)\to \mathbb{R}$ such that (2.14) is satisfied and (2.15) holds for all $x,y\in I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)$ can be uniquely extended to a continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(iii2) Assume $0<{r}_{1}<1\ne \gamma$. If $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ is a continuous solution of equation (1.1), then ${f}_{0}=f{|}_{I\left({x}_{0},f\left({x}_{0}\right)\right)}$ is continuous,

${f}_{0}\left({x}_{0}\right)\ge {f}_{3}\left({x}_{0}\right),\phantom{\rule{2em}{0ex}}{f}_{0}^{2}\left({x}_{0}\right)=\gamma {\left[{f}_{0}\left({x}_{0}\right)\right]}^{\alpha }{x}_{0}^{\beta }$
(2.16)

and

${\left(\frac{y}{x}\right)}^{{r}_{1}}\le \frac{{f}_{0}\left(y\right)}{{f}_{0}\left(x\right)}\le \frac{y}{x}$
(2.17)

for all $x,y\in I\left({x}_{0},f\left({x}_{0}\right)\right)$. Conversely, every continuous function ${f}_{0}:I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)\to \mathbb{R}$ such that (2.16) is satisfied and (2.17) holds for all $x,y\in I\left({x}_{0},{f}_{0}\left({x}_{0}\right)\right)$ can be uniquely extended to a continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(iii3) Assume ${r}_{1}>0$ and $\gamma =1$. Then every continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1) is of the form

with some $0\le a\le b\le +\mathrm{\infty }$.

(iii4) Assume ${r}_{1}<0$ and $\gamma \ne 1$. Then ${f}_{3}$ is the unique continuous solution from $\left(0,+\mathrm{\infty }\right)$ to $\left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(iii5) Assume ${r}_{1}<0$ and $\gamma =1$. Then ${f}_{3}$ is a continuous solution of equation (1.1) and every other continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1) is of the form

$f\left(x\right)=a{x}^{{r}_{1}}$

with some $a\in \left(0,+\mathrm{\infty }\right)$.

1. (iv)

Assume ${r}_{1}={r}_{2}$.

(iv1) Assume ${r}_{1}\ne 1$. Then the formula

$f\left(x\right)={\gamma }^{-\frac{1}{{r}_{1}-1}}{x}^{{r}_{1}}$

defines the unique continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1).

(iv2) Assume ${r}_{1}=1\ne \gamma$. Then equation (1.1) has no continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$.

(iv3) Assume ${r}_{1}=\gamma =1$. Then every continuous solution $f:\left(0,+\mathrm{\infty }\right)\to \left(0,+\mathrm{\infty }\right)$ of equation (1.1) is of the form

$f\left(x\right)=ax$

with some $a\in \left(0,+\mathrm{\infty }\right)$.

## Author’s contributions

The author wrote the whole article, read and approved the final manuscript.

## References

1. Brillouët-Belluot N: Problem posed during The Forty-ninth International Symposium on Functional Equations. Aequ. Math. 2012, 84: 312.

2. Brodsky YS, Slipenko AK: Functional Equations. Vishcha Shkola, Kiev; 1983.

3. Boros Z: Talk given during The Fiftieth International Symposium on Functional Equations. Aequ. Math. 2013, 86: 293.

4. Sablik, M: Talk given during The Fourteenth Debrecen-Katowice Winter Seminar on Functional Equations and Inequalities. Hajdúszoboszló, Hungary, January 29 - February 1 (2014)

5. Jarczyk W, Morawiec J: Note on an equation occurring in a problem of Nicole Brillouët-Belluot. Aequ. Math. 2012, 84: 227-233. 10.1007/s00010-011-0097-7

6. Morawiec J: On a problem of Nicole Brillouët-Belluot. Aequ. Math. 2012, 84: 219-225. 10.1007/s00010-011-0096-8

7. Morawiec J: Around a problem of Nicole Brillouët-Belluot. Aequ. Math. 2014, 88: 175-181. 10.1007/s00010-013-0216-8

8. Morawiec J: Around a problem of Nicole Brillouët-Belluot, II. Aequ. Math. 2014. 10.1007/s00010-013-0249-z

9. Chen L, Shi Y:The real solutions of functional equation ${f}^{\left[m\right]}=1/f$. J. Math. Res. Expo. 2008, 28: 323-330.

10. Ng CT, Zhang W: When does an iterate equal a power? Publ. Math. (Debr.) 2005, 67: 79-91.

11. Shi Y, Chen L:All the meromorphic solutions of the functional equation ${f}^{\left[m\right]}=1/f$. J. Nat. Sci. Nanjing Norm. Univ. 2007, 2: 25-27.

12. Baron K, Jarczyk W: Recent results on functional equations in a single variable, perspectives and open problem. Aequ. Math. 2001, 61: 1-48. 10.1007/s000100050159

13. Kuczma M Monografie Matematyczne 46. In Functional Equations in a Single Variable. PWN, Warsaw; 1968.

14. Babbage C: Essay towards the calculus of functions. Philos. Trans. R. Soc. Lond. 1815, 105: 389-424. 10.1098/rstl.1815.0024

15. Anschuetz R, Scherwood H: When is a function’s inverse equal to its reciprocal? Coll. Math. J. 1997, 27: 388-393.

16. Cheng R, Dasgupta A, Ebanks BR, Kinch LF, Larso LM, McFadden RB:When does ${f}^{-1}=1/f$? Am. Math. Mon. 1998, 105: 704-716. 10.2307/2588987

17. Euler R, Foran J: On functions whose inverse is their reciprocal. Math. Mag. 1981, 54: 185-189. 10.2307/2689629

18. Massera JL, Petracca A:Sobre la ecuación funcional $f\left(f\left(x\right)\right)=1/x$. Rev. Unión Mat. Argent. 1946, 11: 206-211.

19. Kuczma M, Choczewski B, Ger R Encyclopedia of Mathematics and Its Applications 32. In Iterative Functional Equations. Cambridge University Press, Cambridge; 1990.

20. Targonski G: Topics in Iteration Theory. Vandenhoeck & Ruprecht, Göttingen; 1981.

21. Targonski G: Progress of iteration theory since 1981. Aequ. Math. 1995, 50: 50-72. 10.1007/BF01831113

22. Chen J, Zhang W: Leading coefficient problem for polynomial-like iterative equations. J. Math. Anal. Appl. 2009, 349: 413-419. 10.1016/j.jmaa.2008.09.015

23. Jarczyk, W: A Recurrent Method of Solving Iterative Functional Equations. Prace Naukowe Uniwersytetu Śląskiego w Katowicach 1206, Uniwersytet Śląski, Katowice (1991)

24. Matkowski J, Zhang Z: Characteristic analysis for a polynomial-like iterative equation. Chin. Sci. Bull. 1998, 43: 192-196. 10.1007/BF02898908

25. Trif T: Convex solutions to polynomial-like iterative equations on open intervals. Aequ. Math. 2010, 79: 315-325. 10.1007/s00010-010-0020-7

26. Xu B, Zhang W: Construction of continuous solutions and stability for the polynomial like iterative equation. J. Math. Anal. Appl. 2007, 325: 1160-1170. 10.1016/j.jmaa.2006.02.065

27. Yang D, Zhang W: Characteristic solutions of polynomial-like iterative equations. Aequ. Math. 2004, 67: 80-105. 10.1007/s00010-003-2708-4

28. Tabor J, Tabor J: On a linear iterative equation. Results Math. 1995, 27: 412-421. 10.1007/BF03322847

29. Zhang W, Zhang W: On continuous solutions of n -th order polynomial-like iterative equations. Publ. Math. (Debr.) 2010, 76: 117-134.

30. Zhang, P, Gong, X: Continuous solutions of 3-order iterative equation of linear dependence. Manuscript

31. Nabeya S:On the functional equation $f\left(p+qx+rf\left(x\right)\right)=a+bx+cf\left(x\right)$. Aequ. Math. 1974, 11: 199-211. 10.1007/BF01834919

32. Matkowski J: Remark done during The Twenty-sixth International Symposium on Functional Equations. Aequ. Math. 1989, 37: 119.

## Acknowledgements

This research was supported by Silesian University Mathematics Department (Iterative Functional Equations and Real Analysis program).

## Author information

Authors

### Corresponding author

Correspondence to Janusz Morawiec. 