REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT)
|
|
- Arthur Allison
- 5 years ago
- Views:
Transcription
1 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT CONNOR AHLBACH AND JOSH SWANSON Abstract. Thrall [Th] famously considered the computation of certain Schur characters L λ of GL(V -modules L λ coming from the study of free Lie algebras. Kraskiewicz-Weyman [KW] computed the Schur expansion of L (n using major indices of standard Young tableaux. On the other hand, numerous authors have related the L λ to certain group actions involving cyclic groups. Later, Reiner-Stanton-White [RSW] defined the cyclic sieving phenomenon associated to a cyclic group action. Motivated by Kraskiewicz-Weyman s result, we formulate and prove a refined cyclic sieving result for the natural cyclic action on words using the major index generating function and cyclic descents. As an application, this yields a new, nearly bijective proof of Kraskiewicz- Weyman s characterization, as well as Schocker s generalization for higher Lie characters. 1. Introduction We begin by summarizing our main results. Most definitions are deferred to Section 2. Let W n denote the set of words in the alphabet P = {1, 2, 3,...} of length n and let C n := σ denote the cyclic group of order n, which acts on W n via (1 σ w 1 w n 1 w n := w n w 1 w n 1. Let W (α, δ denote the subset of W n consisting of words with content α and cyclic descent type δ. Finally, let W (α, δ maj (q denote the corresponding major index generating function. The Cyclic Sieving Phenomenon was introduced by Reiner, Stanton, and White [RSW]. We show: Theorem 1. The triple (W (α, δ, C n, W (α, δ maj exhibits the cyclic sieving phenomenon. One may deduce the less refined version of Theorem 1 with W (α, δ replaced by δ W (α, δ from Theorem 1.1 in [RSW]; see Theorem 33. We recover Kraskiewicz-Weyman s result from Theorem 1 in Theorem 33, as well as Schocker s generalization in 36. Each step of our argument is bijective, except for the application of Theorem 1. A bijective proof of the following Date: May 19,
2 2 CONNOR AHLBACH AND JOSH SWANSON reformulation of Theorem 1 would make the argument in Theorem 33 fully bijective: the modular major index and the frequency-lex statistics on W (α, δ are equidistributed. Our proof of Theorem 1 is combinatorial and involves an analysis of the sequences of runs and falls of elements in W (α, δ using a version of a standard insertion lemma (Lemma 7. We organize the relevant combinatorial data into certain trees T (α, δ whose edges are labeled by sets and multisets. A key step is the following product formula: Theorem 2. Suppose α = n with α 1 0. Set g := gcd(α 1,... α m, δ 1,... δ m, β x = α α x, and k = k k x. Then W (α, δ maj (q n m ( ( βx 1 k x 1 kx + (α x δ x 1 (mod q g 1 α 1 δ x α x δ x The reduction mod q g 1 is essential. The remainder of the proof of Theorem 1 analyzes cyclic sieving on sets and multisets and essentially builds up Theorem 1 from Theorem 2. In Section 7 we discuss the application of Theorem 1 to Thrall s problem. The paper is organized as follows. In Section 2 we give combinatorial background on cyclic sieving and words. Section 3 summarizes the relevant representation theory and Lie representations for use in Section 7. Section 4 analyzes T (α, δ and proves Theorem 2, while Section 5 deduces Theorem 1. Section 6 discusses a naturally arising instance of cyclic sieving in terms of a method inspired by Wagon-Wilf [WW]. Finally, Section 7 applies the preceding results to Thrall s problem. Section 8 discusses generalizations to higher Lie characters. Section 9 discusses another generalization of Section 7 to arbitrary cyclic sugroups of S n. Section 10 concludes with a few remarks on symmetric and quasisymmetric functions and to the represenation of S n acting on its own conjugacy classes by conjugation. q q 2. Combinatorial Background In this section, we briefly recall or introduce combinatorial notions on words and fix our notation. We use the alphabet of positive integers P throughout. A word w of length n =: w is a sequence w = w 1 w 2... w n of letters w i P. The descent set of w is Des(w := {1 i < n : w i > w i+1 } and the number of descents is des(w := # Des(w. The major index of w is maj(w := i Des(w i. The cyclic descent set of w is CDes(w := {1 i n : w i > w i+1 }, where now the subscripts are taken mod n, and we write cdes(w := # CDes(w. Any position that is not a cyclic descent is a cyclic weak ascent, though we often drop the word weak. The cyclic major
3 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 3 index is cmaj(w := i CDes(w i. Often the distinction between cyclic major index and major index will be irrelevant as cmaj(w n maj(w. We use periods to indicate cyclic descents throughout the paper. For example, if w = , then Des(w = {3, 4, 7}, des(w = 3, CDes(w = {3, 4, 7, 8}, cdes(w = 4, maj(w = 14, and cmaj(w = 22. A necklace is an orbit of the set W n of words of length n under the action of the cyclic group C n by rotation as in (1. The product of two words is their concatenation. A word is primitive if it is not a power of a smaller word. Any word w may be written uniquely as w = v f for f 1 maximal, in which case v is primitive. We call v the period of w, written period(w, and f the frequency of w, written freq(w. The period of a word is the size of its necklace, and freq(w period(w = w. Primitivity, period, frequency, and cdes are all well-defined (constant on necklaces. Example 3. The necklace of w = = ( is [w] := { , , , }, which is not primitive, has period 4, frequency 2. Since the cyclic action on words moves each cyclic descent forward one place modulo n, cmaj forms an arithmetic sequence mod n on [w] with differences k := cdes(w. Hence, cmaj of a necklace is well-defined (constant modulo d := gcd(n, k. The content of a word w, denoted cont(w, is the sequence α whose j-th part is the number of j s in w. Let w (i denote the subsequence of w with all elements larger than i removed. For w W n, cont(w is a weak composition of n. Thus, if m := max(w, we have a filtration 1 α 1 = w (1 w (2 w (m 1 w (m = w, where u v means that u is a subsequence of v. We think of this filtration as building up w by recursively adding all of the copies of the next largest letter where they fit. The cyclic descent type of a word w is the sequence which tracks the number of new cyclic descents at each stage of the filtration. Formally, CDT(w := (cdes w (1, cdes w (2 cdes w (1,..., cdes w (m cdes w (m 1. Note CDT is well-defined (constant on necklaces since rotating w rotates each w (j. We write W (α := {w W n : cont(w = α}, W (α, δ := {w W n : cont(w = α, CDT(w = δ}. We also let N n, N(α, N(α, δ denote the set of necklaces (orbits that appear in W n, W (α, W (α, δ, respectively. Note that we could define W (α, δ more symmetrically by replacing cont with an analogous cyclic weak ascent type
4 4 CONNOR AHLBACH AND JOSH SWANSON which would be the point-wise difference of cont and CDT. However, cont is ubiquitous, so we use it instead. Example 4. Suppose w = We find w (1 = 1111 cdes w (1 = 0, w (2 = cdes w (2 = 2, w (3 = cdes w (3 = 3, w (4 = cdes w (4 = 5. Hence cont(w = (4, 3, 2, 3 and CDT(w = (0, 2 0, 3 2, 5 3 = (0, 2, 1, 2. Given a necklace, consider ordering its elements lexicographically, and assign indexes to these elements starting from 0. The lex statistic lex(w of a word w is the index so assigned to w when lexicographically ordering the necklace of w. For instance, the necklace in Example 3 has lex statistics 0, 3, 2, 1, respectively, so that lex( = 3. The flex statistic of a word w is the product flex(w := freq(w lex(w, so flex( = 2 3 = 6. We briefly recall the cyclic sieving phenomenon (CSP. For more, see [RSW] or the excellent survey in [SaCSP]. Suppose the cyclic group C n of order n acts on a set X, and f(q N[q] is a polynomial in q. In this paper, f(q will be some statistic generating function on X. The triple (X, C n, f(q is said to exhibit the cyclic sieving phenomenon (CSP if for any τ C n, letting X τ := {x X : τ x = x}, #X τ = f(ω p, where p is the order of τ in C n, and ω p is a primitive p-th root of unity. Given a function stat: X Z on a set X, the corresponding generating function is written as X stat (q := x X q stat(x N[q, q 1 ]. We sometimes use the natural multivariable analogue of this notation. For instance, viewing cont: W n Z n, flex: W n Z, we have (cont wn Wn cont,flex (x 1,..., x n, q := x (cont w 1 1 x n q flex w. x X Finally, we assume familiarity with the elementary combinatorics and notation associated to representations of the symmetric group, such as partitions, compositions, (semistandard Young tableaux, descent sets of tableaux, the RSK algorithm, and Schur functions. For details, see [Fu] and/or [Sa].
5 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 5 3. Representation Theory Background We next recall key aspects of Thrall s problem. We assume some familiarity with the representation theory of symmetric and general linear groups over C, though we recall some key points. See [Fu] for more details. For simplicity, all representations are over C, so G-modules are CG-modules. The complex irreducible inequivalent representations of S n are the Specht modules S λ for λ n. The Frobenius character ch is the additive map which sends S λ to the Schur function s λ (x 1, x 2,... := T SSYT(λ xt. Given a subgroup H of S n and a (left H-module M, the induced S n -representation is M Sn H := CS n CH M = Hom CH (CS n, M. Let V be a complex vector space of dimension m. Endow V n with the natural left GL(V -action by linear substitutions and the natural right S n -action by permutation of indexes. Given an S n -module N, define a corresponding GL(V -module E(N := V n CSn N, which we call the Schur module of N. The irreducible inequivalent polynomial representations of GL(V are precisely the Schur modules E λ := E(S λ associated to Specht modules S λ where λ has at most dim(v parts. Pick a finite-dimensional, polynomial representation E of GL(V and a basis, v 1,..., v m for V, and consider the action of a diagonal matrix diag(x 1,..., x m GL(V as a linear endomorphism of E. The Schur character of E, ch(e, is the function which sends (x 1,..., x m to the trace of diag(x 1,..., x m. Indeed, ch(e N[x 1,..., x m ] and ch E(S λ = s λ (x 1,..., x m, 0, 0,.... Given an S n -module N, we then have in the m limit equality of Schur and Frobenius characters, ch E(N = ch N. In light of this, we often leave dependence on m or V implicit. The Littlewood- Richardson rule gives a combinatorial description for the coefficients c λ µν N in ch(e µ E ν = λ c λ µνe λ. We turn to Thrall s problem. As motivation, the Schur module associated to the regular representation is simply V n. The corresponding Schur character is then (2 ch V n = ch E(CS n = s λ = f λ s λ λ n T SYT λ λ n
6 6 CONNOR AHLBACH AND JOSH SWANSON where f λ := # SYT(λ. Our version of Thrall s problem is to find the Schur expansion of certain GL(V -modules arising from the study of free Lie algebras, which we next describe. The tensor algebra of V is T (V := n=0 V n, which is naturally a graded GL(V -representation. Let L(V denote the free Lie algebra on V over C, which is the Lie subalgebra of T (V generated by V. The universal enveloping algebra U(L(V is T (V itself, which is a standard consequence of the PBW theorem. Consider choosing an ordered basis for L(V by successively picking ordered bases for the homogeneous components L n (V := L(V V n. By the PBW theorem, weakly increasing monomials in this basis yield a basis for U(L(V = T (V. It is then straightforward to see that we have a degree-preserving vector space isomorphism (3 U(L(V = Sym m 1 (L 1 (V Sym m 2 (L 2 (V, λ=1 m 1 2 m 2 where the sum is over all partitions of all non-negative integers and Sym m (E denotes the mth symmetric power of a GL(V -representation E. A more careful argument involving Hall bases yields an isomorphism as in (3, except of GL(V -modules. For details, see [Re] (Lemma (Indeed, a comparison of characters gives an abstract isomorphism of GL(V -modules directly. Thus, given λ = 1 m 1 2 m2, we define the higher Lie module To summarize, (4 L λ (V := Sym m 1 (L 1 (V Sym m 2 (L 2 (V. T (V = λ L λ (V = λ ( i L (i m i where L (a b = Sym b (L a (V, L (n (V = L n (V = V n L(V T (V as GL(V -modules. Since our aim is the consideration of the coefficients of s µ in ch L λ, it follows from (4 and the Littlewood-Richardson rule that we may restrict attention to rectangular partitions λ = (a b. The one-row case of this problem was solved by Kraskiewicz-Weyman: Theorem 5 ([KW]. We have ch L (n = λ n a λ,1 s λ. where a λ,1 = #{T SYT(λ : maj(t n 1}. Our primary motivation is giving a proof of Theorem 5 which is as bijective as possible; see Section 7. We next summarize the representation theory of generalized permutation groups. For a finite group G, the group G S n is the semidirect product G n S n where S n acts on G n by permuting the entries, as in σ (g 1,..., g n = (g σ 1 (1,... g σ 1 (n.
7 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 7 More specifically, we multiply two elements in G S n as follows: (g 1,..., g n, σ (h 1,..., h n, τ = (g 1 h σ 1 (1,..., g n h σ 1 (n, στ for all g 1, h 1,... g n, h n G and σ, τ S n. In particular, C a S b is more concretly the subgroup of S ab generated by the cycles c i for i = 1,..., b, and ˆσ for σ S b, where c i = ( ((i 1a + 1 ((i 1a (ia ˆσ( (i 1a + j = (σ(i 1a + j for all i = 1,..., b, and j = 0,... (a 1. That is, c i cycles the interval Ia i := [(i 1a + 1, ia], and ˆσ permutes these b intervals Ia i according to σ. The abelianization of C a S b is C a (S b /A b. For example, at b = 2 we find that #(C a S 2 /#(C a C 2 = 2!a2 2a = a, so 1/a-th of the dimensions of the regular representation of C a S 2 are accounted for by one-dimensional representations. 4. Major Index, Runs, and Falls Throughout this section, w W n, and k = cdes(w. Our definitions may differ slightly from other literature because we are interested in cyclic phenomena. A run of a word w is a maximal sequence of consecutive, weakly increasing entries, allowing wrapping. That is, w i w i+1... w j is a run of w if and only if w i 1 > w i w i+1 w j > w j+1, where the indices are taken modulo n. Similarly, a fall of a word w is a maximal sequence of consecutive, strictly decreasing entries, allowing wrapping. Observe that the runs are separated by the cyclic descents, and the falls are separated by cyclic weak ascents. Example 6. If w = = , then the runs of w are 1126, 5, 346, and the falls are 2, 653, 4, 61,1. As each position is either a cyclic descent or ascent, w has k runs and n k falls. Note the constant word w = a n has n has 0 runs and n falls. (This is consistent with definition since no weakly increasing sequence is maximal as we can wrap around forever. Our next task is to characterize what happens to maj when we insert another element to w. Suppose v is obtained from w by inserting an x in some position. Then, cdes(v = cdes(w, or cdes(v = cdes(w + 1. If cdes(v = cdes(w, we say x is added to a NDA (non-descent-adding position. In this case, the number of runs is unchanged, so x must fit into one of the runs of w - that is, x increases the length of one of the runs
8 8 CONNOR AHLBACH AND JOSH SWANSON of w by 1. Else, cdes(v = cdes(w + 1, and we say x is added to a DA (descent-adding position. In this case, the number of falls is unchanged because v cdes(v = w + 1 (cdes(w + 1 = w cdes(w. Thus, x must fit into one of the falls of w - that is, x increases the length of one of the falls of w by 1. See the example after the proof of Lemma 7. We label the runs of w from left to right with 0 to (k 1 and the falls from 0 to (n k 1 starting from whichever run, fall involves the first letter of the word. We describe a few variations on how maj changes when we add a letter to a word. Lemma 7. Let v be the word w with an x inserted in any position, and let k = cdes(w. Then, x fits into some run or some fall of w, and 0 if x fits into 0-th run of w at end, k i if x fits into i-th run of w, cmaj(v cmaj(w = n + 1 if x fits into 0-th fall of w at end. k j if x fits into j-th fall of w Proof. (a If x fits into the 0-th run of w at the end, then the cyclic descent positions remain the same, so cmaj(w = cmaj(v. (b If x fits into the i-th run of w (at the start if i = 0, then x pushes the k i cyclic descents after this run forward 1 index, so cmaj(v cmaj(w = k i. (c Finally, if x fits into the 0-th fall of w at the end, then we add a cyclic descent at the end, with the rest of the cyclic descent positions unchanged. So, cmaj(v cmaj(w = n + 1. (d If x fits into the j-th fall of w, (at the start if j = 0, x pushes the n k j cyclic ascents after this fall forward 1 index. Consider the comaj, given by comaj(w := i [n]\cdes(w (i = ( n+1 2 cmaj(w. Thus, comaj(v comaj(w = n k j, which we can rewrite as cmaj(v cmaj(w = k j. Example 8. Consider w = , which has cmaj(w = = 21, n = 8, and k = 4. We look at an example of each of the four cases of Lemma 7, underlining the element we add. (a If v = , then cmaj(v = 21 = cmaj(w. (b If v = , then 6 fits into run 1, and cmaj(v = 24 = cmaj(w+ k 1. (c If v = then 8 fits into fall 3, and cmaj(v = 29 = cmaj(w+ k (d If v = , then cmaj(v = 30 = maj(w + n + 1.
9 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 9 Given w and x, there are (n + 1 positions in which we can add x, which correspond bijectively with the possibilities listed in Lemma 7, with (a removed if x > w 1, and (c removed if x w 1. Lemma 7 is the engine of Sections 4 to 6. For our purposes, the distinction between DA versus NDA positions is the critical piece we need to demonstrate cyclic sieving for the triple (W (α, δ, C n, W (α, δ maj (q. Weaker results that characterize the change in major index upon insertion of new element are well-known, such as the following Corollary, see [No] or [Gu], which we deduce from our stronger result. Corollary 9. Suppose w is a word of length n, and x is an element that does not appear in w. Letting w i (x denote the word formed by inserting x after the i-th position of w, and at the start if i = 0. Then, (maj(w 0 (x, maj(w 1 (x,..., maj(w n (x maj(w is a permutation of {0, 1,... n}. Proof. Suppose we always treat position n as a cyclic ascent (whether or not it is according to the definition, which replaces cmaj by maj and excludes (c in Lemma 7. The other cases of Lemma 7 correspond bijectively to positions of w in which to add x, so (maj(w 0 (x, maj(w 1 (x,..., maj(w n (x maj(w is a permutation of {0, 1,... n}. Next, we deduce a special case of Lemma 7 when w ends in a 1, and we forbid x from being added to the end of the word. The motivation for these restrictions is that we can rotate a word with a 1 so it ends in a 1, and more importantly, we need there to be 1 position between any two adjacent elements, as there is in a necklace. Without this cyclic symmetry of 1 position between any two elements, many of our future arguments would fall apart. Conveniently, ending in a 1 means position n is a cyclic ascent, so cmaj reduces to maj. Lemma 10. Suppose w is a word ending in 1. Let v be obtained from w by inserting x in any position but the end. Then, { k i if x fits into the i-th run of w maj(v maj(w = k j if x fits into the j-th fall of w Proof. We can never add to 0th run or 0th fall at the end of the word, so (a and (c of Lemma 7 disappear. We are left with the remaining cases of Lemma 7, and we can replace cmaj by maj, as stated above. Finally, we consider the insertion of multiple x s into w. First, we will need to break down this insertion as inserting x s into certain falls, then
10 10 CONNOR AHLBACH AND JOSH SWANSON inserting x s into certain runs. Let [a, b] denote the set of integers between a and b inclusive, and [a] := [1, a]. We let X Y denote that X is a multisubset of Y, so for example, {1, 1, 1, 2, 2, 4} [4]. Lemma 11. Let w end in a 1. Forbidding insertion at the end, any other insertion of α x x s into w can be viewed uniquely as the insertion of x s into some subset {a 1, a 2,..., a δx } [0, n k 1] of falls of w to obtain u, followed by the insertion of x s into a multisubset {b 1, b 2,..., b αx δ x } [0, k+δ x 1] of the runs of u. Proof. Consider the descent-adding (DA positions of w that at least one x gets added to, and say there are δ x such positions. This accounts for all of the DA x s, say which fit into some subset {a 1, a 2,..., a δx } [0, n k 1] of the falls of w, as no two x s can fit into the same fall, to obtain u. Now, the rest of the x s to be inserted into u are NDA, and since multiple x s could fit into a run, the runs inserted into give a multisubset {b 1, b 2,..., b αx δ x } [0, k + δ x 1] of the runs of u, as u has δ x more runs than w. On the other hand, given any subset {a 1, a 2,..., a δx } [0, n k 1] and multisubset {b 1, b 2,..., b αx δx } [0, k + δ x 1], there is a unique way to insert x s to fit into falls a 1, a 2,..., a δx of w to get u, and then insert x s to fit into runs b 1, b 2,..., b αx δx of u. Example 12. Let w = and v = be w with some 4 s added to it. If we first add DA 4 s, we get u = , which is w with 4 s added to falls {2, 6} [0, 6]. Adding NDA 4 s to u gives v = , which is u with 4 s added to runs {0, 1, 2, 2} [0, 4]. So, if w ends in a 1, each step of the filtration from w (x 1 to w (x consists of the insertion of x s into some subset {a 1, a 2,..., a δx } [0, n k 1] of falls of w (x 1 to obtain v, followed by the insertion of x s into a multisubset {b 1, b 2,..., b αx δ x } [0, k + δ x 1] of runs of v to obtain w (x. A clearer way to present this building up process is in the form of a rooted tree. Let W 1 (α, δ = {w W (α, δ : w ends in a 1}. Definition 13. Form a rooted, vertex- and edge-labeled, directed, tree graph T recursively as follows. Begin at the 0th step with root. At the x-th step, at each terminal vertex w in T, pick a subset A of the falls of w, compute the result w of inserting x into falls A of w, and then pick a multiset B on the runs of w. Let w be the result of inserting x into runs B of w. Add an edge from w to w labeled by (x, A, B. For each word w that ends in a 1, let T w be the subgraph of T consisting of paths to terminal vertices that
11 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 11 are rotations of w ending in a 1. Finally, let T α,δ be the union (in T of T w as w ranges over W 1 (α, δ. The uniqueness in Lemma 11 guarantees that T is a tree. Each w W 1 (α, δ occurs once as a terminal vertex in T (α, δ, there is a unique path from the root to w in T (α, δ. Example 14. If w = , then the graph T w, a subgraph of T ((4, 2, 3, (0, 2, 3 is (1,, {0, 0, 0, 0} 1111 (2, {0, 2}, (2, {1, 3}, (3, {0}, {0, 1} (3, {2}, {1, 2} (3, {1}, {0, 1} (3, {3}, {1, 2} Corollary 15. Let w be a word ending in a 1. If we insert x s into the {a 1, a 2,..., a δx } [0, n k 1] falls of w to obtain v, followed by the insertion of x s into a multisubset {b 1, b 2,..., b αx δx } [0, k + δ x 1] of runs of u to obtain v, forbidding insertion at the end, then maj(v maj(w = (k + 1δ x + δ x i=1 a i + ( δx 2 + α x δ x (k + δ x b j. Proof. We add the DA x s from left to right. Note that after adding (i 1 of these DA x s, the number of runs is now k + (i 1, so when we add to the a i -th fall, the maj is increased by k + (i a i. Hence, δ x δ x ( δx maj(u maj(w = (k + (i a i = (k + 1δ x + a i + 2 i=1 We then add the NDA x s from left to right. Since v has k + δ x runs, when we add an x to run b j, we increase maj by k + δ x b j. Hence, maj(v maj(u = (k + δ x b j. α x δ x i=1
12 12 CONNOR AHLBACH AND JOSH SWANSON Putting these together gives Corollary 15. Suppose we fix content α = (α 1,..., α m and circular descent type δ = (δ 1,..., δ m. We assume δ 1 = 0 as otherwise, W (α, δ =. Also, without loss of generality, we assume α has no parts of size 0. Else, let ˆα denote α with all parts of size 0 removed, and consider the natural bijection Flatten : W (α W (ˆα which replaces all copies of the i-th smallest element by i. Clearly, Flatten preserves CDes, maj, period, and frequency. Define the reduced CDT (RCDT of w W n by RCDT(w := CDT(Flatten(w. Equivalently, for w W (α, we obtain RCDT(w by removing all indices j in CDT(w where α j = 0. Note CDT(w j = 0 at these indices as well. Example 16. If w = , then Flatten(w = , w has content (1, 0, 2, 1, 0, 0, 2, 0, 1, and CDT(w = (0, 0, 1, 1, 0, 0, 1, 0, 0, RCDT(w = (0, 1, 1, 1, 0. Thus, if ˆδ is δ with all parts δ j where α j = 0 removed, then Flatten restricts to a bijecton Flatten : W (α, δ W (ˆα, ˆδ. In summary, if α has a part of size 0, replace α by ˆα, CDT by RCDT, and δ by ˆδ. So, assume α j 0 for j = 1,... m. Let (β 1,..., β m, (k 1,..., k m be the partial sums β x := α 1 + α α x, k x := δ 1 + δ δ x, so that β x = w (x and k x = cdes(w (x. Let k := cdes(w = k m. Next, we calculate maj after adding x s to falls in A x := {a x,1, a x,2,... a x,δx } [0, β x 1 k x 1 1] and runs in B x := {b x,1, b x,2,... b x,αx δx } [0, k x 1] for x = 2,..., m starting with 1 α 1. From Corollary 15, for all x 2, maj(w (x maj(w (x 1 = (k x 1 + 1δ x + Since maj(w (1 = 0, m maj(w = (k x 1 + 1δ x + = m δ x k x m δ x + δ x i=1 δ x i=1 a i + (a i + ( δx 2 + α x r x ( α δx x δ x + (k x b j 2 ( m δx a i + m (α x δ x + m i=1 α x δ x m ( δx 2 (k x 1 b j. (k x b j
13 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 13 Now, let B x := k x 1 B x [0, k x 1], and define the sum function on sets or multisets by sum(a = a A a. Then, ( m m ( δx m m m maj(w = δ x k x α x + sum(a x + sum(b 2 x ( k m = + n α 1 + (sum(a x + sum(b 2 x. where we used that k x = x j=2 δ j to deduce that ( m m δ x k x 1 + ( δx 2 = ( k. 2 We also recall the formulas ( [0, n 1] sum ( n (q = e k (1, q, q 2,... q n 1 = q (k 2 k k q ( ( [0, n 1] n + k 1 sum (q = h k (1, q, q 2,... q n 1 = k k Therefore, by uniqueness of paths in T (α, δ to each terminal vertex in W 1 (α, δ, the maj generating function on W 1 (α, δ is given by m ( W 1 (α, δ maj (q = q (k 2+n α 1 [0, βx 1 k x 1 1] sum ( [0, kx 1] (q sum (q δ x α x δ x m ( ( = q (k 2+n α 1 q (δx 2 βx 1 k x 1 kx + (α x δ x 1 δ x α x δ x This gives the following theorem. Theorem 17. The maj generating function on W 1 (α, δ is given by W 1 (α, δ maj (q = q (k 2+ m ( m ( ( δx 2 +n α 1 βx 1 k x 1 kx + (α x δ x 1 δ x α x δ x Next, each necklace in N(α, δ has α 1 n of its elements in W 1 (α, δ. Since maj on each necklace is well-defined modulo d := gcd(k, n, Theorem 17 tells us the following Theorem about the maj generating function on W (α, δ. Theorem 18. The generating function for maj on the set W (α, δ satisfies W (α, δ maj (q = n q (k 2+ m ( m ( ( δx 2 α 1 βx 1 k x 1 kx + (α x δ x 1 α 1 δ x α x δ x modulo q d 1. q q q. q. q q q
14 14 CONNOR AHLBACH AND JOSH SWANSON Let g := gcd(α, δ := gcd(α 1,... α m, δ 1,... δ m, and notice ( k m ( δx m m + = δ i δ j δ x 0 (mod g. 2 2 i j 2 So, reducing modulo q g 1, we get the following Theorem. Theorem 19. The generating function for maj on the set W (α, δ satisfies W (α, δ maj (q n m ( ( βx 1 k x 1 kx + (α x δ x 1 (mod q g 1 α 1 δ x α x δ x 5. Refining the CSP to fixed content and Circular Descent Type Recall, the triple (X, C n, f(q is said to exhibit the CSP, if for all τ C n, or equivalently, if f(q Orbits R X q #X τ = f(ω τ, R 1 i=0 q i n R (mod q n 1, where the orbits are from the action of C n on X. Such an equivalence makes sense since prescribing values of f(q at the n-th roots of unity determines f(q modulo q n 1. In our case, we can rephrase the CSP as saying that two statistics are equidistributed on W (α, δ. Letting maj n (w = maj(w (mod n we can restate the CSP on (W (α, δ, C n, W (α, δ maj (q as W (α, δ maj n(q = N N(α,δ N 1 j=0 q j freq(n = W (α, δ flex (q The second equality comes from the fact that flex has values j freq(n for j = 0,..., N 1 on necklace N. Hence, the CSP on (W (α, δ, C n, W (α, δ maj (q is equivalent to maj modulo n and flex being equidistributed on W (α, δ. To prove the CSP on (W (α, δ, C n, W (α, δ maj (q, we first show the polynomial W (α, δ maj (q has period gcd(α 1,..., α m, δ 1,... δ m modulo n (Definition 20. Then, we show that the CSP holds for (W (α, δ, C g, W (α, δ maj (q using well-known instances of the CSP on subsets and multisubsets. Putting these results together demonstrates the CSP for (W (α, δ, C n, W (α, δ maj (q. Definition 20. We say a statistic stat : X N has period a modulo b on X if for all i, #{x X : stat(x b i} = #{x X : stat(x b i + a}. Similarly, we say a polynomial f(q has period a modulo b if for all i [0, b 1], h(q q i= h(q q (i+a (mod b. q
15 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 15 where h(q q j is the coefficient of q j in q, and h(q is the unique polynomial of degree at most (b 1 that f(q is equivalent to modulo q b 1. By definition, stat has period a modulo b on X if and only if X stat (q has period a modulo b. Here are some useful facts we will be using frequently: (a If stat has period a modulo b and period b modulo c, then stat has period a modulo c. (b If stat has period a modulo c and period b modulo c, then stat has period gcd(a, b modulo c. The same facts hold if stat is replaced by a polynomial f(q. Theorem 21. maj has period g := gcd(α, δ modulo n on W (α, δ. In order to transfer between W (α, δ and W 1 (α, δ, the following lemma will be useful. Lemma 22. maj has period a modulo n on W (α, δ if and only if maj has period a modulo d on W 1 (α, δ. Proof. By the action of rotation, maj has period d modulo n on W (α, δ. Thus, maj has period a modulo n on W (α, δ if and only if maj has period a modulo d on W (α, δ. As α 1 n of each necklace in N(α, δ shows up in W 1 (α, δ, and maj is well-defined modulo d on necklaces, maj has period a modulo d on W (α, δ if and only if maj has period a modulo d on W 1 (α, δ. We set g := gcd(α 1,... α m, δ 1,... δ m. We prove the result by induction on m, the m = 1 being clear as g = n. For m 2, by Theorem 18, (5 W (α, δ maj (q n α 1 q C G(α, δ, q (mod q d 1 where C = ( k 2 + m ( δx2 α1, and m ( ( βx 1 k x 1 kx + (α x δ x 1 G(α, δ, q = δ x q α x δ x q m ( = q (δx 2 [0, βx 1 k x 1 1] sum ( [0, kx 1] (q δ x α x δ x Let γ x := α x δ x. Notice β x = β x 1 + (k x k x 1 + γ x = β x 1 k x 1 = β x k x γ x sum (q By the action of rotation on multisets in ( ( kx α x δ x and sets in ( [0,β x 1 k x 1 1] δ x, which affects the sum by the size of the multiset or set modulo the size of the interval, we deduce that for all x = 2,..., m,
16 16 CONNOR AHLBACH AND JOSH SWANSON (a G(α, δ, q has period γ x = α x δ x modulo k x. (b G(α, δ, q has period δ x modulo β x 1 k x 1 = β x k x γ x. But by equation 5, if G(α, δ, q has period a modulo b, then maj has period gcd(a, d modulo b on W (α, δ. In particular, for x = m, G(α, δ, q has period γ m modulo k. Because maj has periods d, k modulo n on W (α, δ, maj has period gcd(γ m, d mod n. Also, for x = m, we find that G(α, δ, q has period δ m modulo β m k m γ m = n k γ m. Therefore, W (α, q has period gcd(δ m, d modulo gcd(n k γ m, d on W (α, δ, which reduces to maj having period gcd(δ m, d modulo gcd(γ m, d on W (α, δ. Combining with maj having period gcd(γ m, d mod n on W (α, δ, maj has period gcd(δ m, γ m, d on W (α, δ. Now, inductively assume maj has period g := gcd(α 1,..., α m 1, δ 1,... δ m 1 on W (α, δ, where α = (α 1,... α m 1 and δ = (δ 1,..., δ m 1. By Lemma 22, maj has period g modulo gcd(n α m, k m 1 on W 1 (α, δ. This periodicity extends to the set W 1 (α, δ because a choices of falls then runs to add to and a word in W 1 (α, δ uniquely determines a word in W 1 (α, δ, and the difference of their major indices is independent of the word chosen. Again, by Lemma 22, maj has periods δ m, γ m, d, k modulo d on W 1 (α, δ, so maj has period gcd(n α m, k m 1 modulo d on W 1 (α, δ because (6 n α m = n (δ m + γ m, k m 1 = k δ m. Combining this with maj having period g modulo gcd(n α m, k m 1, we deduce that maj has period g modulo d on W 1 (α, δ. So, by Lemma 22, maj has period g modulo n on W (α, δ. Finally, we already have shown maj has periods α m = δ m + γ m and δ m modulo n on W (α, δ. Therefore, maj has period g modulo n on W (α, δ, proving Theorem 21. We acknowledge the previous proof is the least elegant part of our argument, and we would appreciate improvements toward a more elegant proof. Now, we will prove following Theorem, bringing us closer to proving Theorem 1. Theorem 23. (W (α, δ, C g, W (α, δ maj (q exhibits the CSP. If C n = σ n, then C g = σn n/g C n, and let C g act on W (α, δ as this subgroup of C n. Also, assuming g a, C g acts on ( [a] ( b and [a] b by rotation of values by a/g. Recall that if (X, C n, f(q exhibits the CSP, then so does (X, C g, f(q when g n. And, if (X, C n, f(q, (Y, C n, h(q exhibit the CSP, then so does (X Y, C n, f(qh(q, where C n acts on X Y by τ (x, y = (τ x, τ y [RSW]. It is well-known that ( ([n] k, C n, ( n k q, ( ([n] k, C n, ( n + k 1, k q
17 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 17 each exhibit the CSP from action of rotation on sets and multisets [RSW], so ( ([n] ( ( ([n] ( n n + k 1, C g,,, C g,, k k k k q exhibit the CSP. Because g (β x 1 k x 1, δ x, k x, α x δ x for all x, ( m ( [0, βx 1 k x 1 1] m ( [0, kx 1], C g, G(α, δ, q δ x α x δ x exhibits the CSP. So, letting D x = ( [0,β x 1 k x 1 1] δ x and Ex = m #(Dx#(E τ x τ = G(α, δ, ω τ. for all τ C g. Therefore, by Theorem 19, n α 1 m #(Dx#(E τ x τ = W (α, δ maj (ω τ. q ( ([0,kx 1] α x δ x, Next, we need to relate the cyclic action on sets and multisets above to that on words in W (α, δ. For a subset A [0, r 1] or multisubset A [0, r 1], define period(a to be the number of distinct elements in the orbit of A under rotation of values. The frequency freq(a = r/ period(a counts the maximum number of times A can be written as the same set or multiset repeated one after another. Note the concept of frequency depends on what we view it as a subset of. For example, [0, 3] [0, 5] has frequency 2, but [0, 3] [0, 6] has frequency 1. We relate the concepts of frequency on words, subsets, and multisubsets using the tree decomposition in section 4. Recall each w W 1 (α, δ can be uniquely broken down into a choice of falls A x = {a 1,... a δx } D x and runs B x E x for the x s to fit into. Lemma 24. The frequency of a word w W 1 (α, δ is given by freq(w = gcd(freq(a 2, freq(b 2,..., freq(a m, freq(b m, where A 2, B 2,... A m, B m represent the choices of falls and runs to insert into in order to obtain w. Proof. Since w is a word repeated freq(w times, each A x, B x is a subset or multisubset repeated freq(w times. Conversely, since each A x, B x chosen is a subset repeated h := gcd(freq(a 2, freq(b 2,..., freq(a m, freq(b m, times, w is a word repeated h times. Hence, freq(w = h. Lemma 24 tells us τ C g fixes w W 1 (α, δ if and only if τ fixes the corresponding A 2,... A m, B 2,... B m, so #W 1 (α, δ τ = m #(Dτ x#(e τ x.
18 18 CONNOR AHLBACH AND JOSH SWANSON As α 1 n of each necklace in N(α, δ shows up in W 1 (α, δ, #W (α, δ τ = n m #(D α x#(e τ x τ = W (α, δ maj (ω τ. 1 As this holds for all τ C g, we have proven Theorem 23. Finally, we extend the CSP from C g to C n. By Theorem 23, W (α, δ maj (q N N(α,δ N 1 j=0 q j freq(n (mod q g 1. For any N N(α, δ, freq(n g. Thus, noting freq(n = n/ N, W (α, δ maj (q n g N N(α,δ g N n 1 j=0 q j freq(n (mod q g 1 But now, using that W (α, δ maj (q has period g modulo n, W (α, δ maj (q N N(α,δ N 1 j=0 q j freq(n (mod q n 1, so (W (α, δ, C n, W (α, δ maj (q exhibits the CSP, proving Theorem An Explanation of the CSP for subsets and Multisubsets Fix X = ( [0,n 1] ( k, and f(q = n k q. Though it is known the triple (X, C n, f(q exhibits the CSP, the literature is lacking a more insightful proof that links the elements of X with terms in f(q through a statistic, the subset sum function on X. The idea for our argument comes from Wagon and Wilf in [WW], who studied the distribution of subset sums modulo m, and characterize when the subset sum statistic is equidistributed modulo m. We apply the same rotation within intervals they use, but this time in order to obtain cyclic sieving. Let C n = σ act on X by, for A = {a 1, a 2,..., a k } X, σ A = {a (mod n, a (mod n,... a n + 1 (mod n}. Let d = gcd(k, n. For all x n, and j [0, n x 1], let Ij x := [jx, (j + 1x 1], which we call x-intervals. Let C(a, b = {A X : gcd(a, #(A Ia, 0 #(A Ia, 1..., #(A Ia n/a 1 = b}. In particular, C(a, b is empty unless b a. Define the map Int : X {x : x n} by Int(A is the largest r n such that each r-interval in A is full or empty. Also, let S r = {A X : Int(A = r}.
19 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 19 Example 25. If n = 8, k = 4, Int(0134 = 1, Int(0145 = 2, Int(0123 = 4. We first restrict attention to a particular set of subsets satisfying certain gcd requirements and obtain a result that looks remarkably like cyclic sieving. Theorem 26. Let C := C(n, d 0 C(d 0, d 1 C(d p 1, g C(g, x X. Then, the distribution of sum on C satisfies f C (q := q (k 2 C sum (q r x Proof. We proceed by induction on x. Let Then, (7 #(C S r r g/r 1 q ir (mod q g 1 g B(g, a 0,..., a n/g 1 := {A X : #(A I j g = a j for all j}, C = (a 0,...,a n/g 1 i=0 B(g, a 0,..., a n/g 1, where the union is over all (a 0,..., a n/g 1 satisfying gcd ( d r, a a dr/g 1,..., a (n/g dr/g + + a n/g 1 = dr+1 for all r = 1,..., p, where d p = g, d p+1 = x. In particular, gcd(g, a 0,... a n/g 1 = x, so there exists integers c 0, c 1... c n/g 1 such that c 0 a 0 + c 1 a c n/g 1 a n/g 1 x (mod g Thus, the action on B(g, a 0,..., a n/g 1 of rotating the elements in I j g forward c j times in I j g for all j increases sum statistic by a total of x modulo g. This action extends to C, as C by equation 7. Hence, the polynomial f C (q has period x modulo g. Lemma 27. If C S r, then r x. Proof. Fix A C. First, r n. As A C(n, d 0, and each r-interval is full or empty in A, r d 0. Inductively assuming r d i, as each d i -interval is union of r-intervals, r gcd(d i, #(A Id 0 i, #(A Id 2 i,..., #(A I n/d i 1 d i = d i+1. Hence, r x. For the base case, when x = 1, f C has period 1 modulo g, so f C (q #C 1 g 1 q i = #(C S 1 1 g 1 g g i=0 i=0 (mod q g 1,
20 20 CONNOR AHLBACH AND JOSH SWANSON as C S 1 by Lemma 27. Now, Suppose x 2. By conditioning on the gcd of x and the sizes of the intersections with the x-intervals, C = y x (C C(x, y. First, if y < x, by our induction hypothesis, (8 q (k 2 (C C(x, y sum (q r y #(C C(x, y S r r x/r 1 q jr (mod q x 1 x We claim the same identity holds if y = x as well. First, C C(x, x S x by Lemma 27 and definition of C(x, x. Therefore, the right side of equation (8 with y = x becomes #(C C(x, xq 0. But on the other hand, if A S x, then by shifting x-intervals by multiples of x, we can form the set [0, (k 1], which preserves the sum modulo x. Hence, sum(a ( k 2 x 0. Therefore, j=0 q (k 2 (C C(x, x sum (q #(C C(x, xq 0 (mod q x 1, proving equation (8 when y = x. Summing equation (8 over y x, q (k 2 C sum (q = r x #(C S r r x/r 1 q jr (mod q x 1 x j=0 because C = y x (C C(x, y. However, as f C (q has period x modulo g, f C (q = q (k 2 C sum (q r x proving Theorem 26. #(C S r r g/r 1 q ir (mod q g 1, g i=0 Example 28. Suppose C = C(4, 2 C(2, 1 = {02, 03, 12, 13} S 1. Then, f C (q = q 1 + 2q 2 + q 3 2(q 0 + q 1 (mod q 2 1. Note f C (q is not equivalent to q 0 + q 1 + q 2 + q 3 modulo q 4 1, for Theorem 26 only gives equivalence modulo q 2 1 in this case. Also, if C = C(4, 2 = {01, 02, 03, 12, 13, 23}, then #(C S 1 = 4 and #(C S 2 = 2. Then, f C (q (q 0 + q 1 + q 2 + q 3 + (q 0 + q 2 (mod q 4 1. Both of these calculations agree with Theorem 26. In particular, when C = C(n, d = X, f(q = f X (q r n #S r r n/r 1 q ir (mod q n 1 n Define F r := {A X : freq(a = r}, recalling the the frequency of A is n over the size of the orbit of A under C n. First, x, r x n S x is the subset i=0
21 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 21 of X where the r-intervals are full or empty, so choosing which of the k/r r-intervals are full, ( n/r #S x =. k/r x,r x n Also, x,r x n F x is the subset of X where each n/r-interval is shifted version of the first n/r-interval, which must contain k/r elements. So, ( n/r #F x = k/r x,r x n By Mobius inversion on the divisibility lattice of n, #S r = #F r for all r. Therefore, f(q r n #(F r r n/r 1 q ir = n i=0 Orbits R X R 1 i=0 q i n R (mod q n 1, which demonstrates (X, C n, f(q exhibits the CSP. Furthermore, our proof shows that the terms where the orbits have size n/r correspond to the elements of X in S r under the sum statistic., and h(q = ( n+k 1 k Next, let Y = ( [0,n 1] k multisubsets by, if A = {a 1, a 2... a k } Y, then q. Let C n = σ act on σ A = {a (mod n, a (mod n,... a n + 1 (mod n}. Recall the bijection ϕ : Y Z := ( [0,n+k 2] k given by ϕ({a1, a 2,..., a k } = {a 1, a 2 + 1,..., a k + (k 1}, where a 1 a k. This means sum(ϕ(a = sum(a + ( k 2, so ( n + k 1 h(q = = q (k 2 Z sum (q = Y sum (q. k q Just as before, consider the x-intervals Ix j = [jx, (j + 1x 1], but this time for j = 1, 2,..., ( n+k x 1 so the x-intervals cover [0, n + k 1 x]. Note [n + k x, n + k 2], which we call the x-remainder, is leftover after we remove the x-intervals. In this case, for B Z, define Int(B = max{x : x d, I x j B is full or empty for all j}, which differs from Int defined earlier in that we force the output to divide d = gcd(k, n. Also, let S r = {B Z : Int(B = r}. Thus, S r = unless r d. Then, from the same definitions and reasoning for the proof of Theorem 26, we deduce the following Theorem. Theorem 29. Let C := C(d, d 1 C(d 1, d 2 C(d p 1, g C(g, x Z.
22 22 CONNOR AHLBACH AND JOSH SWANSON Then, the distribution of sum on C is given by f C (q := q (k 2 C sum (q r x In particular, for Z = d 1 d C(d, d 1, h(q = q (k 2 Z sum (q r d #(C S r r g/r 1 q ir (mod q g 1 g i=0 #(S r r d/r 1 q ir (mod q d 1 d The action of rotation of multisubsets shows h(q has period k and thus period d modulo n, so h(q r d i=0 #(S r r n/r 1 q ir (mod q n 1 n i=0 Now, set F r = {A Y : freq(a = r}, recalling the the frequency of A is n over the size of the orbit of A under C n. Note F r = unless r d. Then, r x F x is the set of multisubsets where each n/r-interval is shifted version of the first n/r-interval, which must contain k/r elements. Thus, for all r d, ( n/r + k/r 1 x,r x d F x = ( n/r k/r = But on the other hand, r x S x is the set of subsets where each r-interval is full or empty. For B S x, the condition r d k forces the r-remainder to contain no elements of B. Thus, for all r d, by choosing which k/r of the (n/r + k/r 1 r-intervals is full, ( n/r + k/r 1 S x =. k/r x,r x d By Mobius inversion on the divisibility lattice of d, #S r = #F r. Therefore, h(q r n #(F r r n/r 1 q ir = n i=0 Orbits R Y k/r R 1 i=0. q i n R (mod q n 1 This demonstrates that (Y, C n, h(q exhibits the CSP. Furthermore, our proof shows that the terms where the orbits have size n/r correspond to the inverse image of ϕ of the subsets in S r under the sum statistic. 7. Application to Thrall s Problem In this section, we use the work of Kraskiewicz-Weyman to motivate the cyclic sieving phenomenon on words of a fixed content. Reversing the argument, we also deduce Theorem 5 using Theorem 1 to bridge the cyclic
23 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 23 sieving gap. Every step of the argument except our proof of Theorem 1 is bijective. Since Theorem 1 is significantly more refined than necessary for this application, there is hope for a fully bijective proof of Theorem 5 as well as an underlying representation-theoretic explanation of Theorem 1. We begin by summarizing a classical generating function for ch L n and Klyachko s observation connecting L n to the Schur modules of induced representations of irreducible representations of cyclic groups. Theorem 30 ([Ha], Lemma ; [Re], Theorem 7.5. There is a basis for L(V, called a Hall basis, whose elements are certain iterated bracketings of v 1,..., v m. Moreover, there is a natural, weight-preserving bijection between the Hall basis and the set of primitive necklaces on [m] of length n. Let P N n denote the set of primitive necklaces of length n. Corollary 31. ch L n is the content-generating function of primitive necklaces of length n, i.e. ch L n = P N cont n. Proof. The action of diag(x 1,..., x m on an iterated bracketing of v 1,..., v m multiplies the bracketing by x a 1 1 xam m, where a i is the number of v i s. The trace of the endomorphism induced by diag(x 1,..., x m is the sum of these monomials, so the result follows directly from the theorem. The argument in 31 generalizes to say that if a GL(V -representation M has a basis with simple tensors, then the Schur character of M is the content generating function for the indices which show up in a basis. On the other hand, let χ r be the one-dimensional representation of C n := σ given by letting σ act on the left as multiplication by ωn r := e 2πir/n. Note that CC n = r 1 i=0 χ r. We now summarize and slightly generalize observations due to Klyachko [Kl] (equation (6. Theorem 32. There is a basis for E(χ r Sn C n indexed by necklaces of length n on [m] with frequency dividing r, or equivalently indexed by words of length n on [m] with flex r. Hence n 1 (ch χ r Sn C n q r = Wn cont,flex (x 1, x 2,... ; q. r=0 Proof. The suggested necklace basis comes from a sequence of natural identifications, which we now describe. We may restrict the right S n -action on V n to a C n -action where σ := (12 n S n as needed. Now E(χ r Sn C n = V n CSn χ r Sn C n = V n CCn χ r.
24 24 CONNOR AHLBACH AND JOSH SWANSON We may give χ r a right C n -module structure in addition to its left C n -module structure by declaring that σ acts on the right as multiplication by ωn. r In this fashion, n 1 r=0 χ r = CC n as a (C n, C n -bimodule. Hence we have natural right CC n -module isomorphisms V n CCn CC n = V n = HomCCn (CC n, V n. Applying CC n = n 1 r=0 χ r and linearity gives n 1 n 1 V n CCn χ r = Hom CCn (χ r, V n r=0 i=0 which descend to right CC n -module isomorphisms V n CCn χ r = HomCCn (χ r, V n. Moreover, these isomorphisms respect the (left GL(V -module structures. Since χ r is one-dimensional, we may identify Hom CCn (χ r, V n with the subset of V n of images of 1 C, from which we find E(χ r Sn C n = {v V n : v σ = ω r nv}. = Span C {v i1 v i2 v in + ω r nv in v i1 v in 1 + : i 1 i n W n, 1 i j m}. In the final span, we need only include one word in any fixed necklace. If the necklace has period p, we may factor out a constant 1 + ω rp n + ω 2rp n + + ωn r(n p = ωn n 1 ωn rp 1 which is nonzero if and only if ωn rp = 1, so if and only if rp n 0. Since p = n/f where f is the frequency of the necklace, it follows that rp n 0 if and only if f r. Using this basis, the Schur character E(χ r Sn C n is now clearly the weight generating function of the suggested necklaces. Finally, given a necklace N of frequency f r, flex N = {0, f, 2f,..., n f}, so N contains a unique word of flex r, which completes the proof. At r = 1, we require f = 1, resulting in primitive necklaces. Thus we have arrived at Klyachko s observation [Kl] (Proposition 1, ch E(χ 1 Sn C n = ch L (n. Moreover, the argument used a series of natural bijections. The corresponding Schur expansion is well-known: Theorem 33 ([KW]. We have ch χ r Sn C n = λ n a λ,r s λ where a λ,r = #{Q SYT(λ : maj(q n r}.
25 REFINED CYCLIC SIEVING ON WORDS AND THRALL S PROBLEM (DRAFT 25 We now show that Theorem 33 is equivalent to (W (α, C n, W (α maj(q exhibiting the CSP for all contents α. First, let M n,r = {w W n : maj(w n r}, NF D n,r = {w N n : freq(n r}, and let M n,r (α, NF D n,r (α denote the elements in M n,r, NF D n,r content α, respectively. with Proof. We can restate (W (α, C n, W (α maj(q exhibiting the CSP for all contents α as W (α maj (q N N(α N 1 q j freq(n (mod q n 1, or, by comparing the coefficients of q r, #M n,r (α = #NF D n,r (α for all α, or just Mn,r cont = NF Dn,r cont. First, assuming that (W (α, C n, W (α maj(q exhibits the CSP for all α, and using that under RSK, Des(Q(w = Des(w, ch χ r Sn C n = NF Dn,r cont = Mn,r cont = SSYT(λ cont = a λ,r s λ, λ n λ n Q SYT(λ maj(q nr which proves Theorem 33. Here, {Q SYT(λ : maj(q n r} is the set of possible Q (recording-tableau of a given shape for w M n,r, and for a given Q-tableau, all P (insertion-tableau of the same shape are possible. Secondly, assuming Theorem 33, NF D cont n,r = ch χ r Sn C n = a λ,r s λ = Mn,r cont, λ n so (W (α, C n, W (α maj(q exhibits the CSP for all α. The formula NF Dn,r cont = λ n a λ,rs λ also explains the following symmetry among the a λ,r coefficients. Corollary 34. For all n 1 and λ n, a λ,r := #{T SYT(λ : maj(t n r} only depends on gcd(r, n. Proof. For N N n, freq(n n, so freq(n r if and only if freq(n gcd(n, r. Thus, NF Dn,r cont only depends on gcd(n, r. Hence, the coefficients a λ,r in the Schur expansion of NF Dn,r cont must depend only on gcd(n, r.
26 26 CONNOR AHLBACH AND JOSH SWANSON 8. Application to Higher Lie Modules The one-dimensional irreducible representations of C a S b for b > 1 are as follows. Pick 0 r < a, ɛ {0, 1}. Define χ r,ɛ : C a S b C (i 1,..., i b, τ ω r(i 1+ +i b a ( 1 ɛ sgn τ. Given w W ab, define the standardization of w with size-a intervals, written std b (w, as follows. Put the subwords coming from the size-a intervals of w into lexicographic order, numbered 1, 2,.... Replace each size-a interval with the position of its subword in this order, and then apply the usual standardization map to get an element of S b. For example, if w = with b = 3, then the subwords are 12 < 35, so we compute std b (w by standardizing 121, which yields std 3 ( = 132 S 3. Theorem 35. There is a basis for E(χ r,ɛ S ab C a S b indexed by words w of length ab with the following properties: Each length-a sub-interval of w has flex statistic r. If ɛ = 0, std b (w = id. If ɛ = 1, std b (w = w 0 and the length-a intervals of w are all distinct. Proof. Much of the proof of Theorem 32 may be carried out by replacing C n, S n, χ r with C a S b, S ab, χ r,ɛ. Doing so gives E(χ r,ɛ S ab C a S b = {φ(1 V ab : φ Hom C(Ca S b (χ r,ɛ, V ab }. To obtain a generating set for the right-hand side, first consider the natural action of C a S b on W ab. For convenience, we identify words with their image under the natural map W ab V ab, which is (C a S b -equivariant by definition. We have (C a S b -submodules of V ab spanned by each (C a S b - orbit of W ab, so we may break φ(1 into components arising from these orbits. Now consider w W ab and suppose For g C a S b we have φ(1 = v w c v v v w c v (v g = φ(1 g = φ(1 g = (1 gφ(1 = u w (1 gc u u so we must define c w =: (1 gc w g. This may be inconsistent if the same w g appears multiple times on the right-hand side. Assuming the coefficients are non-zero, consistency is equivalent to requiring Stab(w Stab(1. To describe Stab(w more conveniently, we may assume that any two a-intervals of w with equal necklaces are themselves equal. Letting e i be the ith standard basis vector of Ca, b and letting p i be the period of the ith a-interval
YOUNG TABLEAUX AND THE REPRESENTATIONS OF THE SYMMETRIC GROUP
YOUNG TABLEAUX AND THE REPRESENTATIONS OF THE SYMMETRIC GROUP YUFEI ZHAO ABSTRACT We explore an intimate connection between Young tableaux and representations of the symmetric group We describe the construction
More informationREPRESENTATION THEORY OF S n
REPRESENTATION THEORY OF S n EVAN JENKINS Abstract. These are notes from three lectures given in MATH 26700, Introduction to Representation Theory of Finite Groups, at the University of Chicago in November
More informationDefinitions. Notations. Injective, Surjective and Bijective. Divides. Cartesian Product. Relations. Equivalence Relations
Page 1 Definitions Tuesday, May 8, 2018 12:23 AM Notations " " means "equals, by definition" the set of all real numbers the set of integers Denote a function from a set to a set by Denote the image of
More informationCYCLIC SIEVING FOR CYCLIC CODES
CYCLIC SIEVING FOR CYCLIC CODES ALEX MASON, VICTOR REINER, SHRUTHI SRIDHAR Abstract. These are notes on a preliminary follow-up to a question of Jim Propp, about cyclic sieving of cyclic codes. We show
More informationCombinatorial Structures
Combinatorial Structures Contents 1 Permutations 1 Partitions.1 Ferrers diagrams....................................... Skew diagrams........................................ Dominance order......................................
More informationThe Littlewood-Richardson Rule
REPRESENTATIONS OF THE SYMMETRIC GROUP The Littlewood-Richardson Rule Aman Barot B.Sc.(Hons.) Mathematics and Computer Science, III Year April 20, 2014 Abstract We motivate and prove the Littlewood-Richardson
More informationA combinatorial approach to the q, t-symmetry relation in Macdonald polynomials
A combinatorial approach to the q, t-symmetry relation in Macdonald polynomials Maria Monks Gillespie Department of Mathematics University of California, Berkeley Berkeley, CA, U.S.A. monks@math.berkeley.edu
More informationAdjoint Representations of the Symmetric Group
Adjoint Representations of the Symmetric Group Mahir Bilen Can 1 and Miles Jones 2 1 mahirbilencan@gmail.com 2 mej016@ucsd.edu Abstract We study the restriction to the symmetric group, S n of the adjoint
More informationCharacters, Derangements and Descents for the Hyperoctahedral Group
Characters, Derangements and Descents for the Hyperoctahedral Group Christos Athanasiadis joint with Ron Adin, Sergi Elizalde and Yuval Roichman University of Athens July 9, 2015 1 / 42 Outline 1 Motivation
More informationA NOTE ON TENSOR CATEGORIES OF LIE TYPE E 9
A NOTE ON TENSOR CATEGORIES OF LIE TYPE E 9 ERIC C. ROWELL Abstract. We consider the problem of decomposing tensor powers of the fundamental level 1 highest weight representation V of the affine Kac-Moody
More informationarxiv: v1 [math.co] 2 Dec 2008
An algorithmic Littlewood-Richardson rule arxiv:08.0435v [math.co] Dec 008 Ricky Ini Liu Massachusetts Institute of Technology Cambridge, Massachusetts riliu@math.mit.edu June, 03 Abstract We introduce
More informationRow-strict quasisymmetric Schur functions
Row-strict quasisymmetric Schur functions Sarah Mason and Jeffrey Remmel Mathematics Subject Classification (010). 05E05. Keywords. quasisymmetric functions, Schur functions, omega transform. Abstract.
More informationSkew row-strict quasisymmetric Schur functions
Journal of Algebraic Combinatorics manuscript No. (will be inserted by the editor) Skew row-strict quasisymmetric Schur functions Sarah K. Mason Elizabeth Niese Received: date / Accepted: date Abstract
More informationREPRESENTATIONS OF S n AND GL(n, C)
REPRESENTATIONS OF S n AND GL(n, C) SEAN MCAFEE 1 outline For a given finite group G, we have that the number of irreducible representations of G is equal to the number of conjugacy classes of G Although
More informationA DECOMPOSITION OF SCHUR FUNCTIONS AND AN ANALOGUE OF THE ROBINSON-SCHENSTED-KNUTH ALGORITHM
A DECOMPOSITION OF SCHUR FUNCTIONS AND AN ANALOGUE OF THE ROBINSON-SCHENSTED-KNUTH ALGORITHM S. MASON Abstract. We exhibit a weight-preserving bijection between semi-standard Young tableaux and semi-skyline
More informationNOTES ABOUT SATURATED CHAINS IN THE DYCK PATH POSET. 1. Basic Definitions
NOTES ABOUT SATURATED CHAINS IN THE DYCK PATH POSET JENNIFER WOODCOCK 1. Basic Definitions Dyck paths are one of the many combinatorial objects enumerated by the Catalan numbers, sequence A000108 in [2]:
More informationCourse 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra
Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra D. R. Wilkins Contents 3 Topics in Commutative Algebra 2 3.1 Rings and Fields......................... 2 3.2 Ideals...............................
More informationNOTES ON FINITE FIELDS
NOTES ON FINITE FIELDS AARON LANDESMAN CONTENTS 1. Introduction to finite fields 2 2. Definition and constructions of fields 3 2.1. The definition of a field 3 2.2. Constructing field extensions by adjoining
More informationSince G is a compact Lie group, we can apply Schur orthogonality to see that G χ π (g) 2 dg =
Problem 1 Show that if π is an irreducible representation of a compact lie group G then π is also irreducible. Give an example of a G and π such that π = π, and another for which π π. Is this true for
More informationDUAL IMMACULATE QUASISYMMETRIC FUNCTIONS EXPAND POSITIVELY INTO YOUNG QUASISYMMETRIC SCHUR FUNCTIONS
DUAL IMMACULATE QUASISYMMETRIC FUNCTIONS EXPAND POSITIVELY INTO YOUNG QUASISYMMETRIC SCHUR FUNCTIONS EDWARD E. ALLEN, JOSHUA HALLAM, AND SARAH K. MASON Abstract. We describe a combinatorial formula for
More informationSHIFTED K-THEORETIC POIRIER-REUTENAUER BIALGEBRA
SHIFTED K-THEORETIC POIRIER-REUTENAUER BIALGEBRA ADAM KEILTHY, REBECCA PATRIAS, LILLIAN WEBSTER, YINUO ZHANG, SHUQI ZHOU Abstract We use shifted K-theoretic jeu de taquin to show that the weak K-Knuth
More informationPatterns in Standard Young Tableaux
Patterns in Standard Young Tableaux Sara Billey University of Washington Slides: math.washington.edu/ billey/talks Based on joint work with: Matjaž Konvalinka and Joshua Swanson 6th Encuentro Colombiano
More informationALGEBRA. 1. Some elementary number theory 1.1. Primes and divisibility. We denote the collection of integers
ALGEBRA CHRISTIAN REMLING 1. Some elementary number theory 1.1. Primes and divisibility. We denote the collection of integers by Z = {..., 2, 1, 0, 1,...}. Given a, b Z, we write a b if b = ac for some
More informationQUIVERS AND LATTICES.
QUIVERS AND LATTICES. KEVIN MCGERTY We will discuss two classification results in quite different areas which turn out to have the same answer. This note is an slightly expanded version of the talk given
More informationCitation Osaka Journal of Mathematics. 43(2)
TitleIrreducible representations of the Author(s) Kosuda, Masashi Citation Osaka Journal of Mathematics. 43(2) Issue 2006-06 Date Text Version publisher URL http://hdl.handle.net/094/0396 DOI Rights Osaka
More informationOn certain family of B-modules
On certain family of B-modules Piotr Pragacz (IM PAN, Warszawa) joint with Witold Kraśkiewicz with results of Masaki Watanabe Issai Schur s dissertation (Berlin, 1901): classification of irreducible polynomial
More informationINEQUALITIES OF SYMMETRIC FUNCTIONS. 1. Introduction to Symmetric Functions [?] Definition 1.1. A symmetric function in n variables is a function, f,
INEQUALITIES OF SMMETRIC FUNCTIONS JONATHAN D. LIMA Abstract. We prove several symmetric function inequalities and conjecture a partially proved comprehensive theorem. We also introduce the condition of
More informationMultiplicity-Free Products of Schur Functions
Annals of Combinatorics 5 (2001) 113-121 0218-0006/01/020113-9$1.50+0.20/0 c Birkhäuser Verlag, Basel, 2001 Annals of Combinatorics Multiplicity-Free Products of Schur Functions John R. Stembridge Department
More informationCHROMATIC CLASSICAL SYMMETRIC FUNCTIONS
CHROMATIC CLASSICAL SYMMETRIC FUNCTIONS SOOJIN CHO AND STEPHANIE VAN WILLIGENBURG Abstract. In this note we classify when a skew Schur function is a positive linear combination of power sum symmetric functions.
More informationPOLYNOMIAL BEHAVIOUR OF KOSTKA NUMBERS
POLYNOMIAL BEHAVIOUR OF KOSTKA NUMBERS DINUSHI MUNASINGHE Abstract. Given two standard partitions λ + = (λ + 1 λ+ s ) and λ = (λ 1 λ r ) we write λ = (λ +, λ ) and set s λ (x 1,..., x t ) := s λt (x 1,...,
More informationREPRESENTATION THEORY OF THE SYMMETRIC GROUP (FOLLOWING [Ful97])
REPRESENTATION THEORY OF THE SYMMETRIC GROUP (FOLLOWING [Ful97]) MICHAEL WALTER. Diagrams and Tableaux Diagrams and Tableaux. A (Young) diagram λ is a partition of a natural number n 0, which we often
More informationThe symmetric group action on rank-selected posets of injective words
The symmetric group action on rank-selected posets of injective words Christos A. Athanasiadis Department of Mathematics University of Athens Athens 15784, Hellas (Greece) caath@math.uoa.gr October 28,
More informationNoetherian property of infinite EI categories
Noetherian property of infinite EI categories Wee Liang Gan and Liping Li Abstract. It is known that finitely generated FI-modules over a field of characteristic 0 are Noetherian. We generalize this result
More informationSpectra of Semidirect Products of Cyclic Groups
Spectra of Semidirect Products of Cyclic Groups Nathan Fox 1 University of Minnesota-Twin Cities Abstract The spectrum of a graph is the set of eigenvalues of its adjacency matrix A group, together with
More informationOutline 1. Background on Symmetric Polynomials 2. Algebraic definition of (modified) Macdonald polynomials 3. New combinatorial definition of Macdonal
Algebraic and Combinatorial Macdonald Polynomials Nick Loehr AIM Workshop on Generalized Kostka Polynomials July 2005 Reference: A Combinatorial Formula for Macdonald Polynomials" by Haglund, Haiman, and
More informationLittlewood Richardson polynomials
Littlewood Richardson polynomials Alexander Molev University of Sydney A diagram (or partition) is a sequence λ = (λ 1,..., λ n ) of integers λ i such that λ 1 λ n 0, depicted as an array of unit boxes.
More informationKostka multiplicity one for multipartitions
Kostka multiplicity one for multipartitions James Janopaul-Naylor and C. Ryan Vinroot Abstract If [λ(j)] is a multipartition of the positive integer n (a sequence of partitions with total size n), and
More informationCombinatorics for algebraic geometers
Combinatorics for algebraic geometers Calculations in enumerative geometry Maria Monks March 17, 214 Motivation Enumerative geometry In the late 18 s, Hermann Schubert investigated problems in what is
More informationBoolean Product Polynomials and the Resonance Arrangement
Boolean Product Polynomials and the Resonance Arrangement Sara Billey University of Washington Based on joint work with: Lou Billera and Vasu Tewari FPSAC July 17, 2018 Outline Symmetric Polynomials Schur
More informationLECTURE 3: TENSORING WITH FINITE DIMENSIONAL MODULES IN CATEGORY O
LECTURE 3: TENSORING WITH FINITE DIMENSIONAL MODULES IN CATEGORY O CHRISTOPHER RYBA Abstract. These are notes for a seminar talk given at the MIT-Northeastern Category O and Soergel Bimodule seminar (Autumn
More informationPermutations with Ascending and Descending Blocks
Permutations with Ascending and Descending Blocks Jacob Steinhardt jacob.steinhardt@gmail.com Submitted: Aug 29, 2009; Accepted: Jan 4, 200; Published: Jan 4, 200 Mathematics Subject Classification: 05A05
More informationRamsey Theory. May 24, 2015
Ramsey Theory May 24, 2015 1 König s Lemma König s Lemma is a basic tool to move between finite and infinite combinatorics. To be concise, we use the notation [k] = {1, 2,..., k}, and [X] r will denote
More informationEnumeration on row-increasing tableaux of shape 2 n
Enumeration on row-increasing tableaux of shape 2 n Rosena R. X. Du East China Normal University, Shanghai, China Joint work with Xiaojie Fan and Yue Zhao Shanghai Jiaotong University June 25, 2018 2/38
More informationFormal power series rings, inverse limits, and I-adic completions of rings
Formal power series rings, inverse limits, and I-adic completions of rings Formal semigroup rings and formal power series rings We next want to explore the notion of a (formal) power series ring in finitely
More informationON KRONECKER PRODUCTS OF CHARACTERS OF THE SYMMETRIC GROUPS WITH FEW COMPONENTS
ON KRONECKER PRODUCTS OF CHARACTERS OF THE SYMMETRIC GROUPS WITH FEW COMPONENTS C. BESSENRODT AND S. VAN WILLIGENBURG Abstract. Confirming a conjecture made by Bessenrodt and Kleshchev in 1999, we classify
More informationThe Major Problems in Group Representation Theory
The Major Problems in Group Representation Theory David A. Craven 18th November 2009 In group representation theory, there are many unsolved conjectures, most of which try to understand the involved relationship
More informationGroups and Symmetries
Groups and Symmetries Definition: Symmetry A symmetry of a shape is a rigid motion that takes vertices to vertices, edges to edges. Note: A rigid motion preserves angles and distances. Definition: Group
More informationStirling Numbers of the 1st Kind
Daniel Reiss, Colebrook Jackson, Brad Dallas Western Washington University November 28, 2012 Introduction The set of all permutations of a set N is denoted S(N), while the set of all permutations of {1,
More informationABSTRACT. Department of Mathematics. interesting results. A graph on n vertices is represented by a polynomial in n
ABSTRACT Title of Thesis: GRÖBNER BASES WITH APPLICATIONS IN GRAPH THEORY Degree candidate: Angela M. Hennessy Degree and year: Master of Arts, 2006 Thesis directed by: Professor Lawrence C. Washington
More informationMultiplicity Free Expansions of Schur P-Functions
Annals of Combinatorics 11 (2007) 69-77 0218-0006/07/010069-9 DOI 10.1007/s00026-007-0306-1 c Birkhäuser Verlag, Basel, 2007 Annals of Combinatorics Multiplicity Free Expansions of Schur P-Functions Kristin
More informationIsomorphisms between pattern classes
Journal of Combinatorics olume 0, Number 0, 1 8, 0000 Isomorphisms between pattern classes M. H. Albert, M. D. Atkinson and Anders Claesson Isomorphisms φ : A B between pattern classes are considered.
More information9 - The Combinatorial Nullstellensatz
9 - The Combinatorial Nullstellensatz Jacques Verstraëte jacques@ucsd.edu Hilbert s nullstellensatz says that if F is an algebraically closed field and f and g 1, g 2,..., g m are polynomials in F[x 1,
More informationThe (q, t)-catalan Numbers and the Space of Diagonal Harmonics. James Haglund. University of Pennsylvania
The (q, t)-catalan Numbers and the Space of Diagonal Harmonics James Haglund University of Pennsylvania Outline Intro to q-analogues inv and maj q-catalan Numbers MacMahon s q-analogue The Carlitz-Riordan
More informationREPRESENTATION THEORY WEEK 5. B : V V k
REPRESENTATION THEORY WEEK 5 1. Invariant forms Recall that a bilinear form on a vector space V is a map satisfying B : V V k B (cv, dw) = cdb (v, w), B (v 1 + v, w) = B (v 1, w)+b (v, w), B (v, w 1 +
More informationGENERATING SETS KEITH CONRAD
GENERATING SETS KEITH CONRAD 1 Introduction In R n, every vector can be written as a unique linear combination of the standard basis e 1,, e n A notion weaker than a basis is a spanning set: a set of vectors
More information1 Fields and vector spaces
1 Fields and vector spaces In this section we revise some algebraic preliminaries and establish notation. 1.1 Division rings and fields A division ring, or skew field, is a structure F with two binary
More informationClassification of root systems
Classification of root systems September 8, 2017 1 Introduction These notes are an approximate outline of some of the material to be covered on Thursday, April 9; Tuesday, April 14; and Thursday, April
More informationA proof of the Square Paths Conjecture
A proof of the Square Paths Conjecture Emily Sergel Leven October 7, 08 arxiv:60.069v [math.co] Jan 06 Abstract The modified Macdonald polynomials, introduced by Garsia and Haiman (996), have many astounding
More informationOperators on k-tableaux and the k-littlewood Richardson rule for a special case. Sarah Elizabeth Iveson
Operators on k-tableaux and the k-littlewood Richardson rule for a special case by Sarah Elizabeth Iveson A dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of
More informationMathematics for Cryptography
Mathematics for Cryptography Douglas R. Stinson David R. Cheriton School of Computer Science University of Waterloo Waterloo, Ontario, N2L 3G1, Canada March 15, 2016 1 Groups and Modular Arithmetic 1.1
More informationLecture 1. (i,j) N 2 kx i y j, and this makes k[x, y]
Lecture 1 1. Polynomial Rings, Gröbner Bases Definition 1.1. Let R be a ring, G an abelian semigroup, and R = i G R i a direct sum decomposition of abelian groups. R is graded (G-graded) if R i R j R i+j
More informationTopics in linear algebra
Chapter 6 Topics in linear algebra 6.1 Change of basis I want to remind you of one of the basic ideas in linear algebra: change of basis. Let F be a field, V and W be finite dimensional vector spaces over
More informationSCHUR-WEYL DUALITY FOR U(n)
SCHUR-WEYL DUALITY FOR U(n) EVAN JENKINS Abstract. These are notes from a lecture given in MATH 26700, Introduction to Representation Theory of Finite Groups, at the University of Chicago in December 2009.
More informationSupplement to Multiresolution analysis on the symmetric group
Supplement to Multiresolution analysis on the symmetric group Risi Kondor and Walter Dempsey Department of Statistics and Department of Computer Science The University of Chicago risiwdempsey@uchicago.edu
More informationEquality of P-partition Generating Functions
Bucknell University Bucknell Digital Commons Honors Theses Student Theses 2011 Equality of P-partition Generating Functions Ryan Ward Bucknell University Follow this and additional works at: https://digitalcommons.bucknell.edu/honors_theses
More informationRepresentation Theory. Ricky Roy Math 434 University of Puget Sound
Representation Theory Ricky Roy Math 434 University of Puget Sound May 2, 2010 Introduction In our study of group theory, we set out to classify all distinct groups of a given order up to isomorphism.
More informationMajor Index for 01-Fillings of Moon Polyominoes
Major Index for 0-Fillings of Moon Polyominoes William Y.C. Chen a,, Svetlana Poznanović b, Catherine H. Yan a,b,2 and Arthur L.B. Yang a, a Center for Combinatorics, LPMC-TJKLC Nankai University, Tianjin
More informationϕ : Z F : ϕ(t) = t 1 =
1. Finite Fields The first examples of finite fields are quotient fields of the ring of integers Z: let t > 1 and define Z /t = Z/(tZ) to be the ring of congruence classes of integers modulo t: in practical
More informationEXTERIOR AND SYMMETRIC POWERS OF MODULES FOR CYCLIC 2-GROUPS
EXTERIOR AND SYMMETRIC POWERS OF MODULES FOR CYCLIC 2-GROUPS FRANK IMSTEDT AND PETER SYMONDS Abstract. We prove a recursive formula for the exterior and symmetric powers of modules for a cyclic 2-group.
More informationCOMPLEXITY OF SHORT RECTANGLES AND PERIODICITY
COMPLEXITY OF SHORT RECTANGLES AND PERIODICITY VAN CYR AND BRYNA KRA Abstract. The Morse-Hedlund Theorem states that a bi-infinite sequence η in a finite alphabet is periodic if and only if there exists
More informationGenerating p-extremal graphs
Generating p-extremal graphs Derrick Stolee Department of Mathematics Department of Computer Science University of Nebraska Lincoln s-dstolee1@math.unl.edu August 2, 2011 Abstract Let f(n, p be the maximum
More informationCOMBINATORIAL GROUP THEORY NOTES
COMBINATORIAL GROUP THEORY NOTES These are being written as a companion to Chapter 1 of Hatcher. The aim is to give a description of some of the group theory required to work with the fundamental groups
More informationOn Tensor Products of Polynomial Representations
Canad. Math. Bull. Vol. 5 (4), 2008 pp. 584 592 On Tensor Products of Polynomial Representations Kevin Purbhoo and Stephanie van Willigenburg Abstract. We determine the necessary and sufficient combinatorial
More informationFactorization of the Robinson-Schensted-Knuth Correspondence
Factorization of the Robinson-Schensted-Knuth Correspondence David P. Little September, 00 Abstract In [], a bijection between collections of reduced factorizations of elements of the symmetric group was
More informationCoxeter-Knuth Classes and a Signed Little Bijection
Coxeter-Knuth Classes and a Signed Little Bijection Sara Billey University of Washington Based on joint work with: Zachary Hamaker, Austin Roberts, and Benjamin Young. UC Berkeley, February, 04 Outline
More informationA connection between number theory and linear algebra
A connection between number theory and linear algebra Mark Steinberger Contents 1. Some basics 1 2. Rational canonical form 2 3. Prime factorization in F[x] 4 4. Units and order 5 5. Finite fields 7 6.
More informationDIFFERENTIAL POSETS SIMON RUBINSTEIN-SALZEDO
DIFFERENTIAL POSETS SIMON RUBINSTEIN-SALZEDO Abstract. In this paper, we give a sampling of the theory of differential posets, including various topics that excited me. Most of the material is taken from
More information3.1. Derivations. Let A be a commutative k-algebra. Let M be a left A-module. A derivation of A in M is a linear map D : A M such that
ALGEBRAIC GROUPS 33 3. Lie algebras Now we introduce the Lie algebra of an algebraic group. First, we need to do some more algebraic geometry to understand the tangent space to an algebraic variety at
More informationQUASISYMMETRIC (k, l)-hook SCHUR FUNCTIONS
QUASISYMMETRIC (k, l)-hook SCHUR FUNCTIONS SARAH K. MASON AND ELIZABETH NIESE Abstract. We introduce a quasisymmetric generalization of Berele and Regev s hook Schur functions and prove that these new
More informationMath 210C. The representation ring
Math 210C. The representation ring 1. Introduction Let G be a nontrivial connected compact Lie group that is semisimple and simply connected (e.g., SU(n) for n 2, Sp(n) for n 1, or Spin(n) for n 3). Let
More informationMathematics Course 111: Algebra I Part I: Algebraic Structures, Sets and Permutations
Mathematics Course 111: Algebra I Part I: Algebraic Structures, Sets and Permutations D. R. Wilkins Academic Year 1996-7 1 Number Systems and Matrix Algebra Integers The whole numbers 0, ±1, ±2, ±3, ±4,...
More informationFactorization in Polynomial Rings
Factorization in Polynomial Rings Throughout these notes, F denotes a field. 1 Long division with remainder We begin with some basic definitions. Definition 1.1. Let f, g F [x]. We say that f divides g,
More informationThe Radicans. James B. Wilson. April 13, 2002
The Radicans James B. Wilson April 13, 2002 1 Radices of Integers Computational mathematics has brought awareness to the use of various bases in representing integers. The standard for most number systems
More informationAn Investigation on an Extension of Mullineux Involution
An Investigation on an Extension of Mullineux Involution SPUR Final Paper, Summer 06 Arkadiy Frasinich Mentored by Augustus Lonergan Project Suggested By Roman Bezrukavnikov August 3, 06 Abstract In this
More informationLecture 6 : Kronecker Product of Schur Functions Part I
CS38600-1 Complexity Theory A Spring 2003 Lecture 6 : Kronecker Product of Schur Functions Part I Lecturer & Scribe: Murali Krishnan Ganapathy Abstract The irreducible representations of S n, i.e. the
More informationarxiv: v1 [math.co] 8 Feb 2014
COMBINATORIAL STUDY OF THE DELLAC CONFIGURATIONS AND THE q-extended NORMALIZED MEDIAN GENOCCHI NUMBERS ANGE BIGENI arxiv:1402.1827v1 [math.co] 8 Feb 2014 Abstract. In two recent papers (Mathematical Research
More informationComputing inclusions of Schur modules
JSAG 1 (2009), 5 10 The Journal of Software for Algebra and Geometry Computing inclusions of Schur modules STEVEN V SAM ABSTRACT. We describe a software package for constructing minimal free resolutions
More informationQuivers of Period 2. Mariya Sardarli Max Wimberley Heyi Zhu. November 26, 2014
Quivers of Period 2 Mariya Sardarli Max Wimberley Heyi Zhu ovember 26, 2014 Abstract A quiver with vertices labeled from 1,..., n is said to have period 2 if the quiver obtained by mutating at 1 and then
More informationCombinatorial bases for representations. of the Lie superalgebra gl m n
Combinatorial bases for representations of the Lie superalgebra gl m n Alexander Molev University of Sydney Gelfand Tsetlin bases for gln Gelfand Tsetlin bases for gl n Finite-dimensional irreducible representations
More informationA Pieri rule for key polynomials
Séminaire Lotharingien de Combinatoire 80B (2018) Article #78, 12 pp. Proceedings of the 30 th Conference on Formal Power Series and Algebraic Combinatorics (Hanover) A Pieri rule for key polynomials Sami
More informationAutomorphism groups of wreath product digraphs
Automorphism groups of wreath product digraphs Edward Dobson Department of Mathematics and Statistics Mississippi State University PO Drawer MA Mississippi State, MS 39762 USA dobson@math.msstate.edu Joy
More informationPRIMARY DECOMPOSITION FOR THE INTERSECTION AXIOM
PRIMARY DECOMPOSITION FOR THE INTERSECTION AXIOM ALEX FINK 1. Introduction and background Consider the discrete conditional independence model M given by {X 1 X 2 X 3, X 1 X 3 X 2 }. The intersection axiom
More informationPERMUTED COMPOSITION TABLEAUX, 0-HECKE ALGEBRA AND LABELED BINARY TREES
PERMUTED COMPOSITION TABLEAUX, 0-HECKE ALGEBRA AND LABELED BINARY TREES V. TEWARI AND S. VAN WILLIGENBURG Abstract. We introduce a generalization of semistandard composition tableaux called permuted composition
More informationPermutation groups/1. 1 Automorphism groups, permutation groups, abstract
Permutation groups Whatever you have to do with a structure-endowed entity Σ try to determine its group of automorphisms... You can expect to gain a deep insight into the constitution of Σ in this way.
More informationMath 145. Codimension
Math 145. Codimension 1. Main result and some interesting examples In class we have seen that the dimension theory of an affine variety (irreducible!) is linked to the structure of the function field in
More informationCOUNTING NUMERICAL SEMIGROUPS BY GENUS AND SOME CASES OF A QUESTION OF WILF
COUNTING NUMERICAL SEMIGROUPS BY GENUS AND SOME CASES OF A QUESTION OF WILF NATHAN KAPLAN Abstract. The genus of a numerical semigroup is the size of its complement. In this paper we will prove some results
More informationarxiv: v3 [math.co] 9 Feb 2011
The cyclic sieving phenomenon: a survey Bruce E. Sagan arxiv:1008.0790v3 [math.co] 9 Feb 2011 Abstract The cyclic sieving phenomenon was defined by Reiner, Stanton, and White in a 2004 paper. Let X be
More informationMATH 326: RINGS AND MODULES STEFAN GILLE
MATH 326: RINGS AND MODULES STEFAN GILLE 1 2 STEFAN GILLE 1. Rings We recall first the definition of a group. 1.1. Definition. Let G be a non empty set. The set G is called a group if there is a map called
More informationEQUALITY OF P -PARTITION GENERATING FUNCTIONS
EQUALITY OF P -PARTITION GENERATING FUNCTIONS PETER R. W. MCNAMARA AND RYAN E. WARD Abstract. To every labeled poset (P, ω), one can associate a quasisymmetric generating function for its (P, ω)-partitions.
More informationMaximizing the descent statistic
Maximizing the descent statistic Richard EHRENBORG and Swapneel MAHAJAN Abstract For a subset S, let the descent statistic β(s) be the number of permutations that have descent set S. We study inequalities
More information