Abstract
In this paper, we study the existence of entire solutions of finite-order of non-linear difference equations of the form
and
where q, Q are non-zero polynomials, \(c,\lambda ,p_{i},\alpha _{i}(i=1,2)\) are non-zero constants such that \(\alpha _{1}\ne \alpha _{2}\) and \(\Delta _{c}f(z)=f(z+c)-f(z)\not \equiv 0\). Our results are improvements and complements of Wen et al. (Acta Math Sin 28:1295–1306, 2012), Yang and Laine (Proc Jpn Acad Ser A Math Sci 86:10–14, 2010) and Zinelâabidine (Mediterr J Math 14:1–16, 2017).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction and Main Results
In this paper, we assume that the reader is familiar with the fundamental results and standard notation of Nevanlinna theory [5, 7, 14]. In addition, we use \(\rho (f)\) to denote the order of growth of f and \(\lambda (f)\) to denote the exponent of convergence of zeros sequence of f. For simplicity, we denote by S(r, f) any quantify satisfying \(S(r,f)=o(T(r,f)),\) as \(r\rightarrow \infty \), outside of a possible exceptional set of finite logarithmic measure, we use S(f) to denote the family of all small functions with respect to f.
Recently, many scholars have investigated solvability and existence of solutions of non-linear differential equations or difference equations, see [3, 4, 8,9,10,11,12, 16].
Exponential polynomials are important in complex analysis as they have many interesting properties as mentioned, for example, in the paper [13] due to Wen, Heittokangas and Laine. In this paper, we mainly give exact expressions of exponential polynomial solutions of certain class of non-linear difference equations.
In [16], Yang and Laine proved the following result:
Theorem A
A non-linear difference equation
where q is a non-constant polynomial and b, c are non-zero complex constants, Eq. (1.1) does not admit entire solutions of finite order. If q is a non-zero constant, then Eq. (1.1) possesses three district entire solutions of finite order, provided \(b=3\pi n\) and \(q^{3}=(-1)^{n+1}\frac{27}{4}c^{2}\) for a non-zero integer n.
Given Theorem A, it is natural to ask about the solutions of the following more general form
where q is a non-zero polynomial, \(c,\lambda ,p_{i}(i=1,2)\) are non-zero constants such that \(\Delta _{c}f(z)=f(z+c)-f(z)\not \equiv 0\) and \(n\ge 2\) is an integer.
In this paper, we study this problem and obtain the following result.
Theorem 1.1
Let \(n \ge 2\) be an integer, q be a non-zero polynomial, \(c,\lambda ,p_{1},p_{2}\) be non-zero constants. If there exists some entire solution f of finite order to Eq. (1.2), such that \(\Delta _{c}f(z)=f(z+c)-f(z)\not \equiv 0\), then q is a constant, and \(n=2\) or \(n=3\).
When \(n=2\), then
where \(q^{4}=4p_{1}p_{2}\), \(c_{1}^{2}=p_{1},~c_{2}^{2}=p_{2}\), \(\lambda c=2k\pi i\), \(k\in {\mathbb {Z}}\) and k is an odd.
When \(n=3\), then
where \(q^{3}=\frac{27}{8}p_{1}p_{2}\), \(c_{1}^{3}=p_{1},~c_{2}^{3}=p_{2}\), \(\lambda c=3k\pi i\), \(k\in {\mathbb {Z}}\) and k is an odd.
More recently, Zinelâabidine showed in [17]:
Theorem B
Let q be a polynomial, \(p_{i},\alpha _{i}(i=1,2)\) be non-zero constants such that \(\alpha _{1}\pm \alpha _{2}\ne 0\). If f is an entire solution of finite order of equation
such that \(\Delta f(z)=f(z+1)-f(z)\not \equiv 0\), then q is a constant, and one of the following relations holds:
-
1.
\(f(z)=c_{1}\mathrm{e}^{\frac{\alpha _{1}}{3}z}\) and \(c_{1}(\mathrm{e}^{\frac{\alpha _{1}}{3}}-1)q=p_{2}\), \(\alpha _{1}=3\alpha _{2}\),
-
2.
\(f(z)=c_{2}\mathrm{e}^{\frac{\alpha _{2}}{3}z}\) and \(c_{2}(\mathrm{e}^{\frac{\alpha _{2}}{3}}-1)q=p_{1}\), \(\alpha _{2}=3\alpha _{1}\), where \(c_{1},c_{2}\) are non-zero constants satisfying \(c_{1}^{3}=p_{1},~c_{2}^{3}=p_{2}\).
The aim of this paper is to study the difference equation
where q is a non-zero polynomial, \(c,p_{i},\alpha _{i}(i=1,2)\) are non-zero constants such that \(\alpha _{1}\pm \alpha _{2}\ne 0\) and \(\Delta _{c} f(z)=f(z+c)-f(z)\not \equiv 0\). In fact, we prove the following result.
Theorem 1.2
Let q be a non-zero polynomial, \(c,p_{i},\alpha _{i}(i=1,2)\) be non-zero constants such that \(\alpha _{1}\pm \alpha _{2}\ne 0\). If f is an entire solution of finite order of Eq. (1.4), such that \(\Delta _{c} f(z)=f(z+c)-f(z)\not \equiv 0\), then q is a constant, \(\rho (f)=1\) and one of the following conclusions holds:
-
1.
\(f(z)=c_{1}\mathrm{e}^{\frac{\alpha _{1}}{2}z}\), and \(c_{1}(\mathrm{e}^{\frac{\alpha _{1}}{2}c}-1)q=p_{2}\), \(\alpha _{1}=2\alpha _{2}\);
-
2.
\(f(z)=c_{2}\mathrm{e}^{\frac{\alpha _{2}}{2}z}\), and \(c_{2}(\mathrm{e}^{\frac{\alpha _{2}}{2}c}-1)q=p_{1}\), \(\alpha _{2}=2\alpha _{1}\), where \(c_{1},c_{2}\) are non-zero constants satisfying \(c_{1}^{2}=p_{1},~c_{2}^{2}=p_{2}\);
-
3.
$$\begin{aligned} T(r,\varphi )+S(r,f)= & {} \kappa T(r,f),~0<\kappa \le 1,~\text{ and }\\ N\left( r,\frac{1}{f}\right) +S(r,f)= & {} \iota T(r,f),~1-\frac{\kappa }{2}\le \iota \le 1, \end{aligned}$$
where \(\varphi =\alpha _{1}\alpha _{2}f^{2}-2(\alpha _{1}+\alpha _{2})ff'+2(f')^{2}+2ff''\).
Wen, Heittokangas and Laine [13] studied and classified the finite order entire solutions f of equation
where q, Q, P are polynomials, \(n\ge 2\) is an integer and \(c\in {\mathbb {C}}{\setminus }\{0\}\), and obtained the following Theorem C.
Recall that a function f of the form
where \(P_{j}'s\) and \(Q_{j}'s\) are polynomials in z is called an exponential polynomial. Furthermore, let
Theorem C
(See [13]) Let \(n\ge 2\) be an integer, let \(c\in {\mathbb {C}}{\setminus }\{0\}\), and let q, Q, P be polynomials such that Q is not a constant and \(q\not \equiv 0\). Then, we identify the finite order entire solutions f of equation (1.5) as follows:
-
(a)
Every solution f satisfies \(\rho (f)=\deg Q\) and is of mean type.
-
(b)
Every solution f satisfies \(\lambda (f)=\rho (f)\) if and only if \(P\not \equiv 0\).
-
(c)
A solution belongs to \(\Gamma _{0}\) if and only if \(P\equiv 0\). In particular, this is the case if \(n\ge 3\).
-
(d)
If a solution f belongs to \(\Gamma _{0}\) and if g is any other finite-order entire solution to (1.5), then \(f =\eta g\), where \(\eta ^{n-1}=1\).
-
(e)
If f is an exponential polynomial solution of the form (1.6), then \(f\in \Gamma _{1}\). Moreover, if \(f\in \Gamma _{1}{\setminus }\Gamma _{0}\), then \(\rho (f)=1\).
A natural question to ask is about \(P(z)=p_{1}\mathrm{e}^{\lambda z}+p_{2}\mathrm{e}^{-\lambda z}\) in (1.5), where \(\lambda ,p_{1},p_{2}\in {\mathbb {C}}{\setminus }\{0\}\) are constants. We consider this question and obtain the following result.
Theorem 1.3
Let \(n\ge 3\) be an integer, let \(c,\lambda ,p_{1},p_{2}\in {\mathbb {C}}{\setminus }\{0\}\) be constants and let q, Q be polynomials such that Q is not a constant and \(q\not \equiv 0\). If f is an entire solution of finite order of the equation
then the following conclusions hold.
-
1.
Every solution f satisfies \(\rho (f)=\deg Q=1\).
-
2.
If a solution f belongs to \(\Gamma _{0}\), then \(f(z)=e^{\frac{\lambda }{n}z+B},~Q(z)=-\frac{n+1}{n}\lambda z+b\) or \(f(z)=e^{-\frac{\lambda }{n}z+B},~Q(z)=\frac{n+1}{n}\lambda z+b\), where \(b,B\in {\mathbb {C}}\).
Remark 1.1
We conjecture that if \(n=2\), the conclusions of Theorem 1.3 are still valid, although we have not found a suitable method of proof yet. For example, \(f(z)=\mathrm{e}^{z}\) is an entire solution of finite order of the difference equation
and \(f(z)=\mathrm{e}^{z}+\mathrm{e}^{-z}\) is an entire solution of finite order of the difference equation
The following example shows that our estimates in Theorem 1.3 are accurate.
Example 1.1
If \(f(z)=\mathrm{e}^{z}\) is an entire solution of finite order of the difference equation
where \(\lambda =3,~n=3,~b=B=0\), then \(f(z)=\mathrm{e}^{\frac{\lambda }{n}z+B}=\mathrm{e}^{z}\), \(Q(z)=-\frac{n+1}{n}\lambda z+b=-4z\) and \(\rho (f)=\deg Q=1\).
2 Some Lemmas
Lemma 2.1
\((Clunie's~Lemma)\) (See [1, 7, Lem. 2.4.2]) Let f be a transcendental meromorphic solution of
where P(z, f) and Q(z, f) are polynomials in f and its derivatives with meromorphic coefficients, say \(\{a_{\lambda }|\lambda \in I\}\), such that \(m(r,a_{\lambda })=S(r,f)\) for all \(\lambda \in I\). If the total degree of Q(z, f) as a polynomial in f and its derivatives is at most n, then
Lemma 2.2
(See [6, Cor. 3.3]) Let f be a non-constant finite order meromorphic solution of
where P(z, f) and Q(z, f) are difference polynomials in f with small meromorphic coefficients, and let \(c\in {\mathbb {C}}\), \(\delta <1\). If the total degree of Q(z, f) as a polynomial in f and its shifts at most n, then
for all r outside of a possible exceptional set with finite logarithmic measure.
Remark 2.1
In Lemma 2.2, if f is a transcendental meromorphic function with finite order \(\rho \), and P(z, f), Q(z, f) are differential-difference polynomials in f, then by the same reasoning as in the proof of Lemma 2.1, we also obtain the conclusion (2.2). Furthermore, if the coefficients of P(z, f) and Q(z, f) are polynomials \(A_{j},~j=1,\ldots ,n\), for each \(\varepsilon >0\), then (2.2) can be written as:
where r is sufficiently large.
Lemma 2.3
(See [15, Thm. 1.51]) Suppose that \(f_{1}, f_{2},\ldots ,f_{n}(n\ge 2)\) are meromorphic functions and \(g_{1}, g_{2},\ldots ,g_{n}\) are entire functions satisfying the following conditions:
-
1.
\({\sum \limits _{j=1}^{n} f_{j}\mathrm{e}^{g_{j}}\equiv 0.}\)
-
2.
\(g_{j}-g_{k}\) are not constants for \(1\le j<k\le n\).
-
3.
For \(1\le j\le n, 1\le h<k\le n\),
$$\begin{aligned} T(r,f_{j})=o(T(r,e^{g_{h}-g_{k}}))~(r\rightarrow \infty , r\not \in E), \end{aligned}$$
where \(E\subset [1,\infty )\) is finite linear measure or finite logarithmic measure.
Then \(f_{j}\equiv 0~(j=1,\ldots , n)\).
Lemma 2.4
(See [11, Lem. 6]) Suppose that f is a transcendental meromorphic function, a, b, c, d are small functions with respect to f and \(acd\not \equiv 0\). If
then
Lemma 2.5
(See [2, Cor. 2.6]) Let \(\eta _{1},\eta _{2}\) be two complex numbers such that \(\eta _{1}\ne \eta _{2}\) and let f be a finite order meromorphic function. Let \(\rho \) be the order of f. Then for each \(\varepsilon >0\), we have
Using an argument similar to that used in [3, Lem. 2.3], we get the following result.
Lemma 2.6
Let \(n\ge 1\) be an integer, \(\lambda \) be a non-zero constant and H be a non-zero polynomial. Then, the differential equation
has a special solution \(c_{0}\) which is a non-zero polynomial.
3 Proof of Theorem 1.1
Denote \(P=q\Delta _{c}f\). Suppose that f is a transcendental entire solution of finite order of Eq. (1.2). Differentiating (1.2), we have
Differentiating (3.1) yields
It follows from (1.2) and (3.1) that
It follows from (1.2) and (3.2) that
Eliminating \((f')^{2}\) from (3.3) and (3.4) yields
where
and
Q(z, f) is a differential-difference polynomial in f and the total degree is at most \(n+1\). Note that when \(n\ge 2\), then \(2n-1\ge n+1\), by Lemma 2.2 and Remark 2.1, we have \(m(r,\varphi )=S(r,f)\); therefore, \(T(r,\varphi )=S(r,f)\). We distinguish two cases below:
Case 1 If \(\varphi \equiv 0\), i.e. \(\lambda ^{2}f-n^{2}f''\equiv 0\). Every entire solution \(f(\not \equiv 0)\) of this equation can be expressed as:
where \(c_{1},c_{2}\) are non-zero constants. Otherwise, if one of \(c_{1},c_{2}\) is equal to zero, substituting (3.8) into (1.2) and using Lemma 2.3, we obtain a contradiction.
When \(n=3\), then \(f(z)=c_{1}\mathrm{e}^{\frac{\lambda }{3} z}+c_{2}\mathrm{e}^{-\frac{\lambda }{3} z}\), substituting this into (1.2) yields
It follows from (3.9) and Lemma 2.3 that
Note that \(c_{1},c_{2}\in {\mathbb {C}}{\setminus }\{0\}\), then \(q(z)(\mathrm{e}^{\frac{\lambda }{3} c}-1)=q(z)(\mathrm{e}^{-\frac{\lambda }{3} c}-1)=-3c_{1}c_{2}\ne 0\), q(z) reduces to a constant q and \(\lambda c=3k\pi i\), \(k\in {\mathbb {Z}}{\setminus }\{0\}\) and k is an odd, \(q^{3}=\frac{27}{8}p_{1}p_{2}\).
When \(n=2l(l\ge 1)\) is even, \(f(z)=c_{1}\mathrm{e}^{\frac{\lambda }{2l} z}+c_{2}\mathrm{e}^{-\frac{\lambda }{2l} z}\). Substituting this into (1.2) yields
If \(k=l\), then \(\sum _{k=1}^{2l-1}\left( {\begin{array}{c}k\\ 2l\end{array}}\right) c_{1}^{2l-k}c_{2}^{k}e^{\frac{2l-2k}{2l}\lambda z}\) must have a constant term. That is \(\left( {\begin{array}{c}l\\ 2l\end{array}}\right) c_{1}^{l}c_{2}^{l}=\frac{(2l)!}{l!l!}(c_{1}c_{2})^{l}\). By Lemma 2.3, we obtain \(c_{1}c_{2}=0\), a contradiction.
When \(n=2l+1(l\ge 2)\) is odd, \(f(z)=c_{1}\mathrm{e}^{\frac{\lambda }{2l+1} z}+c_{2}\mathrm{e}^{-\frac{\lambda }{2l+1} z}\). Substituting this into (1.2) yields
i.e.
Since \(l\ge 2\), then \(\sum \nolimits _{\begin{array}{c} k=1\\ k\ne l,l+1 \end{array}}^{2l}\left( {\begin{array}{c}k\\ 2l+1\end{array}}\right) c_{1}^{2l+1-k}c_{2}^{k} \mathrm{e}^{\frac{2l+1-2k}{2l+1}\lambda z}\) contains at least two terms. By Lemma 2.3, we have \(\left( {\begin{array}{c}k\\ 2l+1\end{array}}\right) c_{1}^{2l+1-k}c_{2}^{k}=0,~k\ne l,l+1,~k=1,\ldots ,2l\). Then, \(c_{1}c_{2}=0\), a contradiction.
Case 2 As in the beginning of the proof of Theorem 1.2 below, we obtain \(\rho (f)=1\). If \(\varphi \not \equiv 0\), since f is a transcendental entire function with order \(\rho (f)=1\), we see that (3.5) satisfies the conditions of Lemma 2.2 and Remark 2.1. Thus, we have
which implies that \(\lambda ^{2}f-n^{2}f''\) is a polynomial. Denote
where H is a non-zero polynomial. By Lemma 2.6, we see that (3.12) must have a non-zero polynomial solution, say, \(c_{0}(z)\). Since the differential equation
has two fundamental solutions
the general entire solution \(f(\not \equiv 0)\) of (3.12) can be expressed as:
where \(c_{1},c_{2}\) are non-zero constants, \(c_{0}(z)\) is a non-zero polynomial. Otherwise, if one of \(c_{1},c_{2}\) is equal to zero, substituting (3.13) into (1.2) and using Lemma 2.3, we obtain a contradiction.
When \(n=2\), then \(f(z)=c_{0}(z)+c_{1}\mathrm{e}^{\frac{\lambda }{2}z}+c_{2}\mathrm{e}^{-\frac{\lambda }{2}z}\). Substituting this into (1.2) yields
It follows from (3.14) and Lemma 2.3 that
Note that \(c_{1},c_{2}\in {\mathbb {C}}{\setminus }\{0\}\), then \(q(z)(\mathrm{e}^{\frac{\lambda c}{2}}-1)\equiv q(z)(\mathrm{e}^{-\frac{\lambda c}{2}}-1)\equiv -2c_{0}(z)\not \equiv 0\), that is \(q(z)\equiv c_{0}(z)\) and \(\lambda c=2k\pi i\), \(k\in {\mathbb {Z}}{\setminus }\{0\}\) and k is odd. Substituting \(q(z)\equiv c_{0}(z)\) into \(c_{0}^{2}(z)+q(z)[c_{0}(z+c)-c_{0}(z)]+2c_{1}c_{2}\equiv 0\) yields \(c_{0}(z+c)c_{0}(z)+2c_{1}c_{2}\equiv 0\). Then, \(c_{0}(z)\) must be a non-zero constant. Therefore, \(c_{0}^{4}=q^{4}=4c_{1}^{2}c_{2}^{2}=4p_{1}p_{2}\).
When \(n\ge 3\), then \(2n-2\ge n+1\), it follows from (3.5) that \(f^{2n-2}(f\varphi )=Q(z,f)\). By Lemma 2.2 and Remark 2.1, we have \(m(r,f\varphi )=S(r,f)\). Therefore, \(T(r,f\varphi )=S(r,f)\). Since \(\varphi \not \equiv 0\), then \(T(r,f)=m(r,f)\le m(r,f\varphi )+m\left( r,\frac{1}{\varphi }\right) \le T(r,\varphi )+S(r,f) =S(r,f)\), which gives a contradiction.
This completes the proof of Theorem 1.1.
4 Proof of Theorem 1.2
Clearly, \(\rho (p_{1}\mathrm{e}^{\alpha _{1}z}+p_{2}\mathrm{e}^{\alpha _{2}z})=1\), where \(\alpha _{1}\pm \alpha _{2}\ne 0\). From (1.4) and Lemma 2.5, we have
and
i.e.
thus, \(\rho (f)=1\). Denote \(P=q\Delta _{c}f\). Suppose that f is a transcendental entire solution of finite order of equation (1.4). By differentiating (1.4), we have
Eliminating \(\mathrm{e}^{\alpha _{2}z}\) from (1.4) and (4.1), we have
Differentiating (4.2) yields
It follows from (4.2) and (4.3) that
where
and
We distinguish two cases below:
Case 1 If \(\varphi \equiv 0\), then \(Q(z,f)\equiv 0\). Since \(\alpha _{1}\ne \alpha _{2}\), we see that \(\alpha _{1}P-P'\equiv 0\) and \(\alpha _{2}P-P'\equiv 0\) cannot hold simultaneously. Suppose that \(\alpha _{2}P-P'\not \equiv 0\). By (4.6), we have
where A is a non-zero constant. Substituting (4.7) into (4.2), we have
Since the right-hand side of (4.8) is a differential-difference polynomial in f of degree at most 1, by Lemma 2.2 and Remark 2.1, we have
Denote \(\psi =\alpha _{2}f-2f'\). We consider two subcases as follows.
Subcase 1.1 If \(\psi \equiv 0\), that is \(\alpha _{2}f-2f'\equiv 0\), then \(f^{2}={\widetilde{p}}_{2}\mathrm{e}^{\alpha _{2}z},~{\widetilde{p}}_{2}\in {\mathbb {C}}{\setminus }\{0\}\). Substituting this and (4.7) into (1.4) yields
If \(p_{2}\ne {\widetilde{p}}_{2}\), by Lemma 2.2 and Remark 2.1, we have \(T(r,f)=m(r,f)=S(r,f)\), a contradiction. Therefore, \(p_{2}={\widetilde{p}}_{2}\), \(f(z)=c_{2}\mathrm{e}^{\frac{\alpha _{2}}{2}z}\), \(c_{2}(\mathrm{e}^{\frac{\alpha _{2}}{2}c}-1)q=p_{1}\), \(c_{2}^{2}=p_{2}\), \(\alpha _{2}=2\alpha _{1}\).
Subcase 1.2 If \(\psi \not \equiv 0\), then \(\psi '=\alpha _{2}f'-2f''\), \(f'=\frac{\alpha _{2}}{2}f-\frac{\psi }{2}\), \(f''=\frac{\alpha _{2}^{2}}{4}f-\frac{\alpha _{2}}{4}\psi -\frac{\psi '}{2}\). Note that \(\varphi \equiv 0\) and substitute this into (4.5) gives
Since \(\psi \not \equiv 0\), then \(\left( \alpha _{1}-\frac{\alpha _{2}}{2}\right) \psi -\psi '\not \equiv 0\). From the above equality, we have \(T(r,f)=S(r,f)\), which implies a contradiction.
Similarly, if \(\alpha _{1}P-P'\not \equiv 0\), then we obtain \(f(z)=c_{1}\mathrm{e}^{\frac{\alpha _{1}}{2}z}\), \(c_{1}(\mathrm{e}^{\frac{\alpha _{1}}{2}c}-1)q=p_{2}\), \(c_{1}^{2}=p_{1}\), \(\alpha _{1}=2\alpha _{2}\).
Case 2 If \(\varphi \not \equiv 0\), by applying the lemma of logarithmic derivative and Lemma 2.2, from (4.4)–(4.6), we have
Then
Suppose that there exist \(\kappa ,\iota \ge 0\) such that
where \(0\le \kappa ,~\iota \le 1\). It follows from (4.9) that
and
then we have \(1-\frac{\kappa }{2}\le \iota \le 1\) and \(0\le \kappa \le 1\).
Next, we deduce that \(\kappa \ne 0\).
If \(\kappa =0\), then \(T(r,\varphi )=S(r,f)\). It follows from (4.9) that
By (4.5), if \(z_{0}\) is a multiple zero of f, then \(z_{0}\) must be a zero of \(\varphi \). Hence, \(N_{(2}\left( r,\frac{1}{f}\right) =S(r,f)\). Differentiating (4.5) gives
If \(z_{0}\) is a simple zero of f, it follows from (4.5) and (4.10) that \(z_{0}\) is a zero of \(3\varphi f''-[\varphi '+(\alpha _{1}+\alpha _{2})\varphi ]f'\). Define
then we have \(T(r,\alpha )=S(r,f)\). It follows that
Substituting (4.12) into (4.5) yields
where \(a=\alpha _{1}\alpha _{2}+\frac{2\alpha }{3\varphi }\), \(b=\frac{2}{3}\left[ \frac{\varphi '}{\varphi }-2(\alpha _{1}+\alpha _{2})\right] \). By Lemma 2.4, we have
Now, we distinguish two subcases below.
Subcase 2.1 Suppose that \(b^{2}-8a\not \equiv 0\). It follows from (4.14) that
By integration, we see that there exists a \(B\in {\mathbb {C}}{\setminus }\{0\}\) such that
which implies \(\mathrm{e}^{2(\alpha _{1}+\alpha _{2})z}\in S(f)\), then \(\alpha _{1}+\alpha _{2}=0\), a contradiction.
Subcase 2.2 Suppose that \(b^{2}-8a\equiv 0\). Differentiating (4.13) yields
Suppose \(z_{0}\) is a simple zero of f which is not the zero of a, b. It follows from (4.13) and (4.17) that \(z_{0}\) is a zero of \(2\varphi f''-\left( \varphi '-\frac{b}{2}\varphi \right) f'\). Putting
we have \(T(r,\beta )=S(r,f)\). It follows that
Substituting (4.19) into (4.17) yields
where \(c=a'+\frac{b\beta }{2\varphi }\), \(d=2a+b'+\frac{b}{2}\frac{\varphi '}{\varphi }-\frac{b^{2}}{4}+\frac{2\beta }{\varphi }\). Eliminating \((f')^{2}\) from (4.13) and (4.20), we have
where
Note that A(z) and B(z) are small functions of f. If \(z_{0}\) is a simple zero of f and not the zero of B(z), it follows from (4.21) that \(A(z)=B(z)\equiv 0\). By (4.19), we have
where \(b=\frac{2}{3}\left[ \frac{\varphi '}{\varphi }-2(\alpha _{1}+\alpha _{2})\right] \not \equiv 0\). Otherwise, \(e^{2(\alpha _{1}+\alpha _{2})z}=C\varphi \in S(f)\), then \(\alpha _{1}+\alpha _{2}=0\), a contradiction. Substituting \(b^{2}-8a\equiv 0\) into (4.22) yields
It follows from (4.12) and (4.23) that
We deduce that \(\varphi '\not \equiv 0\). Otherwise, we assume that \(\varphi '\equiv 0\), then \(\frac{\alpha }{\varphi }\equiv 0\). Substituting this into \(b^{2}-8a\equiv 0\) yields
which implies that \(\frac{\alpha _{1}}{\alpha _{2}}=2\) or \(\frac{\alpha _{1}}{\alpha _{2}}=\frac{1}{2}\). By substituting \(\varphi '\equiv 0\) into (4.23), we obtain
then
where \(C_{1},C_{2}\in {\mathbb {C}}{\setminus }\{0\}\). Without loss of generality, when \(\frac{\alpha _{1}}{\alpha _{2}}=2\), then
Substituting this into (1.4) and using Lemma 2.3, we can obtain a contradiction.
Differentiating (4.24) gives
It follows from \(b^{2}-8a\equiv 0\) that \(bb'=4a'\), that is
Putting \(\gamma :=\frac{\varphi '}{\varphi }\) and combining the above two equality yields
If \(\gamma '\equiv 0\), then \(\varphi =C_{3}e^{C_{4}z},~C_{3},C_{4}\in {\mathbb {C}}\). It follows from \(\varphi '\not \equiv 0\) that \(C_{3},C_{4}\ne 0\), which implies that \(\varphi \not \in S(f)\), a contradiction. If \(\gamma '\not \equiv 0\), it follows from (4.25) that
which implies that \(e^{(\alpha _{1}+\alpha _{2})z}\in S(f)\), then \(\alpha _{1}+\alpha _{2}=0\), a contradiction.
This completes the proof of Theorem 1.2.
5 Proof of Theorem 1.3
Suppose that f is a transcendental entire solution of finite order of Eq. (1.7). In what follows, we consider three cases.
Case 1 If \(\rho (f)<1\), using Lemma 2.5, from (1.7) we have
then \(\deg Q\le 1\), note that \(\deg Q\ge 1\), therefore \(\deg Q=1\). Denote \(Q(z)=az+b,~a\in {\mathbb {C}}{\setminus }\{0\},~b\in {\mathbb {C}}\). Rewriting (1.7) in the following form:
and differentiating (5.1) we get
where
Eliminating \(\mathrm{e}^{\lambda z}\) and \(\mathrm{e}^{-\lambda z}\) from (5.1) and (5.2) yields
where
Thus, from Lemma 2.3, we have
It follows from \(D(z)\equiv 0\) that
Using Lemma 2.1, we have
and
We deduce that \(\lambda ^{2}f^{2}(z)-n^{2}(f'(z))^{2}\not \equiv 0\). Otherwise, \(4\lambda ^{2}p_{1}p_{2}=0\), a contradiction. Since f is entire,
a contradiction.
Case 2 If \(\rho (f)>1\). Denote \(P(z)=p_{1}\mathrm{e}^{\lambda z}+p_{2}\mathrm{e}^{-\lambda z}\), \(H(z)=q(z)f(z+c)\). It is clear that \(\rho (P)=1\), then \(T(r,P)=S(r,f)\). Equation (1.7) can be written as:
Differentiating (5.5) yields
where \(L(z)=H'(z)+Q'(z)H(z)\). Eliminating \(\mathrm{e}^{Q(z)}\) from (5.5) and (5.6), we have
Note that \(n-1\ge 2\) and \(P(z)L(z)-P'(z)H(z)\) is a differential-difference polynomial in f and the total degree is at most 1. By Lemma 2.2 and Remark 2.1, we obtain
and
If \(L(z)f(z)-nH(z)f'(z)\not \equiv 0\), then
which yields a contradiction. If \(L(z)f(z)-nH(z)f'(z)\equiv 0\), then
By integration, we see that there exists a \(C\in {\mathbb {C}}{\setminus }\{0\}\) such that
Substituting (5.8) into (1.7) gives
If \(1+C\ne 0\), we have \(T(r,f)=m(r,f)=S(r,f)\), a contradiction. If \(1+C=0\), then \(P(z)=p_{1}\mathrm{e}^{\lambda z}+p_{2}\mathrm{e}^{-\lambda z}\equiv 0\), a contradiction.
Case 3 If \(\rho (f)=1\), by applying Lemma 2.5, from (1.7) we obtain
Note that \(\deg Q\ge 1\), then \(1\le \deg Q=\sigma (\mathrm{e}^{Q(z)})\le \max \{\rho (\mathrm{e}^{\lambda z}),\rho (f)\}=1\), that is \(\rho (f)=\deg Q=1\).
If f belongs to \(\Gamma _{0}\), and noting that \(\rho (f)=\deg Q=1\), we define \(f=\mathrm{e}^{Az+B}\) and \(Q(z)=az+b\), where \(a,A\in {\mathbb {C}}{\setminus }\{0\}\) and \(b,B\in {\mathbb {C}}\). Substituting these into (1.7) yields
We distinguish four cases below.
Case 1 If \(nA+\lambda =0\) and \(A+a+\lambda =0\). By Lemma 2.3 and (5.9), we obtain \(p_{1}=0\), a contradiction.
Case 2 If \(nA+\lambda =0\) and \(A+a+\lambda \ne 0\). If \(A+a+\lambda \ne 2\lambda \), by Lemma 2.3 and (5.9), we have \(p_{1}=q(z)\equiv 0\), a contradiction. If \(A+a+\lambda =2\lambda \), then \(A=-\frac{\lambda }{n},~a=\frac{n+1}{n}\lambda \), substituting these into (5.9) yields
by Lemma 2.3, we have \(p_{2}-\mathrm{e}^{nB}\equiv q(z)\mathrm{e}^{Ac+b+B}-p_{1}\equiv 0\), then q(z) reduces to a non-zero constant, and \(f(z)=\mathrm{e}^{-\frac{\lambda }{n}z+B},~Q(z)=\frac{n+1}{n}\lambda z+b\).
Case 3 If \(nA+\lambda \ne 0\) and \(A+a+\lambda =0\). If \(nA+\lambda \ne 2\lambda \), it follows from Lemma 2.3 that \(p_{1}=0\), a contradiction. If \(nA+\lambda =2\lambda \), then \(A=\frac{\lambda }{n},~a=-\frac{n+1}{n}\lambda \). Substituting these into (5.9) yields
by Lemma 2.3, we have \(\mathrm{e}^{nB}-p_{1}\equiv p_{2}-q(z)\mathrm{e}^{Ac+b+B}\equiv 0\), then q(z) reduces to a non-zero constant, and \(f(z)=\mathrm{e}^{\frac{\lambda }{n}z+B},~Q(z)=-\frac{n+1}{n}\lambda z+b\).
Case 4 If \(nA+\lambda \ne 0\) and \(A+a+\lambda \ne 0\). If \(nA+\lambda \), \(A+a+\lambda \) and \(2\lambda \) are pairwise distinct from each other, by Lemma 2.3 and (5.9), we have \(p_{1}=p_{2}\equiv q(z)\equiv 0\), a contradiction. If only two of \(nA+\lambda \), \(A+a+\lambda \) and \(2\lambda \) coincide, without loss of generality, suppose that \(nA+\lambda =A+a+\lambda \ne 2\lambda \), then (5.9) can be written as:
From the above equality and using Lemma 2.3, we have \(p_{1}=p_{2}=0\), which implies a contradiction. If \(nA+\lambda =A+a+\lambda =2\lambda \), then we write (5.9) as:
It follows from Lemma 2.3 that \(p_{2}=0\), a contradiction.
This completes the proof of Theorem 1.3.
References
Clunie, J.: On integral and meromorphic functions. J. Lond. Math. Soc. 37, 17–27 (1962)
Chiang, Y.M., Feng, S.J.: On the Nevanlinna characteristic of \(f(z+\eta )\) and difference equations in the complex plane. Raman. J. 16, 105–129 (2008)
Chen, Z.X., Yang, C.C.: On entire solutions of certain type of differential-difference equations. Taiwan. J. Math. 18, 677–685 (2014)
Chen, M.F., Gao, Z.S.: Entire solutions of certain type of non-linear differential equations and differential-difference equations. J. Comput. Anal. Appl. 24, 137–147 (2018)
Hayman, W.K.: Meromorphic Function. Clarendon Press, Oxford (1964)
Halburd, R.G., Korhonen, R.J.: Difference analogue of the lemma on the logarithmic derivative with applications to difference equations. J. Math. Anal. Appl. 314, 477–487 (2006)
Laine, I.: Nevanlinna Theory and Complex Differential Equations. Walter. de Gruyter, Berlin (1993)
Laine, I., Yang, C.C.: Entire solutions of some non-linear differential equations. Bull. Soc. Sci. Lett. Łódź. Sér. Rech. Déform. 59, 19–23 (2009)
Li, B.Q.: On certain non-linear differential equations in complex domains. Arch. Math. 91, 344–353 (2008)
Li, P.: Entire solutions of certain type of differential equations. J. Math. Anal. Appl. 344, 253–259 (2008)
Li, P.: Entire solutions of certain type of differential equations II. J. Math. Anal. Appl. 375, 310–319 (2011)
Liao, L.W., Yang, C.C., Zhang, J.J.: On meromorphic solutions of certain type of non-linear differential equations. Ann. Acad. Sci. Fenn. Math. 38, 581–593 (2013)
Wen, Z.T., Heittokangas, J., Laine, I.: Exponential polynomials as solutions of certain non-linear difference equations. Acta Math. Sin. 28, 1295–1306 (2012)
Yang, L.: Value Distribution Theory. Science Press, Beijing (1993)
Yang, C.C., Yi, H.X.: Uniqueness Theory of Meromorphic Functions. Science Press, Kluwer Academic, Beijing, Dordrecht (2003)
Yang, C.C., Laine, I.: On analogies between non-linear difference and differential equations. Proc. Jpn. Acad. Ser. A Math. Sci. 86, 10–14 (2010)
Zinelâabidine, L.: On the existence of entire solutions of certain class of non-linear difference equations. Mediterr. J. Math. 14, 1–16 (2017). https://doi.org/10.1007/s00009-017-0914-x
Acknowledgements
The authors would like to thank the referee for his/her thorough review with constructive suggestions and comments on the paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Risto Korhonen.
This research was supported by the National Natural Science Foundation of China (No: 11371225) and by the Fundamental Research Funds for the Central Universities.
Rights and permissions
About this article
Cite this article
Chen, MF., Gao, ZS. & Zhang, JL. Entire Solutions of Certain Type of Non-Linear Difference Equations. Comput. Methods Funct. Theory 19, 17–36 (2019). https://doi.org/10.1007/s40315-018-0250-6
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40315-018-0250-6