On the Periods of 2-Step General Fibonacci Sequences in the Generalized Quaternion Groups
Abstract
We study 2-step general Fibonacci sequences in the generalized quaternion groups Q4n. In cases where the sequences are proved to be simply periodic, we obtain the periods of 2-step general Fibonacci sequences.
1. Introduction
The study of the Fibonacci sequences in groups began with the earlier work of Wall [1] in 1960, where the ordinary Fibonacci sequences in cyclic groups were investigated. In the mid-eighties, Wilcox [2] extended the problem to the abelian groups. In 1990, Campbell et al. [3] expanded the theory to some classes of finite groups. In 1992, Knox proved that the periods of k-nacci (k-step Fibonacci) sequences in the dihedral groups are equal to 2k + 2, in [4]. In the progress of this study, the article [5] of Aydin and Smith proves that the lengths of the ordinary 2-step Fibonacci sequences are equal to the lengths of the ordinary 2-step Fibonacci recurrences in finite nilpotent groups of nilpotency class 4 and a prime exponent, in 1994.
Since 1994, the theory has been generalized and many authors had nice contributions in computations of recurrence sequences in groups and we may give here a brief of these attempts. In [6, 7] the definition of the Fibonacci sequence has been generalized to the ordinary 3-step Fibonacci sequences in finite nilpotent groups. Then in [8] it is proved that the period of 2-step general Fibonacci sequence is equal to the length of the fundamental period of the 2-step general recurrence constructed by two generating elements of a group of nilpotency class 2 and exponent p. In [9] Karaduman and Yavuz showed that the periods of the 2-step Fibonacci recurrences in finite nilpotent groups of nilpotency class 5 and a prime exponent are p · k(p), for 2 < p ≤ 2927, where p is a prime and k(p) is the period of the ordinary 2-step Fibonacci sequence. The main role of [10, 11] in generalizing the theory was to study the 2-step general Fibonacci sequences in finite nilpotent groups of nilpotency class 4 and exponent p and to the 2-step Fibonacci sequences in finite nilpotent groups of nilpotency class n and exponent p, respectively.
One may consult [12, 13] to see the results of the Fibonacci sequences in the modular groups concerning the periodicity of 2-step Fibonacci sequences constructed by two generating elements.
Going on a detailed literature in this area of research, we have to mention certain essential computation on the Fibonacci lengths of new structures like the semidirect products, the direct products, and the automorphism groups of finite groups which have been studied in [14–19]. Finally, we refer to [20] where Karaduman and Aydin studied the periodicity property of 2-step general Fibonacci sequences in dihedral groups and the goal of this paper is to calculate the periods of 2-step general Fibonacci sequences in the generalized quaternion groups.
Theorem 1.1 (see [4].)Every k-nacci sequence in a finite group is simply periodic.
2. Main Theorems
First we consider the following lemma which will be used frequently without further reference.
Lemma 2.1. For an integer r, where 1 ≤ r < 2n, and for s ∈ {1,3}, the following relations hold in Q4n:
- (i)
, where t is any even integer;
- (ii)
, where t is any odd integer.
Proof. We may use an induction method on t for both parts, simultaneously.
Theorem 2.2. Pk(Q4n; a, b) = 2k + 2.
Proof. Obviously, the order of ab is 4. If k = 2, then
Let G = 〈a, b〉 be a finite nonabelian 2-generated group. A 2-step general Fibonacci sequence in the group G is defined by , for i ≥ 2, and the integers m and l. Now we study this sequence for group Q4n.
Theorem 2.3. Let m and l be integers. If m ≡ 0 (mod 2n) or m ≡ 0 (mod 4), then 2-step general Fibonacci sequence in Q4n is not simply periodic.
Proof. First we consider the case m ≡ 0 (mod 2n). Then the sequence is
Now let m ≡ 0 (mod 4). Four cases occur.
Case 1 (l ≡ 0 (mod 4)). Then
Since the cycle does not begin again with a and b, the sequence is not simply periodic.
Case 2 (l ≡ 1 (mod 4)). Then
Case 3 (l ≡ 2 (mod 4)). Note that b2 is a central element of Q4n. Thus
Case 4 (l ≡ 3 (mod 4)). So
Theorem 2.4. Let n be an even integer.
- (i)
If m ≡ 1 (mod 2n) and l ≡ 0 (mod 2n), then 2-step general Fibonacci sequence in Q4n is simply periodic with period 2.
- (ii)
If m ≡ −1 (mod 2n) and l ≡ −1 (mod 2n), then 2-step general Fibonacci sequence in Q4n is simply periodic with period 3.
- (iii)
If m ≡ 1 (mod 2n) and l ≡ 1 (mod 2n), then 2-step general Fibonacci sequence in Q4n is simply periodic with period 6.
Proof. (i) Since n is even, m ≡ 1 (mod 4) and l ≡ 0 (mod 4). Thus
(ii) Since n is even, m ≡ 3 (mod 4) and l ≡ 3 (mod 4). Note that m and l are odd integers. So the sequence reduces to
(iii) Because n is even, m ≡ 1 (mod 4) and l ≡ 1 (mod 4). Furthermore, m and l are odd integers. Then the sequence reduces to
Theorem 2.5. Let n be an odd integer.
- (i)
If m ≡ 1 (mod 2n), m ≡ 1 (mod 4), l ≡ 0 (mod 4) and (l, n) = 1, then 2-step general Fibonacci sequence in Q4n is simply periodic with period 2n.
- (ii)
If m ≡ 1 (mod 2n), m ≡ 3 (mod 4), l ≡ 0 (mod 4) and (l, n) = 1, then 2-step general Fibonacci sequence in Q4n is simply periodic with period 4n.
Proof. (i) By induction on k ≥ 0, we can show that x2k = a and x2k+1 = a−klb. In particular, the period must be even. Now we have
(ii) By induction on k ≥ 0, it may be shown that
Theorem 2.6. Let n and l be odd integers, m ≡ −1 (mod 2n) and (l + 1, n) = 1.
- (i)
If m ≡ 1 (mod 4), then 2-step general Fibonacci sequence in Q4n is simply periodic with period 6n.
- (ii)
If m ≡ 3 (mod 4) and l ≡ 1 (mod 4), then 2-step general Fibonacci sequence in Q4n is simply periodic with period 6n.
- (iii)
If m ≡ 3 (mod 4) and l ≡ 3 (mod 4), then 2-step general Fibonacci sequence in Q4n is simply periodic with period 3n.
Proof. (i) Two cases occur.
Case 1 (l ≡ 1 (mod 4)). By induction on k ≥ 0, we can prove that the following relations hold:
Case 2 (l ≡ 3 (mod 4)). The proof is similar to Case 1 except that
(ii) It is easily shown that the following relations hold for every k ≥ 0:
(iii) By induction on k ≥ 0, we can prove the following relations:
Acknowledgments
The authors would like to express their appreciations to the referees for the valuable and constructive comments regarding the presentation of this paper.