Cyclic factorization numbers of finite groups
Marius T\u{a}rn\u{a}uceanu, Mihai-Silviu Lazorec

TL;DR
This paper introduces the cyclic factorization number of finite groups, providing explicit computations for key classes using Möbius inversion and subgroup analysis.
Contribution
It defines the cyclic factorization number and develops methods to compute it for various finite groups, advancing understanding of their subgroup structures.
Findings
Explicit formulas for cyclic factorization numbers in specific finite groups
Application of Möbius inversion to subgroup lattice analysis
Enhanced understanding of cyclic subgroup structures in finite groups
Abstract
In this paper we introduce and study the concept of cyclic factorization number of a finite group G. By using the Mobius inversion formula and other methods involving the cyclic subgroup structure, this is explicitly computed for some important classes of finite groups.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsFinite Group Theory Research · graph theory and CDMA systems
Cyclic factorization numbers
of finite groups
Marius Tărnăuceanu and Mihai-Silviu Lazorec
(February 5, 2017)
Abstract
In this paper we introduce and study the concept of cyclic factorization number of a finite group . By using the Möbius inversion formula and other methods involving the cyclic subgroup structure, this is explicitly computed for some important classes of finite groups.
MSC (2010): Primary 20D40; Secondary 20D60.
Key words: cyclic factorization number, cyclic subgroup commutativity degree, Möbius function.
1 Introduction
Let be a finite group, be the subgroup lattice of and be the poset of cyclic subgroups of . One of the most interesting positive integers that can be associated to is the factorization number . This counts the number of all pairs satisfying and has been investigated in many recent papers, such as [5, 12]. On the other hand, in our previous paper [9] (see also [11]) we defined the subgroup commutativity degree of as the proportion of the number of ordered pairs such that by . These two quantities are closely connected. More precisely, we have
[TABLE]
and
[TABLE]
by applying the well-known Möbius inversion formula. We also recall the following theorem due to P. Hall [2] (see also [3]), that permits us to compute explicitly the Möbius function of a finite -group.
Theorem 1.1. Let be a finite -group of order . Then unless is elementary abelian, in which case we have
Then, in [14], we defined the cyclic subgroup commutativity degree of by replacing with in the definition of . This also has a probabilistic significance, namely it measures the probability that two cyclic subgroups of commute. So, it is naturally to introduce a new positive integer that corresponds to in the same way as corresponds to . In the following we will denote by the number of all pairs satisfying and we will call it the cyclic factorization number of . Its study is the purpose of the current paper.
The paper is organized as follows. Some basic properties of cyclic factorization numbers are presented in Section 2. Section 3 deals with the computation of for some classes of finite groups. In the final section some further research directions and a list of open problems are indicated.
Most of our notation is standard and will usually not be repeated here. Elementary notions and results on groups can be found in [4, 7]. For subgroup lattice concepts we refer the reader to [6, 8, 13].
2 Basic properties of cyclic factorization
numbers
Let be a finite group. First of all, we remark that , and we have equality if and only if is cyclic. Consequently, for such a group the number is given by Theorem 1.1. of [5].
Proposition 2.1. Let be a finite cyclic group of order . Then
[TABLE]
Given two finite groups and , if then . By Proposition 2.1. we infer that the converse is not true. Also, we note that the weaker condition does not imply , as shows the following elementary example.
Example 2.2. It is well-known that the subgroup lattices of and are isomorphic. On the other hand, we can easily check that .
By a direct calculation, one obtains
[TABLE]
and therefore in general we don’t have
[TABLE]
A sufficient condition in order to this equality holds is that and be of coprime orders. This remark can naturally be extended to arbitrary finite direct products.
Proposition 2.3. Let be a family of finite groups having coprime orders. Then
[TABLE]
The following immediate consequence of Proposition 2.3. shows that the computation of for a finite nilpotent group can be reduced to -groups.
Corollary 2.4. If is a finite nilpotent group and are the Sylow subgroups of , then
[TABLE]
Remarks 2.5.
- a)
The conclusion of Corollary 2.4. fails if is not nilpotent. For example, we have even if for every Sylow subgroup of and , respectively. Notice also that more can be said about such groups, namely for all .
- b)
Let be a prime and be a finite -group. It is well-known (see e.g. [4], I) that can be written as a product of two cyclic subgroups if and only if it is metacyclic. This can be reformulated in the following nice way: ”A finite -group with is metacyclic if and only if its cyclic factorization number is non-zero”.
Finally, we observe that the connections between and are similar with (1) and (2), more exactly
[TABLE]
and
[TABLE]
The equality (4) is the main ingredient that will be used in Section 3 to calculate the cyclic factorization numbers of certain finite groups.
3 Cyclic factorization numbers for some classes of finite groups
As we already have seen, the computation of cyclic factorization numbers of finite abelian groups is reduced to -groups. By the fundamental theorem of finitely generated abelian groups, such a group is of type
[TABLE]
where . Recall that possesses a unique maximal elementary abelian subgroup . Moreover, all elementary abelian subgroups of are contained in . Notice also that we have , for every . We are now able to compute explicitly .
Theorem 3.1. The cyclic factorization number of the finite abelian -group G\cong{{\mathrel{\mathop{{\buildrel{k}\over{\mbox{\Huge\times}}}}\limits_{{i=1}}}{}\!}{}\!}\mathbb{Z}_{p^{\alpha_{i}}}, , is given by the following equality:
[TABLE]
Proof. By using (4) and Theorem 1.1., one obtains
[TABLE]
[TABLE]
For we have
[TABLE]
while for we have
[TABLE]
[TABLE]
by Theorem 4.2 of [10].
In the case the desired equality follows by a direct calculation from (5) and Theorem 4.3. of [10], or by observing that cannot be generated by two elements (and consequently cannot be written as a product of two cyclic subgroups). This completes the proof.
In the second part of this section we will focus on computing the cyclic factorization number of other classes of finite groups starting with the dihedral groups
[TABLE]
The subgroup structure of is the following: for every divisor or , possesses a subgroup isomorphic to , namely , and subgroups isomorphic to , namely , We easily infer that
[TABLE]
where denotes the number of divisors of . On the other hand, we know that
[TABLE]
by Theorem 3.2.1. of [14]. Also, the values of the Möbius function associated to have been determined in Lemma 19 of [1]:
[TABLE]
for all divisors of and all We are now able to complete the computation of .
Theorem 3.2. The cyclic factorization number of the dihedral group , , is given by the following equality:
[TABLE]
Proof. By using (6)-(8) in (4), one obtains
[TABLE]
[TABLE]
[TABLE]
[TABLE]
where
[TABLE]
We observe that
[TABLE]
and consequently
[TABLE]
as desired.
Remark 3.3. An alternative and more facile way of computing uses its definition. Clearly, we have
[TABLE]
We easily infer that is the product of two cyclic subgroups if and only if one of them is and the other is of type with . Hence .
The computation of the factorization numbers of the dihedral group and of the abelian -group G\cong{{\mathrel{\mathop{{\buildrel{k}\over{\mbox{\Huge\times}}}}\limits_{{i=1}}}{}\!}{}\!}\mathbb{Z}_{p^{\alpha_{i}}}, , will be extremely helpful in our study. These results will be used to find the cyclic factorization numbers of the following 5 groups:
- the generalized quaternion group
[TABLE]
- the quasi-dihedral group
[TABLE]
- the modular -group
[TABLE]
- the dicyclic group
[TABLE]
- the generalized dicyclic group
[TABLE]
where is an abelian group of order isomorphic to
As we saw in the alternative proof concerning the cyclic factorization number of the dihedral group , it is important to know the cyclic subgroup structure of the group we study. Hence, it is useful to recall that the poset of cyclic subgroups of is
[TABLE]
We are ready to compute the factorization number of the generalized quaternion group.
Theorem 3.4. The cyclic factorization number of the generalized quaternion group , , is given by
[TABLE]
Proof. By inspecting the subgroup lattice of , we infer that . Let be a positive integer. The unique minimal subgroup of the generalized quaternion group is its center . Also, it is known that
[TABLE]
Let and be two nontrivial cyclic subgroups of such that . Then the two subgroups contain and by the above isomorphism, the cyclic factorization of corresponds to the cyclic factorization of . Also, it is important to remark that by quotiening non-cyclic subgroups of by the center of the group, we do not obtain corresponding cyclic subgroups of . In fact, by this procedure, we always get a dihedral group or a direct product of two abelian groups, namely . With this being checked, we are sure that there are no cyclic factorizations of that could correspond to non-cyclic factorization of . We add that it is obvious that we can not build another cyclic factorization of the generalized quaternion group using its trivial subgroup as a factor. Hence and our proof is complete.
We will see that it may happen that by quotiening we can find a non-cyclic factorization of a group which corresponds to a cyclic factorization of its factor. The class of quasi-dihedral groups provides such an example. First of all, we recall that the cyclic subgroups of are illustrated by the poset
[TABLE]
Using this information, we can find the factorization number of the quasi-dihedral group.
Theorem 3.5. The cyclic factorization number of the quasi-dihedral group , , is given by
[TABLE]
Proof. Besides the center , the quasi-dihedral group has other minimal subgroups contained in the set . Moreover any nontrivial subgroup of contains the center or it is contained in the set . Let and be two cyclic subgroups such that , and . In this case, to count the cyclic factorizations of , we will use the isomorphism
[TABLE]
This result implies that a cyclic factorization of corresponds to a cyclic factorization of . However, the quasi-dihedral group contains subgroups isomorphic to and their images through the above isomorphism are cyclic subgroups of isomorphic to . Each such subgroup is not contained in the maximal cyclic subgroup of and this leads us to cyclic factorizations of which correspond to non-cyclic factorizations of . We remark that for other non-cyclic subgroups of the quasi-dihedral group , we obtain non-cyclic subgroups of the dihedral group after quotiening by . It is clear that we can not build a cyclic factorization of using two subgroups contained in . Hence, the remaining option is to find cyclic factorizations , where and contains . Since, in this case, is trivial, we have only one choice for , this being , which is the only maximal cyclic subgroup of . By a simple computation, we get
[TABLE]
as desired.
The class of modular -groups provided interesting probabilistic aspects such as . The next purpose of this paper is to find an explicit formula for the cyclic factorization numbers of these groups.
Theorem 3.6. The cyclic factorization number of the modular p-group , , is given by
[TABLE]
Proof. Using Theorem 3.2. and the isomorphism , we infer that . Let and be a prime number and a positive integer, respectively. By inspecting the lattice of subgroups of , we remark that this group has minimal subgroups isomorphic to , one of them being the commutator subgroup which is contained in each subgroup of higher order. Again, we will use an isomorphism to count the cyclic factorization number of the modular -group. Indeed, it is known that
[TABLE]
Let be a cyclic factorization of such that and . Through the above isomorphism, there is a corresponding cyclic factorization of , namely . There are non-cyclic proper subgroups of , these being isomorphic to , where . Quotiening these subgroups by , we obtain only one cyclic subgroup isomorphic to a minimal subgroup of . This subgroup can be used to obtain cyclic factorizations of since it is not contained in any maximal cyclic subgroup of this group. Hence, there are cyclic factorizations of corresponding to non-cyclic factorizations of . The only method left to find other cyclic factorizations of is by using the minimal subgroups different from as a factor and the subgroups that contain as the other factor. In this case is trivial, so the second factor must be one of the maximal cyclic subgroups of which are isomorphic to , namely and , where . Adding up the numbers and using Theorem 3.1., we have
[TABLE]
as stated above.
A simplified formula is obtained for the modular 2-groups.
Corollary 3.7. The cyclic factorization number of the modular 2-group , , is given by
[TABLE]
Some probabilistic aspects of the dicyclic groups were studied in [15]. We know that for each divisor of , the dicyclic group has only one subgroup isomorphic to , namely . Also, for each divisor of , possesses subgroups isomorphic to , namely , where . The poset of cyclic subgroups is
[TABLE]
We mention that, by definition, . Also, it is easy to see that . Then and Finding the factorization numbers for all other dicyclic groups is our next purpose.
Theorem 3.8. The cyclic factorization number of the dicyclic group , , is given by
[TABLE]
Proof. It is known that which is isomorphic to and the following isomorphism holds
[TABLE]
We remark that by quotiening non-cyclic subgroups of we do not obtain cyclic subgroups of , so there will be no need to substract a number of factorizations from to find , like we did in the case of other classes of groups. If , where is a positive integer, since , we have according to Theorem 3.3.. If is even and is not a power of 2, a nontrivial subgroup of contains the center of the group or it is a subgroup of of odd order. Let be a cyclic factorization of such that and . Then there is a corresponding cyclic factorization of the dihedral group . If is a nontrivial cyclic subgroup of odd order contained in the maximal subgroup of , then we can use it to form a cyclic factorization if the other factor, , is a cyclic subgroup which is not contained in . Looking at the poset , the only possible choices for are the cyclic subgroups of order 4 from the set . This leads us to , which is false since is even and is odd. Hence, if , we obtain that
[TABLE]
The same reasoning is used if is an odd number. The only change is that, in this case, we can find other cyclic factorizations of besides those corresponding to the ones of . Let be one of the subgroups contained in the set . Again, if is a nontrivial cyclic subgroup of odd order contained in , then is a cyclic factorization of if . According to the subgroup structure of the dicyclic group, there is only one such subgroup isomorphic to . It follows that
[TABLE]
and our proof is complete.
The last class of groups we study is formed by the generalized dicyclic groups , where . Let and be the generators of the cyclic groups and , respectively. Since and , we infer that . The subgroup strucute and explicit formulas for the (cyclic) subgroup commutativity degree of are provided by [15] and our final purpose in this paper is to compute its cyclic factorization number.
Theorem 3.9. *Let be an abelian group, where , and is a positive odd integer. Then,
(i) if and or if is any odd number and , the cyclic factorization number of the generalized dicyclic group is*
[TABLE]
(ii) if and , the cyclic factorization number of the generalized dicylic group is
[TABLE]
(iii) if and , the generalized dicyclic group is isomorphic to the direct product and its cyclic factorization number is
[TABLE]
(iv) if and , the generalized dicyclic group is isomorphic to the abelian group and its cyclic factorization number is
[TABLE]
Proof. The cyclic subgroups of are those contained in the abelian group and the subgroups contained in the set if , respectively in the set if .
(i) Let a positive odd number and . Since and, in this case, we can not form a cyclic factorization using only cyclic subgroups contained in which are of order 4 and do not intersect trivially. Hence a cyclic factorization of is formed by a cyclic subgroup of and a cyclic subgroup belonging to . This leads us to and since , it is clear that . In other words, is a maximal cyclic subgroup of . It is easy to see that has 2 maximal cyclic subgroups, namely and . We remark that the intersection of each one of these subgroups with any subgroup contained in is trivial. Hence, If and , the reasoning is similar, the only change being that contains 3 maximal cyclic subgroups, namely and . It is easy to see that only two of these subgroups trivially intersect the subgroups belonging to as we choose to be or , so we obtain again that If and , the cyclic factorizations of are formed by the cyclic subgroups contained in and two of the three maximal cyclic subgroups contained in , namely and . Therefore, in all the mentioned cases, the cyclic factorization number of the generalized dicyclic group is
[TABLE]
(ii) Let and . The only possible cyclic factorizations of are formed by the factors such that and . We saw that has two maximal cyclic subgroups, but since their intersection with the cyclic subgroups contained in is , it is clear that
[TABLE]
(iii) Let and . For more details about the mentioned isomorphism, we refer the reader to [15]. Denote by and the elements and , respectively. We get the following isomorphism
[TABLE]
where This isomorphism implies that the group contains the subgroup lattice of obtained by quotiening by . In the same paper, we saw that besides the cyclic subgroups contained in , the group has other cyclic subgroups of order 2, namely and , where . Since a factor of a possible cyclic factorization of is one of the above mentioned cyclic subgroups of order 2, the other factor must be a cyclic subgroup contained in such that . However, , so there is no such . It follows that
[TABLE]
A nontrivial subgroup of contains or it is one of the minimal subgroups or . Assume that is a cyclic factorization of such that and . Then there exists a corresponding cyclic factorization of , which is false since we proved that . It is clear that other possible cyclic factorization of is formed by a factor being or and a factor that contains . Since , it follows that the order of the cyclic subgroup is or . However, the group do not have any element that could generate such a cyclic subgroup. Therefore, we proved that
[TABLE]
(iv) Let and . It is straightforward to check the isomorphism . Also, by Theorem 3.1., we have
[TABLE]
as desired.
4 Conclusions and further research
All our previous results show that the cyclic factorization number constitutes an interesting computational aspect of finite groups. It is clear that the study started here can successfully be extended to other classes of finite groups. This will surely be the subject of some further research.
We end our paper by formulating several open problems concerning this topic.
Problem 4.1. Compute explicitly the cyclic factorization number of a finite Zassenhaus metacyclic group (see e.g. [4], I).
Problem 4.2. Determine the class of finite groups/-groups for which .
Problem 4.3. Study other connections between the cyclic factorization number and the factorization number of a finite group.
Problem 4.4. Given a finite non-nilpotent group , describe the manner in which depends on the cyclic factorization numbers of the Sylow subgroups of .
Problem 4.5. What can be said about two finite groups of order having the same cyclic factorization number ?
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] P.J. Cameron, H.R. Maimani, G.R. Omidi, B. Tayfeh-Rezaie, 3 3 3 -Designs from P S L ( 2 , q ) 𝑃 𝑆 𝐿 2 𝑞 PSL(2,q) , Discrete Math. 306 (2006), 3063-3073.
- 2[2] P. Hall, A contribution to the theory of groups of prime-power order , Proc. London Math. Soc. 36 (1933), 29-95.
- 3[3] T. Hawkes, I.M. Isaacs and M. Özaydin, On the Möbius function of a finite group , Rocky Mountain J. Math. 19 (1989), 1003-1033.
- 4[4] B. Hupert, Endliche Gruppen , I, II, Springer Verlag, Berlin, 1967, 1968.
- 5[5] F. Saeedi and M. Farrokhi, Factorization numbers of some finite groups , Glasgow Math. J. 54 (2012), 345-354.
- 6[6] R. Schmidt, Subgroup lattices of groups , de Gruyter Expositions in Mathematics 14, de Gruyter, Berlin, 1994.
- 7[7] M. Suzuki, Group theory , I, II, Springer Verlag, Berlin, 1982, 1986.
- 8[8] M. Tărnăuceanu, Groups determined by posets of subgroups , Ed. Matrix Rom, Bucureşti, 2006.
