 Research
 Open Access
 Published:
Entire solutions of certain type of difference equations
Journal of Inequalities and Applications volume 2014, Article number: 63 (2014)
Abstract
In this paper, we shall study the conditions regarding the existence of transcendental entire solutions of certain type of difference equations. Our results are either supplements to some results obtained recently, or are relating to the conjecture raised in Yang and Laine (Proc. Jpn. Acad., Ser. A, Math. Sci. 86:1014, 2010). Finally, two relevant conjectures are posed for further studies.
MSC:39B32, 34M05, 30D35.
1 Introduction, notations and main results
Let f denote a transcendental entire or meromorphic function. Assuming the reader is familiar with the basics of Nevanlinna’s value distribution theory, we shall adopt the standard notations associated with the theory, such as the characteristic function T(r,f), the counting function of the poles N(r,f), and the proximity function m(r,f) (see, e.g., [1, 2]).
Among many interesting applications of the Nevanlinna theory, there are studies on the growth and existence of entire or meromorphic solutions of various types of nonlinear differential equations, and one can find prototypes for such equations, e.g., in [3–5] and [6]. Recently, the Nevanlinna theory has been applied to study types of nonlinear difference equations, see, e.g., [7, 8]. Now, we shall utilize Clunie type of theorems for differencedifferential polynomials to study some nonlinear difference equations of more general forms and to obtain some improvements of or supplements to [7, 9], and [10].
Notations Given a meromorphic function f, recall that \alpha \not\equiv 0,\mathrm{\infty} is a small function with respect to f, if T(r,\alpha )=S(r,f), where S(r,f) denotes any quantity satisfying S(r,f)=o\{T(r,f)\} as r\to +\mathrm{\infty}, possibly outside a set of r of finite linear measure. For a constant c\in \mathbb{C}, f(z+c) is called a shift of f. As for a difference product, we mean a difference monomial of type {\prod}_{j=1}^{k}f{(z+{c}_{j})}^{{n}_{j}}, where {c}_{1},\dots ,{c}_{k} are complex constants and {n}_{1},\dots ,{n}_{k} are natural numbers.
Definition 1.1 A difference polynomial, respectively, a differencedifferential polynomial, in f is a finite sum of difference products of f and its shifts, respectively, of products of f, derivatives of f and of their shifts, with all the coefficients of these monomials being small functions of f.
Consider a transcendental meromorphic function f(z) and let
where {a}_{l}(z) (l=1,2,\dots ,n) are small functions of f, and {c}_{lj} (l=1,2,\dots ,n; j=0,1,\dots ,k) are complex constants, and {n}_{lj} (l=1,2,\dots ,n; j=0,1,\dots ,k) are natural numbers.
Two terms {a}_{l}(z){\prod}_{j=0}^{k}{\{{f}^{(j)}(z+{c}_{lj})\}}^{{n}_{lj}} and {a}_{{l}^{\prime}}(z){\prod}_{j=0}^{k}{\{{f}^{(j)}(z+{c}_{lj})\}}^{{n}_{{l}^{\prime}j}} are called similar, if
Group together similar terms of P(z,f), if necessary. In the following, we assume that no two terms of P(z,f) are similar and that {a}_{l}(z)\not\equiv 0 (l=1,2,\dots ,n).
Definition 1.2 We define the total degree d of P(z,f)
For the sake of simplicity, we let \mathrm{\Delta}f(z)=f(z+1)f(z), {\mathrm{\Delta}}^{n}f(z)=\mathrm{\Delta}({\mathrm{\Delta}}^{n1}f(z)) (n\ge 2).
Yang and Laine [10] considered the following difference equation and proved it.
Theorem A A nonlinear difference equation
where q(z) is a nonconstant polynomial and b,c\in \mathbb{C} are nonzero constants, does not admit entire solutions of finite order. If q(z)=q is a nonzero constant, then the above equation possesses three distinct entire solutions of finite order, provided that b=3n\pi and {q}^{3}={(1)}^{n+1}{c}^{2}27/4 for a nonzero integer n.
Now, we shall substitute f(z+1) by \mathrm{\Delta}f(z) in Theorem A and prove the following results.
Theorem 1.1 Let n\ge 4 be an integer, q(z) be a polynomial, and {p}_{1}, {p}_{2}, {\alpha}_{1}, {\alpha}_{2} be nonzero constants such that {\alpha}_{1}\ne {\alpha}_{2}. If there exists some entire solution f of finite order to (1.1) below
then q(z) is a constant, and one of the following relations holds:

(1)
f(z)={c}_{1}{\mathrm{e}}^{\frac{{\alpha}_{1}}{n}z}, and {c}_{1}({\mathrm{e}}^{{\alpha}_{1}/n}1)q={p}_{2}, {\alpha}_{1}=n{\alpha}_{2},

(2)
f(z)={c}_{2}{\mathrm{e}}^{\frac{{\alpha}_{2}}{n}z}, and {c}_{2}({\mathrm{e}}^{{\alpha}_{2}/n}1)q={p}_{1}, {\alpha}_{2}=n{\alpha}_{1},
where {c}_{1}, {c}_{2} are constants satisfying {c}_{1}^{n}={p}_{1}, {c}_{2}^{n}={p}_{2}.
Corollary 1.1 Let n\ge 4 be an integer, q(z)\not\equiv 0 be a polynomial, and p, α be nonzero constants. Then the nonlinear difference equation
has no transcendental entire solutions of finite order provided that \frac{\alpha}{n}\ne 2k\pi \mathrm{i}, where k is an integer.
By some further analysis, we can derive the following result.
Theorem 1.2 A nonlinear difference equation
where q(z) is a nonconstant polynomial and b,c\in \mathbb{C} are nonzero constants, does not admit entire solutions of finite order. If q(z)=q is a nonzero constant, then (1.2) possesses solutions of the form f(z)={c}_{1}{\mathrm{e}}^{\frac{b\mathrm{i}}{3}z}+{c}_{2}{\mathrm{e}}^{\frac{b\mathrm{i}}{3}z}, {c}_{1}^{3}={c}_{2}^{3}=\frac{c}{2\mathrm{i}}, provided that b=3n\pi, n is an odd number, {q}^{3}=\frac{27}{2,048}{c}^{2} or b=6n\pi \pm \pi, {q}^{3}=\frac{27}{4}{c}^{2}.
Examples In the special case of
a finite order entire solution is f(z)=2\mathrm{i}sin\pi z={\mathrm{e}}^{\mathrm{i}\pi z}{\mathrm{e}}^{\mathrm{i}\pi z}.
And {f}^{3}(z)+3{\mathrm{\Delta}}^{3}f(z)=2\mathrm{i}sin7\pi z has the entire solution f(z)={\mathrm{e}}^{\frac{7\pi}{3}\mathrm{i}z}{\mathrm{e}}^{\frac{7\pi}{3}\mathrm{i}z}.
Corollary 1.2 The nonlinear difference equation
where q is a nonzero constant, possesses solutions of the form f={c}_{1}{\mathrm{e}}^{\frac{b\mathrm{i}}{3}z}+{c}_{2}{\mathrm{e}}^{\frac{b\mathrm{i}}{3}z}, if and only if n=3.
Theorem B ([10], Theorem 2.4)
Let p, q be polynomials, then a nonlinear difference equation
has no transcendental entire solutions of finite order.
We shall modify the equation in Theorem B above and derive the following result.
Theorem 1.3 Let l\ge 2, n\ge 1 be integers, a(z) and b(z) be meromorphic functions such that T(r,a)={\lambda}_{1}T(r,f)+S(r,f), T(r,b)={\lambda}_{2}T(r,f)+S(r,f), where nonnegative numbers {\lambda}_{1}, {\lambda}_{2} satisfy {\lambda}_{1}+{\lambda}_{2}<1. Then
has no transcendental entire solutions of hyperorder {\sigma}_{2}(f)<1.
2 Preliminaries
In order to prove our conclusions, we need some lemmas.
Lemma 2.1 ([7])
Let f be a transcendental meromorphic solution of finite order ρ of a difference equation of the form
where H(z,f), P(z,f), Q(z,f) are difference polynomials in f such that the total degree of H(z,f) in f and its shifts is n, and that the corresponding total degree of Q(z,f) is ≤n. If H(z,f) contains just one term of maximal total degree, then for any \epsilon >0,
possibly outside of an exceptional set of finite logarithmic measure.
Remark 2.1 The following result is a Clunie type lemma [11] for the differencedifferential polynomials of a meromorphic function f. It can be proved by applying Lemma 2.1 with a similar reasoning as in [10] and stated as follows.
Let f(z) be a meromorphic function of finite order, and let P(z,f), Q(z,f) be two differencedifferential polynomials of f. If {f}^{n}P(z,f)=Q(z,f) holds and if the total degree of Q(z,f) in f and its derivatives and their shifts is ≤n, then m(r,P(z,f))=S(r,f).
Lemma 2.2 ([12])
Suppose that m, n are positive integers satisfying \frac{1}{m}+\frac{1}{n}<1. Then there exist no transcendental entire solutions f and g satisfying the equation a{f}^{n}+b{g}^{m}=1, with a, b being small functions of f and g, respectively.
Lemma 2.3 ([13])
Assume that c\in \mathbb{C} is a nonzero constant, α is a nonconstant meromorphic function. Then the differential equation {f}^{2}+{(c{f}^{(n)})}^{2}=\alpha has no transcendental meromorphic solutions satisfying T(r,\alpha )=S(r,f).
Let f be a transcendental meromorphic function of finite order ρ, then for any complex numbers {c}_{1}, {c}_{2} and for each \epsilon >0, m(r,\frac{f(z+{c}_{1})}{f(z+{c}_{2})})=O({r}^{\rho 1+\epsilon}).
References [16] and [17] further pointed out the following.
Remark 2.2 If f is a nonconstant finite order meromorphic function and c\in \mathbb{C}, then
outside of a possible exceptional set with finite logarithmic measure.
We also know that Remark 2.2 has been improved by Halburd et al. [18]. They proved that (2.2) is also true when f is a meromorphic function of hyperorder {\sigma}_{2}(f)<1.
3 Proof of Theorem 1.1
Suppose that f is a transcendental entire solution of finite order to (1.1). By differentiating both sides of (1.1), we have
From (1.1) and (3.1), we obtain
Differentiating (3.2) yields
It follows from (3.2) and (3.4) that
where
Next, we shall prove \phi \equiv 0. In fact, since T is a differencedifferential polynomial in f, and its degree at most 1. By (3.5) and Remark 2.1 after Lemma 2.1, we have m(r,\phi )=S(r,f), and T(r,\phi )=S(r,f). On the other hand, we can rewrite (3.5) as {f}^{n3}f\phi =T(f), which implies m(r,f\phi )=S(r,f), and T(r,f\phi )=S(r,f). If \phi \not\equiv 0, then T(r,f)=T(r,\frac{f\phi}{\phi})=S(r,f), and this is impossible. Hence \phi \equiv 0, and T\equiv 0, i.e.
If q(z)\mathrm{\Delta}f(z)\equiv 0, then (1.1) can be rewritten as
which is impossible by Lemma 2.2. Thus, q(z)\mathrm{\Delta}f(z)\not\equiv 0. It is easily seen from {\alpha}_{1}\ne {\alpha}_{2} that
and
cannot hold simultaneously.
First of all, we assume that {\alpha}_{2}q(z)\mathrm{\Delta}f(z){(q(z)\mathrm{\Delta}f(z))}^{\prime}\not\equiv 0, then (3.6) gives
where A is a nonzero constant. Substituting the above expression into (3.2), we obtain
Again, Lemma 2.1 shows that {\alpha}_{2}fn{f}^{\prime}\equiv 0, ({\alpha}_{2}{\alpha}_{1}){p}_{1}=A. So
where B is a constant.
Substituting (3.7) into (1.1), we find
If B\ne {p}_{2}, by Lemma 2.1, we get T(r,f)=m(r,f)=S(r,f), which is absurd. So B={p}_{2}, and f={c}_{2}{\mathrm{e}}^{\frac{{\alpha}_{2}}{n}z}={c}_{2}{\mathrm{e}}^{{\alpha}_{1}z}, {c}_{2}^{n}=B={p}_{2}, {c}_{2}({\mathrm{e}}^{{\alpha}_{2}/n}1)q={p}_{1}.
If {\alpha}_{1}q(z)\mathrm{\Delta}f(z){(q(z)\mathrm{\Delta}f(z))}^{\prime}\not\equiv 0, by (3.3) and using similar arguments as above, we can derive f={c}_{1}{\mathrm{e}}^{\frac{{\alpha}_{1}}{n}z}={c}_{1}{\mathrm{e}}^{{\alpha}_{2}z}, {c}_{1}^{n}={p}_{1}, {c}_{1}({\mathrm{e}}^{{\alpha}_{1}/n}1)q={p}_{2}.
This completes the proof of Theorem 1.1.
Remark 3.1 Suggested by the referee, one can also derive the conclusions of Theorem 1.1 when \phi \equiv 0.
In fact, since φ in (3.5) vanishes identically, dividing with {f}^{2}, and recalling {f}^{\u2033}/f={({f}^{\prime}/f)}^{\prime}+{({f}^{\prime}/f)}^{2}, we get the Riccati equation
where t:={f}^{\prime}/f. This equation has two constant solutions, t={\alpha}_{1}/n, t={\alpha}_{2}/n. By Corollary 5.2 in the paper by Bank et al. [19], all other meromorphic solutions are of infinite order. And from this one can obtain the two entire solutions of exponential type.
4 Proof of Theorem 1.2
Suppose that f is a transcendental entire solution of finite order to (1.2), differentiating (1.2) results in
Combining (1.2) and (4.1), we obtain
where {T}_{4}(f) is a differencedifferential polynomial of f, and its total degree at most 4.
If {T}_{4}(f)\equiv 0, it follows from (4.2) that 9{({f}^{\prime})}^{2}+{b}^{2}{f}^{2}\equiv 0, then {f}^{\prime}=\pm \frac{b\mathrm{i}}{3}f, and there exists a nonzero constant c such that f=c{\mathrm{e}}^{\pm \frac{b\mathrm{i}}{3}z}. Substituting the expression of f into (1.2), we arrive at a contradiction. Therefore, {T}_{4}(f)\not\equiv 0. Set \beta =9{({f}^{\prime})}^{2}+{b}^{2}{f}^{2}, then Lemma 2.1 shows that m(r,\beta )=S(r,f), i.e. β is a small function of f. Moreover, from Lemma 2.3, it is easy to see that β is a constant. By differentiating both sides of \beta =9{({f}^{\prime})}^{2}+{b}^{2}{f}^{2}, we get
It follows from (4.3) that
where {c}_{1}, {c}_{2} are constants. Substituting (4.4) into (1.2), we have
where w={\mathrm{e}}^{\frac{b\mathrm{i}}{3}z}, {a}_{1}={c}_{1}^{3}\frac{c}{2\mathrm{i}}, {a}_{2}=3{c}_{1}^{2}{c}_{2}+{c}_{1}q(z){\mathrm{e}}^{b\mathrm{i}}3q(z){c}_{1}{\mathrm{e}}^{\frac{2b\mathrm{i}}{3}}+3q(z){c}_{1}{\mathrm{e}}^{\frac{b\mathrm{i}}{3}}{c}_{1}q(z), {a}_{3}=3{c}_{1}{c}_{2}^{2}+{c}_{2}q(z){\mathrm{e}}^{b\mathrm{i}}3q(z){c}_{2}{\mathrm{e}}^{\frac{2b\mathrm{i}}{3}}+3q(z){c}_{2}{\mathrm{e}}^{\frac{b\mathrm{i}}{3}}{c}_{2}q(z), {a}_{4}={c}_{2}^{3}+\frac{c}{2\mathrm{i}}.
Since w is transcendental, we must have {a}_{1}={a}_{2}={a}_{3}={a}_{4}=0. It follows from {a}_{2}={a}_{3}=0 that
Let v={\mathrm{e}}^{\frac{b\mathrm{i}}{3}}, then from (4.5) we get
where {u}_{1}=1, {u}_{2}=\frac{1+\sqrt{3}\mathrm{i}}{2}, {u}_{3}=\frac{1\sqrt{3}\mathrm{i}}{2}.
Now we distinguish three cases to discuss.
Case 1. Let {u}_{k}=1, we conclude {\mathrm{e}}^{\frac{2b\mathrm{i}}{3}}=1={\mathrm{e}}^{2n\pi \mathrm{i}}, so b=3n\pi, which and {a}_{2}=0 imply that
and 3{c}_{1}{c}_{2}=4q(z)(1{(1)}^{n}), where {c}_{1}\ne 0, {c}_{2}\ne 0 are constants. Therefore n is odd and {q}^{3}=\frac{27}{2,048}{c}^{2}.
Case 2. Assume that {u}_{k}=\frac{1+\sqrt{3}\mathrm{i}}{2}; by a similar method as Case 1, we get b=6n\pi or b=6n\pi \pi.
If b=6n\pi, combining with {a}_{2}=0, we find {c}_{1}{c}_{2}=0, which is a contradiction.
If b=6n\pi \pi, substituting this expression of b into {a}_{2}=0, we have {q}^{3}=\frac{27}{4}{c}^{2}.
Case 3. While {u}_{k}=\frac{1\sqrt{3}\mathrm{i}}{2}, using a similar way as above, we get b=6n\pi or b=6n\pi +\pi. If b=6n\pi, we can get a contradiction, hence b=6n\pi +\pi, and {q}^{3}=\frac{27}{4}{c}^{2}.
This completes the proof of Theorem 1.2.
5 Proof of Theorem 1.3
Suppose that f is a transcendental entire solutions of hyperorder {\sigma}_{2}(f)<1 to (1.3). Without loss of generality, we assume that a\not\equiv 0. Remark 2.2 after Lemma 2.4 and (1.3) will give
So
For l\ge 2, {\lambda}_{1}+{\lambda}_{2}<1, we get a contradiction, thus (1.3) has no transcendental entire solutions of hyperorder {\sigma}_{2}(f)<1.
This also completes the proof of Theorem 1.3.
6 Final remark and conjectures
The current Clunie types of theorems regarding differencedifferential polynomials are mainly useful or effective to deal with problems relating to entire or meromorphic solutions of finite order for certain types of differencedifferential equations. Thus, it is very natural for us to pose the following two conjectures, for further studies.
Conjecture 6.1 There are no entire solutions of infinite order for any equations (1.1), (1.2), and (1.3).
Conjecture 6.2 Equations of the forms
have no entire solutions of infinite order, where P(z,f(z),f(z+{c}_{1}),\dots ,f(z+{c}_{k})) is a difference polynomial, {q}_{1}, {q}_{2} are polynomials, n\ge 2 and k\ge 1 are integers, and {\alpha}_{1}, {\alpha}_{2}, {c}_{j} (j=1,2,\dots ,k) are constants.
References
Hayman WK: Meromorphic Functions. Clarendon, Oxford; 1964.
Yi HX, Yang CC: Uniqueness Theory of Meromorphic Functions. Science Press, Beijing; 1995.
Yosida K: On algebroid solutions of ordinary equations. Jpn. J. Math. 1934, 10: 199–208.
Wittich H:Eindeutige Lösungen der Differentialgleichungen {\omega}^{\u2033}=P(z,\omega ). Math. Ann. 1953, 125: 355–365.
Wittich H Ann. Acad. Sci. Fenn. Ser. A I 379. Zur Theorie linearer Differentialgleichungen im Komplexen 1966.
Yang CC: On the entire solutions of certain class of nonlinear differential equations. J. Math. Anal. Appl. 1971, 33: 644–649. 10.1016/0022247X(71)900849
Laine I, Yang CC: Clunie theorems for difference and q difference polynomials. J. Lond. Math. Soc. 2007,76(2):556–566.
Laine I, Yang CC: Value distribution of difference polynomials. Proc. Jpn. Acad., Ser. A, Math. Sci. 2007, 83: 148–151. 10.3792/pjaa.83.148
Chen ZX: Growth and zeros of meromorphic solution of some linear difference equations. J. Math. Anal. Appl. 2011, 373: 235–241. 10.1016/j.jmaa.2010.06.049
Yang CC, Laine I: On analogies between nonlinear difference and differential equations. Proc. Jpn. Acad., Ser. A, Math. Sci. 2010, 86: 10–14. 10.3792/pjaa.86.10
Clunie J: On integral and meromorphic functions. J. Lond. Math. Soc. 1962, 37: 17–27.
Yang CC: A generalization of a theorem of P. Montel on entire functions. Proc. Am. Math. Soc. 1970, 26: 332–334. 10.1090/S0002993919700264080X
Yang CC, Li P: On the transcendental solutions of a certain type of nonlinear differential equations. Arch. Math. 2004, 82: 442–448.
Chiang YM, Feng SJ:On the Nevanlinna characteristic of f(z+\eta ) and difference equations in the complex plane. Ramanujan J. 2008, 16: 105–129. 10.1007/s1113900791011
Halburd RG, Korhonen RJ: Difference analogue of the lemma on logarithmic derivative with applications to difference equations. J. Math. Anal. Appl. 2006, 314: 477–487. 10.1016/j.jmaa.2005.04.010
Halburd RG, Korhonen RJ: Nevanlinna theory for the difference operator. Ann. Acad. Sci. Fenn., Math. 2006, 31: 463–478.
Liu K, Qi XG: Meromorphic solutions of q shift difference equations. Ann. Pol. Math. 2011, 101: 215–225. 10.4064/ap10132
arXiv: 0903.3236
Bank BS, Gundersen GG, Laine I: Meromorphic solutions of the Riccati differential equation. Ann. Acad. Sci. Fenn., Ser. A 1 Math. 1981, 6: 369–398.
Acknowledgements
The authors would like to thank the referee for his/her reading of the original version of the manuscript with valuable suggestions and comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors drafted the manuscript, read and approved the final manuscript.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License ( https://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
About this article
Cite this article
Liu, N., Lü, W., Shen, T. et al. Entire solutions of certain type of difference equations. J Inequal Appl 2014, 63 (2014). https://doi.org/10.1186/1029242X201463
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/1029242X201463
Keywords
 transcendental entire solution
 difference equation
 differencedifferential polynomial
 Nevanlinna theory