Subgroups of Symmetric Groups Determined Using Signed

0 downloads 0 Views 1MB Size Report
as the number of negative edges involved to form this edge is even (odd). Similarly, a loop รŽยฒ in 1 2 is so called positive (negative). The multiplication isย ...
Subgroups of Symmetric Groups Determined Using Signed Brauer Diagrams Ram Parkash Sharma and Rajni Parmar Department of Mathematics, Himachal Pradesh University Summerhill, Shimla (India) Email: [email protected]

Abstract The group ๐‘†๐‘› of signed Brauer diagrams with no horizontal edges becomes a subgroup of ๐‘†2๐‘› . Therefore, it is natural to find the elements of ๐‘†๐‘› which are even permutations in ๐‘†2๐‘› . There are 2๐‘›โˆ’1 ๐‘›! such elements that form a subgroup of ๐‘†2๐‘› . Consequently, ๐‘†๐‘› has a class of subgroups of order 1, 22โˆ’1 2!, 23โˆ’1 3!, . . ., 2

๐‘› 2

โˆ’1

๐‘› 2

! of even permutations and hence 2

๐‘› 2

๐‘› 2

! | ๐‘›!. Finally, using these

results it is shown that every group of order ๐‘› ๐‘› โˆ’ 1 โ€ฆ ( ๐‘› 2

๐‘› 2

+ 1) has a

๐‘†๐‘ฆ๐‘™๐‘œ๐‘ค 2 โˆ’ ๐‘ ๐‘ข๐‘๐‘”๐‘Ÿ๐‘œ๐‘ข๐‘ of order 2 . Keywords: Signed Brauer algebras; symmetric groups; even permutations. AMSC: 20B30, 20B35. 1. Introduction Parvathi and Kamaraj [2] defined signed Brauer algebras ๐ต๐‘› ๐‘ฅ over โ„‚, (๐‘ฅ โˆˆ โ„‚) having basis consisting of signed diagrams ๐‘‘, which consist of two rows of ๐‘› points labelled {1,2, โ€ฆ , ๐‘›} in first row and {(๐‘› + 1), (๐‘› + 2), โ€ฆ ,2๐‘›} in second row, with each dot joined to precisely one other dot (distinct from itself). Each edge of the diagram ๐‘‘ has a direction: either positive or negative. A positive vertical (horizontal) edge is denoted by โ†“ (โ†’) and a negative vertical (horizontal) edge is denoted by โ†‘ (โ†). An element in ๐ต5 ๐‘ฅ is given as:

To define multiplication in ๐ต๐‘› (๐‘ฅ), it is enough to define the product of two signed diagrams ๐‘‘1 and ๐‘‘2 ,

to do so place ๐‘‘1

above ๐‘‘2

and join

corresponding points (interior signed loops are deleted). Let ๐‘‘3 be the signed diagram obtained in this way. An edge in ๐‘‘3 is positive (negative) according as the number of negative edges involved to form this edge is even (odd). Similarly, a loop ฮฒ in ๐‘‘1 ๐œŠ๐‘‘2 is so called positive (negative). The multiplication is then defined ๐‘‘1 ๐œŠ๐‘‘2 = ๐‘ฅ 2๐‘› 1 +๐‘› 2 ๐‘‘3 where ๐‘›1 and ๐‘›2 are the numbers of positive and negative loops respectively formed placing ๐‘‘1 above ๐‘‘2 . The set of signed Brauer diagrams ๐ต๐‘› contains the set of Brauer diagrams ๐ต๐‘› (each Brauer diagram corresponds to same underlying diagram with all positive edges). For more detail on Brauer and signed Brauer algebras, one can refer to [3,4,5,6]. The Brauer diagrams with no horizontal edges form a permutation group, by identifying a Brauer diagram with a permutation ๐œ‹ connecting the i th upper vertex to the ๏ฐ (i)th lower vertex. In a similar way, the signed Brauer diagrams with no horizontal edges form a group ๐‘†๐‘› which corresponds to the wreath product of Z 2 by ๐‘†๐‘› [see 1]. The group ๐‘†๐‘› becomes a subgroup of ๐‘†2๐‘› . In order to identify ๐‘†๐‘› as a subgroup of ๐‘†2๐‘› , first divide the set

๏ƒก1, 2, . . . n, . . . , 2n๏ƒข into two equal parts that is the numbers ๏‚ฒ n and ๏€พ n. To distinguish the numbers less than or equal to ๐‘› and greater than ๐‘› , we fix the notation

r 1 , r 2 , . . . , r n for the numbers less than or equal to n and ๐‘ 1, ๐‘ 2, โ€ฆ , ๐‘ ๐‘› for the

numbers greater than n. We fix another notation as:

๏„ For each i ๏€ต ๏ƒก1, 2, . . . , 2n๏ƒข, we denote by i the number given by:

i๏„ ๏€ฝ

i ๏€ซ n,

1 ๏‚ฒ i ๏‚ฒ n,

i ๏€ฟ n,

n ๏€ผ i ๏‚ฒ 2n

.

Thus ๐‘Ÿ๐‘–โˆ— > ๐‘›; ๐‘– = 1,2, โ€ฆ , ๐‘› and ๐‘ ๐‘—โˆ— โ‰ค ๐‘›; ๐‘— = 1,2, โ€ฆ , ๐‘›. The elements of ๐‘†๐‘› are identified in ๐‘†2๐‘› as follows: Let ๐œŽ โˆˆ ๐‘†๐‘› have a negative edge

between ๐‘Ÿ๐‘– and ๐‘ ๐‘˜. Then the number ๐‘Ÿ๐‘– (โ‰ค ๐‘›) โŸผ ๐‘ ๐‘˜ (> ๐‘›) and ๐‘Ÿ๐‘–โˆ— โŸผ ๐‘ ๐‘˜โˆ— under ๐œŽ in ๐‘†2๐‘› and vice-versa. If we have a positive edge

between ๐‘Ÿ๐‘– and ๐‘ ๐‘˜ . Then ๐‘Ÿ๐‘– โŸผ ๐‘ ๐‘˜โˆ— and ๐‘Ÿ๐‘–โˆ— โŸผ ๐‘ ๐‘˜ . Example 1.1. Consider the eight elements of ๐‘†2 given below:

According to the above definition, the elements of ๐‘†2 are identified in ๐‘†4 as follows: ๐‘’, 2 4 , 1 3 , 1 3 2 4 , 1 2 3 4 , 1 2 3 4 , 1 4 3 2 , (1 4 )(2 3) .

As ๐‘†๐‘› is a subgroup of ๐‘†2๐‘› , we started this paper with the aim of finding the elements of ๐‘†๐‘› which correspond to even permutations in ๐‘†2๐‘› and prove that the elements of ๐‘†๐‘› having even number of negative edges (see Theorem 3.1) are even permutations in ๐‘†2๐‘› . Since the product of two negative edges in ๐ต๐‘› gives a positive edge and product of negative and positive edges is a negative edge, it is easy to see that these elements form a subgroup of ๐‘†๐‘› and hence of ๐‘†2๐‘› . The total number of elements with even number of negative directions in ๐‘†๐‘› is given by:

๏ƒฌ( n c0 ๏€ซ n c2 ๏€ซ n c4 ๏€ซ n c6 ๏€ซ ... ๏€ซ n cn )n!, if n is even ๏ƒญn n n n ๏ƒฎ( c0 ๏€ซ c2 ๏€ซ c4 ๏€ซ ... ๏€ซ cn ๏€ญ1 )n!, if n is odd ๏€ฝ 2 n๏€ฟ1n!. Thus, ๐‘†2๐‘› has a subgroup of order 2๐‘›โˆ’1 ๐‘›!. In particular, ๐‘†๐‘› has a subgroup of order 2

๐‘› 2

โˆ’1

๐‘› 2

! for ๐‘› โ‰ฅ 2 and hence in general it has a class of subgroups of ๐‘› 2

order 1, 22โˆ’1 2!, 23โˆ’1 3!, โ€ฆ , 2 ๐‘›

it, 2 2

๐‘› 2

โˆ’1

๐‘› 2

! of even permutations. In consequence of

! | ๐‘›! ; hence ๐‘›

๐‘›

22 |๐‘› ๐‘›โˆ’1 โ€ฆ Further, we observe that (Corollary 3.4) 2

๐‘› 2

+1

โˆค๐‘› ๐‘›โˆ’1 โ€ฆ

Therefore, every group of order ๐‘› ๐‘› โˆ’ 1 โ€ฆ ( ๐‘›

2

๐‘› 2

+1 . ๐‘› 2

+1 .

+ 1) has a ๐‘†๐‘ฆ๐‘™๐‘œ๐‘ค 2 โˆ’ ๐‘ ๐‘ข๐‘๐‘”๐‘Ÿ๐‘œ๐‘ข๐‘

of order 2 2 . 2. Even Permutations of ๐‘†๐‘› in ๐‘†2๐‘› First, we prove that the elements of ๐‘†๐‘› having positive direction of all the edges are even permutations in ๐‘†2๐‘› . We, very often, use ri ๏ก s j for ๏ณ (ri ) ๏€ฝ s j and ri ๏‚ฎ ๏€ฏ s j for ๏ณ (ri ) ๏‚น s j when ๐œŽ is understood. Lemma 2.1. If ๐œŽ โˆˆ ๐‘†๐‘› has positive direction of all the edges, then ๐œŽ is an even permutation in ๐‘†2๐‘› . Proof. Let ๐œŽ โˆˆ ๐‘†๐‘› has all edges of positive direction. Then ri ๏‚ฃ n, goes to some rj ๏€ฝ (s๏€ชj ) ๏‚ฃ n under ๐œŽ. Consequently, ri๏€ช goes to some rj๏€ช ๏€ฝ (s๏€ชj )๏€ช ๏€ฝ s j and both ri๏€ช ,

s j ๏€พ n. In this way, we get

r 1 ๏ƒซ r 2 ๏ก r3 ... ๏ก r i ,

and r ๏„1 ๏ƒซ r ๏„2 ๏ก r3๏€ช ... ๏ก r ๏„i . ๏„ ๏„ Note that r i ๏ƒ r 1 and r i ๏ƒ r 1 because this is only possible if we have a ๏„

negative edge between r i and r 1 . Since direction of all the edges of ๐œŽ is positive, there is only one possibility of having ri ๏ก r1 and hence ri๏€ช ๏ก r1* using a ๏„ positive edge between ๐‘Ÿ๐‘– and r 1 . Therefore, corresponding to every cycle ๏ƒr 1 r 2 r 3 . . . r i ๏ƒž written using the numbers โ‰ค ๐‘›, there is another cycle ๏ƒr ๏„1 r ๏„2 r3๏€ช ...ri๏€ช ) of the numbers greater than n having same length. So ๐œŽ is an even permutation in ๐‘†2๐‘› .

Example 2.2. Consider an element

of ๐‘†5 having all edges of positive direction. This element becomes (1 3 5 2 4)(6 8 10 7 9 ) in ๐‘†10 . For ๐œŽ โˆˆ ๐‘†๐‘› with negative direction of all the edges, any cycle ๐‘๐‘– , in the product of cycles of ๐œŽ, no two numbers ๐‘Ÿ๐‘– , ๐‘Ÿ๐‘— occur together and same is true for ๐‘ ๐‘– , ๐‘ ๐‘— also. With this observation, we have Lemma 2.3. Let ๐œŽ โˆˆ ๐‘†๐‘› has negative direction of all the edges. Then, length of each cycle in the product of ๐œŽ in ๐‘†2๐‘› is always even. Proof. Since all the edges are of negative direction, any number ri ๏‚ฃ n goes to a number s j ๏€พ n. Since s๏€ชj ๏‚ฃ n, s ๏€ชj can be ri . In this case, we get a transposition (ri ri๏€ช ) . Let (r1 s1 ...) be a cycle in the product of

๏ก. In this cycle no

ri , rj or si , s j

occur together as all the edges are of negative direction. Moreover the last number will be some sk ๏€พ n. For, if we have a cycle of the type (r1 s1 r2 s 2 ... rk ) . Then * rk ๏ก r1 gives a positive edge between rk and r1 . Hence any cycle in the product

of ๐œŽ will be of the type (r1 s1 r2 s 2 ...rk ๏€ญ1 s k ) having length 2k. The above result does not depend on whether ๐‘› is even or odd. However, the number of cycles in the product of ๐œŽ โˆˆ ๐‘†๐‘› having negative direction of all the edges depends upon n as observed in the following examples.

Example 2.4. For n to be odd, consider an element

of ๐‘†5 having all edges of negative direction. This element becomes (1 7 3 9)(6 2 8 4)(5 10) in ๐‘†10 . Clearly, length of each cycle in the product of ๐œŽ is even, but number of cycles is 3. Example 2.5. For n to be even, consider an element

๏ก ๏€ฝ

of ๐‘†6 having all edges of negative direction which becomes (1 8 4 9)(7 2 10 3)(5 12)(6 11) in ๐‘†12 . Here, again the length of each cycle is even, but the number of cycles is also even. With these observations, we prove a general result regarding the number of cycles of ๐œŽ โˆˆ ๐‘†๐‘› having negative direction of all the edges. Lemma 2.6. Let ๐œŽ โˆˆ ๐‘†๐‘› has negative direction of all the edges. Then the number of disjoint cycles in the product of ๐œŽ is even if n is even and odd if n is odd. Proof. For given ๐œŽ โˆˆ ๐‘†๐‘› with negative direction of all the edges, note that ๐œŽ

fixes no ri or s j , as ๏ณ (ri ) ๏€ฝ ri if and only if we have a positive edge between ๐‘Ÿ๐‘– and ri๏€ช . Similarly ๏ณ (s j ) ๏€ฝ s j if and only if we have a positive edge between s ๏€ชj and s j . Therefore all the numbers 1, 2, 3,..., n, n ๏€ซ 1,...2n will be involved in the cycles of ๐œŽ. We discuss different cases for involvement of even or odd number of numbers in a cycle of ๐œŽ. Here we also use the fact observed in Lemma 2.3 that in any cycle of ๐œŽ โˆˆ ๐‘†๐‘› , no ri , r j or si , s j occur together as all the edges are of negative direction. Case (i) Suppose we want to involve two numbers from each side of n in the product of cycles of ๐œŽ, then the only possibility is ri ๏ก s j (s j ๏‚น ri๏€ช ) and ri๏€ช ๏ก s ๏€ชj . In this case, we get two cycles (ri s j ) and (ri๏€ช s๏€ชj ). The possibility of having these four numbers in a single cycle is ruled out because these four numbers will be in one cycle if and only if we have a positive edge between s ๏€ชj and r j๏€ช . Further, if we want to use four numbers from each side of n, then r1 ๏ก s1 ๏ก r2 ๏ก s2 ,

and r1๏€ช ๏ก s1๏€ช ๏ก r2๏€ช ๏ก s2๏€ช.

(Note that without loss of generality, we can start with r 1 , because if we start with s 1 then we are writing the second row in place of first). Here we take si ๏‚ฎ ๏€ฏ ri , otherwise we get two cycles of length 2, which we have already discussed). These sequences will terminate if s2 ๏ก r1 and s2๏€ช ๏ก r1๏€ช , because ๏€ช ๏€ช s2 ๏‚ฎ ๏€ฏ r1 , s2 ๏‚ฎ ๏€ฏ r1 being on the same side of n as ๐œŽ has all negative edges. So we get two cycles

๏ƒ r 1 s 1 r 2 s 2 ๏ƒž and ๏ƒr ๏„1 s ๏„1 r ๏„2 s ๏„2 ๏ƒž.

In general, the involvement of an even number of numbers from each side is possible if and only if we take r1 ๏ก s1 ๏ก r2 ๏ก s2 ๏ก ... ๏ก ri ๏ก si

and

r ๏„1 ๏ƒซ s ๏„1 ๏ƒซ

,

r2๏€ช ๏ก s2๏€ช ๏ก ... ๏ก ri๏€ช ๏ก si๏€ช .

๏„ ๏„ Again, in this case, we get two cycles ๏ƒ r 1 s 1 r 2 s 2 . . . . r i s i ๏ƒž and ๏ƒr 1 s 1

r ๏„2 s ๏„2 . . . . . r ๏„i s ๏„i ๏ƒž.

Case (ii) Now suppose that only one number from either side of n is involved in a cycle of ๐œŽ , then ri ๏ก ri๏€ช and we get only one cycle (ri ri๏€ช ) , a negative transposition. Three numbers can be involved from either side of n in a cycle of ๐œŽ if and only if

r 1 ๏ƒซ s 1 ๏ƒซ r 2 , ๏ƒs 1 ๏‚ฎ r ๏„1 and r 2 ๏‚ฎ s ๏„1 ๏ƒž

and

r ๏„1 ๏ƒซ s ๏„1 ๏ƒซ r ๏„2 .

๏„ ๏„ These sequences will terminate if r 2 ๏ƒซ r 1 and r 2 ๏ƒซ r 1 , because r 2 ๏ƒ r 1 ๏„ ๏„ and r 2 ๏ƒ r 1 being on the same side of n. So, we get a single cycle ๏ƒr 1 s 1 r 2

r ๏„1 s ๏„1 r ๏„2 ๏ƒž.

In general, if an odd number of ๐‘Ÿ๐‘– and ๐‘ ๐‘– are to be involved from either side of n in a cycle of ๐œŽ, then

r 1 ๏ƒซ s 1 ๏ƒซ r 2 ๏ƒซ s 2 ๏ƒซ r 3 ๏ƒซ s 3 ๏ƒซ r 4. . . . . ๏ƒซ s i๏€ฟ1 ๏ƒซ r i and

r ๏„1 ๏ƒซ s ๏„1 ๏ƒซ r ๏„2 ๏ƒซ s ๏„2 ๏ƒซ r ๏„3 ๏ƒซ s ๏„3 ๏ƒซ r 4. . . ๏ƒซ s ๏„i๏€ฟ1 ๏ƒซ r ๏„i . ๏„ ๏„ These sequences will terminate if r i ๏ƒซ r 1 and r i ๏ƒซ r 1 , because r i ๏ƒ r 1 ๏„ ๏„ and r i ๏ƒ r 1 being on the same side of n. So we get a single cycle ๏ƒr 1 s 1 r 2 s 2 . . . s i๏€ฟ1 r i r ๏„1 s ๏„1 r ๏„2 s ๏„2 . . . . s ๏„i๏€ฟ1 r ๏„i ๏ƒž of even length. Therefore we conclude that there are two type of cycles in the product of ๏ก : (i) cycles having even number of ๐‘Ÿ๐‘– โ‰ค ๐‘› and (ii) cycles having odd number of ๐‘Ÿ๐‘– โ‰ค ๐‘›. The cycles of first kind exist in pair. A cycle of second kind, having length 2t, corresponds to some odd number t which exists in ๐œ† โŠข ๐‘›. Therefore if n is even, then corresponding to an odd number t in the partition ๐œ† โŠข ๐‘› , there exists ๐‘ก โ€ฒ an odd number in the same partition ๐œ† โŠข ๐‘›. Hence we get a pair of cycles of type 2t and 2tโ€™ (see case (ii)). Hence, the number of cycles in the product of ๐œŽ will be even if n is even and odd if ๐‘› is odd. As an easy consequence of the above result, we have Corollary 2.7. Any cycle in the product of ๐œŽ โˆˆ ๐‘†๐‘› having all the edges of negative direction, ends with some ๐‘ ๐‘— if it starts with ๐‘Ÿ๐‘– .

Corollary 2.8. Let ๐œŽ โˆˆ ๐‘†๐‘› with n even, has negative direction of all the edges. Then, ๐œŽ is not a 2๐‘› โ€“cycle in ๐‘†2๐‘› . Proof. This falls in the case (i) of the above theorem. In this case, for each cycle, there corresponds another cycle of the same length, in the product of ๐œŽ. Hence any such ๐œŽ cannot be a single cycle of length 2n. Using the above results proved for ๐œŽ โˆˆ ๐‘†๐‘› having negative direction of all the edges, we have Proposition 2.9. Let ๐œŽ โˆˆ ๐‘†๐‘› with n even, has negative direction of all the edges. Then ๐œŽ is an even permutation in ๐‘†2๐‘› . Proof. By lemma 2.6, number of disjoint cycles of ๐œŽ in ๐‘†2๐‘› is even if and only if n is even. Also by lemma 2.3, length of each cycle of ๐œŽ is even. Even number of disjoint cycles with even length will give even number of transpositions, hence ๐œŽ is an even permutation in ๐‘†2๐‘›. 3. Main Results Now we prove the main results of this paper. Theorem 3.1. For ๐‘› โ‰ฅ 2, let n ๏€ฝ 2m ๏€ซ r, 0 ๏‚ฃ m ๏‚ฃ ๏›n2 ๏ and 0 ๏‚ฃ r ๏‚ฃ n. Then ๐œŽ โˆˆ ๐‘†๐‘› is an even permutation in ๐‘†2๐‘› if and only if ๐œŽ has 2m edges of negative direction. Proof. First, we analyze the role of positive and negative edges to get a cycle in the product of ๐œŽ in ๐‘†2๐‘›. One negative edge between ri and ri๏€ช gives a transposition (ri ri๏€ช ) in ๐‘†2๐‘› . Thus, if we have such 2m ๏€ญ negative edges, then we get an even number of transpositions (r1 r1๏€ช ), (r2 r2๏€ช ), (r3 r3๏€ช ),..., (r2m r2๏€ชm ) in ๐‘†2๐‘› corresponding to these negative edges. The remaining positive edges give pair of cycles of equal length as observed in Lemma 2.1 and hence ๐œŽ is an even permutation in ๐‘†2๐‘› . Suppose some mixed positive and negative edges give a cycle in the product of ๏ณ . Note that two numbers ri , r j or si , s j in a cycle occur if and only if there is a positive edge and two numbers ri , s j or si , r j occur if and only if there is a negative edge. Case (i) Suppose a sequence of the type r1 ๏ก .... ๏ก si ... ๏ก r j

occurs inside a cycle of ๏ณ . In this sequence, there are even number of changes from ri to si and s j to ri (possibly no change). As observed above this sequence occurs if and only if there exists a sequence of even number of negative edges (possibly no negative edge). But then we also have

r1๏€ช ๏ก ... ๏ก si๏€ช ... ๏ก rj๏€ช .

In order to complete the cycles either rj ๏ก r1 and hence rj๏€ช ๏ก r1๏€ช or rj ๏ก r1๏€ช and hence rj๏€ช ๏ก r1 . Here we are not taking the case rj ๏ก r j๏€ช because that will remove r j and r j๏€ช from this cycle and single cycles have been discussed earlier. Subcase (i) rj ๏ก r1 and rj๏€ช ๏ก r1๏€ช . This is only possible with an additional positive edge between r j and r1๏€ช . Therefore the number of negative edges used remain even. Hence in this case, we have used even number of negative edges with some positive edges and we get two cycles (r1 ....si ... rj ) and (r1๏€ช ....si๏€ช ....r j๏€ช ) contributing an even number of transpositions in the product of ๏ณ . Subcase (ii) rj ๏ก r1๏€ช and rj๏€ช ๏ก r1 . This is only possible with an additional negative edge between r j and r1๏€ช . Hence in this case, number of negative edges used changes to odd and we get a single cycle (r1 ....si ... r j r1๏€ช ....si๏€ช ....r j๏€ช ). Clearly the length of this cycle is even as number of ri๏€ช = number of ri and number of si๏€ช = number of si . This contributes an odd number of transpositions in the product of ๐œŽ. Case (ii) Now consider a sequence of the type r1 ๏ก ... ๏ก ri ... ๏ก

si ๏ก ... ๏ก s j .

Here, there are odd number of changes from ri to si and s j to r j (both included). Hence this sequence occurs inside a cycle of ๐œŽ if and only if there exists a sequence of odd number of negative edges. But then we also have r1๏€ช ๏ก ... ๏ก ri๏€ช ... ๏ก si๏€ช ๏ก ... ๏ก s ๏€ชj .

Again in order to complete the cycle either s j ๏ก r1 and hence s๏€ชj ๏ก r1๏€ช or s j ๏ก r1๏€ช and hence s๏€ชj ๏ก r1. Subcase (i) s j ๏ก r1 and s๏€ชj ๏ก r1๏€ช . This is only possible with an additional negative edge between s ๏€ชj and r1๏€ช . Hence, in this case, number of negative edges used, changes to even and we get two cycles (r1 ....ri ....si ....s j ) and (r1๏€ช ....ri๏€ช ... si๏€ช ...s ๏€ชj ) contributing an even number of transpositions in the product of ๏ก. Subcase (ii) s j ๏ก r1๏€ช and s๏€ชj ๏ก r1. This is possible only with an additional positive edge between s๏€ชj ๏ก r1๏€ช . In this case number of negative edges used remains odd and we get a cycle (r1 ....ri ....si ....s j r1๏€ช ....ri๏€ช ... si๏€ช ....s ๏€ชj )

of even length. This contributes in product of ๏ก an odd number of transpositions.

Conclusion: With the above observations, finally we conclude that an even number of negative edges (possibly zero) together with some positive edges contribute two cycles of same length in the product of ๐œŽ and an odd number of negative edges (possibly one) contribute a product of odd number of transpositions for ๐œŽ . If we take any odd number r in a partition of 2m , then there exists an another odd number r๏‚ข in this partition. Hence, by above discussion it follows that in the product of ๐œŽ, there are pairs of cycles of equal length contributing to even number of transposition and some more even number of transpositions coming from the pairs (r , r๏‚ข) of negative edges ( r and r๏‚ข are odd numbers). Therefore, ๐œŽ is an even permutation in ๐‘†2๐‘›. Let ๐œŽ โˆˆ ๐‘†๐‘› has 2๐‘˜ + 1 (๐‘˜ โ‰ฅ 1) edges with negative direction. Then in any partition of 2๐‘˜ + 1, the number of odd numbers will be odd. Therefore, from the above discussion, the total number of transposition in the product of ๐œŽ in ๐‘†2๐‘› will odd. Theorem 3.2. Let ๐ด๐‘› = {๐œŽ โˆˆ ๐‘†๐‘› | ๐œŽ has even number of negative edges ๏ƒข. Then ๐ด๐‘› is a normal subgroup of ๐‘†๐‘› of order 2๐‘›โˆ’1 ๐‘›!. Further ๐ด๐‘› = ๐‘†๐‘› โˆฉ ๐ด2๐‘› and hence ๐‘†2๐‘› has a subgroup of order 2๐‘›โˆ’1 ๐‘›!. In particular, ๐‘†๐‘› has a subgroup of order 2

๐‘› 2

โˆ’1

๐‘› 2

of even permutations.

Proof. The multiplication defined in ๐ต๐‘› yields that ๐ด๐‘› โ‰ค ๐ต๐‘› . The result ๐ด๐‘› = ๐‘†๐‘› โˆฉ ๐ด2๐‘› follows from Theorem 3.1. The order of ๐ด๐‘› is 2๐‘›โˆ’1 ๐‘›! as observed in the Introduction. Also, [๐‘†๐‘› : ๐ด๐‘› ] = 2 implies that ๐ด๐‘› โŠฒ ๐‘†๐‘› . Since ๐‘†2๐‘› is a subgroup of ๐‘†2๐‘›+1, we have a more general result, ๐‘†๐‘› has subgroups of order 1, 22โˆ’1 2!, 23โˆ’1 3!, . . ., 2

๐‘› 2

โˆ’1

๐‘› 2

! for ๐‘› โ‰ฅ 2.

Example 3.3. For k = 2 and 3, we have 2=2.0+2, 2=2.1+0 and 3=2.0+3, 3=2.1+1. Therefore, ๐‘†6 has two subgroups ๐ด2 and ๐ด3 of order 6

2๐ถ0 2! + 2๐ถ2 2! = 22โˆ’1 2! = 4 and 3๐ถ0 3! + 3๐ถ2 3! = 2 2

โˆ’1 6 2

2 ! ๏€ฝ 2 ๏ƒŸ3๏ƒ ! ๏€ฝ 24

respectively, which are obtained from the elements of ๐‘†2 and ๐‘†3 having even number of negative edges. The subgroup ๐ด2 can be written from Example 1.1; that is, {e,(13)(24),(12)(34),(14)(23)}. Here, we find the subgroup ๐ด3 as follows:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Finally, we obtain a result as an application of Theorem 3.2. Corollary 3.4. Every group of order ๐‘› ๐‘› โˆ’ 1 โ€ฆ ( ๐‘› 2

๐‘› 2

+ 1) has a ๐‘†๐‘ฆ๐‘™๐‘œ๐‘ค 2 โˆ’

๐‘ ๐‘ข๐‘๐‘”๐‘Ÿ๐‘œ๐‘ข๐‘ subgroup of order 2 . Proof. Using theorem 3.2, 2 2

๐‘› 2

โˆ’1

๐‘›

|๐‘› ๐‘›โˆ’1 โ€ฆ

2

2

๐‘› 2

๐‘› 2

! | ๐‘›!, hence

+ 1 . We claim that ๐‘› 2

+1

โˆค๐‘› ๐‘›โˆ’1 โ€ฆ

For, we show that ๐‘› 2

+1

๐‘› 2

๐‘› 2

๐‘› 2

+ 1 , nโ‰ฅ 2.

+ 2 โ€ฆn = 2 ๐‘˜, for some positive odd integer ๐‘˜.

The result is true for ๐‘› = 2.

Suppose that the result is true for all integers ๐‘ก, 2 โ‰ค ๐‘ก โ‰ค ๐‘›. Case(i) If ๐‘› is even, then n +1 2

๐‘›

22 ๐‘˜ =

n + 2 โ€ฆ๐‘› 2

implies 2

๐‘› +1 2

๐‘˜=

๐‘›+1 +1 2

๐‘›+1 + 2 โ€ฆ ๐‘›, 2

which yields 2

๐‘› +1 2

๐‘˜ ๐‘›+1 =

๐‘›+1 2

+1

๐‘›+1 2

+ 2 โ€ฆ (๐‘› + 1),

where ๐‘˜ ๐‘› + 1 is odd. Case(ii) If ๐‘› is odd, then 2

๐‘› 2

๐‘› +1 2

๐‘˜=

๐‘› +2 โ€ฆ๐‘› 2

implies 2

๐‘›โˆ’1 2 ๐‘˜

=

n+1 2

n+3 โ€ฆ ๐‘›, 2

which yields n+1 n+1 +1 +2 โ€ฆ ๐‘›+1 . 2 2 ๐‘› So every group of order ๐‘› ๐‘› โˆ’ 1 โ€ฆ ( + 1) has a ๐‘†๐‘ฆ๐‘™๐‘œ๐‘ค 2 โˆ’ ๐‘ ๐‘ข๐‘๐‘”๐‘Ÿ๐‘œ๐‘ข๐‘ of 2

๐‘› 2

order 2 .

๐‘›+1 2 ๐‘˜

=

2

References [1] C. Musli, Representations of finitie groups, Hindustan book agency , 1993. [2] M. Parvathi and M. Kamaraj, Signed Brauer's algebras, Comm.in Algebra, 26(3) (1998), 839-855. [3] M. Parvathi and C. Selvaraj, Signed Brauer's algebras as Centralizer algebras, Comm. in Algebra, 27(12), 5985-5998 (1999). [4] R.P. Sharma and V.S. Kapil, Irreducible S_{n}-Modules and A cellular structure of the Signed Brauer Algebras, Southeast Asian Bulletin of Mathematics, (2011), 35: 497-522. [5] R. P. Sharma and V. S. Kapil, Generic irreducibles of the Brauer algebras, Contemporary Mathematics (AMS), 456(2008), 189-204. [6] H. Wenzl, On the structure of Brauer's centralizer algebras, Ann. of Math., (1988):173-193.