Noncommutative Double Bruhat Cells and Their Factorizations. Arkady Berenstein and Vladimir Retakh. 1 Introduction

Size: px
Start display at page:

Download "Noncommutative Double Bruhat Cells and Their Factorizations. Arkady Berenstein and Vladimir Retakh. 1 Introduction"

Transcription

1 IMRN International Mathematics Research Notices 2005, No 8 Noncommutative Double Bruhat Cells and Their Factorizations Arkady Berenstein and Vladimir Retakh 1 Introduction This paper is a first attempt to generalize results of A Berenstein, S Fomin, and A Zelevinsky on total positivity of matrices over commutative rings to matrices over noncommutative rings The classical theory of total positivity studies matrices whose minors are all nonnegative Motivated by a surprising connection discovered by Lusztig [10, 11] between total positivity of matrices and canonical bases for quantum groups, Berenstein, Fomin, and Zelevinsky, in a series of papers [1, 2, 3, 4], systematically investigated the problem of total positivity from a representation-theoretic point of view In particular, they showed that a natural framework for the study of totally positive matrices is provided by the decomposition of a reductive group G into the disjoint union of double Bruhat cells G u,v = BuB B vb, where B and B are two opposite Borel subgroups in G, and u and v belong to the Weyl group W of G According to [1, 3, 4] there, exist families of birational parametrizations of G u,v, one for each reduced expression of the element u, v in the Coxeter group W W Every such parametrization can be thought of as a system of local coordinates in G u,v Such coordinates are called the factorization parameters associated to the reduced expression of u, v The coordinates are obtained by expressing a generic element x G as an element of the maximal torus H = B B multiplied by the product of elements of various Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Received 19 July 2004 Revision received 10 October 2004 Communicated by Andrei Zelevinsky

2 478 A Berenstein and V Retakh one-parameter subgroups in G associated with simple roots and their negatives; the reduced expression prescribes the order of factors in this product An explicit formula for these factorization parameters as rational functions on the double Bruhat cell G u,v was given As we said, Berenstein, Fomin, and Zelevinsky came to factorization parameters first, for GL n and then for other classical groups from representation theory For the noncommutative case, our program is to go in the opposite direction; from factorization parameters for GL n to total positivity, canonical bases, and representations This paper is a beginning of the program For G = GL n F, where F is a field of characteristic zero, the explicit formulas for factorization parameters are given through the classical determinant calculus As a first step toward noncommutative representation theory and noncommutative total positivity, we generalize here the results of [3, 4] to G = GL n F where F is a noncommutative skew-field by using the quasideterminantal calculus of matrices over noncommutative rings introduced by I Gelfand and V Retakh [5, 6, 7, 8] The noncommutative point of view has many advantages Here is one Let w o W be the element of the maximal length In the commutative case the factorization parameters for x G u,v, G = GL n, u = id, v = w o are given some times as ratios ab/cd, and other times as a/b, where a, b, c, d are minors of the matrix x see [1] In the noncommutative case, for any u and v = w o, the factorization parameters can be written as f 1 g where f, g are quasiminors of x This paper contains other noncommutative formulas and constructions for GL n that are new, even for the commutative case Our results confirm the Gelfand principle: noncommutative algebra properly understood is simpler than its commutative counterpart The paper is organized as follows In Section 2, we recall some facts about quasideterminants and introduce our main tool positive quasiminors i u,v InSection 3, we study basic factorizations in GL n and its Borel subgroup Section 4 contains examples of such factorizations Section 5 is central for the paper First, we introduce noncommutative SL 2 subgroups in GL n For a generic matrix x, we define special quasiminors i u,vx, where u, v W and show that they satisfy certain Plücker relations We note that i u,vx is always positive for positive real matrices x Section 5 also contains the main result; it gives formulas for factorization coordinates for reduced double Bruhat cells For a matrix x G u,v, these coordinates are written as products of quasiminors i s,ty where the matrix y is the so-called noncommutative twist of x InSection 6, we study relations between quasiminors of x G u,wo and its corresponding twisted matrix y In this case, the quasiminors i, y in the main theorem, can be replaced by quasiminors i, x Studying twisted matrices is an important problem by itself and we present Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011

3 Noncommutative Double Bruhat Cells and Their Factorizations 479 several approaches to computations of such matrices These results are new even in the commutative case 2 Quasideterminants and quasiminors A notion of quasideterminant for matrices over a noncommutative ring was introduced in [6, 7] and developed in [8] It has been effective in many areas see, eg, the survey article [5] Here we recall a few facts about quasideterminants which will be used in this paper 21 Definition of quasideterminants Let A = a ij, i I, j J be a matrix of order n over a ring R Construct the following submatrices of A: submatrix A ij, i I, j J obtained from A by deleting its ith row and jth column; row submatrix r i obtained from ith row of A by deleting the element a ij ; column submatrix c j obtained from jth column of A by deleting the element a ij Definition 21 If n = 1, the quasideterminant A ij equals a ij Ifn>1, the quasideterminant A ij is defined if the submatrix A ij is invertible over the ring R In this case, Inductive properties of quasideterminants are given by the following proposition A ij = a ij r i A ij 1 cj 21 Proposition 22 If all quasideterminants of A ij are defined and invertible, then the quasideterminant A ij is defined and A ij = a ij a iq A ij 1 pq a pj 22 Here the sum is taken over all p I {i}, q J {j} Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Remark 23 If each a ij is an invertible morphism V j V i in an additive category, then the quasideterminant A pq is also a morphism from the object V q to the object V p and where multiplication corresponds to the composition of morphisms This is a key point of view in determining what one can and cannot multiply together If A is an n n matrix, there exist up to n 2 quasideterminants of A By definition, an r quasiminor of a square matrix A is a quasideterminant of an r r submatrix of A

4 480 A Berenstein and V Retakh Sometimes it is convenient to adopt a more graphic notation for the quasideterminant A pq by boxing the element a pq ForA = a ij, i, j = 1,,n, we write a 11 a 1q a pn A pq = a p1 a pq a pn 23 a n1 a nq a nn Example 24 1 For a matrix A = a ij, i, j = 1, 2, there exist four quasideterminants if the corresponding entries are invertible A 11 = a 11 a 12 a 1 22 a 21, A 12 = a 12 a 11 a 1 21 a 22, A 21 = a 21 a 22 a 1 12 a 11, A 22 = a 22 a 21 a 1 11 a ForamatrixA = a ij, i, j = 1, 2, 3, there exist nine quasideterminants but we will write here only A 11 = a 11 a 12 a22 a 23 a 1 33 a 1a21 32 a 12 a32 a 33 a 1 23 a 1a31 22 a 13 a23 a 22 a 1 32 a 1a21 33 a 13 a33 a 32 a 1 22 a 1a31 23 provided all inverses are defined 25 The quasideterminant is not a generalization of the determinant over a commutative ring, but a generalization of a ratio of two determinants Proposition 25 IfA is a matrix over a commutative ring, then A pq = 1 p+q det A 26 det Apq Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 The quasideterminants of matrix A are related to the elements of the inverse matrix A 1 in the following way Proposition 26 If A is invertible and A 1 = b ij, then b 1 ij = A ji 27 if the element b ij is invertible

5 22 Elementary properties of quasideterminants Noncommutative Double Bruhat Cells and Their Factorizations 481 Here is a list of elementary properties of quasideterminants i The quasideterminant A pq does not depend on the permutation of rows and columns in the matrix A if the pth row and the qth column are not changed ii The multiplication of rows and columns Let the matrix B be constructed from the matrix A by multiplication of its ith row by a scalar λ from the left Then λ A ij if k = i, B kj = 28 A kj if k = i, λ is invertible Let the matrix C be constructed from the matrix A by multiplication of its jth column by a scalar µ from the right Then A ij µ if l = j, C il = 29 A il if l j, µ is invertible iii The addition of rows and columns Let the matrix B be constructed by adding to some row of the matrix A its kth row multiplied by a scalar λ from the left Then A ij = B ij, i = 1,,k 1, k + 1,,n, j= 1,,n 210 Let the matrix C be constructed by adding to some column of the matrix A its lth column multiplied by a scalar λ from the right Then A ij = C ij, i = 1,,n, j= 1,,l 1, l + 1,,n 211 The following homological relations play an important role in the theory Theorem 27 The following relations hold a Row homological relations: Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 A ij A il 1 sj = A il A ij 1 sl s = i 212 b Column homological relations: A kj 1 it A ij = A ij 1 kt A kj r j 213

6 482 A Berenstein and V Retakh 23 Noncommutative Sylvester formula The following noncommutative version of the famous Sylvester identity found in [6, 7] is closely related with the fundamental Heredity principle see [5, 8] Let A = a ij, i, j = 1,,nbe a matrix over a skew-field FLetk<n 1 Suppose the k k submatrix A 0 = a ij, i I 0, j J 0 is invertible For each p/ I 0, q/ J 0, construct the k + 1 k + 1 submatrix A pq = a ij, where i I 0 {p}, j J 0 {q} Set b pq = Apq pq 214 and construct the matrix B = b pq, p/ I 0, q/ J 0 We call the submatrix A 0 a pivot for matrix B Theorem 28 For s/ I 0, t/ J 0, A st = B st 215 The particular case of the theorem when I 0 = J 0 = {2,,n 1} is called the noncommutative Lewis-Carroll identity Example 29 Let n = 3, I 0 = J 0 = {2}Then A 11 equals a 11 a 12 a 21 a 22 a 12 a 13 a 21 a 22 a 22 a 23 a 22 a 23 a 31 a 32 a 32 a Quasi-Plücker coordinates and Gauss LDU-factorization Here we recall some definitions and results from [5, 8] Let A = a pq, p = 1,,k, q = 1,,n, k<n, be a matrix over a skew-field FFix 1 i, j, i 1,,i k 1 n such that i/ I = {i 1,,i k 1 }For1 s k, set Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 a 1i a 1i1 a 1ik 1 q I ija = a ki a ki1 a kik 1 1 si a 1j a 1i1 a 1ik 1 a kj a ki1 a kik 1 sj 217 Proposition 210 i q I ij A does not depend on s ii q I ij ga = qi ij A for any invertible k k matrix g over F

7 coordinate Noncommutative Double Bruhat Cells and Their Factorizations 483 We call q I ij A a left quasi-plücker coordinate of the matrix A In the commutative case, q I ij = p ji/p ii, where p α1 α k is the standard Plücker Similarly, one can introduce right quasi-plücker coordinates Consider a matrix B = b ij, i = 1,,n; j = 1,,k, k<nover a skew-field F Fix1 i, j, i 1,,i k 1 n such that j/ I = i 1,,i k 1 Alsofix1 t k and set b i1 b ik r I b i1 1 b i1 k ijb = b ik 1 1 b ik 1 k it b j1 b jk b i1 1 b i1 k b ik 1 1 b ik 1 k Proposition 211 i r I ij B does not depend on t ii r I ij Bg = ri ij B for any invertible k k matrix g over F 1 jt 218 We call r I ij B a right quasi-plücker coordinate of the matrix B To describe the Gauss decomposition we need the following notations Let A = a ij, i, j = 1,,n Set A k = a ij, i, j = k,,n, B k = a ij, i = 1,,n, j = k,,n, and C k = a ij, i = k,,n, j = 1,,n These are submatrices of sizes n k + 1 n k + 1, n n k + 1, and n k + 1 n, respectively Suppose that the quasideterminants y k = A k kk, k = 1,,n, 219 are defined and invertible Theorem y z αβ A =, 220 x βα 1 0 y n 0 1 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 where x βα = r 1,,α 1 βα B α, 1 α<β n, z αβ = q 1,,α 1 αβ C α, 1 α<β n 221

8 484 A Berenstein and V Retakh A noncommutative analog of the Bruhat decomposition was given in [5] Example 213 For n = 2, 1 0 a a 1 11 A = a a 21 a A Positive quasiminors Most of the results in this subsection are new For a given matrix x Mat n R and I, J [1, n] = {1,2,,n}, denote by x I,J the submatrix with the I rows and J columns And, if I = J, that is, when x I,J is a square matrix, for any i I, j J, denote by x I,J i,j the quasideterminant of the submatrix x I,J with the marked position i, j We denote by i x the principal i i quasiminor of x Mat n R, that is, i x = x{1,2,,i},{1,2,,i} i,i 223 The following fact is obvious Lemma 214 For any I, J {1,2,,n} such that I = J = k and any i I, j J, there exist permutations u, v of {1,2,,n} such that I = u{1,,k}, J = v{1,,k}, i = uk, j = vk, and for any x Mat n R, k u 1 x v = xi,j i,j 224 where the permutations u and v are identified with their corresponding n n matrices Definition 215 For I, J [1, n], I = J, i I, j J, define the positive quasiminor i,j I,J as follows: i,j I,J x = 1d ii+d j J xi,j i,j, 225 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 where d i I resp, d j J is the number of those elements of I resp, of J which are greater than i resp, than j The definition is motivated by the fact that for a commutative ring R, one has i,j I,J x = det x I,J det x I,J, 226 where I = I \{i}, J = J \{j}thatis, a positive quasiminor is a positive ratio of minors

9 Noncommutative Double Bruhat Cells and Their Factorizations 485 Let S n be the group of permutations on {1,2,,n} and k [1, n] For any permutations u, v S n, set k u,vx := i,j I,J x = 1d ii+d j J k u 1 x v, 227 where I = u{1,,k}, J = v{1,,k}, i = uk, j = vk Denote by D n = D n R the set of all diagonal n n matrices over R Clearly, positive quasiminors satisfy the relation k u,vhxh = h uk k u,vxh vk 228 for h = diagh 1,,h n,h = diagh 1,,h n D n, and k = 1,2,,n Let σ be an involutive automorphism of Mat n R defined by σx ij = x n+1 i,n+1 j 229 The following fact follows from the elementary properties of quasideterminants see, eg,[5, Proposition 124]Letw o = n, n 1,,1 be the longest permutation in S n Lemma 216 For any u, v S n and x Mat n R, i u,v σx = i wou,w ovx 230 Now we present some less obvious identities for positive quasiminors For each permutation v S n, denote by lv the number of inversions of valsofori = 1,2,,n 1, denote by s i the simple transposition i, i + 1 S n Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Proposition 217 Let u, v S n and i [1, n 1] be such that lus i = lu + 1 and lvs i = lv + 1Then i us i,vs i = i us i,v i 1 i u,v u,vsi + i+1 i usi,v 1 i+1 u,v = i u,v u,v i+1 i+1 1 us i,v = i usi,v u,v, 1 i+1 us i,v, u,v i+1 i 1 u,vsi = i+1 u,vs i i 1, u,v i 1, u,v i+1 1 i+1 u,vs i u,v = i 1 i u,v u,vsi 231

10 486 A Berenstein and V Retakh Proof Clearly, the fourth and the fifth identities follow from the second and the third Using Lemma 214 and the Gauss factorization, it suffices to take u = v = 1, i = 1 in the first three identities, that is, work with 2 2 matrices Then the first three identities will take, respectively, the following obvious forms: x 22 = x 21 x 1 11 x x 11 x x 21 x 22, x 1 x 11 x 12 x 11 x x 21 x 22 = x 1 11 x 21 x 22, x 11 x 12 x 21 x 22 x 1 12 = x 11 x 12 x 21 x 22 x 1 11 One can prove the next proposition presenting some generalized Plücker relations Proposition 218 Let u, v S n and i [1, n 2]Iflus i s i+1 s i = lu + 3, then 232 us i+1 i+1,v = i+1 us i s i+1,v + i us i+1 s i,v i 1 i+1 usi,v u,v 233 If lvs i s i+1 s i = lv + 3, then i+1 u,vs i+1 = i+1 u,vs i s i+1 + i+1 u,v i u,vsi 1 i u,vsi+1 s i Basic factorizations in GL n F For i, j = 1,2,,n, denote by E ij the n n matrix unit in the intersection of the ith row and the jth column Then we abbreviate E i := E i,i+1 for i = 1,,n 1 The matrix units E 1,,E n 1 satisfy the relations E 2 i = 0 for i = 1,,n 1, E i E j = E j E i 31 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 if i j 2, and E i E i±1 E i = 0 32 Let i = i 1,,i m be a sequence of indices i k {1,2,,n 1} and x = x ij, i, j = 1,,nbe an upper triangular n n matrix over a skew-field F For such an i and x,

11 we write the formal factorization Noncommutative Double Bruhat Cells and Their Factorizations 487 x = 1 + t 1 E i1 1 + t2 E i2 1 + tm E im, 33 where all t k belong to the skew-field F Let k ij be the position of the ith occurrence of the index j i in the sequence i = 1,,n 1; 1,,n 2; ; 1, 2; 1Thatis, i + 1 k ij = ni 1 + j 34 2 for 1 i<j n The following fact is obvious Lemma 31 Let i = 1,,n 1; 1,,n 2; ; 1, 2; 1 Set temporarily t ij := t kij for 1 i<j n where t k is as in factorization 33 Then the matrix entries of the product x satisfy 1 i n k n 1 x i,i+k = 1 i 1 i 2 i k n+1 i k t i1,i 1 +it i2,i 2 +i+1t i3,i 3 +i+2 t ik,i k +i+k 1 35 Remark 32 In particular, after the specialization t kij := y j for 1 i<j n in 35, for some elements y 2,,y n, we obtain x i,i+k = i<j 1 <j 2 < <j k n y j1 y j2 y jk 36 That is, each matrix entry of a so-specialized matrix x is an elementary symmetric function in y 2,,y n in the sense of [8] Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Proposition 33 System 35 has a unique solution of the form t ij = x j i,n i+1 I,J x I1,J 1 j i+1,n i+1 37 for 1 i<j n, provided the quasiminors are defined Here I = {j i + 1,,j}, I 1 = {j i + 2,,j}, and J = {n i + 1,,n}

12 488 A Berenstein and V Retakh Proof First of all, we have the relations x i,n = t 1,i+1 t 1,i+2 t 1,n 38 for i = 1,,n 1 Therefore, t 1,i+1 = x i,n x 1 i+1,n 39 for all j = 2,,n, which verifies 37 and We define a sequence x 0,x 1,,x n 1 of matrices inductively by setting x 0 = I x m = I + t n m,n+1 m E 12 I + tn m,n+2 m E 23 I + tn m,n E m,m+1 x m 1 for m = 1,2,,n 1 Clearly, x n 1 = x Lemma 34 For all i j m + 1 n, 310 x m ij = xi, j, m ij, 311 where xi, j, m isthe n m n m submatrix of x withthe rows {i, i+1,, i+n m 1} and the columns {j; m + 2, m + 3,,n} Proof We proceed by induction on n m By definition of x m, we have a recursion for the matrix entries of x m : x m i,j for 1 i j m + 1 = t n m,i+n m x m i+1,j + xm 1 ij 312 Taking j = m + 1, we obtain Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Therefore, t n m,i+n m = x m xm 1 i,m+1 i+1,m x m 1 ij = x m i,j x m xm 1x m i,m+1 i+1,m+1 i+1,j = x m i,j x m i+1,j x m i,m+1 x m i+1,m+1 314

13 by the above, Noncommutative Double Bruhat Cells and Their Factorizations 489 Furthermore, we use the inductive hypotheses precisely in the form 311Then, x m 1 ij = xi, j, m ij xi + 1, j, m i+1,j xi, m + 1, m i,m+1 xi + 1, m + 1, m i+1,m Using the Sylvester formula Theorem 28 with A = xi, j, m 1 and A 0 being a submatrix of A with the rows indexed by {i + 1,,i+ n m 1} and the columns indexed by {m + 2, m + 3,,n}, we obtain x m 1 ij = xi, j, m i,j xi + 1, j, m i+1,j This finishes the induction The lemma is proved xi, m + 1, m i,m+1 xi + 1, m + 1, m i+1,m+1 = x m 1 ij ij 316 Finally, using 311, 313, and the fact that x {m+2 i,,m+1},{n m,,n} = xi, m + 1, m for m = 0,1,,n 1, we obtain 37 The proposition is proved Another natural factorization of matrices is given by the following theorem Amatrixx = x ij, i, j = 1,,n, over a skew-field F is called generic if the sequence of rational functions t m,k = t m,k x, 1 m k n 1, is defined x 1,k m+1 x 1k t m,k = x m,k m+1 x mk Clearly, in terms of positive quasiminors, one has 1 x 1,k m+2 x 1,k+1, 317 x m,k m+2 x m,k+1 t m,k = m,k 1 m,k+1 {1,,m},{k m+1,,k} {1,,m},{k m+2,,k+1} 318 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Then define a sequence of matrices xm, k = x m,k ij, 1 m k n 1, by the inductive formula x1, n 1 = x 1 t 1,n 1 E n 1, xm, k = xm, k t m,k E k, xm + 1, n 1 = xm, m 1 t m+1,n 1 E n 1 319

14 490 A Berenstein and V Retakh In other words, xm, k 1 + ti,j E j = x, 320 i,j m,k where the order on all pairs m, k, 1 m k n 1, is defined by i, j m, k if and only if either i<mor i = m, j>k Theorem 35 a For a generic matrix x = x ij, i, j = 1,,n, over a skew-field F, x m,k ij = for all i, j such that i, j 1 m, k ie, for i<j, i<mand for i = m, j>k In particular, xn 1, n 1 is lower triangular b The entries x m,k ij are given by the following formulas For i m, 2 j k, x m,k ij = for i>m, j>k, x m,k ij = x 1,j m+1 x 1j, 322 x m 1,j m+1 x m 1,j x i,j m+1 x i,j x 1,j m x 1,j, 323 x m,j m x m,j x i,j m x i,j and x m,k ij = x ij otherwise Proof It is enough to show that matrices xm, k satisfy the following conditions: i x m,k ij = 0 for i<j, i<mand for i = m, j>k, ii x1, n 1 = x1 + E n 1 t 1,n 1, iii xm, k = xm, k E k t m,k, iv xm + 1, n 1 = xm, m1 + E n 1 t m+1,n 1 We proceed by induction over a totally ordered set of indices 1, n 1,,1, 1, 2, n 1,,2, 2,,n 1, n 1 It is easy to check that the entries of matrix x1, n 1 satisfy conditions i iv Suppose that these conditions are satisfied for matrix xm, l We consider then two cases: l>mand l = m Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011

15 Noncommutative Double Bruhat Cells and Their Factorizations 491 If l>m, then l = k + 1 for k m Define matrix xm, k by formula iii Evidently, the corresponding entries of matrices xm, k and xm, k + 1 coincide except for the entries with indices i, k for i m which are given by the formula x m,k ik = x m,k+1 ik t m,k + x m,k+1 ik For i m, the product x m,k+1 ik t m,k equals x 1,k m+1 x 1,j x 1,k m+1 x 1k x m 1,k m+1 x m 1,j x x i,k m+1 x i,j m,k m+1 x mk x 1,k m+2 x 1,k+1 x m,k m+2 x m,k According to the homological relations for quasideterminants, the last expression can be written as x 1,k m+1 x 1,j x 1,k m+1 x 1k x m 1,k m+1 x m 1,j x x i,k m+1 x i,j m,k m+1 x mk x 1,k m+2 x 1,k+1 x m,k m+2 x m,k Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 It follows that the element x m,k ik x m,k ij = = x m,k+1 ik tm, k + x m,k+1 ik+1 is zero for i = mifi>m, x 1,k m+1 x 1,k x m,k m+1 x m,k+1 x i,k m+1 x i,k+1

16 492 A Berenstein and V Retakh This follows from the Sylvester identity applied to the corresponding matrix with the pivot equal to x 1,k m+2 x 1,k 328 x m 1,k m+2 x m 1,k This shows that the entries of matrix xm, k satisfy part b of the theorem If l = m, one can check in a similar way that the entries of matrix xm + 1, n 1 satisfy part b of the theorem The theorem is proved Remark 36 It follows from the proof that matrices xm, k and elements t m,k are uniquely defined Example 37 Let n = 3Thent 1,2 = x 1 12 x 13, t 1,1 = x 1 11 x 12, x 11 x 12 t 2,2 = x 21 x 22 1 x 12 x 13 x 22 x 23, x 11 x 12 0 x 12 x 13 x 21 x 22 x1, 2 = x 22 x 23, x 12 x 13 x 31 x 32 x 32 x 33 x x 11 x 12 x 12 x 13 x 21 x1, 1 = x 21 x 22 x 22 x 23, x 11 x 12 x 12 x 13 x 31 x 31 x 32 x 32 x 33 x x 11 x 12 x 21 0 x2, 2 = x 21 x 22 x 11 x 12 x 31 x 33 x 31 x Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011

17 Noncommutative Double Bruhat Cells and Their Factorizations Examples 41 A factorization in the Borel subgroup of GL 3 F We write the formal factorization x 11 x 12 x 13 x t 12 + t 23 t 12 t 13 0 x 22 x 23 = 0 x t x x x t t 23 0 = 0 x t x assuming that all x ij,t ij are elements of a skew-field F Then we can express t ij as follows: t 13 = x 1 22 x 23, t 12 = x 1 11 x 13x 1 23 x 22, t 23 = x 1 11 x 12 x 1 11 x 13 x 1 23 x 22 = x x 12 x 13 x 22 x Remark 41 The above factorization exists and therefore is unique if and only if each of x 11, x 22, x 33, and x 23 is invertible 42 A factorization in GL 3 F We write the formal factorization over a skew-field F: x 11 x 12 x 13 x = x 21 x 22 x 23 = hx 2 t1 x 1 t2 x 2 t3 x2 t4 x1 t5 x2 t6, 43 x 31 x 32 x 33 where h t h = 0 h 2 0, x 1t = 0 1 0, x 2t = 0 1 t, 0 0 h t x 1 t = 1 t 0, x 2t = 0 t t 44 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011

18 494 A Berenstein and V Retakh Then we can express h i and t k as follows: x 21 x x 11 x 12 x h 3 = x 31, h 2 = x 31 x 32, h 1 = x 21 x 22 x 23, x 31 x 32 x 33 t 6 = x 1 12 x 13, t 5 = x 1 11 x 12, t 4 = x 22 x 21 x 1 11 x 1 12 x23 x 22 x 1 12 x x 11 x = x 21 x 22 x 11 x 12 x 13 t 1 = x 1 x 21 x x 31 x 32, t 2 = x 1 11 x 21 x 22 x 23, x 31 x 32 x 33 1 x 11 x x 11 x 12 x t 3 = x 21 x 22 x 21 x 22 x 23 x 31 x 32 x 33 In fact, if we define a sequence of matrices 1 x 12 x 13 x 22 x 23, 45 x 5 = x x 2 t6 1, x 4 = x 5 x 1 t5 1, x 3 = x 4 x 1 t4 1, 46 then x k 1 will have exactly one more zero entry in the upper part than x k : x 11 x 12 0 x x 5 = x 21 x 22 x 23, x4 = x 21 x 22 x 23, x 31 x 32 x 33 x 31 x 32 x 33 x x 3 = x 21 x 22 0 x 31 x 32 x Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 This determines t 6, t 5, and t 4 The rest of the parameters h 1, h 2, h 3, t 1, t 2, t 3 are obtained from the equation h 1 t x 3 = hx 2 t1 x 1 t2 x 2 t3 = h 2 t 1 1 h 2 t 1 1 t 2t h 3 h 3 t1 + t 2 t 1 3 h 3 t 1 t 3

19 Noncommutative Double Bruhat Cells and Their Factorizations A factorization in the unipotent subgroup of GL 4 F We write the formal factorization 1 x 12 x 13 x 14 1 t 12 + t 23 + t 34 t 12 t 13 + t 12 t 24 + t 23 t 24 t 12 t 13 t x 23 x x 34 = 0 1 t 13 + t 24 t 13 t t t = t t t t t assuming that all x ij,t ij are elements of a skew-field F Then we can express t k as follows: t 14 = x 34, t 13 = x 24 x 1 34, t 12 = x 14 x 1 24, t 24 = x 23 x 24 x 1 34 = x 23 x 24 1 x 34, t 23 = x 13 x 14 x 1 24 x 23 x23 x 24 x x 13 x 14 x 23 x =, x 23 x 24 1 x 34 x 12 x 13 x 14 t 34 =x 12 x 13 x23 x 24 x x14 x 1 34 x23 x 24 x = 1 x 23 x x Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Remark 42 The above factorization exists and therefore is unique if and only if x 24, x 34, and x 23 x 24 are invertible in F 1 x 34

20 496 A Berenstein and V Retakh 5 Double Bruhat cells in GL n F and their factorizations 51 Structure of GL n F Throughout this and the next section, we denote G := GL n F and will use the abbreviation for a, b Z {a, a + 1,,b} if a b, [a, b] = 51 otherwise Let U resp, U be the upper resp, lower unitriangular subgroup of GFori [1, r], we define the elementary unitriangular matrices x i t and y i t by x i t = I + te i, y i t = I + tf i 52 for i [1, n 1], where E i = E i,i+1, F i = E i+1,i are the corresponding matrix units in the notation of Section 3 Let H denote the subgroup of all diagonal matrices in G LetB resp, B be the subgroup of all upper resp, lower triangular matrices in G Clearly, B = HU, B = HU, and H = B B We denote by G 0 = B U the subset of elements x G that have a Gaussian LDUdecomposition This unique decomposition will be written as x = [x] [x] + where [x] B, but not necessarily in U For any x in the Gauss cell G 0 = B U, denote by [x] 0 the diagonal component of the Gauss LDU-factorization In particular, [x] 0 = [[x] ] 0 for any x G 0 For each i [1, n 1], let ϕ i : GL 2 F G denote the embedding corresponding to the 2 2 block at the intersection of the ith and i + 1st rows and the ith and i + 1st columns Thus, we have 1 t 1 0 x i t = ϕ i, y i t = ϕ i t 1 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 We also set t 0 t 1 0 h i t = ϕ i H, x i t = ϕ i 0 t 1 1 t 54 for any i and any t F By definition, x i t = y i th i t 1 = h i t 1 y i t 1 55

21 Noncommutative Double Bruhat Cells and Their Factorizations 497 More generally, it is easy to see that for each i [1, n 1] and any diagonal matrix h = diagh 1,,h n H, one has Hence hx i th 1 = x i hi th 1 i+1, h 1 y i th = y i h 1 i+1 th i 56 h j sx i t = x i s ε ji ts ε ij hj s, y i th j s = h j sy i s ε ij ts ε ji 57 for any i, j [1, n 1], where ε ij = δ ij δ i,j 1 Lemma 51 i For each i [1, n 1], x i sx i t = x i s 1 ts + t 1 x i s + t ii For each i [1, n 2], x i sx i+1 t = x i+1 stx i s iii For each i [2, n 1], x i sx i 1 t = x i 1 tsx i s iv For any i, j [1, n 1] such that i j >1, x i sx j t = x j tx i s 58 Proof Part i follows from the obvious identity s t s 1 s 1 t 1 s 1 ts + t 1 s + t 1 0 = = 1 s s+ t s+ t for s, t F Part ii follows from for s, t F 59 s s s s t = 1 s st = 0 1 st 1 s Part iii follows from t 0 1 t 0 1 ts s = 0 s 1 0 = s s s s Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 for s, t F Part iv is obvious

22 498 A Berenstein and V Retakh The symmetric group S n of G is naturally embedded into G via 0 1 i, i + 1 ϕ i for i [1, n 1] We also define a representative s i of the transposition i, i + 1 by 0 1 s i = ϕ i The elements s i satisfy the braid relations in W; thus the representative w can be unambiguously defined for any w W by requiring that uv = u v whenever luv = lu + lv 52 Bruhat cells and double Bruhat cells The group G has two Bruhat decompositions, with respect to opposite Borel subgroups B and B : G = BuB = B vb 514 u S n v S n Now define the Schubert cell Uw := wu w 1 U for w S n Then the following obvious fact demonstrates that the Bruhat cells BuB and B vb behave similarly to their commutative counterparts Lemma 52 a For each u S n, BuB = UuuB = BuUu, UuU = UuuU = UuU u b For each v S n, B vb = B Uvv = B Uvv 1 1 = vu v 1 B = v 1 1 U v 1 B 516 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Definition 53 For any permutations u, v S n, define the double Bruhat cell G u,v by G u,v = BuB B vb In this section, we will concentrate on the following subset L u,v G u,v, which we call a reduced double Bruhat cell: L u,v = UuU B vb 517

23 Noncommutative Double Bruhat Cells and Their Factorizations 499 Remark 54 In the commutative case, the reduced double Bruhat cells are symplectic leafs of the Poisson-Lie structure on GL n C see, eg,[9] These cells also emerge in the study of total positivity see [3, Section 4] on GL n The equations defining L u,v inside G u,v are as follows Proposition 55 An element x G u,v belongs to L u,v if and only if [u 1 x] 0 = 1, or equivalently if i u,ex = 1 for each i [1, n] The maximal torus H acts freely on G u,v by left or right translations, and L u,v is a section of this action Thus, L u,v is naturally identified with G u,v /H, and all properties of G u,v can be translated in a straightforward way into the corresponding properties of L u,v A double reduced word forapairu, v S n is a reduced word for an element u, v of the group S n S n To avoid confusion, we will use the indices 1,, r for the simple reflections in the first copy of W, and 1,,rfor the second copy A double reduced word for u, v is simply a shuffle of a reduced word i for u written in the alphabet [ 1, r] we will denote such a word by i and a reduced word i for v written in the alphabet [1, r] We denote the set of double reduced words for u, v by Ru, v For any sequence i = i 1,,i m of indices from the alphabet [1, r] [ 1, r], we define the product map x i : F m G by x i t1,,t m = xi1 t1 xim tm Factorization problem for reduced double Bruhat cells In this section, we address the following factorization problem for L u,v : for any double reduced word i Ru, v, find explicit formulas for the inverse birational isomorphism x 1 i between L u,v and F m, thus expressing the factorization parameters t k in terms of the product x = x i t 1,,t m L u,v Definition 56 Let x x ι be the involutive antiautomorphism of G given by Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 x ι = J n x 1 J n 519 for any x G, where J n = diag 1, 1, 1,, 1 n We will refer to the antiautomorphism x x ι as the positive inverse in G Itis easy to see that a ι = a 1 a H, x i t ι = x i t, y i t ι = y i t 520

24 500 A Berenstein and V Retakh The following fact is a direct noncommutative analogue of [4, Theorem 16 ] Lemma 57 For any u, v S n, BuB ι = Bu 1 B, UuU ι = Uu 1 U, B vb ι = B v 1 B 521 In particular, G u,v ι = G u 1,v 1 Definition 58 For any u, v W, the twist map ψ u,v : L u,v G is defined by ψ u,v x = [ xv 1] ι x ι 1[ u 1 x ] ι The following result is a noncommutative analogue of [3, Theorem 47] Theorem 59 The twist map ψ u,v is an isomorphism between L u,v and L v,u The inverse isomorphism is ψ v,u Proof The proof essentially follows the pattern of the commutative case from [3, 4] We need the following obvious fact Lemma 510 The twist map ψ u,v satisfies ψ u,v x = [ vx ι 1] + v[ u 1 x] + ι = [ xv 1 ] ιu 1 1[ u 1 x ι 1] 523 The restriction of ψ u,v to L u,v B U isamapl u,v B U L u,v B U given by the formula ψ u,v x x + = [ x+ v 1] ι [ u 1 ] ι x 524 In particular, the twist map ψ u,e : L u,e L e,u is given by + Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 ψ u,e x = [ u 1 x ] + ι 525 And ψ e,v : L e,v L v,e is given by ψ e,v x = [ xv 1] ι 526

25 Noncommutative Double Bruhat Cells and Their Factorizations 501 The formula 523 guarantees that ψ u,v L u,v UvU B ub = L v,u, that is, ψ u,v is a well-defined map L u,v L v,u Finally, we prove that ψ v,u is the inverse of ψ u,v, that is, ψ v,u ψ u,v = id Given x L u,v, denote y = ψ u,v x By definition 522, we have y = [ xv 1] ι x ι 1[ u 1 x ] + ι 527 Or, equivalently, y ι 1 [ ] = xv 1 1x [ u 1 1, x] + x = [ xv 1] y ι 1[ u 1 x ] Since ψ v,u y = [ yu 1] ι y ι 1[ v 1 y ] ι, in order to prove that ψ v,u y = x, it suffices to show that [ ] yu 1 ι [ ] = xv 1, [ v 1 y ] ι [ + = u 1 x ], or, equivalently, [ ] yu 1 = [ xv 1] ι, [ v 1 y] + = [ u 1 x ] ι We prove the first identity 531 Denote temporarily z = [xv 1 ] ι Then523 implies that Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 yu 1 = z u 1 1[ u 1 x ι 1] u According to Lemma 57, for any x UuU, we have x ι 1 Uu 1 1 U, and furthermore, by Lemma 52a, u 1 x ι u 1 Uuu 1 1 U U U and [u 1 x ι 1 ] u 1 Uuu 1 1 Hence, u 1 1 [u 1 x ι 1 ] u 1 U Therefore, [ ] yu 1 = [ z u 1[ 1 u 1 x ι 1] u 1] = [z] = z 533

26 502 A Berenstein and V Retakh This proves the first identity in 531 Now we prove the second identity in 531 Again denote temporarily t = [u 1 x] + ι Then523 implies that v 1 y = v 1[ vx ι v t 534 1]+ According to Lemma 52b, for any x B vb, one has xv 1 B Uv and [vx ι 1 ] + Uv Hence, v 1 [vx ι 1 ] + v v 1 Uvv U Therefore, [ v 1 y] + = [ v 1[ vx ι 1] v t] = [t] = t 535 This proves the second identity in 531 Theorem 59 is proved Nowwefixapairu, v S n S n and a double reduced word i = i 1,,i m Ru, v Recall that i is a shuffle of a reduced word for u written in the alphabet [ 1, r] and a reduced word for v written in the alphabet [1, r] In particular, the length m of i is equal to lu + lv We will use the convention that s i = 1 for each i [1, n 1]Fork [1, m], denote u k = s im s im 1 s ik, u >k = s im s im 1 s ik+1, 536 v k = s i1 s i2 s ik, v <k = s i1 s i2 s ik For example, if i = 2, 1, 3, 3, 2, 1, 2, 1, 1, then, say, u 7 = s 1 s 2 and v <7 = s 1 s 3 s 2 Now we are ready to state our solution to the factorization problem Theorem 511 Let i = i 1,,i m be a double reduced word for u, v, and suppose an element x L u,v can be factored as x = x i1 t 1 x im t m, with all t k nonzero elements of F Then the factorization parameters t k are determined by the following formula: Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 i v t k = <k,u >k y 1 i v <k,u k y = v i+1 <k,u k y 1 v i+1 <k,u >k y if i k <0, i v k,u >k y 1 v i+1 <k,u >k y = i v <k,u >k y 1 v i+1 k,u >k y if i k >0, 538 where y = ψ u,v x and i = i k

27 Noncommutative Double Bruhat Cells and Their Factorizations 503 Proof First, we list some important properties of positive quasiminors Recall that in Section 25, for i [1, n], we defined the principal quasiminor i by i x = x[1,i],[1,i] i,i 539 for any x G, where x [1,i],[1,i] denotes the principal i i submatrix of x In particular, 1 x = x 11 and n x = x n,n The following fact is obvious Lemma 512 The principal quasiminors are invariant under the left multiplication by U and the right multiplication by U, that is, i x xx + = i x 540 for any x + U, x U, x G in particular, i x = i [x] 0 = [x] 0 ii Furthermore, for any u, v S n, Also i u,vx = i u 1 xv 541 i u,v x ι = n+1 i w ovw o,w ouw o x We will prove 538 by induction on lu + lv The base of the induction with u = v = e is obvious We will consider the following four cases Case 1 u = e, v = e and i is separated, that is, i 1,, i l [1, n 1] and i l+1,,i m [1, n 1] for some l, or, equivalently, u = s i1 s il and v = s il+1 s im Case 2 u e, v e and i is not separated Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Case 3 u = e, v e Case 4 u e, v = e Consider Case 1 first Denote x := x i1 t1 xil tl, x+ := x il+1 tl+1 xim tm 543

28 504 A Berenstein and V Retakh Clearly, x L u,e, x + L e,v, and x = x x + L u,v Furthermore, the inductive hypothesis 538 for x says that t k = i e,u >k y+ 1 i v<k,u k y+ = i+1 e,u k y+ 1 i+1 e,u >k y+ 544 for k [1, l], where y + = ψ u,e x, i = i k And the inductive hypothesis 538 for x + says that t k = i 1 i+1 v k,e y v <k,e y = i 1 i+1 v<k,e y v k,e y 545 for k [l + 1, m], where y = ψ e,v x +, i = i k According to 524, 525, and 526, ψ u,v x = [ x + v 1] ι [ u 1 x ] + ι = y y Note also that j e,wy + = j e,wy y + and j w,ey = j w,ey y + for any w S n and j [1, n] Finally, taking into the account that v k = v <k = e for each k l, and u k = u >k = e for each k>l, we obtain 538 for x = x x + This finishes Case 1 Now consider Case 2 Wesaythatgiveni, apairi l,i l+1 is an inversion if i l >0 and i l+1 <0 Clearly, i has no inversions if and only if i is separated Here we will proceed by induction on the number of inversions The base of the induction is the already considered Case 1 no inversions Assume that i has an inversion i l,i l+1 = i, j, where i, j [1, n 1] Leti be obtained from i by switching i l and i l+1, that is, i has one inversion less than i According to the inductive hypothesis,538 holds for the factorization relative to i x = x i1 t1 xil 1 tl 1 x j tl xi tl+1 xil+2 tl+2 xim tm 547 Note that, according to Lemma 51, x j tl xi tl+1 = xi t l x j t l+1, 548 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 where tl+1,t l t l,t l+1 tl t l+1,t l = tl+1 t l,t l t 1 1,tl l t l+1 tl + t l+1 + t l+1 if i j >1, if i j = 1, if i j = 1, if i = j 549

29 Noncommutative Double Bruhat Cells and Their Factorizations 505 We have to prove that each of the parameters t 1,,t l 1,t l,t l+1,t l+2,,t m in the factorization relative to i x = x i1 t1 xil 1 tl 1 xi t l x j t l+1 xil+2 tl+2 xim tm 550 is given by 538 for i Obviously, if k l, l + 1, then v i <k = vi <k, vi k = vi k, ui >k = ui >k, ui k = ui k Therefore, each t k, k = l, l + 1 in the latter decomposition is given by 538 for i It remains to prove that t l and t l+1 are both given by 538 for i Denote temporarily u = u >l, v = v <l so that taking into account that i l = i l+1 j, i l+1 = i l = i we have v i l = vi <l+1 = v, v i l+1 = v s i, u i l+1 = ui >l+1 = u, u i l = u s j Therefore,538 for i with k = l and k = l + 1 becomes with the convention y = ψ u,v x, y = v 1 yu t l = j e,ey 1 j e,s j y = j+1 e,s j y 1 j+1 e,e y, t l+1 = i s i,ey 1 i+1 e,e y = i e,ey 1 i+1 s i,ey 551 Taking into account that v i <l = v, v i l = vi <l+1 = vi l+1 = v s i, u i >l+1 = u, u i l+1 = u i >l+1 = ui l = u s j, we only have to prove that t l = i s i,s j y 1 i+1 e,s j y = i e,s j y 1 i+1 s i,s j y, 552 t l+1 = j s i,ey 1 j s i,s j y = j+1 s i,s j y 1 j+1 s i,ey 553 Consider the following four subcases 1 i j > 1 Then clearly, i s i,s j y = i s i,ey, i+1 e,s j y = i+1 e,e y, and j s i,s j y = j e,s j y, j s i,ey = j e,ey Finally, by 549, t l = t lt l+1 and t l+1 = t l All these immediately imply 552 and j = i 1 According to 549, t l = i e,s j y 1 i e,ey, t l+1 = i e,ey 1 i+1 s i,ey, t l = t lt l+1 = i e,s j y 1 i+1 s i,ey = i e,s j y 1 i+1 s i,s j y, 554 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 which proves 552 Similarly, we obtain t l+1 = t l = j e,ey 1 j e,s j y = j s i,ey 1 j s i,s j y, 555 which proves 553

30 506 A Berenstein and V Retakh 3 j = i + 1 According to 549, t l = t l+1t l = i s i,ey 1 i+1 e,e y j e,ey 1 j e,s j y = i s i,ey 1 j e,s j y, 556 which proves 552 because i s i,ey = i s i,s i+1 y Similarly, we obtain t l+1 = t l = j+1 e,s j y 1 j+1 e,e y = j+1 s i,s j y 1 j+1 s i,ey, 557 which proves i = j According to 549, t l+1 = t l + t l+1 = i e,ey 1 i e,s i y + i s i,ey 1 e,e i+1 y = i s i,ey 1 i s i,ey i e,ey 1 i e,s i y + e,e i+1 y = i s i,ey 1 i s i,s i y by 231 This proves 553 Furthermore, according to 549, t lt l+1 = t 1 l t l+1 = i e,s i y 1 i e,ey i e,ey 1 s i+1 i,ey = i e,s i y 1 i+1 s i,ey Therefore, using the already proved formula 553, we obtain t l = i e,s i y 1 s i+1 i,ey t l 1 = i e,si y 1 s i+1 i,ey s i+1 i,ey 1 s i+1 i,s i y = i e,s i y 1 i+1 s i,s i y, which proves 552 This finishes Case Now we consider Case 3: i = i 1,,i m, where all i k >0, that is, i is a reduced word for v And let i = i m so that v = v s i and lv = lv + 1Let Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 x = x i t1 xim tm, x = x i t1 xim 1 tm 1 x i t 1 m 561 It is easy to see that xs i x i t 1 m = x 562

31 Noncommutative Double Bruhat Cells and Their Factorizations 507 Indeed, this follows from x i t 1 = x i ts i x i t 1, 563 which in turn follows from the obvious identity 1 t 0 1 t 1 t 0 1 t 1 = = t Note that x is factored along the reduced word i = i 1,,i m 1 ; i for s i,v Therefore, we can use the already proved Case 2 for the i -factorization of x Formula 538 for the factorization parameters t 1,,t m 1,t 1 m of x takes the form t k = i k v k,s i y 1 i k+1 v <k,s i y = i k v<k,s i y 1 i k+1 v k,s i y 565 for k [1, m 1], and t 1 m = i v,ey 1 i v,s i y = i+1 v,s i y 1 i+1 v,e y, 566 where y = ψ s i,v x Clearly, in order to finish Case 3, that is, to verify formula 538 for the i-factorization parameters t 1,,t m of x L e,v, it will suffice to prove that for any w S n, j [1, n], one has j w,s i y = j w,ey, 567 where y = ψ e,v xnotethat j w,s i y = j w,ey s i Thus, it will suffice to prove [ y s i ] = y 568 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Taking into account that xs i x i t 1 = x, 569 all we need to prove is the following fact

32 508 A Berenstein and V Retakh Lemma 513 Let v = v s i for some i such that lv = lv + 1 Then for any x L e,v and any t F, ψ e,v x x i t = [ ψ s i,v x x i t 1 s i ] 570 Proof Indeed, by Lemma 510, ψ e,v x x i t = [ x x i tv 1] ι 571 Using 563, we obtain x x i tv 1 = x x i ts i v 1 = x x i t 1 x i t 1 v 1 = x x i t 1 v 1 u for some u + U Therefore, [ x x i tv 1] = [ x x i t 1 v 1 ] u + = [ x x i t 1 v 1 ] 573 Summarizing, we obtain ψ e,v x x i t = [ x x i t 1 v 1 ] ι 574 On the other hand, by the second identity of 523 we have for any x L s i,v, [ ψ s i,v x ] ιsi 1 s [ i s i x ι 1] ] = [[ x v 1 ] + s i = [ x v 1 ] ι 575 because z = [s i x ι 1 ] + U ϕ i GL 2 and, therefore, s 1 i zs i B Thus, taking x = x x i t 1, we obtain Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 [ ψ s i,v x x i t 1 ] s i = [ x x i t 1 v ] 1 ι = ψ e,v x x i t 576 The lemma is proved This finishes Case 3 Case 4 is almost identical to Case 3 Therefore, Theorem 511 is proved

33 Remark 514 The commutative version of 538 is Noncommutative Double Bruhat Cells and Their Factorizations 509 v<k ω i,u k ω i y v<k ω i,u >k ω i y t k = v<k ω i 1,u k ω i 1 y v<k ω i+1,u k ω i+1 y v<k ω i,u k ω i y v k ω i,u >k ω i y if i k <0, if i k > Factorizations of G u,v In this section, we extend the result of Theorem 511 to factorizations in G u,v In order to do so, we first have to extend the twist ψ u,v to an isomorphism G u,v Gv,u which we will denote in the same way by ψ u,v hx = hψ u,v x 578 for any h H and any x L u,v This means that the twist ψ u,v is a left H-equivariant map G u,v G v,u Recall that for any g in the Gauss cell G 0 = B U, we denote by [g] 0 the diagonal component of the Gauss factorization Lemma 515 The general twist ψ u,v : G u,v Gv,u is given by ψ u,v g = u [ u 1 g ] [ 0 ] gv 1 ι g ι 1[ u 1 g ] ι, for any g G u,v Other formulas for ψ u,v are ψ u,v g = u [ u 1 g ] [ 0 vg ι 1] v[ u 1 g ] ι, + + ψ u,v g = u [ u 1 g ] [ ] 0 gv 1 ιu 1 1[ u 1 g ι 580 1] Also ψ u,v is symmetric: ψ u,v 1 = ψ v,u In particular, for u = v, the twist ψ v,v is an involution on G v,v Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Proof Clearly, for any h H and x UuU, we have [ u 1 hx ] 0 = [ u 1 hu u 1 hx ] 0 = u 1 hu [u 1 hx ] 0 = u 1 hu = u 1 h 581 Therefore, taking g = hx, where h H and x L u,v, and taking into account 522 and 523, we obtain the desirable formulas

34 510 A Berenstein and V Retakh Theorem 511 admits the following obvious generalization Theorem 516 Let i = i 1,,i m be a double reduced word for u, v, and suppose an element x G u,v can be factored as x = hx i1 t 1 x im t m, with all t k nonzero elements of F, and h = diagh 1,,h n H Then the factorization parameters h 1,,h n, t 1,,t m are determined by formula 538 and the following formula: h i = u 1 i u,e x for i [1, n] 582 The following two special cases of Theorem 516 will be of particular importance: u, v = e, w 0 and u, v = w 0,e, where w 0 is the longest element in S n In these cases, Definition 58 and Theorem 59 can be simplified as follows Formula 538 now takes the following form Corollary 517 Let i = i 1,,i m be a reduced word for w S n, and let t 1,,t m be nonzero elements of F i If x = x i1 t 1 x im t m, then the factorization parameters h 1,,h n and t 1,,t m are given by h i = i e,ex = x ii 583 for i [1, n], and t k = i s i1 s ik,ey 1 i+1 s i1 s ik 1,ey = i s i1 s ik 1,ey 1 i+1 s i1 s ik,ey, where y = ψ e,w x is given by 526, and i = i k 584 ii If x = hx i1 t 1 x im t m, then the factorization parameters h 1,,h n and t 1,,t m are given by h i = w 1 i w,e x 585 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 for i [1, n], and t k = i e,s im s ik+1 y 1 i e,s im s ik y = i+1 e,s im s ik y 1 i+1 e,s im s ik+1 y, 586 where y = ψ w,e x is given by 525

35 Noncommutative Double Bruhat Cells and Their Factorizations Other factorizations in GL n F and the maximal twist ψ wo,wo In this section, we will provide some explicit factorizations in G u,wo and G wo,v We consider a factorization of x G u,wo of the form x = x x n 1 x n 2 x 1, 61 where x G u,e and x m L e,sms m+1 s n 1 is given by x m = x m tm,m xm+1 tm,m+1 xn 1 tm,n 1 for m [1, n 1] Lemma 61 In the notation of 61, 62 t m,k = m,k [1,m],[k m+1,k] x 1 m,k+1 [1,m],[k m+2,k+1]x 63 for all 1 m k n 1 Proof The proof follows immediately from Theorem 35 Lemma 62 In the notation of 61, t ij = i,j [1,i] [n+i+1 j,n],[1,j] y 1 i,j+1 [1,i] [n+i j,n],[1,j+1]y 64 for all 1 i j<n, where y = ψ u,wo x Proof Denote by i 0 the following standard reduced word for w o : i 0 = n 1; n 2, n 1; ; 1,2,,n 1 65 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 It is convenient to identify i 0 with the sequence of pairs n 1, n 1; n 2, n 2, n 2, n 1; ; 1, 1, 1, 2,,1, n 1 66 Let i be any reduced word for u S n Then we put i and i 0 into a separated word i = i, i 0 for the element u, w o S n S n Denote by w i,n o the longest element of the subgroup of S n generated by the simple transpositions s i,s i+1,,s n 1

36 512 A Berenstein and V Retakh pair i, j, formula: Then in the notation of 537, we have for the position k of i corresponding to the i v k = w i+1,n o s i s i+1 s j, v <k = w i+1,n o s i s i+1 s j 1, ii v k j + 1 = w i+1,n o s i s i+1 s j j + 1 = w i+1,n o i = i, iii v <k j = w i+1,n o s i s i+1 s j 1 j = w i+1,n o i = i, iv v k [1, j+1] = w i+1,n o s i s i+1 s j 1 [1, j+1] = w i+1,n o [1, j+1] = [1, i] [n+i j, n], v v <k [1, j] = w i+1,n o s i s i+1 s j 1 [1, j] = w i+1,n o [1, j] = [1, i] [n + i + 1 j, n] On the other hand, taking 538 for i = i, i 0 with i k = j yields the following t k = j v <k,ey 1 j+1 v k,ey 67 Combining this with the results of the above computations, we obtain the desired formula for t k = t ij The lemma is proved The above facts imply an immediate corollary Corollary 63 For any u S n, the twist map ψ u,wo satisfies i,j [1,i] [n+i+1 j,n],[1,j] ψ u,w o x 1 i,j+1 [1,i] [n+i j,n],[1,j+1] ψ u,w o x 68 = i,j [1,i],[j+1 i,j] x 1 i,j+1 [1,i],[j i+2,j+1] x for all 1 i j n 1 We consider a factorization of x G wo,v of the form x = h x n 1 x n 2 x 1 x +, 69 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 where x + L e,v, h H, and x m L sm s n 1s m,e is of the form x m = x m τm,m x m 1 τm,m+1 x n 1 τm,n for m [1, n 1]

37 Noncommutative Double Bruhat Cells and Their Factorizations 513 The following result generalizes the factorization from Section 42 Proposition 64 In the notation of 69, that is, x n 1,1 h n = x n1,h n 1 = x n,1 x n,2 x n 1,2,, h 1 = 1 n 1 x 1n, 611 h m = m,n+1 m [m,n],[1,n+1 m]x 612 for m [1, n], and x m,1 x m,k+1 m τ m,k = 1 k m x k,1 x k,k+1 m for all 1 m k<n, that is, 1 h m 613 τ m,k = m,k+1 m [m,k],[1,k+1 m] x 1 m,n+1 m [m,n],[1,n+1 m]x 614 The proof is similar to the proof of Theorem 35 Example 65 Let n = 3 Then in the factorization x = h x 2 τ22 x 1 τ11 x 2 τ12 x+, 615 where h H and x + U, we have τ 11 = x ,3 123,123x, τ 12 = 1,2 12,12x 1 1,3 123,123x, τ 22 = x ,1 23,12x 616 Downloaded from imrnoxfordjournalsorg at Mass Inst of Technology on January 5, 2011 Our next result is a direct consequence of Theorem 516 Lemma 66 In the notation of 69, τ ij = j,j+1 i [1,j],[n+2 i,n] [1,j+1 i] y 1 j,n+1 i [1,j],[n+1 i,n] [1,j i]y 617 for all 1 i j<n, where y = ψ wo,e x

THREE LECTURES ON QUASIDETERMINANTS

THREE LECTURES ON QUASIDETERMINANTS THREE LECTURES ON QUASIDETERMINANTS Robert Lee Wilson Department of Mathematics Rutgers University The determinant of a matrix with entries in a commutative ring is a main organizing tool in commutative

More information

Math 249B. Geometric Bruhat decomposition

Math 249B. Geometric Bruhat decomposition Math 249B. Geometric Bruhat decomposition 1. Introduction Let (G, T ) be a split connected reductive group over a field k, and Φ = Φ(G, T ). Fix a positive system of roots Φ Φ, and let B be the unique

More information

arxiv:math/ v3 [math.rt] 21 Jan 2004

arxiv:math/ v3 [math.rt] 21 Jan 2004 CLUSTER ALGEBRAS III: UPPER BOUNDS AND DOUBLE BRUHAT CELLS arxiv:math/0305434v3 [math.rt] 21 Jan 2004 ARKADY BERENSTEIN, SERGEY FOMIN, AND ANDREI ZELEVINSKY Abstract. We develop a new approach to cluster

More information

Fundamentals of Engineering Analysis (650163)

Fundamentals of Engineering Analysis (650163) Philadelphia University Faculty of Engineering Communications and Electronics Engineering Fundamentals of Engineering Analysis (6563) Part Dr. Omar R Daoud Matrices: Introduction DEFINITION A matrix is

More information

ELEMENTARY LINEAR ALGEBRA

ELEMENTARY LINEAR ALGEBRA ELEMENTARY LINEAR ALGEBRA K R MATTHEWS DEPARTMENT OF MATHEMATICS UNIVERSITY OF QUEENSLAND First Printing, 99 Chapter LINEAR EQUATIONS Introduction to linear equations A linear equation in n unknowns x,

More information

1 Multiply Eq. E i by λ 0: (λe i ) (E i ) 2 Multiply Eq. E j by λ and add to Eq. E i : (E i + λe j ) (E i )

1 Multiply Eq. E i by λ 0: (λe i ) (E i ) 2 Multiply Eq. E j by λ and add to Eq. E i : (E i + λe j ) (E i ) Direct Methods for Linear Systems Chapter Direct Methods for Solving Linear Systems Per-Olof Persson persson@berkeleyedu Department of Mathematics University of California, Berkeley Math 18A Numerical

More information

Schubert Varieties. P. Littelmann. May 21, 2012

Schubert Varieties. P. Littelmann. May 21, 2012 Schubert Varieties P. Littelmann May 21, 2012 Contents Preface 1 1 SMT for Graßmann varieties 3 1.1 The Plücker embedding.................... 4 1.2 Monomials and tableaux.................... 12 1.3 Straightening

More information

Lemma 8: Suppose the N by N matrix A has the following block upper triangular form:

Lemma 8: Suppose the N by N matrix A has the following block upper triangular form: 17 4 Determinants and the Inverse of a Square Matrix In this section, we are going to use our knowledge of determinants and their properties to derive an explicit formula for the inverse of a square matrix

More information

Yangian Flags via Quasideterminants

Yangian Flags via Quasideterminants Yangian Flags via Quasideterminants Aaron Lauve LaCIM Université du Québec à Montréal Advanced Course on Quasideterminants and Universal Localization CRM, Barcelona, February 2007 lauve@lacim.uqam.ca Key

More information

a 11 x 1 + a 12 x a 1n x n = b 1 a 21 x 1 + a 22 x a 2n x n = b 2.

a 11 x 1 + a 12 x a 1n x n = b 1 a 21 x 1 + a 22 x a 2n x n = b 2. Chapter 1 LINEAR EQUATIONS 11 Introduction to linear equations A linear equation in n unknowns x 1, x,, x n is an equation of the form a 1 x 1 + a x + + a n x n = b, where a 1, a,, a n, b are given real

More information

Lecture Notes Introduction to Cluster Algebra

Lecture Notes Introduction to Cluster Algebra Lecture Notes Introduction to Cluster Algebra Ivan C.H. Ip Updated: April 14, 017 Total Positivity The final and historically the original motivation is from the study of total positive matrices, which

More information

II. Determinant Functions

II. Determinant Functions Supplemental Materials for EE203001 Students II Determinant Functions Chung-Chin Lu Department of Electrical Engineering National Tsing Hua University May 22, 2003 1 Three Axioms for a Determinant Function

More information

Boolean Inner-Product Spaces and Boolean Matrices

Boolean Inner-Product Spaces and Boolean Matrices Boolean Inner-Product Spaces and Boolean Matrices Stan Gudder Department of Mathematics, University of Denver, Denver CO 80208 Frédéric Latrémolière Department of Mathematics, University of Denver, Denver

More information

A matrix over a field F is a rectangular array of elements from F. The symbol

A matrix over a field F is a rectangular array of elements from F. The symbol Chapter MATRICES Matrix arithmetic A matrix over a field F is a rectangular array of elements from F The symbol M m n (F ) denotes the collection of all m n matrices over F Matrices will usually be denoted

More information

Representations and Linear Actions

Representations and Linear Actions Representations and Linear Actions Definition 0.1. Let G be an S-group. A representation of G is a morphism of S-groups φ G GL(n, S) for some n. We say φ is faithful if it is a monomorphism (in the category

More information

Parametrizations of k-nonnegative Matrices

Parametrizations of k-nonnegative Matrices Parametrizations of k-nonnegative Matrices Anna Brosowsky, Neeraja Kulkarni, Alex Mason, Joe Suk, Ewin Tang October 2, 2017 Abstract Totally nonnegative (positive) matrices are matrices whose minors are

More information

Direct Methods for Solving Linear Systems. Simon Fraser University Surrey Campus MACM 316 Spring 2005 Instructor: Ha Le

Direct Methods for Solving Linear Systems. Simon Fraser University Surrey Campus MACM 316 Spring 2005 Instructor: Ha Le Direct Methods for Solving Linear Systems Simon Fraser University Surrey Campus MACM 316 Spring 2005 Instructor: Ha Le 1 Overview General Linear Systems Gaussian Elimination Triangular Systems The LU Factorization

More information

Upper triangular matrices and Billiard Arrays

Upper triangular matrices and Billiard Arrays Linear Algebra and its Applications 493 (2016) 508 536 Contents lists available at ScienceDirect Linear Algebra and its Applications www.elsevier.com/locate/laa Upper triangular matrices and Billiard Arrays

More information

NEW CONSTRUCTION OF THE EAGON-NORTHCOTT COMPLEX. Oh-Jin Kang and Joohyung Kim

NEW CONSTRUCTION OF THE EAGON-NORTHCOTT COMPLEX. Oh-Jin Kang and Joohyung Kim Korean J Math 20 (2012) No 2 pp 161 176 NEW CONSTRUCTION OF THE EAGON-NORTHCOTT COMPLEX Oh-Jin Kang and Joohyung Kim Abstract The authors [6 introduced the concept of a complete matrix of grade g > 3 to

More information

Linear Algebra M1 - FIB. Contents: 5. Matrices, systems of linear equations and determinants 6. Vector space 7. Linear maps 8.

Linear Algebra M1 - FIB. Contents: 5. Matrices, systems of linear equations and determinants 6. Vector space 7. Linear maps 8. Linear Algebra M1 - FIB Contents: 5 Matrices, systems of linear equations and determinants 6 Vector space 7 Linear maps 8 Diagonalization Anna de Mier Montserrat Maureso Dept Matemàtica Aplicada II Translation:

More information

Problem 1. Let f n (x, y), n Z, be the sequence of rational functions in two variables x and y given by the initial conditions.

Problem 1. Let f n (x, y), n Z, be the sequence of rational functions in two variables x and y given by the initial conditions. 18.217 Problem Set (due Monday, December 03, 2018) Solve as many problems as you want. Turn in your favorite solutions. You can also solve and turn any other claims that were given in class without proofs,

More information

CHANGE OF BASIS AND ALL OF THAT

CHANGE OF BASIS AND ALL OF THAT CHANGE OF BASIS AND ALL OF THAT LANCE D DRAGER Introduction The goal of these notes is to provide an apparatus for dealing with change of basis in vector spaces, matrices of linear transformations, and

More information

ELEMENTARY SUBALGEBRAS OF RESTRICTED LIE ALGEBRAS

ELEMENTARY SUBALGEBRAS OF RESTRICTED LIE ALGEBRAS ELEMENTARY SUBALGEBRAS OF RESTRICTED LIE ALGEBRAS J. WARNER SUMMARY OF A PAPER BY J. CARLSON, E. FRIEDLANDER, AND J. PEVTSOVA, AND FURTHER OBSERVATIONS 1. The Nullcone and Restricted Nullcone We will need

More information

Cluster Algebras and Compatible Poisson Structures

Cluster Algebras and Compatible Poisson Structures Cluster Algebras and Compatible Poisson Structures Poisson 2012, Utrecht July, 2012 (Poisson 2012, Utrecht) Cluster Algebras and Compatible Poisson Structures July, 2012 1 / 96 (Poisson 2012, Utrecht)

More information

MATRIX ALGEBRA AND SYSTEMS OF EQUATIONS. + + x 1 x 2. x n 8 (4) 3 4 2

MATRIX ALGEBRA AND SYSTEMS OF EQUATIONS. + + x 1 x 2. x n 8 (4) 3 4 2 MATRIX ALGEBRA AND SYSTEMS OF EQUATIONS SYSTEMS OF EQUATIONS AND MATRICES Representation of a linear system The general system of m equations in n unknowns can be written a x + a 2 x 2 + + a n x n b a

More information

Infinite-Dimensional Triangularization

Infinite-Dimensional Triangularization Infinite-Dimensional Triangularization Zachary Mesyan March 11, 2018 Abstract The goal of this paper is to generalize the theory of triangularizing matrices to linear transformations of an arbitrary vector

More information

Math 249B. Nilpotence of connected solvable groups

Math 249B. Nilpotence of connected solvable groups Math 249B. Nilpotence of connected solvable groups 1. Motivation and examples In abstract group theory, the descending central series {C i (G)} of a group G is defined recursively by C 0 (G) = G and C

More information

Intrinsic products and factorizations of matrices

Intrinsic products and factorizations of matrices Available online at www.sciencedirect.com Linear Algebra and its Applications 428 (2008) 5 3 www.elsevier.com/locate/laa Intrinsic products and factorizations of matrices Miroslav Fiedler Academy of Sciences

More information

INVARIANT POLYNOMIALS IN THE FREE SKEW FIELD. Robert Lee Wilson

INVARIANT POLYNOMIALS IN THE FREE SKEW FIELD. Robert Lee Wilson INVARIANT POLYNOMIALS IN THE FREE SKEW FIELD Robert Lee Wilson Introduction The free associative algebra on {x 1,..., x n } over a field F, denoted F < x 1,..., x n >, has a universal quotient ring F (

More information

Matrix Factorization and Analysis

Matrix Factorization and Analysis Chapter 7 Matrix Factorization and Analysis Matrix factorizations are an important part of the practice and analysis of signal processing. They are at the heart of many signal-processing algorithms. Their

More information

Math 121 Homework 5: Notes on Selected Problems

Math 121 Homework 5: Notes on Selected Problems Math 121 Homework 5: Notes on Selected Problems 12.1.2. Let M be a module over the integral domain R. (a) Assume that M has rank n and that x 1,..., x n is any maximal set of linearly independent elements

More information

STAT 309: MATHEMATICAL COMPUTATIONS I FALL 2018 LECTURE 13

STAT 309: MATHEMATICAL COMPUTATIONS I FALL 2018 LECTURE 13 STAT 309: MATHEMATICAL COMPUTATIONS I FALL 208 LECTURE 3 need for pivoting we saw that under proper circumstances, we can write A LU where 0 0 0 u u 2 u n l 2 0 0 0 u 22 u 2n L l 3 l 32, U 0 0 0 l n l

More information

Classical Lie algebras and Yangians

Classical Lie algebras and Yangians Classical Lie algebras and Yangians Alexander Molev University of Sydney Advanced Summer School Integrable Systems and Quantum Symmetries Prague 2007 Lecture 1. Casimir elements for classical Lie algebras

More information

Parametrizations of Canonical Bases and Totally Positive Matrices

Parametrizations of Canonical Bases and Totally Positive Matrices advances in mathematics 122, 49149 (1996) article no. 0057 Parametrizations of Canonical Bases and Totally Positive Matrices Arkady Berenstein Department of Mathematics, Northeastern University, Boston,

More information

MATH 326: RINGS AND MODULES STEFAN GILLE

MATH 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 information

The 4-periodic spiral determinant

The 4-periodic spiral determinant The 4-periodic spiral determinant Darij Grinberg rough draft, October 3, 2018 Contents 001 Acknowledgments 1 1 The determinant 1 2 The proof 4 *** The purpose of this note is to generalize the determinant

More information

Primitive ideals and automorphisms of quantum matrices

Primitive ideals and automorphisms of quantum matrices Primitive ideals and automorphisms of quantum matrices S Launois and T H Lenagan Abstract Let q be a nonzero complex number that is not a root of unity. We give a criterion for 0 to be a primitive ideal

More information

Groups. 3.1 Definition of a Group. Introduction. Definition 3.1 Group

Groups. 3.1 Definition of a Group. Introduction. Definition 3.1 Group C H A P T E R t h r e E Groups Introduction Some of the standard topics in elementary group theory are treated in this chapter: subgroups, cyclic groups, isomorphisms, and homomorphisms. In the development

More information

Linear Algebra Notes. Lecture Notes, University of Toronto, Fall 2016

Linear Algebra Notes. Lecture Notes, University of Toronto, Fall 2016 Linear Algebra Notes Lecture Notes, University of Toronto, Fall 2016 (Ctd ) 11 Isomorphisms 1 Linear maps Definition 11 An invertible linear map T : V W is called a linear isomorphism from V to W Etymology:

More information

1 Differentiable manifolds and smooth maps. (Solutions)

1 Differentiable manifolds and smooth maps. (Solutions) 1 Differentiable manifolds and smooth maps Solutions Last updated: March 17 2011 Problem 1 The state of the planar pendulum is entirely defined by the position of its moving end in the plane R 2 Since

More information

5.6. PSEUDOINVERSES 101. A H w.

5.6. PSEUDOINVERSES 101. A H w. 5.6. PSEUDOINVERSES 0 Corollary 5.6.4. If A is a matrix such that A H A is invertible, then the least-squares solution to Av = w is v = A H A ) A H w. The matrix A H A ) A H is the left inverse of A and

More information

Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra

Course 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 information

Polynomial Properties in Unitriangular Matrices 1

Polynomial Properties in Unitriangular Matrices 1 Journal of Algebra 244, 343 351 (2001) doi:10.1006/jabr.2001.8896, available online at http://www.idealibrary.com on Polynomial Properties in Unitriangular Matrices 1 Antonio Vera-López and J. M. Arregi

More information

QUATERNIONS AND ROTATIONS

QUATERNIONS AND ROTATIONS QUATERNIONS AND ROTATIONS SVANTE JANSON 1. Introduction The purpose of this note is to show some well-known relations between quaternions and the Lie groups SO(3) and SO(4) (rotations in R 3 and R 4 )

More information

Exercises on chapter 1

Exercises on chapter 1 Exercises on chapter 1 1. Let G be a group and H and K be subgroups. Let HK = {hk h H, k K}. (i) Prove that HK is a subgroup of G if and only if HK = KH. (ii) If either H or K is a normal subgroup of G

More information

THE S 1 -EQUIVARIANT COHOMOLOGY RINGS OF (n k, k) SPRINGER VARIETIES

THE S 1 -EQUIVARIANT COHOMOLOGY RINGS OF (n k, k) SPRINGER VARIETIES Horiguchi, T. Osaka J. Math. 52 (2015), 1051 1062 THE S 1 -EQUIVARIANT COHOMOLOGY RINGS OF (n k, k) SPRINGER VARIETIES TATSUYA HORIGUCHI (Received January 6, 2014, revised July 14, 2014) Abstract The main

More information

Matrices and systems of linear equations

Matrices and systems of linear equations Matrices and systems of linear equations Samy Tindel Purdue University Differential equations and linear algebra - MA 262 Taken from Differential equations and linear algebra by Goode and Annin Samy T.

More information

Littlewood Richardson coefficients for reflection groups

Littlewood Richardson coefficients for reflection groups Littlewood Richardson coefficients for reflection groups Arkady Berenstein and Edward Richmond* University of British Columbia Joint Mathematical Meetings Boston January 7, 2012 Arkady Berenstein and Edward

More information

Linear Equations and Matrix

Linear Equations and Matrix 1/60 Chia-Ping Chen Professor Department of Computer Science and Engineering National Sun Yat-sen University Linear Algebra Gaussian Elimination 2/60 Alpha Go Linear algebra begins with a system of linear

More information

RANK AND PERIMETER PRESERVER OF RANK-1 MATRICES OVER MAX ALGEBRA

RANK AND PERIMETER PRESERVER OF RANK-1 MATRICES OVER MAX ALGEBRA Discussiones Mathematicae General Algebra and Applications 23 (2003 ) 125 137 RANK AND PERIMETER PRESERVER OF RANK-1 MATRICES OVER MAX ALGEBRA Seok-Zun Song and Kyung-Tae Kang Department of Mathematics,

More information

Determinants. Recall that the 2 2 matrix a b c d. is invertible if

Determinants. Recall that the 2 2 matrix a b c d. is invertible if Determinants Recall that the 2 2 matrix a b c d is invertible if and only if the quantity ad bc is nonzero. Since this quantity helps to determine the invertibility of the matrix, we call it the determinant.

More information

GAUSSIAN ELIMINATION AND LU DECOMPOSITION (SUPPLEMENT FOR MA511)

GAUSSIAN ELIMINATION AND LU DECOMPOSITION (SUPPLEMENT FOR MA511) GAUSSIAN ELIMINATION AND LU DECOMPOSITION (SUPPLEMENT FOR MA511) D. ARAPURA Gaussian elimination is the go to method for all basic linear classes including this one. We go summarize the main ideas. 1.

More information

Linear Algebra March 16, 2019

Linear Algebra March 16, 2019 Linear Algebra March 16, 2019 2 Contents 0.1 Notation................................ 4 1 Systems of linear equations, and matrices 5 1.1 Systems of linear equations..................... 5 1.2 Augmented

More information

Chapter 4 - MATRIX ALGEBRA. ... a 2j... a 2n. a i1 a i2... a ij... a in

Chapter 4 - MATRIX ALGEBRA. ... a 2j... a 2n. a i1 a i2... a ij... a in Chapter 4 - MATRIX ALGEBRA 4.1. Matrix Operations A a 11 a 12... a 1j... a 1n a 21. a 22.... a 2j... a 2n. a i1 a i2... a ij... a in... a m1 a m2... a mj... a mn The entry in the ith row and the jth column

More information

MAT 2037 LINEAR ALGEBRA I web:

MAT 2037 LINEAR ALGEBRA I web: MAT 237 LINEAR ALGEBRA I 2625 Dokuz Eylül University, Faculty of Science, Department of Mathematics web: Instructor: Engin Mermut http://kisideuedutr/enginmermut/ HOMEWORK 2 MATRIX ALGEBRA Textbook: Linear

More information

arxiv: v4 [math.rt] 9 Jun 2017

arxiv: v4 [math.rt] 9 Jun 2017 ON TANGENT CONES OF SCHUBERT VARIETIES D FUCHS, A KIRILLOV, S MORIER-GENOUD, V OVSIENKO arxiv:1667846v4 [mathrt] 9 Jun 217 Abstract We consider tangent cones of Schubert varieties in the complete flag

More information

Irreducible subgroups of algebraic groups

Irreducible subgroups of algebraic groups Irreducible subgroups of algebraic groups Martin W. Liebeck Department of Mathematics Imperial College London SW7 2BZ England Donna M. Testerman Department of Mathematics University of Lausanne Switzerland

More information

1 Determinants. 1.1 Determinant

1 Determinants. 1.1 Determinant 1 Determinants [SB], Chapter 9, p.188-196. [SB], Chapter 26, p.719-739. Bellow w ll study the central question: which additional conditions must satisfy a quadratic matrix A to be invertible, that is to

More information

Definition 2.3. We define addition and multiplication of matrices as follows.

Definition 2.3. We define addition and multiplication of matrices as follows. 14 Chapter 2 Matrices In this chapter, we review matrix algebra from Linear Algebra I, consider row and column operations on matrices, and define the rank of a matrix. Along the way prove that the row

More information

Linear Systems and Matrices

Linear Systems and Matrices Department of Mathematics The Chinese University of Hong Kong 1 System of m linear equations in n unknowns (linear system) a 11 x 1 + a 12 x 2 + + a 1n x n = b 1 a 21 x 1 + a 22 x 2 + + a 2n x n = b 2.......

More information

1 Differentiable manifolds and smooth maps. (Solutions)

1 Differentiable manifolds and smooth maps. (Solutions) 1 Differentiable manifolds and smooth maps Solutions Last updated: February 16 2012 Problem 1 a The projection maps a point P x y S 1 to the point P u 0 R 2 the intersection of the line NP with the x-axis

More information

arxiv: v1 [math.rt] 4 Dec 2017

arxiv: v1 [math.rt] 4 Dec 2017 UNIFORMLY COLUMN SIGN-COHERENCE AND THE EXISTENCE OF MAXIMAL GREEN SEQUENCES PEIGEN CAO FANG LI arxiv:1712.00973v1 [math.rt] 4 Dec 2017 Abstract. In this paper, we prove that each matrix in M m n (Z 0

More information

Linear Algebra: Lecture notes from Kolman and Hill 9th edition.

Linear Algebra: Lecture notes from Kolman and Hill 9th edition. Linear Algebra: Lecture notes from Kolman and Hill 9th edition Taylan Şengül March 20, 2019 Please let me know of any mistakes in these notes Contents Week 1 1 11 Systems of Linear Equations 1 12 Matrices

More information

School of Mathematics and Statistics. MT5824 Topics in Groups. Problem Sheet I: Revision and Re-Activation

School of Mathematics and Statistics. MT5824 Topics in Groups. Problem Sheet I: Revision and Re-Activation MRQ 2009 School of Mathematics and Statistics MT5824 Topics in Groups Problem Sheet I: Revision and Re-Activation 1. Let H and K be subgroups of a group G. Define HK = {hk h H, k K }. (a) Show that HK

More information

Linear Algebra. Linear Equations and Matrices. Copyright 2005, W.R. Winfrey

Linear Algebra. Linear Equations and Matrices. Copyright 2005, W.R. Winfrey Copyright 2005, W.R. Winfrey Topics Preliminaries Systems of Linear Equations Matrices Algebraic Properties of Matrix Operations Special Types of Matrices and Partitioned Matrices Matrix Transformations

More information

MAT 445/ INTRODUCTION TO REPRESENTATION THEORY

MAT 445/ INTRODUCTION TO REPRESENTATION THEORY MAT 445/1196 - INTRODUCTION TO REPRESENTATION THEORY CHAPTER 1 Representation Theory of Groups - Algebraic Foundations 1.1 Basic definitions, Schur s Lemma 1.2 Tensor products 1.3 Unitary representations

More information

Math113: Linear Algebra. Beifang Chen

Math113: Linear Algebra. Beifang Chen Math3: Linear Algebra Beifang Chen Spring 26 Contents Systems of Linear Equations 3 Systems of Linear Equations 3 Linear Systems 3 2 Geometric Interpretation 3 3 Matrices of Linear Systems 4 4 Elementary

More information

. =. a i1 x 1 + a i2 x 2 + a in x n = b i. a 11 a 12 a 1n a 21 a 22 a 1n. i1 a i2 a in

. =. a i1 x 1 + a i2 x 2 + a in x n = b i. a 11 a 12 a 1n a 21 a 22 a 1n. i1 a i2 a in Vectors and Matrices Continued Remember that our goal is to write a system of algebraic equations as a matrix equation. Suppose we have the n linear algebraic equations a x + a 2 x 2 + a n x n = b a 2

More information

LECTURE 3: REPRESENTATION THEORY OF SL 2 (C) AND sl 2 (C)

LECTURE 3: REPRESENTATION THEORY OF SL 2 (C) AND sl 2 (C) LECTURE 3: REPRESENTATION THEORY OF SL 2 (C) AND sl 2 (C) IVAN LOSEV Introduction We proceed to studying the representation theory of algebraic groups and Lie algebras. Algebraic groups are the groups

More information

CHAPTER 6. Representations of compact groups

CHAPTER 6. Representations of compact groups CHAPTER 6 Representations of compact groups Throughout this chapter, denotes a compact group. 6.1. Examples of compact groups A standard theorem in elementary analysis says that a subset of C m (m a positive

More information

Some notes on linear algebra

Some notes on linear algebra Some notes on linear algebra Throughout these notes, k denotes a field (often called the scalars in this context). Recall that this means that there are two binary operations on k, denoted + and, that

More information

Chapter 5: Matrices. Daniel Chan. Semester UNSW. Daniel Chan (UNSW) Chapter 5: Matrices Semester / 33

Chapter 5: Matrices. Daniel Chan. Semester UNSW. Daniel Chan (UNSW) Chapter 5: Matrices Semester / 33 Chapter 5: Matrices Daniel Chan UNSW Semester 1 2018 Daniel Chan (UNSW) Chapter 5: Matrices Semester 1 2018 1 / 33 In this chapter Matrices were first introduced in the Chinese Nine Chapters on the Mathematical

More information

Lecture 12 (Tue, Mar 5) Gaussian elimination and LU factorization (II)

Lecture 12 (Tue, Mar 5) Gaussian elimination and LU factorization (II) Math 59 Lecture 2 (Tue Mar 5) Gaussian elimination and LU factorization (II) 2 Gaussian elimination - LU factorization For a general n n matrix A the Gaussian elimination produces an LU factorization if

More information

Margulis Superrigidity I & II

Margulis Superrigidity I & II Margulis Superrigidity I & II Alastair Litterick 1,2 and Yuri Santos Rego 1 Universität Bielefeld 1 and Ruhr-Universität Bochum 2 Block seminar on arithmetic groups and rigidity Universität Bielefeld 22nd

More information

0.1 Rational Canonical Forms

0.1 Rational Canonical Forms We have already seen that it is useful and simpler to study linear systems using matrices. But matrices are themselves cumbersome, as they are stuffed with many entries, and it turns out that it s best

More information

Matrices. Chapter Definitions and Notations

Matrices. Chapter Definitions and Notations Chapter 3 Matrices 3. Definitions and Notations Matrices are yet another mathematical object. Learning about matrices means learning what they are, how they are represented, the types of operations which

More information

Equality: Two matrices A and B are equal, i.e., A = B if A and B have the same order and the entries of A and B are the same.

Equality: Two matrices A and B are equal, i.e., A = B if A and B have the same order and the entries of A and B are the same. Introduction Matrix Operations Matrix: An m n matrix A is an m-by-n array of scalars from a field (for example real numbers) of the form a a a n a a a n A a m a m a mn The order (or size) of A is m n (read

More information

Chapter 3. Basic Properties of Matrices

Chapter 3. Basic Properties of Matrices 3.1. Basic Definitions and Notations 1 Chapter 3. Basic Properties of Matrices Note. This long chapter (over 100 pages) contains the bulk of the material for this course. As in Chapter 2, unless stated

More information

Non-commutative arithmetic circuits with division

Non-commutative arithmetic circuits with division Non-commutative arithmetic circuits with division P. Hrubeš, A. Wigderson. December 2, 2013 Abstract We initiate the study of the complexity of arithmetic circuits with division gates over non-commuting

More information

MULTI-RESTRAINED STIRLING NUMBERS

MULTI-RESTRAINED STIRLING NUMBERS MULTI-RESTRAINED STIRLING NUMBERS JI YOUNG CHOI DEPARTMENT OF MATHEMATICS SHIPPENSBURG UNIVERSITY SHIPPENSBURG, PA 17257, U.S.A. Abstract. Given positive integers n, k, and m, the (n, k)-th m- restrained

More information

On Systems of Diagonal Forms II

On Systems of Diagonal Forms II On Systems of Diagonal Forms II Michael P Knapp 1 Introduction In a recent paper [8], we considered the system F of homogeneous additive forms F 1 (x) = a 11 x k 1 1 + + a 1s x k 1 s F R (x) = a R1 x k

More information

CANONICAL FORMS FOR LINEAR TRANSFORMATIONS AND MATRICES. D. Katz

CANONICAL FORMS FOR LINEAR TRANSFORMATIONS AND MATRICES. D. Katz CANONICAL FORMS FOR LINEAR TRANSFORMATIONS AND MATRICES D. Katz The purpose of this note is to present the rational canonical form and Jordan canonical form theorems for my M790 class. Throughout, we fix

More information

ON MULTI-AVOIDANCE OF RIGHT ANGLED NUMBERED POLYOMINO PATTERNS

ON MULTI-AVOIDANCE OF RIGHT ANGLED NUMBERED POLYOMINO PATTERNS INTEGERS: ELECTRONIC JOURNAL OF COMBINATORIAL NUMBER THEORY 4 (2004), #A21 ON MULTI-AVOIDANCE OF RIGHT ANGLED NUMBERED POLYOMINO PATTERNS Sergey Kitaev Department of Mathematics, University of Kentucky,

More information

CIRCULAR CHROMATIC NUMBER AND GRAPH MINORS. Xuding Zhu 1. INTRODUCTION

CIRCULAR CHROMATIC NUMBER AND GRAPH MINORS. Xuding Zhu 1. INTRODUCTION TAIWANESE JOURNAL OF MATHEMATICS Vol. 4, No. 4, pp. 643-660, December 2000 CIRCULAR CHROMATIC NUMBER AND GRAPH MINORS Xuding Zhu Abstract. This paper proves that for any integer n 4 and any rational number

More information

The Spectral Theorem for normal linear maps

The Spectral Theorem for normal linear maps MAT067 University of California, Davis Winter 2007 The Spectral Theorem for normal linear maps Isaiah Lankham, Bruno Nachtergaele, Anne Schilling (March 14, 2007) In this section we come back to the question

More information

DETERMINANTS. , x 2 = a 11b 2 a 21 b 1

DETERMINANTS. , x 2 = a 11b 2 a 21 b 1 DETERMINANTS 1 Solving linear equations The simplest type of equations are linear The equation (1) ax = b is a linear equation, in the sense that the function f(x) = ax is linear 1 and it is equated to

More information

CHAPTER 6. Direct Methods for Solving Linear Systems

CHAPTER 6. Direct Methods for Solving Linear Systems CHAPTER 6 Direct Methods for Solving Linear Systems. Introduction A direct method for approximating the solution of a system of n linear equations in n unknowns is one that gives the exact solution to

More information

SERRE FINITENESS AND SERRE VANISHING FOR NON-COMMUTATIVE P 1 -BUNDLES ADAM NYMAN

SERRE FINITENESS AND SERRE VANISHING FOR NON-COMMUTATIVE P 1 -BUNDLES ADAM NYMAN SERRE FINITENESS AND SERRE VANISHING FOR NON-COMMUTATIVE P 1 -BUNDLES ADAM NYMAN Abstract. Suppose X is a smooth projective scheme of finite type over a field K, E is a locally free O X -bimodule of rank

More information

G1110 & 852G1 Numerical Linear Algebra

G1110 & 852G1 Numerical Linear Algebra The University of Sussex Department of Mathematics G & 85G Numerical Linear Algebra Lecture Notes Autumn Term Kerstin Hesse (w aw S w a w w (w aw H(wa = (w aw + w Figure : Geometric explanation of the

More information

Topics in linear algebra

Topics 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 information

Twists for positroid cells

Twists for positroid cells Twists for positroid cells Greg Muller, joint with David Speyer December 0, 204 Objects of study: positroid cells The matroid of a k n matrix is the set of subsets of columns which form a basis, written

More information

Counting matrices over finite fields

Counting matrices over finite fields Counting matrices over finite fields Steven Sam Massachusetts Institute of Technology September 30, 2011 1/19 Invertible matrices F q is a finite field with q = p r elements. [n] = 1 qn 1 q = qn 1 +q n

More information

Cell decompositions and flag manifolds

Cell decompositions and flag manifolds Cell decompositions and flag manifolds Jens Hemelaer October 30, 2014 The main goal of these notes is to give a description of the cohomology of a Grassmannian (as an abelian group), and generalise this

More information

Category O and its basic properties

Category O and its basic properties Category O and its basic properties Daniil Kalinov 1 Definitions Let g denote a semisimple Lie algebra over C with fixed Cartan and Borel subalgebras h b g. Define n = α>0 g α, n = α

More information

GENERATING SETS KEITH CONRAD

GENERATING 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 information

M.6. Rational canonical form

M.6. Rational canonical form book 2005/3/26 16:06 page 383 #397 M.6. RATIONAL CANONICAL FORM 383 M.6. Rational canonical form In this section we apply the theory of finitely generated modules of a principal ideal domain to study the

More information

Linear Algebra and Matrix Inversion

Linear Algebra and Matrix Inversion Jim Lambers MAT 46/56 Spring Semester 29- Lecture 2 Notes These notes correspond to Section 63 in the text Linear Algebra and Matrix Inversion Vector Spaces and Linear Transformations Matrices are much

More information

Math 203, Solution Set 4.

Math 203, Solution Set 4. Math 203, Solution Set 4. Problem 1. Let V be a finite dimensional vector space and let ω Λ 2 V be such that ω ω = 0. Show that ω = v w for some vectors v, w V. Answer: It is clear that if ω = v w then

More information

Formal power series rings, inverse limits, and I-adic completions of rings

Formal 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 information

Canonical lossless state-space systems: staircase forms and the Schur algorithm

Canonical lossless state-space systems: staircase forms and the Schur algorithm Canonical lossless state-space systems: staircase forms and the Schur algorithm Ralf L.M. Peeters Bernard Hanzon Martine Olivi Dept. Mathematics School of Mathematical Sciences Projet APICS Universiteit

More information