* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

# Download Graded decomposition numbers for the

Survey

Document related concepts

Polynomial ring wikipedia , lookup

Signal-flow graph wikipedia , lookup

Structure (mathematical logic) wikipedia , lookup

Eisenstein's criterion wikipedia , lookup

Factorization wikipedia , lookup

System of polynomial equations wikipedia , lookup

Linear algebra wikipedia , lookup

Boolean algebras canonically defined wikipedia , lookup

Heyting algebra wikipedia , lookup

Universal enveloping algebra wikipedia , lookup

Corecursion wikipedia , lookup

Geometric algebra wikipedia , lookup

Exterior algebra wikipedia , lookup

Clifford algebra wikipedia , lookup

Vertex operator algebra wikipedia , lookup

Transcript

Graded decomposition numbers for the diagrammatic Cherednik algebra Liron Speyer Osaka University, Suita, Osaka 565-0871, Japan [email protected] 1 Motivation Fix F a field of characteristic p > 0 throughout and e ∈ {3, 4, . . . } ∪ {∞}. Let q ∈ F be a primitive eth root of unity, (if e = ∞ then q is not a root of unity) and κ = (κ1 , . . . , κl ) ∈ (Z/eZ)l . We denote by Hn = Hn (q, κ) the cyclotomic Hecke algebra (of type G(l, 1, n)) of degree n with parameters q and κ. Brundan and Kleshchev have shown that Hn is a Z-graded algebra, and Hu and Mathas have shown that it is in fact a graded cellular algebra. The cellular structure agrees with that of the Dipper–James–Mathas construction. The cell modules are indexed by the set Pnl of l-multipartitions of n and the simple modules by a certain subset Θ ⊂ Pnl of these. Ariki’s theorem tells us that in fact there are many different parameterisations Θ ⊂ l Pn for the simple modules, and the Dipper–James–Mathas setup sees only one of these. We would like different cellular structures for each one of these. Aim Study graded decomposition numbers for Hn corresponding to various parameterisations of the simple modules. In order to do this, we must lift to the setting of quasi-hereditary covers of Hn . We will see that the diagrammatic Cherednik algebra depends on a weighting, θ, which in turn determines which parameterisation Θ we are seeing. (For any θ we get a quasi-hereditary cover of Hn , and different covers may yield different parameterisations of simples.) 2 The diagrammatic Cherednik algebra We will now discuss the combinatorics of Webster’s diagrammatic Cherednik algebra. DefinitionP2.1. A partition of n is a weakly decreasing sequence λ = (λ1 , λ2 , . . . , λk ) such that |λ| = P λi = n. An l-multipartition of n is an l-tuple of partitions (λ(1) , λ(2) , . . . , λ(l) ) such that |λ(i) | = n. We take a mirrored-Russian convention for drawing our Young diagrams. For example, the Young diagram for the partition (4, 1) is drawn as 1 2 Liron Speyer Definition 2.2. The residue of a node (r, c, m) ∈ [λ] (i.e. the node in the rth row and cth column of the Young diagram [λ(m) ]) is defined to be κm + c − r mod e. Example. Suppose κ = (0), e = 4 and λ = (4, 1). Then the residues are given below. 3 2 1 0 3 Given a weighting θ ∈ Rl , we have an associated diagrammatic Cherednik algebra A(n, θ, κ) which is a quasi-hereditary cover of the cyclotomic Hecke algebra Hn . Definition 2.3. Given a weighting θ and some λ ∈ Pnl , we draw the Young diagram [λ] of λ by placing the first node of the jth component at point θj on the x-axis, with all boxes have diagonals of length 2. We tilt our Young diagrams ever so slightly clockwise, to ensure that the top vertex of each box is at a different x-coordinate. (The loading iλ is the n-tuple of real numbers given by projecting the top vertices of boxes of [λ] onto the real line, along with the residue associated to each box.) Example. Suppose θ = (0, 0.5). If λ = ((3), (12 )) and µ = ((2), (2, 1)), then we draw [λ] and [µ] as below, with the loadings given by projections onto the real line. -2 -1 0 0.5 1.5 -1 -0.5 0 0.5 1.5 There is a partial order on Pnl , which we denote by Qθ and call the θ-dominance order, which is a coarsening of the usual dominance order when l = 1. For l > 1, there is some subtle dependence on θ. Let λ, µ ∈ Pnl . We have a notion of semistandard tableaux of shape λ and weight µ, which has a technical definition but is analogous to the usual notion. We denote the set of semistandard tableaux of shape λ and weight µ by SStd(λ, µ). Theorem 2.4 (Webster). The diagrammatic Cherednik algebra A(n, θ, κ) is a graded cellular algebra with respect to the θ-dominance order and a basis indexed by SStd(λ, µ) as λ and µ range over Pnl . In particular we have graded standard modules ∆(λ) = hCT | T ∈ SStd(λ, −)iF with graded simple heads L(λ) forming a complete set of graded simple modules, up to grading shift. Over C, the module category of A(n, θ, κ) is equivalent to category O for the rational cyclotomic Cherednik algebra. If θ is well-separated (that is, θj − θk >> 0 for all j and k), then A(n, θ, κ) is Morita equivalent to the q-Schur algebra of Dipper–James–Mathas over arbitrary fields. We would like to compute the graded decomposition numbers dλµ (v) for A(n, θ, κ). Graded decomposition numbers for the diagrammatic Cherednik algebra 3 3 Decomposition number results for A(n, θ, κ) Definition 3.1. We call a weighting θ ∈ Rl FLOTW if 0 < θj − θi < 1 for all i < j. Lemma 3.2. If θ ∈ Rl is a FLOTW weighting, then the set of one-column l-multipartitions (m) (i.e. the set of l-multipartitions λ such that λi = 0 or 1 for all 1 6 m 6 l and for all i) is saturated in the θ dominance order. Therefore, for any such λ, we have that the graded decomposition number [∆(λ) : L(µ)]v 6= 0 only if µ is also a one-column multipartition. (We may thus restrict our attention to a subcategory of modules whose simple constituents are labelled by one-column multipartitions.) Theorem 3.3 (Bowman, Cox, S.). Suppose F = C and let λ and µ be one-column l-multipartitions of n. The graded decomposition number [∆(λ) : L(µ)]v is an affine Kazhdan–Lusztig polynomial of affine type Âl−1 . We can compute these efficiently using an algorithm of Soergel’s in an alcove geometry. Note that when l = 2, we recover a result for the Temperley–Lieb algebra of type B, sometimes also called the blob algebra. In this case, we have also recovered the submodule structure of all cell modules ∆(λ). Now let F be a field of arbitrary characteristic again. We have the following result, which is a generalisation of results of Kleshchev, Chuang–Miyachi–Tan and Tan–Teo. Theorem 3.4 (Bowman, S.). Suppose κ contains i with multiplicity at most one, and that λ and µ are l-multipartitions of n such that λ Qθ µ and one can be obtained from the other by moving nodes of residue i. Then [∆(λ) : L(µ)]v is a Kazhdan–Lusztig polynomial of type Ar ×As ⊆ Ar+s . Moreover, we can explicitly compute [∆(λ) : L(µ)]v using a result of Tan–Teo. Remark. • In fact, we do better than this. So long as S = {i1 , . . . , ij } is an adjacency free set of residues, λ may differ from µ by moving nodes of any residues in S. • The computation of [∆(λ) : L(µ)]v relies only on the order of addable and removable i-nodes in λ and µ, reading from left to right. • Our results are at the level of isomorphisms of graded algebras (certain subquotients of A(n, θ, κ)), and we are able to reduce the problem to that solved by Tan and Teo. Example. Let e = 4. In level 1, we could have λ = (9, 8, 7, 5), µ = (8, 72 , 6, 1). 0 3 2 2 1 1 0 0 0 3 3 3 2 2 2 3 1 1 1 1 2 0 0 0 0 1 3 3 3 1 0 0 2 2 0 1 3 3 3 2 2 2 2 1 1 1 1 0 0 0 0 3 3 3 2 2 1 0 Definition 3.5. Suppose for some a ∈ R we can draw a vertical line at x = a such that it intersects the same nodes in both λ and µ, and the number of nodes to the left (and 4 Liron Speyer therefore right) of the line x = a is the same in both λ and µ. Then we say that the pair (λ, µ) admits a θ-diagonal cut at x = a. In this case we define µL to be the smallest multipartition containing all nodes which intersect the line x = a as well all nodes to the left. Similarly, we define µR to be the smallest multipartition containing all nodes which intersect the line x = a as well all nodes to the right. Likewise we may define λL and λR . Theorem 3.6 (Bowman, S.). Suppose λ Qθ µ and (λ, µ) admits a diagonal cut at x = a for some a ∈ R. Then dλµ (v) = dλL µL (v) × dλR µR , and, furthermore, ExtkA(n,θ,κ) (∆(λ), ∆(µ)) ∼ = M ExtiA(nL ,θ,κ) (∆(λL ), ∆(µL )) ⊗ ExtjA(nR ,θ,κ) (∆(λR ), ∆(µR )), i+j=k where nL = |λL | = |µL | and nR = |λR | = |µR |. Examples. 1. Fix e = 3 and θ = 0. The partitions (5, 4, 3, 2, 1) and (43 , 13 ) admit a diagonal cut at x = 0.5. We have λL = (5, 4, 3), µL = (43 ), λR = (33 , 2, 1), µR = (33 , 13 ). These are depicted below. 2. Let θ = (0, 50), λ = ((9, 3, 23 , 14 ), (2, 12 )) and µ = ((6, 42 , 22 , 1), (32 , 1)). Then (λ, µ) admits a θ-diagonal cut at x = 1.5. Graded decomposition numbers for the diagrammatic Cherednik algebra 5 Then we have λL = ((9, 3, 22 ), (∅)), λR = ((25 , 14 ), (2, 12 )), µL = ((6, 42 , 2), (∅)), µR = ((25 , 1), (32 , 1)). ∅ ∅ 3. Let θ = (0, 0.5), λ = ((11, 9, 7, 32 , 2, 13 ), (9, 4, 2, 14 )) and µ = ((10, 9, 8, 4, 3, 15 ), (8, 4, 2, 14 )). Then (λ, µ) admits a θ-diagonal cut at x = 2.6. Then we have λL = ((11, 9, 7, 32 ), (9, 4, 2)), λR = ((35 , 2, 13 ), (17 )), µL = ((10, 9, 8, 4, 3), (8, 4, 2)), µR = ((35 , 15 ), (17 )). 6 Liron Speyer