Department of Mathematics and Applied Mathematics

J1과 J2는 두 개의 유니터리 대칭군입니다. 이 논문에서는 J1과 J2가 특정 세 쌍에 대해 유일하게 생성되는지 여부를 증명합니다. 또한, J1이 주어진 세 쌍에 대한 생성자로 사용할 수 있는 가능한 세 쌍을 모두 찾습니다.

J1은 (2, 3, 5), (2, 3, 7), (2, 3, 11), (2, 3, 19), (2, 5, 7), (2, 5, 11), (2, 5, 19), (2, 7, 11), (2, 7, 19), (2, 11, 19), (3, 5, 7), (3, 5, 11), (3, 5, 19), (3, 7, 11), (3, 7, 19), (3, 11, 19), (5, 7, 11), (5, 7, 19), (5, 11, 19), (7, 11, 19)로 생성될 수 있는 세 쌍이 모두 존재한다는 것을 보여줍니다.

J1은 (2, 3, 19)는 유일하게 생성되는 세 쌍입니다.

J2의 경우, J1과 마찬가지로 다양한 세 쌍에 대해 유니크한 생성자를 찾을 수 있습니다.

논문에서는 이 문제를 해결하기 위해 여러 가지 방법을 사용합니다. 예를 들어, J1의 특정 두 개의 대역에 대한 유일한 생성자를 찾는 데 사용된 방법은, J1이 특정 세 쌍에 대해 유니크하게 생성될 수 있음을 보여주는 것을 포함합니다.

논문에서 제공하는 증명들은 이 문제를 해결하기 위해 사용된 기법과 결과에 대한 명확한 이해를 제공하며, 이러한 정보는 다른 관련 문제에 적용할 수 있습니다.

Department of Mathematics and Applied Mathematics

arXiv:math/9306201v1 [math.GR] 10 Jun 1993(p, q, r)-GENERATIONS FOR THE JANKO GROUPS J1AND J2JAMSHID MOORIDepartment of Mathematics and Applied MathematicsUniversity of NatalP.O.Box 375, Pietermaritzburg3200 South Africa1. IntroductionSuppose that G is a group which is generated by two elements a andb such that o(a) = l and o(b) = m and their product ab has order n.We say that G is a (l, m, n)-generated group and it is obvious that inthis case G must be a quotient of the triangle group ∆(l, m, n) givenby the presentation∆(l, m, n) =< a, b, c | al = bm = cn = abc = 1 > .If G is a (l, m, n)-generated group, then it is also a (r, s, t)-generatedgroup whenever (r, s, t) is a rearrangement of (l, m, n).

Thus we mayassume that l ≤m ≤n. It is well-known that ∆(l, m, n) is finite if andonly if 1/l + 1/m + 1/n > 1.

(see [11] and [4]). Finite ∆(l, m, n) are∆(1, n, n) the cyclic group of order n, ∆(2, 2, n) the dihedral group oforder 2n, ∆(2, 3, 3) the alternating group A4, ∆(2, 3, 4) the symmetricgroup S4, and ∆(2, 3, 5) the alternating group A5.

If 1/l+1/m+1/n =1, namely in the cases of ∆(2, 3, 6), ∆(2, 4, 4) and ∆(3, 3, 3), then thetriangle group is infinite but soluble. In the case when 1/l+1/m+1/n <1, then the triangle group is infinite and insoluble.These trianglegroups have a remarkable wealth of interesting finite quotient groups.We encourage reader to consult [4], [14] and [15] for discussion andbackground material on the triangle groups.

A (2, 3, 7)-generated groupG which give rise to compact Riemann surfaces of genus greater than 2with automorphism groups of maximal order, are called Hurwitz groups([7] and [15]).Woldar in [16] determined that J1, J2, He, Ru, Co3, HN, Ly are Hur-witz, while M11, M12, M22, M23, HS, J3, M24, McL, Suz, O’N, Co1, Co21991 Mathematics Subject Classification. 20D08, 20F05.supported by a research grant from the University of Natanl.1

2JAMSHID MOORIare non-Hurwitz. In [8] Kleidman, Parker and Wilson proved that F23is not Hurwitz, while Linton [9] and Wilson [10] showed that Th andF ′24 are Hurwitz.

In [17] Woldar completes the problem, except for Band M where the question is unresolved, by proving that J4 andF22 areHurwitz.If G is a finite group, C1, C2, C3 conjugate classes of G, and c a fixedrepresentative of C3, then ξG(C1, C2, C3) denotes the structure constantof the group algebra Z(C[G]) which is equal to the number of distinctordered pairs (a, b) with a ∈C1, b ∈C2 and ab = c. Using the charactertable of G, the number ξG(C1, C2, C3) can be calculated by the formulaξG(C1, C2, C3) = |C1|.|C2||G|kXi=1χi(a)χi(b)χi(c)χi(1)where χ1, χ2, . .

. , χk are the irreducible complex characters of G. Letξ∗G(C1, C2, C3) denote the number of distinct ordered pairs (a, b) witha ∈C1, b ∈C2, ab = c and G =< a, b >.

Obviously G is a (l, m, n)-generated group if and only if there exist conjugacy classes C1, C2, C3with representatives a, b, c, respectively such that o(a) = l, o(b) = mand o(c) = n, for which ξ∗G(C1, C2, C3) > 0. In this case we say that Gis (C1, C2, C3)-generated.

If H is a subgroup of G containing c and Bis a conjugacy class of H such that c ∈B, then σH(C1, C2, B) denotesthe number of distinct ordered pairs (a, b) such that a ∈C1, b ∈C2,ab = c and < a, b >≤H.It is well-known that every finite non-abelian simple group can begenerated by two suitable elements. (For details see [5]).

If G is non-abelian finite simple group and l, m, n are divisors of |G| such that1/l+1/m+1/n < 1, then the following question arises: Is G a (l, m, n)-generated group? We are a great distance away from a complete answerto the question, however we are aiming at a partial answer to it byconsidering the case when l = p, m = q and n = r are primes and thegroup G is isomorphic to one of the sporadic simple groups.In the present paper we investigate (p, q, r)-generations for the Jankgroups J1 and J2 where p, q and r are distinct primes satisfying p < q < r.We prove the following results.Theorem 2.8 .

The group J1 is (p, q, r)-generated for p, q, r ∈{2, 3, 5, 7, 11, 19}with p < q < r, except when (p, q, r) = (2, 3, 5).Theorem 2.9 . The group J1 is generated by three involutions a, b, c ∈2A such that abc ∈11A.Theorem 3.5 .

(a) The group J2 is (2B, 3B, 7A)-, (2A, 5C, 7A)-, (2B, 5A, 7A)-, (2B, 5C, 7A)-,(3A, 5C, 7A)-, (3B, 5A, 7A)-, and (3B, 5C, 7A)-generated. (b) The group J2 is not (2A, 3A, 7A)-, (2A, 3B, 7A)-, (2B, 3A, 7A)-,(2A, 5A, 7A)-, or (3A, 5A, 7A)-generated.

GENERATIONS FOR J1 AND J23The remaining cases , namely when p = q = r, p ̸= q = r orp = q ̸= r will be treated separately in another paper by the author.For the description of the conjugacy classes, the character tables andinformation on the maximal subgroups readers are referred to ATLAS[3]. Computations were performed with the aid of CAYLEY and GAP(see [1] and [13]) running on a SUN GX2 computer.1.1.

Acknowledgments. I am very grateful to the referee for his valu-able and constructive remarks, and to Jonathan Hall for several of thearguments in this paper.

I also thank Andrew Woldar for his privatecommunication on the Lemma 2.1.2. (p, q, r)-generations for J1We list bellow the structure constants for the Janko group J1.ξJ1(C1, C2, C3)C1C2C3492A3A7A552A3A11A382A3A19A492A5A7A442A5A11A572A5A19A2092A7A11A2092A7A19A1332A11A19A1893A5A7A1983A5A11A1713A5A19A8583A7A11A8363A7A19A4943A11A19A8585A7A11A8365A7A19A5135A11A19A22997A11A19AWe would also like to mention here that if G is a (p, q, r)-generatedgroup where p, q and r are distinct primes, then G has no solublequotient.

This is an obvious consequence of the following elementaryresult:Lemma 2.1. Let G be a (l, m, n)-generated group with l, m, n pairwiseco-prime.

Then G has no soluble quotient.

4JAMSHID MOORIProof. Assume that G is generated by a, b such that o(a) = l, o(b) = mand o(ab) = n. If G has soluble quotient, then it also has a non-trivialabelian quotient A.

Let ¯a,¯b and ab be the respective images of a, b andab in A. Then o(ab) divides lm and n. Since lm and n are relativelyprime, ab = ¯e, where e is the identity of G. This implies that ¯a = (¯b)−1.So if ¯b ̸= ¯e then o(b) divides both l and m, a contradiction to co-primeness.But if ¯b = ¯e, then ¯a = ¯e.Hence A =< ¯a,¯b >= {¯e},contradicting the non-triviality of A.

This completes the proof of thelemma.We now deal with the (p, q, r)-generations of J1.Lemma 2.2. The group J1 is (2, 3, 7)-generated.Proof.

The fact that J1 is (2A, 3A, 7A)-generated is given by Woldarin [16].Lemma 2.3. The group J1 is (2, 3, 11)-generated.Proof.

The only non-soluble maximal subgroups of J1 with order di-visible by 11 are isomorphic to the group L2(11).If H ≤J1 withH ∼= L2(11), thenσH(2A, 3A, 11A) = ξH(2A, 3A, 11A) = 11.We also observe that NH(11A) = 11:5 and NJ1(11A) = (11:5):2. Let2 =< α >.

Then a fixed element in 11A is contained in H and Hg forg ∈J1 −H if and only if g = h−1βα for some h ∈H and β ∈NH(11A).Thus a fixed element of order 11 in J1 is contained in precisely twocopies of H, namely H and Hα. Hence H and its conjugates contribute11×2 = 22 to the ξJ1(2A, 3A, 11A).

Since ξJ1(2A, 3A, 11A) = 55 > 22,ξ∗J1(2A, 3A, 11A) ≥33. Therefore J1 is (2A, 3A, 11A)-generated.Lemma 2.4.

The group J1 is (2, 3, 19)-generated.Proof. The only maximal subgroup of J1 with order divisible by 19, upto isomorphism, is the group 19 : 6.

Since a (2, 3, 19)-generated groupcannot have a soluble quotient, J1 posses no proper (2, 3, 19)-generatedsubgroup. Now the result follows from the fact that ξJ1(2A, 3A, 19A) =ξ∗J1(2A, 3A, 19A) = 38.Lemma 2.5.

The group J1 is (2, 5, 11)-generated.Proof. Up to isomorphism, L2(11) is the only non-soluble maximal sub-group of J1 with order divisible by 11 × 10 = 110.We also haveσL2(11)(2A, 5A, 11A) = ξL2(11)(2A, 5A, 11A) = 11, and as in Lemma 2.3,L2(11) and it’s conjugates contribute 11×2 = 22 to the ξJ1(2A, 5A, 11A).Since ξJ1(2A, 5A, 11A) = 55, result follows.

GENERATIONS FOR J1 AND J25Lemma 2.6. The group J1 is (2, 5, 7)-, (2, 5, 19)-, (2, 7, 11)-, (2, 7, 19)-,(2, 11, 19)-, (5, 7, 11)-, (5, 7, 19)-, (5, 11, 19)-, (3, 5, 7)-, (3, 5, 19)-,(3, 7, 11)-, (3, 7, 19)-, (3, 11, 19)-, and (7, 11, 19)-generated.Proof.

The group J1 has no proper subgroups of order divisible by2 × 5 × 7, 2 × 5 × 19, 2 × 7 × 11, 2 × 7 × 19, 2 × 11 × 19, 5 × 7 × 11,5 × 7 × 19 , 5 × 11 × 19, 3 × 5 × 7, 3 × 5 × 19, 3 × 7 × 11, 3 × 7 × 19,3 × 11 × 19, and 7 × 11 × 19 respectively. Since ξJ1(C1, C2, C3) for thecorresponding conjugacy classes are positive, the result follows.Lemma 2.7.

The group J1 is (3, 5, 11)-generated.Proof. Up to isomorphism, L2(11) is the only maximal subgroup ofJ1 with order divisible by 3 × 5 × 11.Now σL2(11)(3A, 5A, 11A) =ξL2(11)(3A, 5A, 11A) = 22 implies that L2(11) and its conjugates con-tribute 22 × 2 = 44 to the ξJ1(3A, 5A, 11A).

Since ξJ1(3A, 5A, 11A) =198, the result follows.Theorem 2.8. The group J1 is (p, q, r)-generated for p, q, r ∈{2, 3, 5, 7, 11, 19}with p < q < r, except when (p, q, r) = (2, 3, 5).Proof.

This follows from the Lemmas 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8 andthe fact ∆(2, 3, 5) ∼= A5.If C1, C2, C3, C4 are conjugacy classes of a group G, and d a fixedrepresentative of C4 then ξG(C1, C2, C3, C4) denotes the number of dis-tinct triples (a, b, c) with a ∈C1, b ∈C2 and c ∈C3 such that abc = d.This number is computed using the formulaξG(C1, C2, C3, C4) = |C1||C2||C3||G|kXi=1χi(a)χi(b)χi(c)χi(d)(χi(1))2when χ1, χ2, . .

. , χk are the irreducible complex characters of G. Weuse this fact to prove the following theorem.Theorem 2.9.

The group J1 is generated by three involutions a, b, c ∈2A such that abc ∈11A.Proof. Using the character table of J1 we have ξJ1(2A, 2A, 2A, 11A) =17908.

In J1 we have two maximal subgroups, up to isomorphism, withorder divisible by 11, namely 11:10 and L2(11). Here NJ1(11A) = 11:10.

We also haveσL2(11)(2A, 2A, 2A, 11A) = ξL2(11)(2A, 2A, 2A, 11A) = 242.A fixed element of order 11 in J1 lies in two conjugates of L2(11). HenceL2(11) contributes 242 × 2 = 484 to the number ξJ1(2A, 2A, 2A, 11A).

6JAMSHID MOORIIt is routine to compute the character table of the group 11 : 10. Werepresent a part of this character table, giving the values of irreduciblecharacters on the classes 1A, 2A, and 11A in Table 1.Centralizer1101011Class1A2A11Aχ1111χ21-11χ3111χ41-11χ5111χ61-11χ7111χ81-11χ9111χ101-11χ111001Table 1.

Partial character table of 11:10Using Table 1 we haveσ11:10(2A, 2A, 2A, 11A) =|2A|311 × 10kXi=1(χi(2A))3χi(11A)(χi(1))2=|2A|311 × 1010Xi=1(−1)i+1= 0.Hence 11 : 10 does not contribute to the number ξJ1(2A, 2A, 2A, 11A).Since ξJ1(2A, 2A, 2A, 11A) = 17908 > 484, the group J1 is (2A, 2A, 2A, 11A)-generated. This completes the theorem.3.

(p, q, r)-generations for J2We list below the structure constants for the group J2.

GENERATIONS FOR J1 AND J27ξJ2(C1, C2, C3)C1C2C302A3A7A72A3B7A02B3A7A702B3B7A02A5A7A72A5C7A72B5A7A492B5C7A03A5A7A143A5C7A563B5A7A3433B5C7ALemma 3.1. The group J2 is not (2A, 3A, 7A)-, (2B, 3A, 7A)-, (2A, 5A, 7A)-, or (3A, 5A, 7A)-generated.Proof.

This follows trivially since ξJ2(C1, C2, C3) = 0 for the corre-sponding conjugacy classes of J2.Lemma 3.2. The group J2 is (2A, 5C, 7A)-, (2B, 5A, 7A)-, (2B, 5C, 7A)-,(3A, 5C, 7A)-, (3B, 5A, 7A)-, (3B, 5C, 7A)-generated.Proof.

The only maximal subgroups of J2 with order divisible by 7, upto isomorphism, are the groups U3(3) and L3(2) : 2. Since |U3(3)| =25.32.7 and |L3(2) : 2| = 24.

3. 7 and 5 does not divide neither |U3(3)|nor |L2(3) : 2|, we can say that no proper subgroup of J2 is (p, 5, 7)-generated when p ∈{2, 3}.

AsξJ2(2A, 5C, 7A) = ξJ2(2B, 5A, 7A) = 7,ξJ2(2B, 5C, 7A) = 49ξJ2(3A, 5C, 7A) = 14,ξJ2(3B, 5A, 7A) = 56,ξJ2(3B, 5C, 7A) = 343the result will immediately follow.Lemma 3.3. The group J2 is (2B, 3B, 7A)-generated.Proof.

This is given in [6].Lemma 3.4. The group J2 is not (2A, 3B, 7A)-generated.Proof.

Here we use a theorem of Ree (see [12] and [2]) which states: If Gis a transitive permutation group generated by permutations g1, g2, . .

. , gsacting on a set of n elements such that g1g2 · · · gs is the identity per-mutation, and if generator gi has exactly ci cycles for 1 ≤i ≤s, thenc1 + c2 + · · · + cs ≤(s −2)n + 2.

8JAMSHID MOORIThe group J2 acts as a transitive rank-3 group on a set X of 100elements. The point stabilizer is isomorphic to the group U3(3) withorbits of lengths 1, 36, and 63.If we denote the character of thepermutation representation of J2 on the set X by χ, then by referringto the character table of J2 we obtain χ = 1a + 36a + 63a where ma isthe first irreducible character of degree m in the ATLAS character tableof J2.

It is easy to verify that χ(2A) = 20, χ(3B) = 4, and χ(7A) = 2.This implies that, in the action of J2 on the set X, the elements in2A, 3B and 7A induce permutations with cycle types 120240, 14332 and12714 respectively. Nowc1 + c2 + c3 = (20 + 40) + (4 + 32) + (2 + 14) = 112and(s −2)n + 2 = (3 −2) × 100 + 2 = 102imply that c1 + c2 + c3 > (s −2)n + 2.

This contradicts Ree’s theoremstated above. Hence J2 is not (2A, 3B, 7A)-generated.Theorem 3.5.

For the Janko group J2 we have the following. (a) J2 is (2B, 3B, 7A)-, (2A, 5C, 7A)-, (2B, 5A, 7A)-, (2B, 5C, 7A)-,(3A, 5C, 7A)-, (3B, 5A, 7A)-, and (3B, 5C, 7A)-generated.

(b) J2 is not (2A, 3A, 7A)-, (2A, 3B, 7A)-, (2B, 3A, 7A)-, (2A, 5A, 7A)-,or (3A, 5A, 7A)-generated.Proof. This follows from the Lemmas 3.1, 3.2, 3.3 and 3.4.References[1] J. J. Cannon, An introduction to the group theory language CAYLEY, Compu-tational Group Theory (M. Atkinson, ed.

), Academic Press, San Diego, 1984,pp. 145–183.

[2] M. D. E. Conder and J. Mckay, A necessary condition for transitivity of a finitepermutation group, Bull. London Math.

Soc. 20 (1988), 235–238.

[3] J. H. Conway, R. T. Curtis, S. P. Norton, R. A. Wilson, Atlas of Finite Groups,Oxford Univ. Press (Clarendon), Oxford, 1985.

[4] H. S. M. Coxeter and W. O. J. Moser, Generators and Relations for DiscreteGroups, Springer-Verlag, 1980. [5] L. Di Martino and M. C. Tamburini, 2-Generation of finite simple groups andsome related topics, Generators and Relations in Groups and Geometries (A. Barlotti et al, eds.

), Kluwer Academic Publishers, Dordrecht, 1991, pp. 195–233.

[6] L. Finkelstein and A. Rudvalis, Maximal subgroups of the Hall-Janko-Walesgroup, J. Algebra 24 (1973), 486–493. [7] A. Hurwitz, ¨Uber algebraische Gebilde mit eindeutigen Transformationen insich, Math.

Ann. 41 (1893), 408–442.

GENERATIONS FOR J1 AND J29[8] P. B. Kleidman, R. A. Parker and R. A. Wilson, The maximal subgroups ofFi23, J. London Math. Soc.

(2) 39 (1989), 89–101. [9] S. A. Linton, The maximal subgroups of the Thompson group, J. London Math.Soc.

(2) 39 (1989), 79–88. [10] S. A. Linton and R. A. Wilson, The maximal subgroups of the Fischer groupsFi24 and Fi′24, Proc.

London Math. Soc.

(3) 63 (1991), 113–164. [11] G. A. Miller, Groups defined by the orders of two generators and the order oftheir product, Amer.

J. Math.

24 (1902), 96–100. [12] R. Ree, A theorem on permutations, J. Comb.

Theory A 10 (1971), 174–175. [13] M. Sch¨onert et al., GAP—Groups, Algorithms and Programming, Lehrstul DF¨ur Mathematik, RWTH, Aachen, 1992.

[14] D. Singerman, Subgroups of Fuchsian groups and finite permutation groups,Bull. London Math.

Soc. 2 (1970), 319–323.

[15] T. W. Tucker, Finite groups acting on surfaces and the genus of a group, J.Comb. Theory B 34 (1983), 82–98.

[16] A. J. Woldar, On Hurwitz generation and genus actions of sporadic groups,Illinois Math. J.

(3) 33 (1989), 416–437. [17], Sporadic simple groups which are Hurwitz, J. Algebra 144 (1991),443–450.


출처: arXiv:9306.201원문 보기

Subscribe to koineu.com

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe