Next Article in Journal
Toward Bulk Synchronous Parallel-Based Machine Learning Techniques for Anomaly Detection in High-Speed Big Data Networks
Next Article in Special Issue
Bounded Solutions to Nonhomogeneous Linear Second-Order Difference Equations
Previous Article in Journal
Chiral Analysis of Pesticides and Drugs of Environmental Concern: Biodegradation and Enantiomeric Fraction
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Solvable Three-Dimensional Product-Type System of Difference Equations with Multipliers

1
Mathematical Institute of the Serbian Academy of Sciences, Knez Mihailova 36/III, 11000 Beograd, Serbia
2
Operator Theory and Applications Research Group, Department of Mathematics, King Abdulaziz University, P.O. Box 80203, Jeddah 21589, Saudi Arabia
Symmetry 2017, 9(9), 195; https://doi.org/10.3390/sym9090195
Submission received: 17 August 2017 / Revised: 1 September 2017 / Accepted: 4 September 2017 / Published: 16 September 2017

Abstract

:
The solvability of the following three-dimensional product-type system of difference equations
x n + 1 = α y n a z n 1 b , y n + 1 = β z n c x n 1 d , z n + 1 = γ x n f y n 1 g , n N 0 ,
where a , b , c , d , f , g Z , α , β , γ C   \   { 0 } and x i , y i , z i C   \   { 0 } , i { 0 , 1 } , is shown. This is the first three-dimensional system of the type with multipliers for which formulas are presented for their solutions in closed form in all the cases.

1. Introduction

There has been a huge interest on difference equations and systems of difference equations (see, for example, [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33] and the references therein). The classical problem of solving difference equations, among other topics, has re-attracted some recent interest (see, for example, [3,20,23,24,25,26,27,28,30,31,32,33] and the references therein). The solvability problem is of interest, since it is expected that obtained formulas for solutions to some equations and systems can help in studying the long-term behavior of the solutions. In 2004, we presented a method for solving a generalization of a nonlinear difference equation for which some formulas for their solutions had been given without any explanation how they were obtained, which have had some impact in the interest, and the methods and ideas therein have been applied and developed later in a series of papers (see, for example, [3,20,30] and the references therein). The main feature of these papers is that the equations and systems therein have been solved by using some changes of variables that transform them into the linear solvable ones (for the classical theory of linear difference equations and systems, see, for example, [6,8,9,10]). In [1], a solvable equation is studied in another way. Motivated, essentially, by papers [11,12,13], several experts started investigating symmetric, cyclic and other closely related systems of difference equations, which we frequently call close-to-symmetric/cyclic (see, for example, [3,14,15,16,17,21,22,23,24,25,27,28,29,30,31,32,33] and the references therein). As it can be seen, many of the above quoted papers belong to both areas, that is, they deal with some solvable close-to-symmetric systems of difference equations.
In a series of papers, we have studied some equations and systems, which, for some values of parameters, are reduced to product-type ones (see, for example, [29], as well as [4,5,19] and the references therein). The product-type equations and systems have helped to some extent in the study of the original/general equations and systems.
If the constant in the system in [29] is taken to be zero, and only positive solutions are considered, then the following system is obtained:
x n + 1 = y n p z n 1 p , y n + 1 = z n p x n 1 p , z n + 1 = x n p y n 1 p , n N 0 .
When the initial values are positive, then system (1) can be solved by using a well-known method (see, for example, the explanation in [24]). However, if they are complex, then we immediately see that there is a problem with multi-valuedness of functions
f p ( z ) = z p ,
for non-integer p-s.
These facts have motivated us to study the solvability of product-type systems in the complex domain. The first paper in the area was [28] where we have studied a two-dimensional relative of system (1), while the solvability of the following generalization of system (1)
x n + 1 = y n a z n 1 b , y n + 1 = z n c x n 1 d , z n + 1 = x n f y n 1 g ,
n N 0 , where a , b , c , d , f , g Z and x i , y i , z i C   \ { 0 } , i { 0 , 1 } , was studied in [24]. The next paper on the topic was [31], where we further developed our methods for solving product-type systems. In [26], we came across some product-type equations with multipliers, which motivated us to add some multipliers in the study of product-type systems, which had not been previously done in [24,28,31]. The first system with multipliers was studied in [23]. Having published [23], we got the idea of studying the solvability of product-type systems of the form
z n = α z n k a w n l b , w n = β w n m c z n s d , n N ,
where k , l , m , s N , a , b , c , d Z and α , β C . The study has been conducted in a series of papers (see [25,27,32,33], where some special cases of system (2) were studied in detail), with the aim to find all solvable systems of this type and to find closed-form formulas for their solutions. It should be pointed out that there are only finitely many solvable systems of the type, which is connected to the well known fact that there are polynomials of degree greater than or equal to five, which are not solvable by radicals. To the equations treated in [25,27,32,33] are essentially associated some polynomials of degree less than or equal to four, which helped in solving them. A detailed analysis of the structure of solutions to an equation whose associated polynomial is of the fourth order can be found in [27].
Since the study of solvable two-dimensional product-type systems is about finishing, it is a natural problem to find all three-dimensional product-type systems with multipliers, which are solvable in closed form. Here, we start considering the problem by investigating the solvability of the following three-dimensional close-to-cyclic product-type system of difference equations:
x n + 1 = α y n a z n 1 b , y n + 1 = β z n c x n 1 d , z n + 1 = γ x n f y n 1 g ,
for n N 0 , where a , b , c , d , f , g Z , α , β , γ C   \ { 0 } and x i , y i , z i C   \ { 0 } , i { 0 , 1 } . As far as we know, this is the first paper devoted to investigating of a three-dimensional product-type system of difference equations with multipliers in the complex domain. Our main results extend several of the results on the solvability in the literature.
Note that the case when one of the multipliers α , β , γ or initial values x i , y i , z i , i { 0 , 1 } is zero is excluded from the investigation, since, in the case, trivial or not well-defined solutions are obtained.

2. Main Result

This section presents our main result. Before we do this, we give a list of first several members of sequences x n , y n and z n defined by (3), which will be used in the proof of the result. We have
x 1 = α y 0 a z 1 b , x 2 = α β a x 1 a d z 0 a c + b , x 3 = α β a γ a c + b y 1 a c g + b g x 0 a c f + a d + b f , x 4 = α 1 + a c f + a d + b f β a γ a c + b z 1 a b c f + a b d + b 2 f y 0 a 2 c f + a c g + a 2 d + a b f + b g , y 1 = β z 0 c x 1 d , y 2 = β γ c y 1 c g x 0 f c + d , y 3 = β γ c α c f + d z 1 b c f + b d y 0 a c f + c g + a d , y 4 = α c f + d β 1 + a c f + c g + a d γ c x 1 a c d f + c d g + a d 2 z 0 a c 2 f + b c f + c 2 g + a c d + b d , z 1 = γ x 0 f y 1 g , z 2 = γ α f z 1 b f y 0 a f + g , z 3 = γ α f β a f + g x 1 a d f + d g z 0 a c f + b f + c g , z 4 = α f β a f + g γ 1 + a c f + b f + c g y 1 a c f g + b g f + c g 2 x 0 a c f 2 + a d f + b f 2 + c g f + d g .
Theorem 1.
Assume that a , b , c , d , f , g Z , α , β , γ , x i , y i , z i C   \ { 0 } , i { 0 , 1 } . Then, system (3) is solvable.
Proof of Theorem 1.
From (3) and since α , β , γ C   \ { 0 } and x i , y i , z i C   \ { 0 } , i { 0 , 1 } , it easily follows that x n y n z n 0 for every n 1 .
We have
x n + 1 = α y n a z n 1 b = α ( β z n 1 c x n 2 d ) a z n 1 b = α β a z n 1 a c + b x n 2 a d
= α β a ( γ x n 2 f y n 3 g ) a c + b x n 2 a d = α β a γ a c + b x n 2 a c f + b f + a d y n 3 a c g + b g ,
y n + 1 = β z n c x n 1 d = β ( γ x n 1 f y n 2 g ) c x n 1 d = β γ c x n 1 c f + d y n 2 c g
= β γ c ( α y n 2 a z n 3 b ) c f + d y n 2 c g = β γ c α c f + d y n 2 a c f + a d + c g z n 3 b c f + b d ,
z n + 1 = γ x n f y n 1 g = γ ( α y n 1 a z n 2 b ) f y n 1 g = γ α f y n 1 a f + g z n 2 b f
= γ α f ( β z n 2 c x n 3 d ) a f + g z n 2 b f = γ α f β a f + g z n 2 a c f + c g + b f x n 3 a d f + d g ,
for n 2 .
From Equations (5), (7) and (9), we get, respectively,
z n 1 a c + b = x n + 1 α β a x n 2 a d ,
x n 1 c f + d = y n + 1 β γ c y n 2 c g ,
y n 1 a f + g = z n + 1 γ α f z n 2 b f ,
for n N .
Taking Equation (10) to the a c + b -th power and using (11),
x n + 3 α β a x n a d = ( γ α f β a f + g ) a c + b x n α β a x n 3 a d a c f + c g + b f x n 3 ( a d f + d g ) ( a c + b ) ,
is obtained, that is,
x n + 3 = α 1 c g β b g + a γ a c + b x n a c f + b f + a d + c g x n 3 b d g , n 2 .
Taking Equation (6) to the c f + d -th power and using (12),
y n + 3 β γ c y n c g = ( α β a γ a c + b ) c f + d y n β γ c y n 3 c g a c f + b f + a d y n 3 ( a c g + b g ) ( c f + d ) ,
is obtained, that is,
y n + 3 = α c f + d β 1 b f γ b d + c y n a c f + b f + a d + c g y n 3 b d g , n 2 .
Taking Equation (8) to the a f + g -th power and using (13) is obtained
z n + 3 γ α f z n b f = ( β γ c α c f + d ) a f + g z n γ α f z n 3 b f a c f + a d + c g z n 3 ( b c f + b d ) ( a f + g ) ,
that is,
z n + 3 = α d g + f β a f + g γ 1 a d z n a c f + b f + a d + c g z n 3 b d g , n 2 .
Let δ 1 : = α 1 c g β b g + a γ a c + b , δ 2 : = α c f + d β 1 b f γ b d + c , δ 3 : = α d g + f β a f + g γ 1 a d ,
a 1 ( i ) = a c f + b f + a d + c g , b 1 ( i ) = b d g , u 1 ( i ) = 1 ,
for i = 1 , 0 , 1 . Then, Equations (14)–(16) can be written in the form
x 3 m + i = δ 1 u 1 ( i ) x 3 ( m 1 ) + i a 1 ( i ) x 3 ( m 2 ) + i b 1 ( i ) ,
y 3 m + i = δ 2 u 1 ( i ) y 3 ( m 1 ) + i a 1 ( i ) y 3 ( m 2 ) + i b 1 ( i ) ,
z 3 m + i = δ 3 u 1 ( i ) z 3 ( m 1 ) + i a 1 ( i ) z 3 ( m 2 ) + i b 1 ( i ) ,
for m 2 and i = 1 , 0 , 1 .
By using Equation (18) with m m 1 , into Equation (18), we have
x 3 m + i = δ 1 u 1 ( i ) x 3 ( m 1 ) + i a 1 ( i ) x 3 ( m 2 ) + i b 1 ( i ) = δ 1 u 1 ( i ) ( δ 1 x 3 ( m 2 ) + i a 1 ( i ) x 3 ( m 3 ) + i b 1 ( i ) ) a 1 ( i ) x 3 ( m 2 ) + i b 1 ( i ) = δ 1 u 1 ( i ) + a 1 ( i ) x 3 ( m 2 ) + i a 1 ( i ) a 1 ( i ) + b 1 ( i ) x 3 ( m 3 ) + i b 1 ( i ) a 1 ( i ) = δ 1 u 2 ( i ) x 3 ( m 2 ) + i a 2 ( i ) x 3 ( m 3 ) + i b 2 ( i ) ,
for m 3 and i = 1 , 0 , 1 , where
a 2 ( i ) : = a 1 ( i ) a 1 ( i ) + b 1 ( i ) , b 2 ( i ) : = b 1 ( i ) a 1 ( i ) , u 2 ( i ) : = u 1 ( i ) + a 1 ( i ) .
Assume that, for some k 2 , we have proved that
x 3 m + i = δ 1 u k ( i ) x 3 ( m k ) + i a k ( i ) x 3 ( m k 1 ) + i b k ( i ) ,
for m k + 1 and i = 1 , 0 , 1 , where
a k ( i ) = a 1 ( i ) a k 1 ( i ) + b k 1 ( i ) , b k ( i ) = b 1 ( i ) a k 1 ( i ) , u k ( i ) = u k 1 ( i ) + a k 1 ( i ) .
Then, by using Equation (18) with m m k into Equation (23), we get
x 3 m + i = δ 1 u k ( i ) ( δ 1 x 3 ( m k 1 ) + i a 1 ( i ) x 3 ( m k 2 ) + i b 1 ( i ) ) a k ( i ) x 3 ( m k 1 ) + i b k ( i ) = δ 1 u k ( i ) + a k ( i ) x 3 ( m k 1 ) + i a 1 ( i ) a k ( i ) + b k ( i ) x 3 ( m k 2 ) + i b 1 ( i ) a k ( i ) = δ 1 u k + 1 ( i ) x 3 ( m k 1 ) + i a k + 1 ( i ) x 3 ( m k 2 ) + i b k + 1 ( i ) ,
for m k + 2 and i = −1, 0, 1, where
a k + 1 ( i ) : = a 1 ( i ) a k ( i ) + b k ( i ) , b k + 1 ( i ) : = b 1 ( i ) a k ( i ) , u k + 1 ( i ) : = u k ( i ) + a k ( i ) .
Relations (21), (22), (25), (26) along with the induction show that (23) and (24) hold for every k and m such that 2 k m 1 and each i = 1 , 0 , 1 .
From the first two equations in (26), it follows that
a k + 1 ( i ) ( a c f + b f + a d + c g ) a k ( i ) b d g a k 1 ( i ) = 0 , k 2 .
The third equalities in (17) and (24) imply
u k ( i ) = 1 + j = 1 k 1 a j ( i ) ,
for k N and i = 1 , 0 , 1 .
Similarly is obtained
y 3 m + i = δ 2 u k ( i ) y 3 ( m k ) + i a k ( i ) y 3 ( m k 1 ) + i b k ( i ) ,
z 3 m + i = δ 3 u k ( i ) z 3 ( m k ) + i a k ( i ) z 3 ( m k 1 ) + i b k ( i ) ,
for m k + 1 and i = 1 , 0 , 1 .
By taking k = m 1 in Equations (23), (29) and (30), using the second relation in (24), and the fact that due to (17) and (24), sequences ( a k ( i ) ) k N , as well as ( b k ( i ) ) k N and ( u k ( i ) ) k N , are the same for i { 1 , 0 , 1 } , and denoting them by a k , b k , u k , we get
x 3 m + i = δ 1 u m 1 x 3 + i a m 1 x i b 1 a m 2 ,
y 3 m + i = δ 2 u m 1 y 3 + i a m 1 y i b 1 a m 2 ,
z 3 m + i = δ 3 u m 1 z 3 + i a m 1 z i b 1 a m 2 ,
for m 3 and i = −1, 0, 1.
Employing (4) and (24), in (31)–(33), we have
x 3 m 1 = δ 1 u m 1 x 2 a m 1 x 1 b 1 a m 2 = ( α 1 c g β b g + a γ a c + b ) u m 1 ( α β a x 1 a d z 0 a c + b ) a m 1 x 1 b 1 a m 2 = α u m c g u m 1 β b g u m 1 + a u m γ ( a c + b ) u m 1 x 1 a d a m 1 + b d g a m 2 z 0 ( a c + b ) a m 1 ,
x 3 m = δ 1 u m 1 x 3 a m 1 x 0 b 1 a m 2 = ( α 1 c g β b g + a γ a c + b ) u m 1 ( α β a γ a c + b y 1 a c g + b g x 0 a c f + a d + b f ) a m 1 x 0 b 1 a m 2 = α u m c g u m 1 β b g u m 1 + a u m γ ( a c + b ) u m y 1 ( a c + b ) g a m 1 x 0 a m c g a m 1 ,
x 3 m + 1 = δ 1 u m 1 x 4 a m 1 x 1 b 1 a m 2 = ( α 1 c g β b g + a γ a c + b ) u m 1 ( α y 0 a z 1 b ) b 1 a m 2 × ( α 1 + a c f + a d + b f β a γ a c + b z 1 a b c f + a b d + b 2 f y 0 a 2 c f + a c g + a 2 d + a b f + b g ) a m 1 = α u m + 1 c g u m β b g u m 1 + a u m γ ( a c + b ) u m z 1 b ( a m c g a m 1 ) y 0 a a m + b g a m 1 ,
y 3 m 1 = δ 2 u m 1 y 2 a m 1 y 1 b 1 a m 2 = ( α c f + d β 1 b f γ b d + c ) u m 1 ( β γ c y 1 c g x 0 f c + d ) a m 1 y 1 b 1 a m 2 = α ( c f + d ) u m 1 β u m b f u m 1 γ b d u m 1 + c u m y 1 c g a m 1 + b d g a m 2 x 0 ( f c + d ) a m 1 ,
y 3 m = δ 2 u m 1 y 3 a m 1 y 0 b 1 a m 2 = ( α c f + d β 1 b f γ b d + c ) u m 1 ( β γ c α c f + d z 1 b c f + b d y 0 a c f + c g + a d ) a m 1 y 0 b 1 a m 2 = α ( c f + d ) u m β u m b f u m 1 γ b d u m 1 + c u m z 1 b ( c f + d ) a m 1 y 0 a m b f a m 1 ,
y 3 m + 1 = δ 2 u m 1 y 4 a m 1 y 1 b 1 a m 2 = ( α c f + d β 1 + a c f + c g + a d γ c x 1 a c d f + c d g + a d 2 z 0 a c 2 f + b c f + c 2 g + a c d + b d ) a m 1 × ( α c f + d β 1 b f γ b d + c ) u m 1 ( β z 0 c x 1 d ) b 1 a m 2 = α ( c f + d ) u m β u m + 1 b f u m γ b d u m 1 + c u m x 1 d ( a m b f a m 1 ) z 0 c a m + b d a m 1 ,
z 3 m 1 = δ 3 u m 1 z 2 a m 1 z 1 b 1 a m 2 = ( α d g + f β a f + g γ 1 a d ) u m 1 ( γ α f z 1 b f y 0 a f + g ) a m 1 z 1 b 1 a m 2 = α d g u m 1 + f u m β ( a f + g ) u m 1 γ u m a d u m 1 z 1 b f a m 1 + b d g a m 2 y 0 ( a f + g ) a m 1 ,
z 3 m = δ 3 u m 1 z 3 a m 1 z 0 b 1 a m 2 = ( α d g + f β a f + g γ 1 a d ) u m 1 ( γ α f β a f + g x 1 a d f + d g z 0 a c f + b f + c g ) a m 1 z 0 b 1 a m 2 = α d g u m 1 + f u m β ( a f + g ) u m γ u m a d u m 1 x 1 d ( a f + g ) a m 1 z 0 a m a d a m 1 ,
z 3 m + 1 = δ 3 u m 1 z 4 a m 1 z 1 b 1 a m 2 = ( α f β a f + g γ 1 + a c f + b f + c g y 1 a c f g + b g f + c g 2 x 0 a c f 2 + a d f + b f 2 + c g f + d g ) a m 1 × ( α d g + f β a f + g γ 1 a d ) u m 1 ( γ x 0 f y 1 g ) b 1 a m 2 = α d g u m 1 + f u m β ( a f + g ) u m γ u m + 1 a d u m y 1 g ( a m a d a m 1 ) x 0 f a m + d g a m 1 ,
for m 3 .
Case b = 0 . In this case, we have
a n + 1 = ( a c f + a d + c g ) a n , n 2 ,
from which along with (17) and (24), it follows that
a n = ( a c f + a d + c g ) n ,
for n N .
From (28) and (43), it follows that
u n = j = 0 n 1 ( a c f + a d + c g ) j , n N ,
and consequently
u n = ( a c f + a d + c g ) n 1 a c f + a d + c g 1 , n N ,
when a c f + a d + c g 1 , while
u n = n , n N ,
when a c f + a d + c g = 1 .
On the other hand, from (34)–(42), we get
x 3 m 1 = α u m c g u m 1 β a u m γ a c u m 1 x 1 a d a m 1 z 0 a c a m 1 ,
x 3 m = α u m c g u m 1 β a u m γ a c u m y 1 a c g a m 1 x 0 a m c g a m 1 ,
x 3 m + 1 = α u m + 1 c g u m β a u m γ a c u m y 0 a a m ,
y 3 m 1 = α ( c f + d ) u m 1 β u m γ c u m y 1 c g a m 1 x 0 ( f c + d ) a m 1 ,
y 3 m = α ( c f + d ) u m β u m γ c u m y 0 a m ,
y 3 m + 1 = α ( c f + d ) u m β u m + 1 γ c u m x 1 d a m z 0 c a m ,
z 3 m 1 = α d g u m 1 + f u m β ( a f + g ) u m 1 γ u m a d u m 1 y 0 ( a f + g ) a m 1 ,
z 3 m = α d g u m 1 + f u m β ( a f + g ) u m γ u m a d u m 1 x 1 d ( a f + g ) a m 1 z 0 a m a d a m 1 ,
z 3 m + 1 = α d g u m 1 + f u m β ( a f + g ) u m γ u m + 1 a d u m y 1 g ( a m a d a m 1 ) x 0 f a m + d g a m 1 ,
for m 3 .
Subcase a c f + a d + c g 1 . Employing (43) and (44) in (46)–(54),
x 3 m 1 = α a ( c f + d ) ( a c f + a d + c g ) m 1 + c g 1 a c f + a d + c g 1 β a ( a c f + a d + c g ) m 1 a c f + a d + c g 1 γ a c ( a c f + a d + c g ) m 1 1 a c f + a d + c g 1 × x 1 a d ( a c f + a d + c g ) m 1 z 0 a c ( a c f + a d + c g ) m 1 ,
x 3 m = α a ( c f + d ) ( a c f + a d + c g ) m 1 + c g 1 a c f + a d + c g 1 β a ( a c f + a d + c g ) m 1 a c f + a d + c g 1 γ a c ( a c f + a d + c g ) m 1 a c f + a d + c g 1 × y 1 a c g ( a c f + a d + c g ) m 1 x 0 a ( c f + d ) ( a c f + a d + c g ) m 1 ,
x 3 m + 1 = α a ( c f + d ) ( a c f + a d + c g ) m + c g 1 a c f + a d + c g 1 β a ( a c f + a d + c g ) m 1 a c f + a d + c g 1 γ a c ( a c f + a d + c g ) m 1 a c f + a d + c g 1 × y 0 a ( a c f + a d + c g ) m ,
y 3 m 1 = α ( c f + d ) ( a c f + a d + c g ) m 1 1 a c f + a d + c g 1 β ( a c f + a d + c g ) m 1 a c f + a d + c g 1 γ c ( a c f + a d + c g ) m 1 a c f + a d + c g 1 × y 1 c g ( a c f + a d + c g ) m 1 x 0 ( f c + d ) ( a c f + a d + c g ) m 1 ,
y 3 m = α ( c f + d ) ( a c f + a d + c g ) m 1 a c f + a d + c g 1 β ( a c f + a d + c g ) m 1 a c f + a d + c g 1 γ c ( a c f + a d + c g ) m 1 a c f + a d + c g 1 y 0 ( a c f + a d + c g ) m ,
y 3 m + 1 = α ( c f + d ) ( a c f + a d + c g ) m 1 a c f + a d + c g 1 β ( a c f + a d + c g ) m + 1 1 a c f + a d + c g 1 γ c ( a c f + a d + c g ) m 1 a c f + a d + c g 1 × x 1 d ( a c f + a d + c g ) m z 0 c ( a c f + a d + c g ) m ,
z 3 m 1 = α ( d g + a c f 2 + a d f + c g f ) ( a c f + a d + c g ) m 1 d g f a c f + a d + c g 1 β ( a f + g ) ( a c f + a d + c g ) m 1 1 a c f + a d + c g 1 × γ c ( a f + g ) ( a c f + a d + c g ) m 1 + a d 1 a c f + a d + c g 1 y 0 ( a f + g ) ( a c f + a d + c g ) m 1 ,
z 3 m = α ( d g + a c f 2 + a d f + c g f ) ( a c f + a d + c g ) m 1 d g f a c f + a d + c g 1 β ( a f + g ) ( a c f + a d + c g ) m 1 a c f + a d + c g 1 × γ c ( a f + g ) ( a c f + a d + c g ) m 1 + a d 1 a c f + a d + c g 1 x 1 d ( a f + g ) ( a c f + a d + c g ) m 1 × z 0 c ( a f + g ) ( a c f + a d + c g ) m 1 ,
z 3 m + 1 = α ( d g + a c f 2 + a d f + c g f ) ( a c f + a d + c g ) m 1 d g f a c f + a d + c g 1 β ( a f + g ) ( a c f + a d + c g ) m 1 a c f + a d + c g 1 × γ c ( a f + g ) ( a c f + a d + c g ) m + a d 1 a c f + a d + c g 1 y 1 c g ( a f + g ) ( a c f + a d + c g ) m 1 × x 0 ( a c f 2 + a d f + c g f + d g ) ( a c f + a d + c g ) m 1 ,
is obtained for m 3 .
Subcase a c f + a d + c g = 1 . Employing (43) and (45) in (46)–(54),
x 3 m 1 = α ( 1 c g ) m + c g β a m γ a c ( m 1 ) x 1 a d z 0 a c ,
x 3 m = α ( 1 c g ) m + c g β a m γ a c m y 1 a c g x 0 1 c g ,
x 3 m + 1 = α ( 1 c g ) m + 1 β a m γ a c m y 0 a ,
y 3 m 1 = α ( c f + d ) ( m 1 ) β m γ c m y 1 c g x 0 f c + d ,
y 3 m = α ( c f + d ) m β m γ c m y 0 ,
y 3 m + 1 = α ( c f + d ) m β m + 1 γ c m x 1 d z 0 c ,
z 3 m 1 = α ( d g + f ) m d g β ( a f + g ) ( m 1 ) γ ( 1 a d ) m + a d y 0 a f + g ,
z 3 m = α ( d g + f ) m d g β ( a f + g ) m γ ( 1 a d ) m + a d x 1 d ( a f + g ) z 0 1 a d ,
z 3 m + 1 = α ( d g + f ) m d g β ( a f + g ) m γ ( 1 a d ) m + 1 y 1 g ( 1 a d ) x 0 f + d g ,
is obtained for m 3 .
Case d = 0 . In this case, we have
a n + 1 = ( a c f + b f + c g ) a n , n 2 ,
from which, along with (17) and (24), it follows that
a n = ( a c f + b f + c g ) n ,
for n N .
From (28) and (73), it follows that
u n = j = 0 n 1 ( a c f + b f + c g ) j , n N ,
and consequently
u n = ( a c f + b f + c g ) n 1 a c f + b f + c g 1 , n N ,
when a c f + b f + c g 1 , while
u n = n , n N ,
when a c f + b f + c g = 1 .
On the other hand, from (34)–(42), we get
x 3 m 1 = α u m c g u m 1 β b g u m 1 + a u m γ ( a c + b ) u m 1 z 0 ( a c + b ) a m 1 ,
x 3 m = α u m c g u m 1 β b g u m 1 + a u m γ ( a c + b ) u m y 1 ( a c + b ) g a m 1 x 0 a m c g a m 1 ,
x 3 m + 1 = α u m + 1 c g u m β b g u m 1 + a u m γ ( a c + b ) u m z 1 b ( a m c g a m 1 ) y 0 a a m + b g a m 1 ,
y 3 m 1 = α c f u m 1 β u m b f u m 1 γ c u m y 1 c g a m 1 x 0 f c a m 1 ,
y 3 m = α c f u m β u m b f u m 1 γ c u m z 1 b c f a m 1 y 0 a m b f a m 1 ,
y 3 m + 1 = α c f u m β u m + 1 b f u m γ c u m z 0 c a m ,
z 3 m 1 = α f u m β ( a f + g ) u m 1 γ u m z 1 b f a m 1 y 0 ( a f + g ) a m 1 ,
z 3 m = α f u m β ( a f + g ) u m γ u m z 0 a m ,
z 3 m + 1 = α f u m β ( a f + g ) u m γ u m + 1 y 1 g a m x 0 f a m ,
for m 3 .
We have to consider two subcases.
Subcase a c f + b f + c g 1 . Employing (73) and (74) in (76)–(84),
x 3 m 1 = α ( a c + b ) f ( a c f + b f + c g ) m 1 + c g 1 a c f + b f + c g 1 β ( a 2 c f + a b f + a c g + b g ) ( a c f + b f + c g ) m 1 a b g a c f + b f + c g 1 × γ ( a c + b ) ( a c f + b f + c g ) m 1 1 a c f + b f + c g 1 z 0 ( a c + b ) ( a c f + b f + c g ) m 1 ,
x 3 m = α ( a c + b ) f ( a c f + b f + c g ) m 1 + c g 1 a c f + b f + c g 1 β ( a 2 c f + a b f + a c g + b g ) ( a c f + b f + c g ) m 1 a b g a c f + b f + c g 1 × γ ( a c + b ) ( a c f + b f + c g ) m 1 a c f + b f + c g 1 y 1 ( a c + b ) g ( a c f + b f + c g ) m 1 × x 0 ( a c + b ) f ( a c f + b f + c g ) m 1 ,
x 3 m + 1 = α ( a c + b ) f ( a c f + b f + c g ) m + c g 1 a c f + b f + c g 1 β ( a 2 c f + a b f + a c g + b g ) ( a c f + b f + c g ) m 1 a b g a c f + b f + c g 1 × γ ( a c + b ) ( a c f + b f + c g ) m 1 a c f + b f + c g 1 z 1 b f ( a c + b ) ( a c f + b f + c g ) m 1 × y 0 ( a 2 c f + a b f + a c g + b g ) ( a c f + b f + c g ) m 1 ,
y 3 m 1 = α c f ( a c f + b f + c g ) m 1 1 a c f + b f + c g 1 β c ( a f + g ) ( a c f + b f + c g ) m 1 + b f 1 a c f + b f + c g 1 γ c ( a c f + b f + c g ) m 1 a c f + b f + c g 1 × y 1 c g ( a c f + b f + c g ) m 1 x 0 f c ( a c f + b f + c g ) m 1 ,
y 3 m = α c f ( a c f + b f + c g ) m 1 a c f + b f + c g 1 β c ( a f + g ) ( a c f + b f + c g ) m 1 + b f 1 a c f + b f + c g 1 γ c ( a c f + b f + c g ) m 1 a c f + b f + c g 1 × z 1 b c f ( a c f + b f + c g ) m 1 y 0 c ( a f + g ) ( a c f + b f + c g ) m 1 ,
y 3 m + 1 = α c f ( a c f + b f + c g ) m 1 a c f + b f + c g 1 β c ( a f + g ) ( a c f + b f + c g ) m + b f 1 a c f + b f + c g 1 γ c ( a c f + b f + c g ) m 1 a c f + b f + c g 1 × z 0 c ( a c f + b f + c g ) m ,
z 3 m 1 = α f ( a c f + b f + c g ) m 1 a c f + b f + c g 1 β ( a f + g ) ( a c f + b f + c g ) m 1 1 a c f + b f + c g 1 γ ( a c f + b f + c g ) m 1 a c f + b f + c g 1 × z 1 b f ( a c f + b f + c g ) m 1 y 0 ( a f + g ) ( a c f + b f + c g ) m 1 ,
z 3 m = α f ( a c f + b f + c g ) m 1 a c f + b f + c g 1 β ( a f + g ) ( a c f + b f + c g ) m 1 a c f + b f + c g 1 γ ( a c f + b f + c g ) m 1 a c f + b f + c g 1 × z 0 ( a c f + b f + c g ) m ,
z 3 m + 1 = α f ( a c f + b f + c g ) m 1 a c f + b f + c g 1 β ( a f + g ) ( a c f + b f + c g ) m 1 a c f + b f + c g 1 γ ( a c f + b f + c g ) m + 1 1 a c f + b f + c g 1 × y 1 g ( a c f + b f + c g ) m x 0 f ( a c f + b f + c g ) m ,
is obtained for m 3 .
Subcase a c f + b f + c g = 1 . Employing (73) and (75) in (76)–(84),
x 3 m 1 = α ( 1 c g ) m + c g β ( b g + a ) m b g γ ( a c + b ) ( m 1 ) z 0 a c + b ,
x 3 m = α ( 1 c g ) m + c g β ( b g + a ) m b g γ ( a c + b ) m y 1 ( a c + b ) g x 0 1 c g ,
x 3 m + 1 = α ( 1 c g ) m + 1 β ( b g + a ) m b g γ ( a c + b ) m z 1 b ( 1 c g ) y 0 a + b g ,
y 3 m 1 = α c f ( m 1 ) β ( 1 b f ) m + b f γ c m y 1 c g x 0 f c ,
y 3 m = α c f m β ( 1 b f ) m + b f γ c m z 1 b c f y 0 1 b f ,
y 3 m + 1 = α c f m β ( 1 b f ) m + 1 γ c m z 0 c ,
z 3 m 1 = α f m β ( a f + g ) ( m 1 ) γ m z 1 b f y 0 a f + g ,
z 3 m = α f m β ( a f + g ) m γ m z 0 ,
z 3 m + 1 = α f m β ( a f + g ) m γ m + 1 y 1 g x 0 f ,
is obtained for m 3 .
Case g = 0 . In this case, we have
a n + 1 = ( a c f + a d + b f ) a n , n 2 ,
from which along with (17) and (24), it follows that
a n = ( a c f + a d + b f ) n , n N .
From (28) and (103), it follows that
u n = j = 0 n 1 ( a c f + a d + b f ) j , n N ,
and consequently
u n = ( a c f + a d + b f ) n 1 a c f + a d + b f 1 , n N ,
when a c f + a d + b f 1 , while
u n = n , n N ,
when a c f + a d + b f = 1 .
On the other hand, from (34)–(42), we get
x 3 m 1 = α u m β a u m γ ( a c + b ) u m 1 x 1 a d a m 1 z 0 ( a c + b ) a m 1 ,
x 3 m = α u m β a u m γ ( a c + b ) u m x 0 a m ,
x 3 m + 1 = α u m + 1 β a u m γ ( a c + b ) u m z 1 b a m y 0 a a m ,
y 3 m 1 = α ( c f + d ) u m 1 β u m b f u m 1 γ b d u m 1 + c u m x 0 ( f c + d ) a m 1 ,
y 3 m = α ( c f + d ) u m β u m b f u m 1 γ b d u m 1 + c u m z 1 b ( c f + d ) a m 1 y 0 a m b f a m 1 ,
y 3 m + 1 = α ( c f + d ) u m β u m + 1 b f u m γ b d u m 1 + c u m x 1 d ( a m b f a m 1 ) z 0 c a m + b d a m 1 ,
z 3 m 1 = α f u m β a f u m 1 γ u m a d u m 1 z 1 b f a m 1 y 0 a f a m 1 ,
z 3 m = α f u m β a f u m γ u m a d u m 1 x 1 a d f a m 1 z 0 a m a d a m 1 ,
z 3 m + 1 = α f u m β a f u m γ u m + 1 a d u m x 0 f a m ,
for m 3 .
Subcase a c f + a d + b f 1 . Employing (103) and (104) in (106)–(114),
x 3 m 1 = α ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a ( a c f + a d + b f ) m 1 a c f + a d + b f 1 γ ( a c + b ) ( a c f + a d + b f ) m 1 1 a c f + a d + b f 1 × x 1 a d ( a c f + a d + b f ) m 1 z 0 ( a c + b ) ( a c f + a d + b f ) m 1 ,
x 3 m = α ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a ( a c f + a d + b f ) m 1 a c f + a d + b f 1 γ ( a c + b ) ( a c f + a d + b f ) m 1 a c f + a d + b f 1 × x 0 ( a c f + a d + b f ) m ,
x 3 m + 1 = α ( a c f + a d + b f ) m + 1 1 a c f + a d + b f 1 β a ( a c f + a d + b f ) m 1 a c f + a d + b f 1 γ ( a c + b ) ( a c f + a d + b f ) m 1 a c f + a d + b f 1 × z 1 b ( a c f + a d + b f ) m y 0 a ( a c f + a d + b f ) m ,
y 3 m 1 = α ( c f + d ) ( a c f + a d + b f ) m 1 1 a c f + a d + b f 1 β a ( c f + d ) ( a c f + a d + b f ) m 1 + b f 1 a c f + a d + b f 1 × γ ( a c 2 f + a c d + b c f + b d ) ( a c f + a d + b f ) m 1 b d c a c f + a d + b f 1 x 0 ( f c + d ) ( a c f + a d + b f ) m 1 ,
y 3 m = α ( c f + d ) ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a ( c f + d ) ( a c f + a d + b f ) m 1 + b f 1 a c f + a d + b f 1 × γ ( a c 2 f + a c d + b c f + b d ) ( a c f + a d + b f ) m 1 b d c a c f + a d + b f 1 z 1 b ( f c + d ) ( a c f + a d + b f ) m 1 × y 0 a ( c f + d ) ( a c f + a d + b f ) m 1 ,
y 3 m + 1 = α ( c f + d ) ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a ( c f + d ) ( a c f + a d + b f ) m + b f 1 a c f + a d + b f 1 × γ ( a c 2 f + a c d + b c f + b d ) ( a c f + a d + b f ) m 1 b d c a c f + a d + b f 1 x 1 a d ( c f + d ) ( a c f + a d + b f ) m 1 × z 0 ( a c 2 f + a c d + b c f + b d ) ( a c f + a d + b f ) m 1 ,
z 3 m 1 = α f ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a f ( a c f + a d + b f ) m 1 1 a c f + a d + b f 1 γ f ( a c + b ) ( a c f + a d + b f ) m 1 + a d 1 a c f + a d + b f 1 × z 1 b f ( a c f + a d + b f ) m 1 y 0 a f ( a c f + a d + b f ) m 1 ,
z 3 m = α f ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a f ( a c f + a d + b f ) m 1 a c f + a d + b f 1 γ f ( a c + b ) ( a c f + a d + b f ) m 1 + a d 1 a c f + a d + b f 1 × x 1 a d f ( a c f + a d + b f ) m 1 z 0 f ( a c + b ) ( a c f + a d + b f ) m 1 ,
z 3 m + 1 = α f ( a c f + a d + b f ) m 1 a c f + a d + b f 1 β a f ( a c f + a d + b f ) m 1 a c f + a d + b f 1 γ f ( a c + b ) ( a c f + a d + b f ) m + a d 1 a c f + a d + b f 1 × x 0 f ( a c f + a d + b f ) m ,
is obtained for m 3 .
Subcase a c f + a d + b f = 1 . Employing (103) and (105) in (106)–(114),
x 3 m 1 = α m β a m γ ( a c + b ) ( m 1 ) x 1 a d z 0 a c + b ,
x 3 m = α m β a m γ ( a c + b ) m x 0 ,
x 3 m + 1 = α m + 1 β a m γ ( a c + b ) m z 1 b y 0 a ,
y 3 m 1 = α ( c f + d ) ( m 1 ) β ( 1 b f ) m + b f γ ( b d + c ) m b d x 0 f c + d ,
y 3 m = α ( c f + d ) m β ( 1 b f ) m + b f γ ( b d + c ) m b d z 1 b ( c f + d ) y 0 1 b f ,
y 3 m + 1 = α ( c f + d ) m β ( 1 b f ) m + 1 γ ( b d + c ) m b d x 1 d ( 1 b f ) z 0 c + b d ,
z 3 m 1 = α f m β a f ( m 1 ) γ ( 1 a d ) m + a d z 1 b f y 0 a f ,
z 3 m = α f m β a f m γ ( 1 a d ) m + a d x 1 a d f z 0 1 a d ,
z 3 m + 1 = α f m β a f m γ ( 1 a d ) m + 1 x 0 f ,
is obtained for m 3 .
Case b d g 0 . Let λ 1 , 2 be the roots of the characteristic polynomial
P ( λ ) = λ 2 ( a c f + a d + b f + c g ) λ b d g ,
associated with the equation
u n + 2 ( a c f + a d + b f + c g ) u n + 1 b d g u n = 0 , n N .
It is known that general solution of Equation (133) has the following form
u n = α 1 λ 1 n + α 2 λ 2 n , n N ,
if ( a c f + a d + b f + c g ) 2 4 b d g , where
λ 1 , 2 = a c f + a d + b f + c g ± ( a c f + a d + b f + c g ) 2 + 4 b d g 2 ,
and α 1 and α 2 are arbitrary constants, while in the case ( a c f + a d + b f + c g ) 2 = 4 b d g , the general solution has the following form
u n = ( β 1 n + β 2 ) λ 1 n , n N ,
where λ 1 = ( a c f + a d + b f + c g ) / 2 , and β 1 , β 2 are arbitrary constants.
Since b d g 0 , from (133), we have
u n = u n + 2 ( a c f + a d + b f + c g ) u n + 1 b d g ,
from which it follows that u n can be calculated also for every n 0 . Hence, it is easily seen that, for a n ,
a 1 = 0 , a 0 = 1 ,
holds, from which it follows that
a n = λ 1 n + 1 λ 2 n + 1 λ 1 λ 2 ,
when ( a c f + a d + b f + c g ) 2 4 b d g , and consequently
u n = 1 + j = 1 n 1 λ 1 j + 1 λ 2 j + 1 λ 1 λ 2 = ( λ 2 1 ) λ 1 n + 1 ( λ 1 1 ) λ 2 n + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) .
By using (135) and (136) in (34)–(42), and after some calculations, we get
x 3 m 1 = α ( λ 2 1 ) ( λ 1 c g ) λ 1 m ( λ 1 1 ) ( λ 2 c g ) λ 2 m + ( λ 1 λ 2 ) ( 1 c g ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( λ 2 1 ) ( a λ 1 + b g ) λ 1 m ( λ 1 1 ) ( a λ 2 + b g ) λ 2 m + ( λ 1 λ 2 ) ( b g + a ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( a c + b ) ( λ 2 1 ) λ 1 m ( λ 1 1 ) λ 2 m + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × x 1 d ( a λ 1 + b g ) λ 1 m 1 ( a λ 2 + b g ) λ 2 m 1 λ 1 λ 2 z 0 ( a c + b ) λ 1 m λ 2 m λ 1 λ 2 ,
x 3 m = α ( λ 2 1 ) ( λ 1 c g ) λ 1 m ( λ 1 1 ) ( λ 2 c g ) λ 2 m + ( λ 1 λ 2 ) ( 1 c g ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( λ 2 1 ) ( a λ 1 + b g ) λ 1 m ( λ 1 1 ) ( a λ 2 + b g ) λ 2 m + ( λ 1 λ 2 ) ( b g + a ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( a c + b ) ( λ 2 1 ) λ 1 m + 1 ( λ 1 1 ) λ 2 m + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × y 1 ( a c + b ) g λ 1 m λ 2 m λ 1 λ 2 x 0 ( λ 1 c g ) λ 1 m ( λ 2 c g ) λ 2 m λ 1 λ 2 ,
x 3 m + 1 = α ( λ 2 1 ) ( λ 1 c g ) λ 1 m + 1 ( λ 1 1 ) ( λ 2 c g ) λ 2 m + 1 + ( λ 1 λ 2 ) ( 1 c g ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( λ 2 1 ) ( a λ 1 + b g ) λ 1 m ( λ 1 1 ) ( a λ 2 + b g ) λ 2 m + ( λ 1 λ 2 ) ( b g + a ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( a c + b ) ( λ 2 1 ) λ 1 m + 1 ( λ 1 1 ) λ 2 m + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) z 1 b ( λ 1 c g ) λ 1 m ( λ 2 c g ) λ 2 m λ 1 λ 2 × y 0 ( a λ 1 + b g ) λ 1 m ( a λ 2 + b g ) λ 2 m λ 1 λ 2 ,
y 3 m 1 = α ( c f + d ) ( λ 2 1 ) λ 1 m ( λ 1 1 ) λ 2 m + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( λ 2 1 ) ( λ 1 b f ) λ 1 m ( λ 1 1 ) ( λ 2 b f ) λ 2 m + ( λ 1 λ 2 ) ( 1 b f ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( λ 2 1 ) ( c λ 1 + b d ) λ 1 m ( λ 1 1 ) ( c λ 2 + b d ) λ 2 m + ( λ 1 λ 2 ) ( b d + c ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × y 1 g ( c λ 1 + b d ) λ 1 m 1 ( c λ 2 + b d ) λ 2 m 1 λ 1 λ 2 x 0 ( f c + d ) λ 1 m λ 2 m λ 1 λ 2 ,
y 3 m = α ( c f + d ) ( λ 2 1 ) λ 1 m + 1 ( λ 1 1 ) λ 2 m + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( λ 2 1 ) ( λ 1 b f ) λ 1 m ( λ 1 1 ) ( λ 2 b f ) λ 2 m + ( λ 1 λ 2 ) ( 1 b f ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( λ 2 1 ) ( c λ 1 + b d ) λ 1 m ( λ 1 1 ) ( c λ 2 + b d ) λ 2 m + ( λ 1 λ 2 ) ( b d + c ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × z 1 b ( c f + d ) λ 1 m λ 2 m λ 1 λ 2 y 0 ( λ 1 b f ) λ 1 m ( λ 2 b f ) λ 2 m λ 1 λ 2 ,
y 3 m + 1 = α ( c f + d ) ( λ 2 1 ) λ 1 m + 1 ( λ 1 1 ) λ 2 m + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( λ 2 1 ) ( λ 1 b f ) λ 1 m + 1 ( λ 1 1 ) ( λ 2 b f ) λ 2 m + 1 + ( λ 1 λ 2 ) ( 1 b f ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( λ 2 1 ) ( c λ 1 + b d ) λ 1 m ( λ 1 1 ) ( c λ 2 + b d ) λ 2 m + ( λ 1 λ 2 ) ( b d + c ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × x 1 d ( λ 1 b f ) λ 1 m ( λ 2 b f ) λ 2 m λ 1 λ 2 z 0 ( c λ 1 + b d ) λ 1 m ( c λ 2 + b d ) λ 2 m λ 1 λ 2 ,
z 3 m 1 = α ( λ 2 1 ) ( f λ 1 + d g ) λ 1 m ( λ 1 1 ) ( f λ 2 + d g ) λ 2 m + ( λ 1 λ 2 ) ( d g + f ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( a f + g ) ( λ 2 1 ) λ 1 m ( λ 1 1 ) λ 2 m + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( λ 2 1 ) ( λ 1 a d ) λ 1 m ( λ 1 1 ) ( λ 2 a d ) λ 2 m + ( λ 1 λ 2 ) ( 1 a d ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × z 1 b ( f λ 1 + d g ) λ 1 m 1 ( f λ 2 + d g ) λ 2 m 1 λ 1 λ 2 y 0 ( a f + g ) λ 1 m λ 2 m λ 1 λ 2 ,
z 3 m = α ( λ 2 1 ) ( f λ 1 + d g ) λ 1 m ( λ 1 1 ) ( f λ 2 + d g ) λ 2 m + ( λ 1 λ 2 ) ( d g + f ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( a f + g ) ( λ 2 1 ) λ 1 m + 1 ( λ 1 1 ) λ 2 m + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( λ 2 1 ) ( λ 1 a d ) λ 1 m ( λ 1 1 ) ( λ 2 a d ) λ 2 m + ( λ 1 λ 2 ) ( 1 a d ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × x 1 d ( a f + g ) λ 1 m λ 2 m λ 1 λ 2 z 0 ( λ 1 a d ) λ 1 m ( λ 2 a d ) λ 2 m λ 1 λ 2 ,
z 3 m + 1 = α ( λ 2 1 ) ( f λ 1 + d g ) λ 1 m ( λ 1 1 ) ( f λ 2 + d g ) λ 2 m + ( λ 1 λ 2 ) ( d g + f ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × β ( a f + g ) ( λ 2 1 ) λ 1 m + 1 ( λ 1 1 ) λ 2 m + 1 + λ 1 λ 2 ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × γ ( λ 2 1 ) ( λ 1 a d ) λ 1 m + 1 ( λ 1 1 ) ( λ 2 a d ) λ 2 m + 1 + ( λ 1 λ 2 ) ( 1 a d ) ( λ 1 λ 2 ) ( λ 1 1 ) ( λ 2 1 ) × y 1 g ( λ 1 a d ) λ 1 m ( λ 2 a d ) λ 2 m λ 1 λ 2 x 0 ( f λ 1 + d g ) λ 1 m ( f λ 2 + d g ) λ 2 m λ 1 λ 2 ,
for m N 0 .
If ( a c f + a d + b f + c g ) 2 = 4 b d g , then
λ 1 = λ 2 = a c f + a d + b f + c g 2 .
Then, from (134), it follows that
a n = ( n + 1 ) λ 1 n ,
from which it follows that
u n = 1 + j = 1 n 1 ( j + 1 ) λ 1 j = 1 ( n + 1 ) λ 1 n + n λ 1 n + 1 ( 1 λ 1 ) 2 .
By using (146) and (147) in (34)–(42), and, after some calculations, we get
x 3 m 1 = α 1 c g + c g m λ 1 m 1 ( m ( 1 + c g ) + 1 c g ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × β a + b g b g m λ 1 m 1 + ( m ( b g a ) b g a ) λ 1 m + a m λ 1 m + 1 ( 1 λ 1 ) 2 γ ( a c + b ) 1 m λ 1 m 1 + ( m 1 ) λ 1 m ( 1 λ 1 ) 2 × x 1 d ( ( a λ 1 + b g ) m b g ) λ 1 m 2 z 0 ( a c + b ) m λ 1 m 1 ,
x 3 m = α 1 c g + c g m λ 1 m 1 ( m ( 1 + c g ) + 1 c g ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × β a + b g b g m λ 1 m 1 + ( m ( b g a ) b g a ) λ 1 m + a m λ 1 m + 1 ( 1 λ 1 ) 2 γ ( a c + b ) 1 ( m + 1 ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × y 1 g ( a c + b ) m λ 1 m 1 x 0 ( m ( λ 1 c g ) + λ 1 ) λ 1 m 1 ,
x 3 m + 1 = α 1 c g + c g ( m + 1 ) λ 1 m ( m ( 1 + c g ) + 2 ) λ 1 m + 1 + ( m + 1 ) λ 1 m + 2 ( 1 λ 1 ) 2 × β a + b g b g m λ 1 m 1 + ( m ( b g a ) b g a ) λ 1 m + a m λ 1 m + 1 ( 1 λ 1 ) 2 γ ( a c + b ) 1 ( m + 1 ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × z 1 b ( m ( λ 1 c g ) + λ 1 ) λ 1 m 1 y 0 ( ( a λ 1 + b g ) m + a λ 1 ) λ 1 m 1 ,
y 3 m 1 = α ( c f + d ) 1 m λ 1 m 1 + ( m 1 ) λ 1 m ( 1 λ 1 ) 2 × β 1 b f + b f m λ 1 m 1 ( m ( 1 + b f ) + 1 b f ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × γ c + b d b d m λ 1 m 1 + ( m ( b d c ) b d c ) λ 1 m + c m λ 1 m + 1 ( 1 λ 1 ) 2 × y 1 g ( ( c λ 1 + b d ) m b d ) λ 1 m 2 x 0 ( f c + d ) m λ 1 m 1 ,
y 3 m = α ( c f + d ) 1 ( m + 1 ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × β 1 b f + b f m λ 1 m 1 ( m ( 1 + b f ) + 1 b f ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × γ c + b d b d m λ 1 m 1 + ( m ( b d c ) b d c ) λ 1 m + c m λ 1 m + 1 ( 1 λ 1 ) 2 × z 1 b ( c f + d ) m λ 1 m 1 y 0 ( ( λ 1 b f ) m + λ 1 ) λ 1 m 1 ,
y 3 m + 1 = α ( c f + d ) 1 ( m + 1 ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × β 1 b f + b f ( m + 1 ) λ 1 m ( m ( 1 + b f ) + 2 ) λ 1 m + 1 + ( m + 1 ) λ 1 m + 2 ( 1 λ 1 ) 2 × γ c + b d b d m λ 1 m 1 + ( m ( b d c ) b d c ) λ 1 m + c m λ 1 m + 1 ( 1 λ 1 ) 2 × x 1 d ( ( λ 1 b f ) m + λ 1 ) λ 1 m 1 z 0 ( ( c λ 1 + b d ) m + c λ 1 ) λ 1 m 1 ,
z 3 m 1 = α f + d g d g m λ 1 m 1 + ( m ( d g f ) d g f ) λ 1 m + f m λ 1 m + 1 ( 1 λ 1 ) 2 × β ( a f + g ) 1 m λ 1 m 1 + ( m 1 ) λ 1 m ( 1 λ 1 ) 2 γ 1 a d + a d m λ 1 m 1 ( m ( 1 + a d ) + 1 a d ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × z 1 b ( ( f λ 1 + d g ) m d g ) λ 1 m 2 y 0 ( a f + g ) m λ 1 m 1 ,
z 3 m = α f + d g d g m λ 1 m 1 + ( m ( d g f ) d g f ) λ 1 m + f m λ 1 m + 1 ( 1 λ 1 ) 2 × β ( a f + g ) 1 ( m + 1 ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 γ 1 a d + a d m λ 1 m 1 ( m ( 1 + a d ) + 1 a d ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 × x 1 d ( a f + g ) m λ 1 m 1 z 0 ( ( λ 1 a d ) m + λ 1 ) λ 1 m 1 ,
z 3 m + 1 = α f + d g d g m λ 1 m 1 + ( m ( d g f ) d g f ) λ 1 m + f m λ 1 m + 1 ( 1 λ 1 ) 2 × β ( a f + g ) 1 ( m + 1 ) λ 1 m + m λ 1 m + 1 ( 1 λ 1 ) 2 γ 1 a d + a d ( m + 1 ) λ 1 m ( m ( 1 + a d ) + 2 ) λ 1 m + 1 + ( m + 1 ) λ 1 m + 2 ( 1 λ 1 ) 2 × y 1 g ( ( λ 1 a d ) m + λ 1 ) λ 1 m 1 x 0 ( ( f λ 1 + d g ) m + f λ 1 ) λ 1 m 1 ,
for m N 0 .
It is time-consuming, but not difficult to see, that all formulas above really present closed-form formulas for solutions to system (3) in the corresponding cases, which was checked by the author by hand. ☐
Corollary 1.
Assume that a , b , c , d , f , g Z , α , β , γ C   \ { 0 } , and x i , y i , z i , C   \ { 0 } , i { 0 , 1 } . Then, the following statements are true:
(a) 
If b = 0 and a c f + a d + c g 1 , then the general solution to system (3) is given by (55)–(63).
(b) 
If b = 0 and a c f + a d + c g = 1 , then the general solution to system (3) is given by (64)–(72).
(c) 
If d = 0 and a c f + b f + c g 1 , then the general solution to system (3) is given by (85)–(93).
(d) 
If d = 0 and a c f + b f + c g = 1 , then the general solution to system (3) is given by (94)–(102).
(e) 
If g = 0 and a c f + a d + b f 1 , then the general solution to system (3) is given by (115)–(123).
(f) 
If g = 0 and a c f + a d + b f = 1 , then the general solution to system (3) is given by (124)–(132).
(g) 
If b d g 0 and ( a c f + a d + b f + c g ) 2 4 b d g , then the general solution to system (3) is given by (137)–(145).
(h) 
If b d g 0 and ( a c f + a d + b f + c g ) 2 = 4 b d g , then the general solution to system (3) is given by (148)–(156).
Remark. 
Bearing in mind that system (3) is obtained from the one in [24] by adding some mulitipliers only, it is a natural question if (3) is its real generalization, that is, whether there are cases when the systems are not equivalent. If these two systems are equivalent, it is highly expected that the equivalence can be obtained by using scaling of dependent variables, that is, by using the transformation
( x n , y n , z n ) = ( λ 1 x ^ n , λ 2 y ^ n , λ 3 z ^ n ) , n N 0 ,
where λ j , j = 1 , 3 ¯ , are some nonzero numbers.
If we use (157) in (3), we get
λ 1 x ^ n + 1 = α λ 2 a λ 3 b y ^ n a z ^ n 1 b , λ 2 y ^ n + 1 = β λ 3 c λ 1 d z ^ n c x ^ n 1 d , λ 3 z ^ n + 1 = γ λ 1 f λ 2 g x ^ n f y ^ n 1 g ,
from which it follows that, for the equivalence of the systems, it must hold
λ 1 λ 2 a λ 3 b = α , λ 1 d λ 2 λ 3 c = β , λ 1 f λ 2 g λ 3 = γ .
However, the nonlinear system (158) need not have solutions if the determinant
Δ : = 1 a b d 1 c f g 1
is equal to zero (if λ j > 0 , j = 1 , 3 ¯ , and α , β , γ R + , the statement is immediately obtained by transforming system (158) to a linear one, by taking the logarithm; the case when some of these parameters are complex is dealt with in another way). It is easy to see that the equation Δ = 0 has many solutions in the set Z 6 . Hence, the systems, in general cases, are not equivalent.

Acknowledgments

I would like to express my sincere thanks to the referees and the editor of the paper for their comments, which improved the presentation of this paper.

Conflicts of Interest

The author declares that he has no conflicts of interest.

References

  1. Andruch-Sobilo, A.; Migda, M. On the rational recursive sequence xn+1 = axn−1/(b + cxnxn−1). Tatra Mt. Math. Publ. 2009, 43, 1–9. [Google Scholar]
  2. Berezansky, L.; Braverman, E. On impulsive Beverton-Holt difference equations and their applications. J. Differ. Equ. Appl. 2004, 10, 851–868. [Google Scholar] [CrossRef]
  3. Berg, L.; Stević, S. On some systems of difference equations. Appl. Math. Comput. 2011, 218, 1713–1718. [Google Scholar]
  4. Iričanin, B. On a higher-order nonlinear difference equation. Abstr. Appl. Anal. 2010, 2010. Article 418273. [Google Scholar] [CrossRef]
  5. Iričanin, B. The boundedness character of two Stević-type fourth-order difference equations. Appl. Math. Comput. 2010, 217, 1857–1862. [Google Scholar] [CrossRef]
  6. Jordan, C. Calculus of Finite Differences; Chelsea Publishing Company: New York, NY, USA, 1956. [Google Scholar]
  7. Karakostas, G. Asymptotic 2-periodic difference equations with diagonally self-invertible responces. J. Differ. Equ. Appl. 2000, 6, 329–335. [Google Scholar] [CrossRef]
  8. Krechmar, V.A. A Problem Book in Algebra; Mir Publishers: Moscow, Russia, 1974. [Google Scholar]
  9. Levy, H.; Lessman, F. Finite Difference Equations; Dover Publications: New York, NY, USA, 1992. [Google Scholar]
  10. Mitrinović, D.S.; Kečkić, J.D. Methods for Calculating Finite Sums; Naučna Knjiga: Belgrade, Serbia, 1984. (In Serbian) [Google Scholar]
  11. Papaschinopoulos, G.; Schinas, C.J. On a system of two nonlinear difference equations. J. Math. Anal. Appl. 1998, 219, 415–426. [Google Scholar] [CrossRef]
  12. Papaschinopoulos, G.; Schinas, C.J. On the behavior of the solutions of a system of two nonlinear difference equations. Comm. Appl. Nonlinear Anal. 1998, 5, 47–59. [Google Scholar]
  13. Papaschinopoulos, G.; Schinas, C.J. Invariants for systems of two nonlinear difference equations. Differ. Equ. Dyn. Syst. 1999, 2, 181–196. [Google Scholar] [CrossRef]
  14. Papaschinopoulos, G.; Schinas, C.J. Invariants and oscillation for systems of two nonlinear difference equations. Nonlinear Anal. TMA 2001, 46, 967–978. [Google Scholar] [CrossRef]
  15. Papaschinopoulos, G.; Schinas, C.J. Oscillation and asymptotic stability of two systems of difference equations of rational form. J. Differ. Equ. Appl. 2001, 7, 601–617. [Google Scholar] [CrossRef]
  16. Papaschinopoulos, G.; Schinas, C.J. On the system of two difference equations x n + 1 = i = 0 k A i / y n i p i , y n + 1 = i = 0 k B i / x n i q i . J. Math. Anal. Appl. 2002, 273, 294–309. [Google Scholar] [CrossRef]
  17. Papaschinopoulos, G.; Schinas, C.J. On the dynamics of two exponential type systems of difference equations. Comput. Math. Appl. 2012, 64, 2326–2334. [Google Scholar] [CrossRef]
  18. Papaschinopoulos, G.; Schinas, C.J.; Hatzifilippidis, V. Global behavior of the solutions of a max-equation and of a system of two max-equations. J. Comput. Anal. Appl. 2003, 5, 237–254. [Google Scholar]
  19. Papaschinopoulos, G.; Schinas, C.J.; Stefanidou, G. On the nonautonomous difference equation x n + 1 = A n + ( x n 1 p / x n q ) . Appl. Math. Comput. 2011, 217, 5573–5580. [Google Scholar]
  20. Papaschinopoulos, G.; Stefanidou, G. Asymptotic behavior of the solutions of a class of rational difference equations. Inter. J. Differ. Equ. 2010, 5, 233–249. [Google Scholar]
  21. Stefanidou, G.; Papaschinopoulos, G.; Schinas, C.J. On a system of max difference equations. Dyn. Contin. Discret. Impuls. Syst. Ser. A 2007, 14, 885–903. [Google Scholar]
  22. Stefanidou, G.; Papaschinopoulos, G.; Schinas, C.J. On a system of two exponential type difference equations. Commun. Appl. Nonlinear Anal. 2010, 17, 1–13. [Google Scholar]
  23. Stević, S. First-order product-type systems of difference equations solvable in closed form. Electron. J. Differ. Equ. 2015, 2015, 1–11, Article 308. [Google Scholar] [CrossRef]
  24. Stević, S. Product-type system of difference equations of second-order solvable in closed form. Electron. J. Qual. Theory Differ. Equ. 2015, 2015. Article 56. [Google Scholar] [CrossRef]
  25. Stević, S. New solvable class of product-type systems of difference equations on the complex domain and a new method for proving the solvability. Electron. J. Qual. Theory Differ. Equ. 2016, 2016. Article 120. [Google Scholar] [CrossRef]
  26. Stević, S. Solvable subclasses of a class of nonlinear second-order difference equations. Adv. Nonlinear Anal. 2016, 5, 147–165. [Google Scholar] [CrossRef]
  27. Stević, S. Solvable product-type system of difference equations whose associated polynomial is of the fourth order. Electron. J. Qual. Theory Differ. Equ. 2017, 2017. Article 13. [Google Scholar] [CrossRef]
  28. Stević, S.; Alghamdi, M.A.; Alotaibi, A.; Elsayed, E.M. Solvable product-type system of difference equations of second order. Electron. J. Differ. Equ. 2015, 2015, 1–20, Article 169. [Google Scholar] [CrossRef]
  29. Stević, S.; Alghamdi, M.A.; Alotaibi, A.; Shahzad, N. On positive solutions of a system of max-type difference equations. J. Comput. Anal. Appl. 2014, 16, 906–915. [Google Scholar]
  30. Stević, S.; Diblik, J.; Iričanin, B.; Šmarda, Z. On some solvable difference equations and systems of difference equations. Abstr. Appl. Anal. 2012, 2012. Article 541761. [Google Scholar] [CrossRef]
  31. Stević, S.; Iričanin, B.; Šmarda, Z. On a product-type system of difference equations of second order solvable in closed form. J. Inequal. Appl. 2015, 2015. Article 327. [Google Scholar] [CrossRef]
  32. Stević, S.; Iričanin, B.; Šmarda, Z. Solvability of a close to symmetric system of difference equation. Electron. J. Differ. Equ. 2016, 2016, 1–13, Article 159. [Google Scholar]
  33. Stević, S.; Iričanin, B.; Šmarda, Z. Two-dimensional product-type system of difference equations solvable in closed form. Adv. Differ. Equ. 2016, 2016. Article 253. [Google Scholar] [CrossRef]

Share and Cite

MDPI and ACS Style

Stević, S. Solvable Three-Dimensional Product-Type System of Difference Equations with Multipliers. Symmetry 2017, 9, 195. https://doi.org/10.3390/sym9090195

AMA Style

Stević S. Solvable Three-Dimensional Product-Type System of Difference Equations with Multipliers. Symmetry. 2017; 9(9):195. https://doi.org/10.3390/sym9090195

Chicago/Turabian Style

Stević, Stevo. 2017. "Solvable Three-Dimensional Product-Type System of Difference Equations with Multipliers" Symmetry 9, no. 9: 195. https://doi.org/10.3390/sym9090195

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop