The Algebra of Block Permutations

1 2 1
Ryan Contreras , Isabel Corona , Matt Sarmiento
1 2
Columbia University Metropolitan State University of Denver

Abstract The Planar Rook Algebra PRn Shuffle Product and Breaking Points
We show that the set of block permutations of [n], denoted BPn, is We show that the planar rook algebra PRn, which is the set of n ⇥ n matrices with A (p, q)-shuffle is a permutation ⇠ 2 Sp+q 3:◆The following
I I Example
✓ ✓ ◆are S1+2
✓ shuffles.
◆
closed under multiplication, and is therefore a monoid. We study the entries from {0, 1} having at most one 1 in each row and column, of Halverson et al. is such that 123 123 123
structure of BPn as an algebra and the subalgebra of planar diagrams, ⇠1 = ⇠2 = ⇠3 =
isomorphic to the planar subalgebra Pn+1. Consequently, we can classify the irreducible ⇠(1) < · · · < ⇠(p) 123 213 312
Pn, give presentations for both, and investigate their representation representations of Pn using the results of [3]. and
theory. ⇠
I Theorem: PRn = Pn+1.
L n
⇠(p + 1) < · · · < ⇠(p + q).
n
I Proposition: RPn acts on V = spanC{vs |S ⇢ {1, . . . , n}} and V =
n ⇠ n
k=0 Vk , where
Background n ILet f 2 BPp and g 2 BPq . We obtain the I Example 4: If b1 2 BP2 and t̃1 2 BP3, then
Vk = spanC{vs |S ⇢ {1, . . . , n}, |S| = k} and each submodule has multiplicity 1.
concatenation of f and g, f ⇥ g 2 BPp+q , by b1 ⇥ t̃1 2 BP5
I We have the following Bratteli Diagram. Lines represent the restriction of irreducible
I A set partition of [n] = {1, 2, . . . , n} is a collection of non-empty disjoint adding p to every vertex in g and placing g
representations of RPn to those of RPn 1, which arises from the fact that b1 ⇥ t̃1 =
subsets of [n], called blocks, whose union is [n]. n ⇠ n 1 n 1 to the right of f . .
Vk = Vk 1 Vk as an RPn 1 module.
I A block permutation of [n] is a bijection f : A ! B between two set I i 2 [n] is a breaking point of f 2 BPn if one can place a vertical line between the vertices i
0
partitions A, B ` [n] both having k blocks, and can be represented as a V0 and i + 1, and if the sum of the sizes of blocks of A mapping to blocks of B up to i equals i.
diagram as follows. We denote the set of breaking points of f by B(f ).
1 1
I Example 1: I Example 2: V0 V1 I In Example 4, the red line at i = 2 is a breaking point, but the green one at i = 4 is not.
i
... i . . . = ti si = ... ... 2 2 2
V0 V1 V2 Hopf Algebra Structure of BPn
i +1
... i ... = t̃i 3 3 3 3
V0 V1 V2 V3 Theorem: The graded vector space
.
. .
. .
. .
. I
M
... i ... = t̃i ti = bi n BP = K[BPn]
i +1
I The dimensions of each module is given by . If we replace each module with its
k n 0
corresponding dimension, we obtain Pascal’s triangle. with the product ⇤, coproduct , unit ;, and counit ✏ is a graded connected Hopf algebra.
I Example 1 depicts diagram multiplication i.e. t̃i ti = bi . I Proposition: For 0  k  n and f 2 Pn+1, the irreducible characters are given by
( I The unit is ;, the empty diagram, and the counit is the map ✏ : BP ! K given by ✏(f ) = ;,f .
I g is planar if g’s diagram can be drawn without edges crossing inside of
i
the rectangle formed by its vertices. In Example 2, si is non-planar. n k if k  i I We define the product ⇤ on BP to be
k (f ) = X
I The symmetric group Sn is contained in BPn .
0 if k > i, f ⇤g = ⇠ · (f ⇥ g) 2 K[BPp+q ]
where i is the number of vertical edges in d 2 RPn. ⇠
I The character table for
n n 8 f 2 BPp , g 2 BPq and all (p, q)-shuffles ⇠.
The Monoid Algebra BPn k , the irreducible character of Vk , is given by Pascal’s triangle.
I If we take ⇠1, ⇠2, ⇠3 as in Example 3, then the shuffle product of two diagrams f 2 BP1 and

I Proposition: The dimension of I Proposition: The dimension of the Presentation of BPn g 2 BP2 is just the diagram multiplication of f ⇥ g distributed over ⇠1 + ⇠2 + ⇠3.
BPn is planar subalgebra Pn is I We define the coproduct on BP to be
n ✓ ◆ ✓ ◆ X
X
2
n 1
X n 1 2
2(n 1) I Proposition: BPn is generated by ti , e ti , bi , and the elementary transpositions si . (f ) = 0
f(i) ⌦ f(n i).
|BPn| = k!S(n, k) , |Pn| = = e
I Proposition: BPn = hsi , bi , ti , ti | R i, where R consists of R and the following relations:
0 0
k 1 n 1 i2B(f )
k=1 k=1
I si sj = sj si if |i j| > 1 I si si+1si = si+1si si+1 I ti = ti si = si+1ti
where S(n, k) is the Stirling I This is also the dimension of the I Let f = b1 ⇥ t̃1 as in Example 4. Then (f ) is given by
I si bi+1si = si+1bi si+1 2
Is = 1
number of the second kind. I bi sj = sj bi if |i j| > 1
planar rook algebra. i (f ) = ; ⌦ f + ⌦ +f ⌦ ;.
I si+2ti si+2 = si+1ti+1ti
I ti sj = sj ti if j i > 2 or i j > 1 I si ti+1si = ti+1ti si+1

Presentation of Pn I ti 1si ti = si 1si bi 1bi+1si si 1

I Proposition: BPn factors as BPn = Sn Pn Sn . Acknowledgements
I Proposition: Pn is generated by ti , t̃i , bi as in Example 1. I Example: t1s2s1t1 = s1s2b1.
I Proposition: Pn = hbi , ti , e ti | Ri, where R is the following set of I We would like to thank Dr. Rosa Orellana, Dr. Ivelisse Rubio, Mr. Alexander Diaz-Lopez, and the MSRI staff.
I Our work was conducted during the 2013 MSRI-UP, which is supported by the National Security Agency (NSA) grant
relations. They are defined for 1  i  n unless otherwise noted. Representation Theory of BPn H-98230-13-1-0262 and the National Science Foundation (NSF) grant 11566499.
I bj ti = ti bj if j < i or ˜ ˜ ˜
I ti ti+1 = bi+2ti = ti bi+2 ˜ 2
I ti ti = b = bi
i
j i >1 ˜
I ti+1ti = bi ti+1 = ti+1bi I t ˜
i i = bi+1
t I Proposition: I = BPnbn 1BPn = {f · bn 1 · g | f , g 2 BPn } is an ideal of BPn. References
I ti tj = tj ti for |i I bi+1ti = ti bi =
⇠
I Corollary: BPn = BPn bn 1BPn C[Sn].
j| > 1 I ti+1ti ti+1 = ti ti+1ti = ti ti+1 1. M. Aguiar and R. Orellana. The Hopf Algebra of Uniform Block Permutations. Journal of Algebraic Combinatorics. 28(1):115-138,
2 ti
⇠
I Proposition: BPn 1 = bn 1BPn bn 1. 2008.
I t = bi ti = ti bi+1 =
i 2. A. H. Clifford and G. B. Preston. The Algebraic Theory of Semigroups: Vol.: 1. American Mathematical Society Providence, Rhode
I We use the above propositions to show that the irreducible representations of BPn can
bi bi+1 Island, 1961.
be indexed by Young Diagrams. 3. D. Flath, T. Halverson, and K. Herbig. The Planar Rook Algebra and Pascal’s Triangle. L’Enseignement Mathematique. (2), 55
(2009), 77-92.
4. T. Halverson. Characters of the Partition Algebras. Journal of Algebra. 238(2):502-533, 2001.
5. T. Halverson and A. Ram, Partition Algebras. European Journal of Combinatorics. 26(6):869-921, 2005.

Mathematical Sciences Research Institute Undergraduate Program http://www.msri.org/web/msri/education/for-undergraduates/msri-up