ebook img

symplectic modules over overrings of polynomial rings PDF

20 Pages·2012·0.14 MB·English
by  
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 symplectic modules over overrings of polynomial rings

IndianJ.PureAppl. Math.,43(4): 371-390,August2012 (cid:176)c IndianNationalScienceAcademy SYMPLECTICMODULESOVEROVERRINGSOFPOLYNOMIALRINGS AlpeshM.Dhorajia DepartmentofMathematics,IITMumbai,Mumbai400076,India e-mail: [email protected] (Received13January2011;afterfinalrevision9June2012; accepted20June2012) Let B be a commutative Noetherian ring of dimension d and let S be a set of all monic polynomials in B[X]. Let A be a subring of S−1B[X] which contains B[X]. Let P be a symplectic A-module of rank 2n ≥ d, n > 0. ThenweprovethatESp(A2 ⊥P,(cid:104),(cid:105))actstransitivelyonUm(A2⊕P). Keywords: Projectivemodule;unimodularelement;cancellationproblem. 1. INTRODUCTION LetAbeacommutativeNoetherianring. AsymplecticA-moduleisapair(P,(cid:104),(cid:105)), where P is a finitely generated projective A-module and (cid:104),(cid:105) : P ×P → A is a non-degeneratealternatingbilinearform. Let A be a ring of dimension d and P be a symplectic A-module of rank 2n with2n ≥ d. ThenSwan[11]hasprovedthatESp(A2 ⊥ P,(cid:104),(cid:105))actstransitively on Um(A2 ⊕ P). Further, the above result of Swan is extended to polynomial 372 ALPESHM.DHORAJIA rings by Bhatwadekar and Laurent polynomial rings by Keshari in [2] and [6] re- spectively. In this paper we prove the following result see (3.9). This generalizes a result ofBhatwadekar([2],Theorem4.8). Theorem A — Let B be a ring of dimension d and let S be a multiplicative closed subset of B[X] containing monic polynomials. Let A be a ring such that B[X] ⊆ A ⊆ S−1B[X]. LetP beasymplecticA-moduleofrank2nwith2n ≥ d, n > 0. ThenESp(A2 ⊥ P,(cid:104),(cid:105))actstransitivelyonUm(A2⊕P). Weprovethefollowingresult(3.13),whichgeneralizesaresultofKeshari([6], TheoremA.7). TheoremB—LetB bearingofdimensiondandA = B[X ,...,X ,Y, 1], 1 m g whereg ∈ B[Y]ismonic. LetP beasymplecticA-moduleofrank2n ≥ d. Then ESp(A2 ⊥ P,(cid:104),(cid:105))actstransitivelyonUm(A2⊕P). Thefollowingresult(4.3),generalizesaresultofKeshari([6],TheoremA.8). TheoremC—LetB bearingofdimension2andA = B[X ,...,X ,Y, 1], 1 n g whereg isamonicpolynomialinB[Y]. LetP beaprojectiveA-moduleofrank2 withtrivialdeterminant. IfA2 iscancellative,thenP iscancellative. 2. PRELIMINARIES All rings considered in this paper are assumed to be commutative Noetherian and allmodulesarefinitelygenerated. Let A be a ring and let P be a projective A-module. Recall that p ∈ P is called a unimodular element if there exists a φ ∈ P∗ = Hom (P,A), such that A φ(p) = 1. ThesetofallunimodularelementsofP isdenotedbyUm(P). Arow (a ,...,a ) ∈ An is called a unimodular row if there exists (b ,...,b ) ∈ An 1 n 1 n such that a b +...+a b = 1. The set of all unimodular rows of length n is 1 1 n n denotedbyUm (A). n SYMPLECTICMODULESOVEROVERRINGS 373 Let n be a positive integer. Let GL (A) be the set of all n × n invertible n matricesoverA. SL (A) := {M ∈ GL (A)|det(M) = 1}. LetI beanidealin n n A, we define SL (A,I) to be kernel of the natural map SL (A) → SL (A/I). n n n WhenI = (a),wewriteSL (A,a)forSL (A,I). n n Werecallsomepreliminaryfactsaboutsymplecticmodules. Fornotationsand terminologywefollow[2]. Formoredetail,see([2],Section4). Let A be a ring and let P be a finitely generated projective A-module. A bilinear form (cid:104),(cid:105) : P ×P → A is called alternating if (cid:104)p,p(cid:105) = 0 ∀p ∈ P. An alternating bilinear form (cid:104),(cid:105) induces a homomorphism α : P → P∗ (defined as α(p)(q) = (cid:104)p,q(cid:105)) such that α+α∗ = 0. An alternating form (cid:104),(cid:105) on P is called non-degenerateiftheinducedhomomorphismfromP toP∗ isanisomorphism. A symplectic A-module is a pair (P,(cid:104),(cid:105)), where P is a finitely generated projectiveA-moduleand(cid:104),(cid:105) : P ×P → Aisanon-degeneratealternatingbilinear form. If (P,(cid:104),(cid:105)) is a symplectic A-module then the rank of P is even and P has trivialdeterminant. If (P,(cid:104),(cid:105)) and (Q,(cid:104),(cid:105)) are two symplectic modules, then the non-degenerate alternating bilinear forms on P and Q will give rise (in a canonical manner) to a non-degenerate alternating bilinear form onP ⊕Q and we denote the symplectic module thus obtained by (P ⊥ Q,(cid:104),(cid:105)). The standard alternating form on A2 (in thenaturalbasis{e ,e })isgivenby(0 1). 1 2 1 0 Twosymplecticmodules(P,(cid:104),(cid:105))and(Q,(cid:104),(cid:105))areisomorphicifthereexistsan isomorphismτ : P → Qsuchthat (cid:104)p ,p (cid:105) = (cid:104)τ(p ),τ(p )(cid:105),∀p ,p ∈ P. 1 2 1 2 1 2 Anisometryofthesymplecticmodule(P,(cid:104),(cid:105))isanautomorphismof(P,(cid:104),(cid:105)). We denote by Sp(P,(cid:104),(cid:105)) the group of isometries of (P,(cid:104),(cid:105)). It is easy to see that SL (A) ⊆ Sp(A2,(cid:104),(cid:105)). 2 Let(P,(cid:104),(cid:105))beasymplecticA-moduleandletu,v ∈ P besuchthat(cid:104)u,v(cid:105) = 0. 374 ALPESHM.DHORAJIA Leta ∈ Aandletτ : P → P beamapdefinedby (a,u,v) τ (p) = p+(cid:104)p,v(cid:105)u+(cid:104)p,u(cid:105)v+a(cid:104)p,u(cid:105)u. (a,u,v) Thenitiseasytoseethatτ ∈ Sp(P,(cid:104),(cid:105)). Anisometryτ iscalleda (a,u,v) (a,u,v) symplectic transvectionifeitheruorvisunimodular. WedenotebyESp(P,(cid:104),(cid:105)) the subgroup of Sp(P,(cid:104),(cid:105)), generated by symplectic transvections. ESp(P,(cid:104),(cid:105)) isanormalsubgroupofSp(P,(cid:104),(cid:105)). Let (P,(cid:104),(cid:105)) be a symplectic A-module. Let c,d ∈ A, q ∈ P. If u = (0,1,0) and v = (0,c,q) ∈ A2 ⊕P, then τ is a symplectic transvection of (A2 ⊥ (−c,u,v) P,(cid:104),(cid:105))suchthat τ ((a,b,p)) = (a,b+ca+(cid:104)p,q(cid:105),p+aq). (−c,u,v) Similarly,ifu = (1,0,0)andv = (−d,0,−q),then τ ((a,b,p)) = (a+bd+(cid:104)q,p(cid:105),b,p+bq). (d,u,v) E(A2 ⊥ P,(cid:104),(cid:105))denotesthesubgroupofSp(A2 ⊥ P,(cid:104),(cid:105))generatedbyΘ and (c,q) σ forc,d ∈ Aandq ∈ P,whereΘ ,σ aredefinedasfollows: (d,q) (c,q) (d,q) Θ (a,b,p) = (a,b+ca+(cid:104)p,q(cid:105),p+aq) (c,q) σ (a,b,p) = (a+bd+(cid:104)q,p(cid:105),b,p+bq) (d,q) for(a,b,p) ∈ A2⊕P. Let B be a ring and let A = B[X]. Let F be a free A-module with a basis (cid:80) {e ,...,e }. Let p(X) = r γ (X)e ∈ F, where γ (X) ∈ A for 1 ≤ i ≤ r. 1 r i=1 i i (cid:80) i Then,forb ∈ B,wedenotebyp(bX)theelement r γ (bX)e ofF. i=1 i i WestatethefollowingresultwhichisduetoBhatwadekar([2],Theorem4.8). Theorem 2 — Let R be a ring of dimension d. Let A be a polynomial ring in r (≥ 0)variablesoverR. Let(P,(cid:104),(cid:105))beasymplecticA-moduleofrank2n > 0. If2n ≥ dthenESp(A2 ⊥ P,(cid:104),(cid:105))actstransitivelyonUm(A2⊕P). SYMPLECTICMODULESOVEROVERRINGS 375 Remark 2.2 : We have observed that in the above theorem we can replace ESp(A2 ⊥ P,(cid:104),(cid:105)) by E(A2 ⊥ P,(cid:104),(cid:105)). Let I be an ideal in A, then clearly the naturalmapE(A2 ⊥ P,(cid:104),(cid:105)) → E((A/I)2 ⊥ P/IP,(cid:104),(cid:105))issurjective. The following result is due to Keshari ([6], Theorem A.7), which extends the aboveresultofBhatwadekartoLaurentpolynomialring. Theorem 3 — Let B be a ring of dimension d and A = B[Y ,...,Y ,X±1, 1 r(cid:48) 1 ...,X±1]. Let(P,(cid:104),(cid:105))beasymplecticA-moduleofrank2n ≥ d,thenESp(A2 ⊥ r P,(cid:104),(cid:105))actstransitivelyonUm(A2⊕P). 3. MAIN THEOREM WebeginthissectionbyprovingthefollowingresultofBhatwadekarandRoy([3], Lemma4.1). Lemma 3.1 — Let B ⊂ C be rings of dimension d and x ∈ B such that B = C . Then x x (i)B/(1+xb) = C/(1+xb)forallb ∈ B. (ii)IfI isanidealofC suchthatht(I) ≥ dandI+xC = C,thenthereexists anelementb ∈ B suchthat1+xb ∈ I. PROOF: (i)SinceBx = Cx,bygoingmodulotheideal(1+xb)Bxwithb ∈ B, theprooffollows. (ii) Since ht(I) = d, without loss of generality we may assume that I is a maximal ideal of height d. By hypothesis, localizing C at x(1 + xB), we have B = C . SincedimB < dimB,wehavedimC < x(1+xB) x(1+xB) x(1+xB) x(1+xB) d. SinceI ismaximalidealofheightd,wegetIC = C . Therefore x(1+xB) x(1+xB) I contains an element of the form xm(1 + xa)n for some a ∈ B and for some positiveintegersmandn. SinceI+xC = C andI isaprimeideal,I containsan elementoftheform1+xbforsomeb ∈ B. Thiscompletestheproofof(ii). ThefollowingresultisaconsequenceofatheoremofEisenbud-Evansasstated 376 ALPESHM.DHORAJIA in([9],p.1420). (cid:50) Lemma 3.2 — Let A be a ring and let P be a projective A-module of rank r. Let(α,a) ∈ (P∗⊕A). Thenthereexistsanelementβ ∈ P∗ suchthatht(I ) ≥ r, a whereI = (α+aβ)(P). Inparticular,iftheideal(α(P),a)hasheight≥ r,then ht(I) ≥ r. Further,if(α(P),a)isanidealofheight≥ r andI isaproperidealof R,thenht(I) = r. ThefollowinglemmaisduetoSuslin([10],Lemma2.1). Lemma 3.3 — Let B be a ring and let A = B[X, F1,..., Fr], where F ∈ g g i B[X] for 1 ≤ i ≤ r and g ∈ B[X] is monic. Suppose f ,f ∈ B[X] and 1 2 c ∈ (f ,f )B[X] ∩ B. Then for any ideal I of A and g,h ∈ A ∩ B[X] with 1 2 h−g ∈ cI,thereexists∆ ∈ SL (A,I)suchthat[f (g),f (g)]∆ = [f (h),f (h)]. 2 1 2 1 2 Wenowprovethefollowingextensionlemma,whichforpolynomialring(i.e. A = B[X])isprovedin[7]. Ourprooffollowsusingthesameargumentasinthe proof([7],Lemma1.1,Chap.3). Lemma 3.4 — Let B be a ring and let A = B[X, F1,..., Fr], where F ∈ g g i B[X] for 1 ≤ i ≤ r and g ∈ B[X] is monic. If I is an ideal of A containing 1+gH for some monic polynomial H ∈ B[X] and J is an ideal of B such that I +JA = A,thenB∩I +J = B. PROOF: LetR = A/I ⊇ B/(B∩I)andletJ¯betheimageofJ inB/(B∩I). Byhypothesis,wehaveJR = R. SinceI containsapolynomial1+gH,itiseasy to seethatR isintegraloverB/(B ∩I). Therefore the“Going-Up” Theorem for integral extensions ([8], p.34) implies that J = B/(B ∩ I), therefore B ∩ I + J = B. (cid:50) Definition3.5—LetAbearingandsbeanon-zerodivisorinA. Let(P,(cid:104),(cid:105)) beasymplecticA-moduleofrank2n. Aset{e ,...,e ,f ,...,f } ⊂ P iscalled 1 n 1 n ans-symplecticbasisofP ifthefollowingholds: (i)(cid:104)e ,e (cid:105) = 0 = (cid:104)f ,f (cid:105)for1 ≤ i,j ≤ n. i j i j SYMPLECTICMODULESOVEROVERRINGS 377 (ii)(cid:104)e ,f (cid:105) = sfor1 ≤ i ≤ nand(cid:104)e ,f (cid:105) = 0foreveryi (cid:54)= j. i i i j Remark 3.6 : Let P be a symplectic A-module of rank 2n. If {e ,...,e ,f ,...,f } ⊆ P is s-symplectic basis of P then the module F := 1 n 1 n (cid:80) (cid:80) Ae + Af isafreeA-submoduleofP andsP ⊆ F. Fortheproofsee([2], i j Lemma4.2). Lemma3.7—LetB beareducedringofdimensiondandletg ∈ B[X]bea monicpolynomial. Assumethateither (i)A = B[X, f1,..., fn],wheref ∈ B[X]fori = 1,...,nor g g i (ii)A = B[X ,...,X ,X, 1]. 1 m g Let (P,(cid:104),(cid:105)) be a symplectic A-module of rank 2r ≥ d, r > 0. Then P has an s-symplectic basis {e ,...,e ,f ,...,f } such that sP ⊆ F, where F := 1 n 1 n (cid:80) (cid:80) Ae + Af isafreeA-submoduleofP ofrank2r. i j PROOF : Let S be the set of all non-zerodivisors in B. Since rank of P is constant, S−1A is a PID if A is as in (i) else S−1A = k[X ,...,X ,X, 1] for 1 m g some field k. Using [5], in both the cases S−1P is a free S−1A-module of rank 2r. Let p˜, q˜ ∈ S−1P such that (cid:104)p˜,q˜(cid:105) = 1, where (cid:104),(cid:105) is a induced form on 1 1 1 1 S−1P. WriteS−1P = p˜S−1A⊕q˜S−1A⊕Q,whereQisaS−1A-submoduleof 1 1 S−1P ofrank2r−2. ApplythesameargumenttoQ,byinductively,threreexists p˜,...,p˜ and q˜,...,q˜ such that (cid:104)p˜,q˜(cid:105) = 1 and (cid:104)p˜,q˜(cid:105) = 0 for i (cid:54)= j. Also, 2 r 2 r i i i j wehave(cid:104)p˜,q˜(cid:105) = 0and(cid:104)p˜,q˜(cid:105) = 0for2 ≤ i ≤ r. 1 i i 1 Chooset ∈ S suchthatp˜ = ei,q˜ = fi,wheree ,f ∈ P for1 ≤ i ≤ r. Let i t i t i i s = t2. Claim: {e ,...,e ,f ,...,f }isans-symplecticbasisofP. 1 r 1 r ProofofClaim : Weshowthat(i)(cid:104)e ,e (cid:105) = 0 = (cid:104)f ,f (cid:105)for1 ≤ i,j ≤ nand i j i j (ii)(cid:104)e ,f (cid:105) = sand(cid:104)e ,f (cid:105) = 0foreveryi (cid:54)= j. i i i j Foranyi,j for1 ≤ i,j ≤ r,wehave(cid:104)e ,e (cid:105) = (cid:104)tp˜,tp˜(cid:105). Sinces(cid:104)p˜,p˜(cid:105) = 0 i j i j i j in S−1A, we have (cid:104)e ,e (cid:105) = 0. Similarly, we can show that (cid:104)f ,f (cid:105) = 0 for i j i j 378 ALPESHM.DHORAJIA 1 ≤ j ≤ r and(cid:104)e ,f (cid:105) = 0foreveryi (cid:54)= j. i j Foranyi,1 ≤ i ≤ r,wehave e f (cid:104)e ,f (cid:105) = (cid:104)t i,t i(cid:105) = t2(cid:104)p˜,p˜(cid:105) = s(cid:104)p˜,p˜(cid:105) = s. i i i j i j t t Note that the above expression is computed in S−1A. This proves the claim. Nowby(3.6),theprooffollows. (cid:50) Now we prove the following result which is inspired by and the proof closely follows,Bhatwadekar’sresult([2],Proposition4.7). Proposition 3.8 — Let B be a ring and let s ∈ B be a non-zerodivisor. Let A = B[X, F1,...,Fr], where F ∈ B[X] and g ∈ B[X] is monic and let g g i (P,(cid:104),(cid:105)) be a symplectic A-module of rank 2n. Let {e ,...,e ,f ,...,f } ⊆ P 1 n 1 n be an s-symplectic basis of P. Let (α(X),β(X),p(X)) ∈ Um(A2 ⊕P), where (cid:80) (cid:80) α(X),β(X) ∈ B[X], p(X) ∈ n B[X]e + n B[X]f with α(X) ≡ 1 i=1 i j=1 j modulo (sXg) and β(X) = 1+gH with H ∈ B[X] is monic. Let b,b(cid:48) ∈ B be suchthatb−b(cid:48) ∈ sB. ThenthereexistsΨ ∈ SL (A,sX)ESp(A2 ⊥ P,(cid:104),(cid:105))such 2 thatΨ(α(bX),β(bX),p(bX)) = (α(b(cid:48)X),β(b(cid:48)X),p(b(cid:48)x)). PROOF : Since ESp(A2 ⊥ P,(cid:104),(cid:105)) is a normal subgroup ofSp(A2 ⊥ P,(cid:104),(cid:105)), G = SL (A,sX)ESp(A2 ⊥ P,(cid:104),(cid:105))isagroup. LetJ bethesetofelementsc ∈ B 2 havingfollowingproperty, b−b(cid:48) ∈ csB ⇒ ∃Φ ∈ G suchthatΦ(α(bX),β(bX), p(bX)) = (α(b(cid:48)X),β(b(cid:48)X),p(b(cid:48)x)). ClearlyJ isnonempty. SinceGisagroup, J isanidealofB. WeprovethatJ = B. Let t ∈ B ∩(α(X)2,β(X)2)A and let b,b(cid:48) ∈ B be such that b−b(cid:48) ∈ tsB. (cid:80) (cid:80) Sincep(X) ∈ F = B[X]e + B[X]f ,p(bX) = p(b(cid:48)X)−tsq(X)forsome i j q(X) ∈ F. Claim: B∩(α(X)2,β(X)2)A = B∩(α(X)2,β(X)2)B[X]. ProofofClaim: ItisenoughtoproveB∩(α(X)2,β(X)2)A ⊂ B∩(α(X)2, β(X)2)B[X]. Lett ∈ B∩(α(X)2,β(X)2)Athent = aα(X)2+bβ(X)2,where SYMPLECTICMODULESOVEROVERRINGS 379 a,b ∈ Aandβ(X)2 = 1+gK. MultiplyingbygK toexpressionoft,wegetthe following gKt = agKα(X)2+bgKβ(X)2. Addingbothsides−tandsimplifying,weget −t = −t(1+gK)+agKα(X)2+bgKβ(X)2. Sinceβ(X)2 = 1+gK,weget t = −agKα(X)2+(t−bgK)β(X)2. Applyingthisargumentfurtherifneeded,weseethattbelongstotheidealof B[X]generatedbyα(X)2 andβ(X)2. Thisprovestheclaim. Nowsincet ∈ (α(X)2,β(X)2)∩B,talsobelongstotheidealofB[X]gen- eratedbyα(bX)2 andβ(bX)2. By([2],Lemma4.3),thereexistsψ ∈ ESp(A2 ⊥ P,(cid:104),(cid:105)) such that ψ(α(bX),β(bX),p(bX)) = (α(bX),β(bX),p(b(cid:48)X)). Since bX − b(cid:48)X ∈ tsXB[X], by (3.3), the element (α(bX),β(bX),p(b(cid:48)X)) can be transformedto(α(b(cid:48)X),β(b(cid:48)X),p(b(cid:48)X)),byanelementoftheSL (A,sX),hence 2 t ∈ J. Sinceβ(X) = 1+gH withH ∈ B[X]monic,B/(B∩(α(X)2,β(X)2)) (cid:44)→ A/(α(X)2,β(X)2)) is an integral extension. Since (α(X)2,β(X)2)+sA = A, by(3.4),wehaveB∩(α(X)2,β(X)2)+sB = B. Thereforetheaboveargument showsthatJ +sB = B. Let m be a maximal ideal of B. If s ∈ m then m + J = B. Now as- sume that s ∈/ m. To complete the proof, it is enough to show that m + J = B. Since α(X) ≡ 1 modulo (sXg) and (α(X),β(X),p(X)) ∈ Um(A2 ⊕ P), we have (α(X),β(X),sXp(X)) ∈ Um(A2 ⊕ P). Let “−” denote reduc- tion modulo the ideal (mA,β(X)). Note that (α(X),sXp(X)) ∈ Um(A¯ ⊕ P¯). Then it is easy to see that g ≡ 1 modulo (mA,β(X)), where g is of 1 1 the form α(X) + (cid:104)p(cid:48),sXp(X)(cid:105) for some p(cid:48) ∈ P. This shows that the ide- 1 1 als (α(X) + (cid:104)p(cid:48),sXp(X)(cid:105),β(X)) and mA are comaximal. Further α(X) + 1 380 ALPESHM.DHORAJIA (cid:104)p(cid:48),sXp(X)(cid:105) + dβ(X) ∈ B[X] for some suitable d ∈ A. Clearly the ideals 1 (α(X)+(cid:104)p(cid:48),sXp(X)(cid:105)+dβ(X),β(X))and(α(X)+(cid:104)p(cid:48),sXp(X)(cid:105),β(X))are 1 1 equal. Therefore we can assume that g = α(X) + (cid:104)p(cid:48),sXp(X)(cid:105) ∈ B[X] for 1 some p(cid:48) ∈ P with (cid:104)p(cid:48),sXp(X)(cid:105) ∈ B[X]. Also we can take q(X) = sXp(cid:48) ∈ (cid:80) (cid:80) B[X]e + B[X]f . i j Letp (X) = p(X)+β(X)q(X)andη(X) = α(X)+(cid:104)q(X),p(X)(cid:105). Clearly 1 (cid:80) (cid:80) η(X) ∈ B[X] and p (X) ∈ B[X]e + B[X]f . Then η(X) ≡ 1 modulo 1 i j (sX)andmA+(η(X),β(X)) = A. Moreover,forb ∈ B σ (α(bX),β(bX),p(bX)) = (α(bX)+(cid:104)q(bX),p(bX)(cid:105),β(X), (0,q(bX)) p(bX)+β(bX)q(bX)) = (η(bX),β(bX),p (bX)). 1 LetJ = B∩(η(X)2,β(X)2). SincemA+(η(X)2,β(X)2) = Aandβ(X)2 1 ismonic,asbeforewegetm+(η(X)2,β(X)2)∩B = B. Thereforem+J = B. 1 Let t ∈ J . As before we see that b,b(cid:48) ∈ B such that b−b(cid:48) ∈ st B then there 1 1 1 existsΦ ∈ Gsuch thatΦ(η(bX),β(bX),p (bX)) = (η(b(cid:48)X),β(b(cid:48)X),p (b(cid:48)X)). 1 1 Therefore σ −1◦Φ◦σ (α(bX),β(bX),p(bX)) = (α(b(cid:48)X),β(b(cid:48)X),p(b(cid:48)X)). (0,q(b(cid:48)X)) (0,q(bX)) Hence J ⊂ J and thus m+J = B. Therefore J = B. This completes the 1 proof. (cid:164) Let B be a commutative Noetherian ring of dimension d. Let A be a ring such that B[X] ⊆ A ⊆ S−1B[X], where S denotes the multiplicatively closed set of monic polynomials in B[X]. Let P be any finitely generated projective A- module then P is regarded as a finitely generated projective D-module, where D is a subring of A generated by B[X] and some finite number of elements of A. Moreover,wemayassumethatD ⊆ B[X, 1]forsomeg ∈ S. Thereforetoprove g Theorem A, we may assume that (P,(cid:104),(cid:105)) is a symplectic A = B[X, f1,..., fl]- g g module.

Description:
Department of Mathematics, IIT Mumbai, Mumbai 400 076, India e-mail: [email protected]. (Received 13 January 2011; after final revision 9 June
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.