ebook img

Polynomiality of some hook-content summations for doubled distinct and self-conjugate partitions PDF

0.28 MB·
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Polynomiality of some hook-content summations for doubled distinct and self-conjugate partitions

POLYNOMIALITY OF SOME HOOK-CONTENT SUMMATIONS FOR DOUBLED DISTINCT AND SELF-CONJUGATE 6 PARTITIONS 1 0 2 GUO-NIUHANANDHUANXIONG n a Abstract. In 2009, the first author proved the Nekrasov-Okounkov formula J onhooklengthsforintegerpartitionsbyusinganidentityofMacdonaldinthe e 7 frameworkof type A affine rootsystems, and conjectured that somesumma- 1 tions overthesetofallpartitionsofsizenarealwayspolynomialsinn. This conjecturewasgeneralizedandprovedbyStanley. Recently,P´etr´eollederived e e ] two Nekrasov-Okounkov type formulas for C and Cˇwhich involve doubled O distinct and self-conjugate partitions. Inspired by all those previous works, C weestablishthepolynomialityofsomehook-content summationsfordoubled distinctandself-conjugatepartitions. . h t a m 1. Introduction [ 1 The following so-called Nekrasov-Okounkovformula v z q|λ| 1− = (1−qk)z−1, 9 h2 6 λX∈P h∈YH(λ)(cid:0) (cid:1) kY≥1 3 where P is the set of all integer partitions λ with |λ| denoting the size of λ and 4 H(λ) the multiset of hook lengths associated with λ (see [6]), was discovered in- 0 . dependently several times: First, by Nekrasov and Okounkov in their study of the 1 theory of Seiberg-Witten on supersymmetric gauges in particle physics [16]; Then, 0 6 proved by Westbury using D’Arcais polynomials [28]; Finally, by the first author 1 using an identity of Macdonald [15] in the framework of type A affine root sys- : tems [6]. Moreover,he asked to find Nekrasov-Okounkovtype formulas associated v i with other root systems [7, Problem 6.4], and conjectured that e X 1 r n! h2k a H(λ)2 |λX|=n h∈XH(λ) is always a polynomial in n for any k ∈ N, where H(λ) = h. This con- h∈H(λ) jecture was proved by Stanley in a more general form. In particular, he showed Q that 1 n! F (h2 :h∈H(λ))F (c:c∈C(λ)) H(λ)2 1 2 |λX|=n is a polynomial in n for any symmetric functions F and F , where C(λ) is the 1 2 multiset of contents associated with λ (see [24]). For some special functions F 1 Date:December25,2015. 2010 Mathematics Subject Classification. 05A15,05A17,05A19,05E05,05E10, 11P81. Key words and phrases. strict partition, doubled distinct partition, self-conjugate partition, hooklength,content, shiftedYoungtableau, differenceoperator. 1 2 GUO-NIUHANANDHUANXIONG Figure 1. From strict partitions to doubled distinct partitions. and F the latter polynomial has explicit expression, as shown by Fujii, Kanno, 2 Moriyama,Okada and Panova [4, 19]. A strict partition is a finite strict decreasing sequence of positive integers λ¯ = (λ¯ ,λ¯ ,...,λ¯ ). Theinteger|λ¯|= λ¯ iscalledthesize andℓ(λ¯)=ℓiscalled 1 2 ℓ 1≤i≤ℓ i thelengthofλ¯.Forconvenience,letλ¯ =0fori>ℓ(λ¯). Astrictpartitionλ¯couldbe Pi identicalwithitsshiftedYoungdiagram,whichmeansthatthei-throwoftheusual Young diagram is shifted to the right by i boxes. We define the doubled distinct partition of λ¯, denoted by λ¯λ¯, to be the usual partition whose Young diagram is obtained by adding λ¯ boxes to the i-th column of the shifted Young diagram of λ¯ i for 1≤ i≤ ℓ(λ¯) (see [5, 20, 21]). For example, (6,4,4,1,1) is the doubled distinct partition of (5,2,1) (see Figure 1). For each usual partition λ, let λ′ denote the conjugate partition of λ (see [5, 15, 20, 21]). A usual partition λ is called self-conjugate if λ = λ′. The set of all doubled distinct partitions and the set of all self-conjugate partitions are denoted by DD and SC respectively. For each positive integer t, let H (λ)={h∈H(λ): h≡0 (mod t)} t be the multiset of hook lengths of multiples of t. Write H (λ)= h. t h∈Ht(λ) Recently, P´etr´eolle derived two Nekrasov-Okounkov type formulas for C and Q Cˇwhich involve doubled distinct and self-conjugate partitions. In particular, he obtained the following two formulas [20, 21]. e e Theorem 1.1 (P´etr´eolle[20, 21]). For positive integers n and t we have 1 1 (1.1) = , if t is odd; H (λ) (2t)nn! t λ∈DDX,|λ|=2nt #Ht(λ)=2n 1 1 (1.2) = , if t is even. H (λ) (2t)nn! t λ∈SCX,|λ|=2nt #Ht(λ)=2n Inspired by all those previous works, we establish the polynomiality of some hook-content summations for doubled distinct and self-conjugate partitions. Our main result is stated next. Theorem 1.2. Let t be a given positive integer. The following two summations for the positive integer n F (h2 :h∈H(λ))F (c:c∈C(λ)) (1.3) (2t)nn! 1 2 (t odd) H (λ) t λ∈DDX,|λ|=2nt #Ht(λ)=2n POLYNOMIALITY OF SOME HOOK-CONTENT SUMMATIONS 3 and F (h2 :h∈H(λ))F (c:c∈C(λ)) (1.4) (2t)nn! 1 2 (t even) H (λ) t λ∈SCX,|λ|=2nt #Ht(λ)=2n are polynomials in n for any symmetric functions F and F . 1 2 In fact, the degrees of the two polynomials in Theorem 1.2 can be estimated explicitly in terms of F and F (see Corollary 4.8 and Theorem 5.3). When F 1 2 1 and F are two constant symmetric functions, we derive Theorem 1.1. Other spe- 2 cializations are listed as follows. Corollary 1.3. We have 1 1 (1.5) (2t)nn! h2 =6t2n2+ (t2−6t+2)tn (t odd), H (λ) 3 t λ∈DDX,|λ|=2nt h∈XH(λ) #Ht(λ)=2n 1 1 (1.6) (2t)nn! h2 =6t2n2+ (t2−6t−1)tn (t even), H (λ) 3 t λ∈SCX,|λ|=2nt h∈XH(λ) #Ht(λ)=2n 1 1 (1.7) (2t)nn! c2 =2t2n2+ (t2−6t+2)tn (t odd), H (λ) 3 t λ∈DDX,|λ|=2nt c∈XC(λ) #Ht(λ)=2n 1 1 (1.8) (2t)nn! c2 =2t2n2+ (t2−6t−1)tn (t even). H (λ) 3 t λ∈SCX,|λ|=2nt c∈XC(λ) #Ht(λ)=2n TherestofthepaperisessentiallydevotedtocompletetheproofofTheorem1.2. The polynomiality of summations in (1.3) for t = 1 with F = 1 or F = 1 has an 1 2 equivalentstatementintermsofstrictpartitions,whoseproofisgiveninSection2. AfterrecallingsomebasicdefinitionsandpropertiesofLittlewooddecompositionin Section 3, the doubled distinct and self-conjugate cases of Theorem 1.2 are proved in Sections 4 and 5 respectively. Finally, Corollary 1.3 is proved in Section 6. 2. Polynomiality for strict and doubled distinct partitions In this section we prove an equivalent statement of the polynomiality of (1.3) for t=1 with F =1 or F =1, which consists a summation over the set of strict 1 2 partitions. Let λ¯ = (λ¯ ,λ¯ ,...,λ¯ ) be a strict partition. Therefore the leftmost 1 2 ℓ box in the i-th row of the shifted Young diagram of λ¯ has coordinate (i,i+1). The hook length of the (i,j)-box, denoted by h , is defined to be the number of (i,j) boxesexactlytotheright,orexactlyabove,ortheboxitself,plusλ¯ . Forexample, j consider the box (cid:3) = (i,j) = (1,3) in the shifted Young diagram of the strict partition (7,5,4,1). There are 1 and 5 boxes below and to the right of the box (cid:3) respectively. Since λ¯ =4, the hook length of (cid:3) is equal to 1+5+1+4 =11, as 3 illustrated in Figure 1. The content of (cid:3) = (i,j) is defined to be c(cid:3) = j −i, so that the leftmost box in each row has content 1. Also, let H(λ¯) be the multi-set of hook lengths of boxes and H(λ¯) be the product of all hook lengths of boxes in λ¯. The hook length and content multisets of the doubled distinct partition λ¯λ¯ can be obtained from H(λ¯) and C(λ¯) by the following relations: (2.1) H(λ¯λ¯)=H(λ¯)∪H(λ¯)∪{2λ¯ ,2λ¯ ,...,2λ¯ }\{λ¯ ,λ¯ ,...,λ¯ }, 1 2 ℓ 1 2 ℓ 4 GUO-NIUHANANDHUANXIONG 1 1 5 4 2 1 1 2 3 4 9 6 5 3 2 1 2 3 4 5 1211 8 7 5 4 1 1 2 3 4 5 6 7 Figure 2. The shifted Young diagram, the hook lengths and the contents of the strict partition (7,5,4,1). Figure 3. The skew shifted Young diagram of the skew strict partition (7,5,4,1)/(4,2,1). (α ,β ) · 1 1 (α ,β ) · 2 2 · · (α ,β ) · m m · Figure 4. A strict partition and its corners. The outer corners are labelled with (α ,β ) (i = 1,2,...,m). The inner corners are i i indicated by the dot symbol “·”. (2.2) C(λ¯λ¯)=C(λ¯)∪{1−c|c∈C(λ¯)}. For two strict partitions λ¯ and µ¯, we write λ¯ ⊇ µ¯ if λ¯ ≥ µ¯ for any i ≥ 1. In i i this case, the skew strict partition λ¯/µ¯ is identical with the skew shifted Young diagram. For example, the skew strict partition (7,5,4,1)/(4,2,1) is represented by the white boxes in Figure 2. Let fλ¯ (resp. fλ¯/µ¯) be the number of standard shifted Young tableaux of shape λ¯ (resp. λ¯/µ¯). The following formulas for strict partitions are well-known (see [2, 23, 27]): (2.3) fλ¯ = H|λ¯(|λ¯!) and n1! 2n−ℓ(λ¯)fλ¯2 =1. |λ¯X|=n Identity (1.1) with t=1, obtained by P´etr´eolle,becomes 1 1 = , H(λ) 2nn! λ∈DDX,|λ|=2n which is equivalent to the second identity of (2.3) in view of (2.1). POLYNOMIALITY OF SOME HOOK-CONTENT SUMMATIONS 5 For a strict partition λ¯, the outer corners (see [11]) are the boxes which can be removed in such a way that after removal the resulting diagram is still a shifted Young diagram of a strict partition. The coordinates of outer corners are denoted by (α ,β ),...,(α ,β ) such that α > α > ··· > α . Let y := β − α 1 1 m m 1 2 m j j j (1≤j ≤m) be the contents of outer corners. We set α =0, β =ℓ(λ¯)+1 and m+1 0 call(α ,β ),(α ,β ),...,(α ,β )the inner corners ofλ¯. Letx =β −α be 1 0 2 1 m+1 m i i i+1 the contents of inner corners for 0 ≤ i ≤ m (see Figure 3). The following relation of x and y are obvious. i j (2.4) x =1≤y <x <y <x <···<y <x . 0 1 1 2 2 m m Notice that x0 = y1 = 1 iff λ¯ℓ(λ¯) = 1. Let λ¯i+ = λ¯ {(cid:3)i} such that c(cid:3)i = xi for 0≤i≤m. Here λ¯0+ does not exist if y =1. The set of contents of inner corners 1 S andthesetofcontentsofoutercornersofλ¯aredenotedbyX(λ¯)={x ,x ,...,x } 0 1 m andY(λ¯)={y ,y ,...,y }respectively. Thefollowingrelationsbetweenthehook 1 2 m lengths of λ¯ and λ¯i+ are established in [11]. Theorem 2.1 (Theorem 3.1 of [11]). Let λ¯ be a strict partition with X(λ¯) = {x ,x ,...,x } and Y(λ¯)={y ,y ,...,y }. For 1≤i≤m, we have 0 1 m 1 2 m H(λ¯)∪{1,x ,2x −2}∪{|x −x |:1≤j ≤m,j 6=i} i i i j ∪{x +x −1:1≤j ≤m,j 6=i} i j =H(λ¯i+)∪{|x −y |:1≤j ≤m}∪{x +y −1:1≤j ≤m} i j i j and xi − yj H(λ¯) 1 1≤j≤m 2 2 = · . H(λ¯i+) 2 Q (cid:0)(cid:0)xi(cid:1)−(cid:0)xj(cid:1)(cid:1) 2 2 0≤j≤m jQ6=i (cid:0)(cid:0) (cid:1) (cid:0) (cid:1)(cid:1) If y >1, we have 1 H(λ¯)∪{1,x ,x −1,x ,x −1,··· ,x ,x −1} 1 1 2 2 m m =H(λ¯0+) ∪{y ,y −1,y ,y −1,··· ,y ,y −1} 1 1 2 2 m m and x0 − yj H(λ¯) 1≤j≤m 2 2 = . H(λ¯0+) Q (cid:0)(cid:0)x0(cid:1)−(cid:0)xj(cid:1)(cid:1) 2 2 1≤j≤m Q (cid:0)(cid:0) (cid:1) (cid:0) (cid:1)(cid:1) Let k be a nonnegative integer, and ν = (ν ,ν ,...,ν ) be a usual partition. 1 2 ℓ(ν) For arbitrary two finite alphabets A and B, the power sum of the alphabet A−B is defined by [14, p.5] (2.5) Ψk(A,B):= ak− bk, a∈A b∈B X X ℓ(ν) (2.6) Ψν(A,B):= Ψνj(A,B). j=1 Y Let λ¯ be a strict partition. We define x y (2.7) Φν(λ¯):=Ψν { i },{ i } . 2 2 (cid:16) (cid:18) (cid:19) (cid:18) (cid:19) (cid:17) 6 GUO-NIUHANANDHUANXIONG Theorem 2.2 (Theorem 3.5 of [11]). Let k be a given nonnegative integer. Then, there exist some ξ ∈Q such that j k−1 j x Φk(λ¯i+)−Φk(λ¯)= ξ i j 2 j=0 (cid:18) (cid:19) X for every strict partition λ¯ and 0≤i≤m, where x ,x ,...,x are the contents of 0 1 m inner corners of λ¯. Lemma 2.3. Let k be agiven nonnegative integer. Then, thereexist some a such ij that i j x y (x−y)2k+(x+y−1)2k = a ij 2 2 i+j≤k (cid:18) (cid:19) (cid:18) (cid:19) X for every x,y ∈C. Proof. The claim follows from x y (x−y)2+(x+y−1)2 =4 +4 +1 2 2 (cid:18) (cid:19) (cid:18) (cid:19) and x y (x−y)2(x+y−1)2 = 2 −2 2. (cid:3) 2 2 (cid:18) (cid:19) (cid:18) (cid:19) (cid:0) (cid:1) Lemma 2.4 (Theorem 3.2 of [11]). Let k be a nonnegative integer. Then, there exist some ξ ∈Q indexed by usual partitions ν such that ν (a −b ) i j 1≤j≤m ak = ξ Ψν({a },{b }) Q (a −a ) i ν i i i j 0≤Xi≤m0≤j≤m |νX|≤k jQ6=i for arbitrary complex numbers a <a <···<a and b <b <···<b . 0 1 m 1 2 m We define the difference operator D¯ for strict partitions by m (2.8) D¯ g(λ¯) :=2 g(λ¯i+)+g(λ¯0+)− g(λ¯), i=1 (cid:0) (cid:1) X where λ¯ is a strict partition and g is a function of strict partitions. In the above definition,thesymbolg(λ¯0+)takesthevalue0ifλ¯0+ doesnotexist,orequivalently if λ¯ℓ(λ¯) =1. By Theorem 2.1, we have 1 (2.9) D¯ =0. H(λ¯) (cid:16) (cid:17) Theorem 2.5 (Theorem2.3 of [11]). Let g be a function of strict partitions and µ¯ be a given strict partition. Then we have n n (2.10) 2|λ¯|−|µ¯|−ℓ(λ¯)+ℓ(µ¯)fλ¯/µ¯g(λ¯)= k D¯kg(µ¯) |λ¯/Xµ¯|=n Xk=0(cid:18) (cid:19) and n n (2.11) D¯ng(µ¯)= (−1)n+k k 2|λ¯|−|µ¯|−ℓ(λ¯)+ℓ(µ¯)fλ¯/µ¯g(λ¯). Xk=0 (cid:18) (cid:19)|λ¯/Xµ¯|=k POLYNOMIALITY OF SOME HOOK-CONTENT SUMMATIONS 7 In particular, if there exists some positive integer r such that D¯rg(λ¯) = 0 for every strict partition λ¯, then the left-hand side of (2.10) is a polynomial of n with degree at most r−1. For each usual partition δ let pδ(λ¯):=Ψδ({h2 :h∈H(λ¯λ¯)},∅). By (2.1), we have ℓ(λ¯) pk(λ¯)= h2k =2 h2k+(4k−1) λ¯2k i h∈XH(λ¯λ¯) h∈XH(λ¯) Xi=1 for a nonnegative integer k. Theorem 2.6. Suppose that ν and δ are two given usual partitions. Then, pδ(λ¯)Φν(λ¯) (2.12) D¯r =0 H(λ¯) (cid:16) (cid:17) for every strict partition λ¯, where r =|δ|+ℓ(δ)+|ν|+1. Consequently, for a given strict partition µ¯, (2.13) 2|λ¯|−ℓ(λ¯)fλ¯/µ¯pδ(λ¯) H(λ¯) |λ¯/Xµ¯|=n is a polynomial in n of degree at most |δ|+ℓ(δ). Proof. Let X(λ¯) = {x ,x ,...,x } and Y(λ¯) = {y ,y ,...,y }. First, we show 0 1 m 1 2 m that the difference pk(λ¯i+)−pk(λ¯) can be written as the following form k j x η (λ¯) i j 2 j=0 (cid:18) (cid:19) X for 0≤i≤ m and a nonnegative integer k, where each coefficient η (λ¯) is a linear j combination of some Φτ(λ¯) for some usual partition τ of size |τ| ≤ k. Indeed, by Lemma 2.3 and Theorem 2.1, m m pk(λ¯0+)−pk(λ¯)=2 (x2k+(x −1)2k)−2 (y2k+(y −1)2k)+22k+1 j j j j j=1 j=1 X X k x j =η0(λ¯)= ηj(λ¯) 20 [ if i=0 and λ¯ℓ(λ¯) ≥2] j=0 (cid:18) (cid:19) X and pk(λ¯i+)−pk(λ¯) m m =2 ((x −x )2k+(x +x −1)2k)−2 (x −y )2k+(x +y −1)2k i j i j i j i j j=1 j=1 X X(cid:0) (cid:1) +2x2k+2(2x −2)2k+2−2(2x −1)2k+(22k−1) x2k−(x −1)2k i i i i i k x j (cid:0) (cid:1) = η (λ¯) i [ if 1≤i≤m ]. j 2 j=0 (cid:18) (cid:19) X 8 GUO-NIUHANANDHUANXIONG Next, let A=Φν(λ¯) and B =pδ(λ¯). We have ∆iA:=Φν(λ¯i+)−Φν(λ¯)= Φνs(λ¯) Φνs′(λ¯i+)−Φνs′(λ¯) , X(∗) sY∈U sY′∈V(cid:0) (cid:1) ∆iB :=pδ(λ¯i+)−pδ(λ¯)= pδs(λ¯) pδs′(λ¯i+)−pδs′(λ¯) , X(∗∗)sY∈U sY′∈V(cid:0) (cid:1) where the sum (∗) (resp. (∗∗)) ranges over all pairs (U,V) of positive integer sets such that U ∪V ={1,2,...,ℓ(ν)} (resp. U ∪V ={1,2,...,ℓ(δ)}), U ∩V =∅ and V 6=∅. Finally, it follows from (2.9) and Theorem 2.1 that pδ(λ¯)Φν(λ¯) H(λ¯)D¯ H(λ¯) (cid:16) (cid:17) H(λ¯) = pδ(λ¯0+)Φν(λ¯0+)−pδ(λ¯)Φν(λ¯) H(λ¯0+) (cid:0)m H(λ¯) (cid:1) +2 pδ(λ¯i+)Φν(λ¯i+)−pδ(λ¯)Φν(λ¯) H(λ¯i+) i=1 X (cid:0) (cid:1) xi − yj 2 2 = 1≤j≤m pδ(λ¯i+)Φν(λ¯i+)−pδ(λ¯)Φν(λ¯) Q (cid:0)(cid:0)xi(cid:1)−(cid:0)xj(cid:1)(cid:1) 0≤i≤m 2 2 X 0≤j≤m (cid:0) (cid:1) jQ6=i (cid:0)(cid:0) (cid:1) (cid:0) (cid:1)(cid:1) xi − yj 2 2 1≤j≤m = A·∆ B+B·∆ A+∆ A·∆ B . Q (cid:0)(cid:0)xi(cid:1)−(cid:0)xj(cid:1)(cid:1) i i i i 0≤i≤m 2 2 X 0≤j≤m (cid:0) (cid:1) jQ6=i (cid:0)(cid:0) (cid:1) (cid:0) (cid:1)(cid:1) ByTheorems2.4and2.2,eachoftheabovethreetermscouldbewrittenasalinear combination of some pδ(λ¯)Φν(λ¯) satisfying |δ|+ℓ(δ)+|ν| ≤ |δ|+ℓ(δ)+|ν|−1. Then the claim follows by induction on |δ|+ℓ(δ)+|ν|. (cid:3) When µ¯=∅, the summation (2.13) in Theorem 2.6 becomes 2n−ℓ(λ¯)n! (2.14) pδ(λ¯) H(λ¯)2 |λ¯X|=n or 1 (2.15) 2nn! Ψδ({h2 :h∈H(λ¯λ)},∅) H(λ¯λ¯) |λ¯λX¯|=2n by (2.1). The above summation is a polynomial in n. Consequently, Theorem 1.2 is true when t=1 and F =1. Other specializations are listed as follows. 2 Theorem 2.7. Let µ¯ be a given strict partition. Then, (2.16) 2|λ¯|−ℓ(λ¯)−|µ¯|+ℓ(µ¯)fλ¯/µ¯Hµ¯ p1(λ¯)−p1(µ¯) =12 n +(12|µ¯|+5)n. |λ¯/Xµ¯|=n Hλ¯ (cid:0) (cid:1) (cid:18)2(cid:19) POLYNOMIALITY OF SOME HOOK-CONTENT SUMMATIONS 9 Let µ¯ =∅. We obtain 1 n (2.17) 2nn! h2 =12 +5n. H(λ¯λ¯) 2 |λ¯λX¯|=2n h∈XH(λ¯λ¯) (cid:18) (cid:19) Proof. We have m m p1(λ¯0+)−p1(λ¯)=2 (x2+(x −1)2)−2 (y2+(y −1)2)+22+1 j j j j j=1 j=1 X X =η0(λ¯)=8|λ¯|+5 [ if i=0 and λ¯ℓ(λ¯) ≥2] and p1(λ¯i+)−p1(λ¯) m m =2 ((x −x )2+(x +x −1)2)−2 ((x −y )2+(x +y −1)2) i j i j i j i j j=1 j=1 X X +2x2+2(2x −2)2+2−2(2x −1)2+(22−1)(x2−(x −1)2) i i i i i x =4 i +8|λ¯|+5 [ if 1≤i≤m ]. 2 (cid:18) (cid:19) So that xi − yj Hλ¯D(cid:16)pH1(λλ¯¯)(cid:17)=0≤Xi≤m01≤≤Qjj≤≤mm(cid:0)(cid:0)x22i(cid:1)−(cid:0)x22j(cid:1)(cid:1)(4(cid:18)x2i(cid:19)+8|λ¯|+5) jQ6=i (cid:0)(cid:0) (cid:1) (cid:0) (cid:1)(cid:1) =4Φ1(λ¯)+8|λ¯|+5 =12|λ¯|+5. Therefore we have p1(λ¯) Hλ¯D2 Hλ¯ =12, (cid:16)p1(λ¯)(cid:17) Hλ¯D3 Hλ¯ =0. (cid:16) (cid:17) Identity (2.16) follows from Theorem 2.5. By (2.1), we derive (2.17). (cid:3) Recall the following results obtained in [11] involving the contents of strict par- titions. Theorem 2.8. Suppose that Q is a given symmetric function, and µ¯ is a given strict partition. Then 2|λ¯|−|µ¯|−ℓ(λ¯)+ℓ(µ¯)fλ¯/µ¯Q c :c∈C(λ¯) H(λ¯) 2 |λ¯/Xµ¯|=n (cid:16)(cid:18) (cid:19) (cid:17) is a polynomial in n. Theorem 2.9. Suppose that k is a given nonnegative integer. Then 2|λ¯|−ℓ(λ¯)fλ¯ c+k−1 = 2k n . H(λ¯) 2k (k+1)! k+1 |λ¯X|=n c∈XC(λ¯)(cid:18) (cid:19) (cid:18) (cid:19) 10 GUO-NIUHANANDHUANXIONG Theorem 2.10. Let µ¯ be a strict partition. Then, 2|λ¯|−ℓ(λ¯)−|µ¯|+ℓ(µ¯)fλ¯/µ¯Hµ¯ c c n (2.18) − = +n|µ¯|. H(λ¯) 2 2 2 |λ¯/Xµ¯|=n (cid:0)c∈XC(λ¯)(cid:18) (cid:19) c∈XC(µ¯)(cid:18) (cid:19)(cid:1) (cid:18) (cid:19) The above results can be interpreted in terms of doubled distinct partitions. In particular, we obtain Theorem 1.2 when t=1 and F =1. 1 Theorem 2.11. For each usual partition δ, the summation 1 (2.19) 2nn! Ψδ(C(λ¯λ¯),∅) H(λ¯λ¯) |λ¯λX¯|=2n is a polynomial in n. Proof. Since c+(1−c) = 1 and c(1−c) = −2 c , there exists some a such that 2 i ck+(1−c)k = s a c i. By (2.2), we obtain i=1 i 2 (cid:0) (cid:1) s i P (cid:0) (cid:1) c ck = ck+(1−c)k = a . i 2 c∈XC(λ¯λ¯) c∈XC(λ¯)(cid:0) (cid:1) Xi=1 c∈XC(λ¯)(cid:18) (cid:19) The claim follows from Theorem 2.8. (cid:3) The following results are corollaries of Theorems 2.9 and 2.10. Theorem 2.12. Suppose that k is a given nonnegative integer. Then, 1 c+k−1 2k+1 n (2.20) 2nn! = , H(λ¯λ¯) 2k (k+1)! k+1 |λ¯λX¯|=2n c∈XC(λ¯λ¯)(cid:18) (cid:19) (cid:18) (cid:19) 1 n n (2.21) 2nn! c2 =4 + . H(λ¯λ¯) 2 1 |λ¯λX¯|=2n c∈XC(λ¯λ¯) (cid:18) (cid:19) (cid:18) (cid:19) 3. The Littlewood decomposition and corners of usual partitions Inthissectionwerecallsomebasicdefinitionsandpropertiesforusualpartitions (see[9],[15,p.12],[25,p.468],[12, p.75],[5]). LetW be the setofbi-infinite binary sequences beginning with infinitely many 0’s and ending with infinitely many 1’s. Each element w of W can be represented by (a′) = ···a′ a′ a′ a′a′a′a′ ···. i i −3 −2 −1 0 1 2 3 However,therepresentationisnotunique,sinceforanyfixedintegerkthesequence (a′ ) alsorepresentsw. Thecanonical representationofw isthe uniquesequence i+k i (a ) =···a a a a a a a ··· such that i i −3 −2 −1 0 1 2 3 #{i≤−1,a =1}=#{i≥0,a =0}. i i Itwillbefurtherdenotedby···a a a .a a a a ··· withadotsymbolinserted −3 −2 −1 0 1 2 3 between the letters a and a . There is a natural one-to-one correspondence −1 0 between P and W (see, e.g. [25, p.468], [1] for more details). Let λ be a partition. We encode each horizontal edge of λ by 1 and each vertical edge by 0. Reading these (0,1)-encodings from top to bottom and from left to right yields a binary word u. By adding infinitely many 0’s to the left and infinitely many 1’s to the right of u we get an element w = ···000u111···∈ W. Clearly, the map λ 7→ w is a one-to-one correspondence between P and W. For example, take λ=(6,3,3,1). Thenu=0100110001,sothatw =(a ) =···1110100.110001000··· (seeFigure5). i i

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.