ebook img

A polynomial with Galois group SL2(F16) PDF

0.12 MB·English
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 A polynomial with Galois group SL2(F16)

A polynomial with Galois group SL (F ) 2 16 7 0 JohanBosman∗ 0 2 February 2, 2008 n a J 6 1 Abstract ] Inthispaperweshowthatthepolynomialx17−5x16+12x15−28x14+72x13−132x12+116x11− T 74x9+90x8−28x7−12x6 +24x5 −12x4 −4x3 −3x−1 ∈ Q[x]hasGaloisgroupSL2(F16), N fillinginagapinthetablesofJu¨rgenKlu¨nersandGuntherMalle(see[12]). Thecomputationofthis . polynomialusesmodularformsandtheirGaloisrepresentations. h t a m 1 Introduction [ ItisacomputationalchallengetoconstructpolynomialswithaprescribedGaloisgroup,see[12]formeth- 1 odsandexamples. Here, bythe Galois groupofa polynomialf ∈ Q[x] we meanthe Galoisgroupofa v splitting field of f overQ togetherwith its naturalaction onthe rootsof f in this splitting field. Ju¨rgen 2 Klu¨nersinformedmeaboutaninterestinggroupforwhichapolynomialhadnotbeenfoundyet, namely 4 4 SL2(F16)withitsnaturalactiononP1(F16). Thisactionisfaithfulbecauseofchar(F16) = 2. Itmustbe 1 notedthattheexistenceofsuchapolynomialwasalreadyknowntoMestre(unpublished).Inthispaperwe 0 willgiveanexplicitexample. 7 0 Proposition1. Thepolynomial / h t P(x):= x17−5x16+12x15−28x14+72x13−132x12+116x11 a m −74x9+90x8−28x7−12x6+24x5−12x4−4x3−3x−1∈Q[x] : v hasGaloisgroupisomorphictoSL (F )withitsnaturalactiononP1(F ). 2 16 16 i X WhatisstillunknowniswhetherthereexistsaregularextensionofQ(T)withGaloisgroupisomorphicto r a SL2(F16);regularheremeansthatitcontainsnoalgebraicelementsoverQapartfromQitself. Insection 2wewillsaysomewordsaboutthecalculationofthepolynomialandtheconnectionwithmodularforms. We’ll indicate how one can verify that it has the claimed Galois group in section 3 using computational Galoistheory. We willshowinsection4thatthispolynomialgivesa Galoisrepresentationassociatedto anexplicitlygivenmodularform. 2 Computation of the polynomial Inthissectionwewillbrieflyindicatehowonecanfindapolynomialasinproposition1. Wewillmakeuse ofmodularforms. Foranoverviewaswellasmanyfurtherreferencesonthissubjectthereaderisrefered to[6]. Let N be a positive integer and consider the space S (Γ (N)) of holomorphic cusp forms of weight 2 2 0 for Γ (N). A newform f ∈ S (Γ (N)) has a q-expansion f = a qn where the coefficients a 0 2 0 n≥1 n n are in a number field. The smallest number field containing all thPe coefficients is denoted by Kf. To a ∗PartiallysupportedbytheDutchscientificorganisationNWO. E-mail:[email protected] 1 given prime number ℓ and a place λ of K above ℓ one can attach a semi-simple Galois representation f ρ : Gal(Q/Q) → GL (F ) unramified outside Nℓ satisfying the following property: for each prime f 2 λ p∤NℓandanyFrobeniuselementFrob ∈Gal(Q/Q)attachedtopwehave p Tr(ρ (Frob ))≡a modλ and Det(ρ (Frob ))≡pmodλ. (1) f p p f p Therepresentationρ isuniqueuptoisomorphism.ThefixedfieldofKer(ρ )inQisGaloisoverQwith f f GaloisgroupisomorphictoIm(ρ ). Forℓ=2andanyλaboveℓequation(1)togetherwithChebotarev’s f density theoremimply that Im(ρ ) is containedin SL (F ). So to show that there is an extensionof Q f 2 λ with Galoisgroupisomorphicto SL (F )it sufficesto findan N anda newformf ∈ S (Γ (N)) such 2 16 2 0 thatthereisaprimeλofdegree4above2inK andIm(ρ )isthefullgroupSL (F ). Usingmodular f f 2 λ symbols we can calculate the coefficients of f, hence traces of matrices that occur in the image of ρ . f For a survey paper on how this works, see [18]. A subgroupΓ of SL (F ) contains elements of every 2 16 traceifandonlyifΓequalsSL (F );thiscanbeshowninseveralways,eitherbyadirectcalculationor 2 16 byinvokinga moregeneralclassificationresultlike[20, Thm. III.6.25]. Withthisinmind,afterasmall computersearch in which we check the occurringvaluesof Tr(ρ (Frob )) up to some moderatebound f p ofp,onefindsthatasuitablemodularformf existsinS2(Γ0(137)). ItturnsoutthatwehaveKf ∼=Q(α) withtheminimalpolynomialofαequaltox4+3x3−4x−1andthatf istheformwhoseq-expansion startswith f =q+αq2+(α3+α2−3α−2)q3+(α2−2)q4+··· . Now the next question comes in: knowing this modular form, how does one produce a polynomial? In general, one can use the Jacobian J (N) to construct ρ . In this particular case we can do that in the 0 f followingway. WeobservethatK isofdegree4andthattheprime2isinertinit. Furthermorewecan f verifythatthesubspaceofS (Γ (137))fixedbytheAtkin-Lehneroperatorw isexactlythesubspace 2 0 137 generatedbyallthecomplexconjugatesoff. Theseobservationsimplythatρ isisomorphictotheaction f ofGal(Q/Q)onJac(X (137)/hw i)[2], wherewe givethislatter space anF -vectorspace structure 0 137 16 viatheactionoftheHeckeoperators.NotethatIm(ρ )=SL (F )impliessurjectivityofthenaturalmap f 2 16 T → OK,f/(2) ∼= F16,whereTistheHeckealgebraattachedtoS2(Γ0(N)). Themethodsdescribedin [8]allowusnowtogivecomplexapproximationsofthe2-torsionpointsofJac(X (137)/hw i)toahigh 0 137 precision. Thispartofthecalculationtookbyfarthemosteffort;theauthorwillwritemoredetailsabout howthisworksinafuturepaper(orthesis). Weusethistogivearealapproximationofapolynomialwith GaloisgroupisomorphictoSL (F ). Thepaper[8]does,atleastimplicitly,giveatheoreticalupperbound 2 16 fortheheightofthecoefficientsofthepolynomialhenceanupperboundforthecalculationprecisiontoget anexactresult. Thoughthisupperboundissmallinthesensethatitleadstoapolynomialtimealgorithm, itisstillfartoohightobeofuseinpractice.Howeveritturnsoutthatwecanuseamuchsmallerprecision toobtainourpolynomial,theonlydrawbackbeingthatthisdoesnotgiveusaproofofitscorrectness,so wehavetoverifythisafterwards. ThepolynomialP′obtainedinthiswayhascoefficientsofabout200digitssowewanttofindapolynomial ofsmallerheightdefiningthesamenumberfieldK. Todothis,wefirstcomputetheringofintegersO K ofK. In[2]analgorithmtodothisisdescribed,providedthatoneknowsthesquarefreefactorisationof Disc(f)andevenifwedon’tknowthesquafrefreefactorisationofthediscriminant,thealgorithmproduces a ’good’orderin K. AssumingthatourpolynomialP′ is correctwe knowthatK is unramifiedoutside 2·137so we caneasily calculatethesquarefreefactorisationofDisc(f)andhenceapplythealgorithm. HavingdonethisweobtainanorderinK withadiscriminantsmallenoughtobeabletofactorandhence weknowthatthisisindeedthemaximalorderO . Explicitly,thediscriminantisequalto K Disc(O )=230·1378. (2) K We embed O as a lattice into C[K:Q] in the natural way and use lattice basis reduction, see [13], to K computeashortvectorα∈O −Z.Theminimalpolynomialofαhassmallcoefficients.Inourparticular K case [K : Q] is equal to 17, which is a prime number, hence this new polynomial must define the full fieldK. ThismethodgivesusalsoawayofexpressingαasanelementofQ(x)/(P′(x)). 2 3 Verification of the Galois group NowthatwehavecomputedapolynomialP(x),wewanttoverifythatitsGaloisgroupGal(P)isreally isomorphictoSL (F )andthatwecanidentifythesetΩ(P)ofrootsofP withP1(F )insucha way 2 16 16 thattheactionofGal(P)onΩ(P)isidentifiedwiththeactionofSL (F )onP1(F ). 2 16 16 ForcompletenessletusremarkthatitiseasytoverifythatP(x)isirreduciblesinceitisirreduciblemodulo 5.TheirreducibilityofP impliesthatGal(P)isatransitivepermutationgroupofdegree17.Thetransitive permutationgroupsof degree17havebeenclassified, see forexample[17]. From[20, Thm. III.6.25]it follows that up to conjugacy there is only one subgroup of index 17 in SL (F ), namely the group of 2 16 uppertriangularmatrices. Thisimpliesthatupto conjugacythereisexactlyonetransitiveG < S that 17 is isomorphicto SL (F ). Hence if Gal(P) ∼= SL (F ) is an isomorphismof groupsthen there is an 2 16 2 16 identificationof Ω(P)with P1(F ) suchthatthe groupactionsbecomecompatible. Itfollowsfromthe 16 classification in [17] that if the order of a transitive G < S is divisible by 5, then G must contain a 17 transitivesubgroupisomorphictoSL (F ). Toshowthat5 | #Gal(P)weusethefactthatforaprime 2 16 p ∤ Disc(P)wethedecompositiontypeofP modulopisequaltothecycletypeofanyFrob ∈ Gal(P) p attachedtop. Onecanverifythatmodulo7wegetthefollowingdecompositionintoirreducibles: P =(x−3)(x−5)(x15+3x14+4x12+6x11+3x10+x9+5x8+x6+x5+2x4+4x3+4x2+3x+6), showingthatindeed5|#Gal(P)henceGal(P)containsSL (F )asasubgroup. 2 16 ToshowthatGal(P)cannotbebiggerthanSL (F )itseemsinevitabletouseheavycomputercalcula- 2 16 tions. Itwouldbeinterestingtoseeamethodwhichdoesnotusethis. NotethattheactionofSL (F )onP1(F )issharply3-transitive. So firstweshowthatGal(P) isnot 2 16 16 4-transitivetoprovethatitdoesnotcontainA . Todothiswestartwithcalculatingthepolynomial 17 Q(x):= (X −α −α −α −α ), (3) 1 2 3 4 {α1,α2,αY3,α4}⊂Ω(P) wheretheproductrunsoverallsubsetsof{1,...,17}consistingofexactly4elements. Thisimpliesthat deg(Q) = 2380. OnecancalculateQ(x)bysymbolicmethods,see[5]. SupposethatGal(P)actingon Ω(P)is4-transitive. ThentheactiononΩ(Q)istransitivehenceQ(x)isirreducible. Soifwecanshow thatQ(x)isreducible,wehaveshownthatGal(P)isnot4-transitive. We have two ways to find a nontrivialfactor of Q(x): the first way is use a factorisation algorithm and thesecondwayistoproduceacandidatefactorourselves. Analgorithmthatworksverywellforourtype of polynomial is Van Hoeij’s algorithm, see [10]. One finds that Q(x) is the product of 3 polynomials of degrees 340, 1020 and 1020 respectively. A more direct way to produce a candidate factorisation is as follows. The methodfrom[8] givesa bijectionbetweenthe set ofapproximatedcomplexrootsof P′ and the set P1(F ) such thatthe action of Gal(P′) on Ω(P′) correspondsto the action of SL (F ) on 16 2 16 P1(F ), assumingthe outcomeis correct. From theprevioussectionwe knowhowto expressthe roots 16 ofP asrationalexpressionsintherootsofP hencethisgivesusabijectionbetweenΩ(P)andP1(F ), 16 conjecturallycompatiblewiththegroupactionsofGal(P)andSL (F )respectively.Acalculationshows 2 16 thattheactionofSL (F )onthesetofunorderedfour-tuplesofelementsofP1(F )has3orbits,ofsize 2 16 16 340,1020and1020respectively.Usingapproximationstoahighprecisionoftheroots,weusetheseorbits toproducesub-productsof(3), roundoffthecoefficientstothenearestintegerandverifyafterwardsthat theobtainedpolynomialsareindeedfactorsofQ(x). Let us remark that SL (F )⋊Aut(F ) with its natural action on P1(F ) is a transitive permutation 2 16 16 16 groupofdegree17,hencealsoitssubgroupG:=SL (F )⋊hFrob2i. Furthermore,itiswell-knownthat 2 16 2 SL (F )⋊Aut(F )isisomorpictoAut(SL (F ))(whereSL (F )actsbyconjugationandAut(F ) 2 16 16 2 16 2 16 16 acts on matrix entries) and actually inside S this group is the normaliser of both SL (F ) and itself. 17 2 16 3 Accordingto the classifiation of transitive permutationgroupsof degree17in [17] these two groupsare theonlyonesthatliestrictlybetweenSL (F )andA . OncewehavefixedSL (F )insideS ,these 2 16 17 2 16 17 twogroupsareactuallyuniquesubgroupsofS ,notjustuptoconjugacy. 17 Notethattheindex[A :Aut(SL (F ))]ishuge,namely10897286400.IfwecanverifythatGisnota 17 2 16 subgroupofGal(P),thenwearedone. ThefactthattheindexissmallandtheuniquenessofGmakean algorithmofGeisslerandKlu¨ners,see[9],verysuitabletodecidewhetherG < Gal(P). Itturnsoutthat thisisnotthecase,henceGal(P)∼=SL (F ). 2 16 4 Does P indeed define ρ ? f SonowthatwehaveshownthatGal(P)∼=SL (F )wecanwonderwhetherwecanprovethatP comes 2 16 fromthemodularformf weusedtoconstructitwith. OnceanisomorphismofGal(P)withSL (F )is 2 16 given,thepolynomialP definesarepresentationρ :Gal(Q/Q)→SL (F ). Abovewementionedthat P 2 16 thatOut(SL (F ))isisomorphictoAut(F )actingonmatrixentries.Hence,uptoanautomorphismof 2 16 16 F ,themapsendingσ ∈Gal(Q/Q)tothecharacteristicpolynomialofρ inF [x]isdeterminedbyP 16 P 16 andinfacttheisomorphismclassofρ iswell-defineduptoanautomorphismofF . Moreconcretely, P 16 wehavetoshowthatthesplittingfieldofP,whichwewilldenotebyL,isthefixedfieldofKer(ρ ). In f thissectionwewillbeusingbasicpropertiesoflocalfieldsascanbefoundin[15]. A continuousrepresentationρ : Gal(Q/Q) → GL (F ) hasa levelN(ρ) anda weight k(ρ). Insteadof 2 ℓ repeatingthefulldefinitionshere,whicharelengthy(atleastfortheweight)andcanbefoundin[16](see also[7]foradiscussiononthedefinitionoftheweight),wewilljustsaythattheyaredefinedintermsof the local representationsρ : Gal(Q /Q ) → GL (F ) obtained from ρ. The level is defined in terms p p p 2 ℓ of the representationsρ with p 6= ℓ andthe weightis definedin termsof ρ . Serre states the following p ℓ conjecturein[16]. Conjecture1. Letℓbeaprimeandletρ:Gal(Q/Q)→GL (F )beacontinuousoddirreducibleGalois 2 ℓ representation(arepresentationiscalledoddiftheimageofacomplexconjugationhasdeterminant−1). Thenthereexistsamodularformf oflevelN(ρ)andweightk(ρ)whichisanormalisedeigenformanda primeλ|ℓofK suchthatρandρ becomeisomorphicafterasuitableembeddingofF intoF . f f,λ λ ℓ Recently,KhareandWintenbergerprovedin[11]thefollowingpartofconjecture1. Theorem1. Conjecture1holdsineachofthefollowingcases: • N(ρ)isoddandℓ>2. • ℓ=2andk(ρ)=2. Withtheorem1inminditissufficienttoprovethatarepresentationρ = ρ attachedtoP haslevel137 P andweight2,whicharethelevelandweightofthemodularformf weusedtoconstructitwithandthat ofalleigenformsinS (Γ (137)),theformf isonewhichgivesrisetoρ . Therefore,intheremainderof 2 1 P thissectionwewillverifythefollowingproposition. Proposition2. Letf bethecuspformfromsection2. UptoanautomorphismofF ,therepresentations 16 ρ andρ areisomorphic.Inparticular,therepresentationρ hasSerre-level137andSerre-weight2. P f,(2) P 4.1 Verification ofthelevel Thelevelisthe easiest ofthe two to verify. Herewe have to dolocalcomputationsin p-adicfieldswith p 6= 2. AccordingtothedefinitionofN(ρ)in[16]itsufficestoverifythatρisunramifiedoutside2and 137, tamely ramified at 137 and the local inertia subgroupI at 137 leaves exactly one pointof P1(F ) 16 fixed.Thatρ isunramifiedoutside2and137followsimmediatelyfrom(2). P 4 From(2)andthefactthat1378kDisc(P)itfollowsthatthemonogeneousorderdefinedbyP ismaximal at137.Modulo137,thepolynomialP factorsas P =(x+14)(x2+6x+101)2(x2+88x+97)2(x2+106x+112)2(x2+133x+110)2. Letvbeanyprimeabove137inL.Fromtheabovefactorisationitfollowsthattheprime137decomposes in K as a product of 5 primes; one of them has its inertial and ramification degree equal to 1 and the other four ones have their inertial and ramification degrees equal to 2. Thus deg(v) is a power of 2, as L is obtained by succesively adjoining roots of P and in each step the relative inertial and ramification degreesoftheprimebelowv arebothatmost2. Inparticular,Gal(L /Q )isasubgroupofSL (F ) v 137 2 16 whose order is a power of 2. Now, { 1 ∗ } is a Sylow 2-subgroup of SL (F ), so Gal(L /Q ) is, 0 1 2 16 v 137 up to conjugacy, a subgroup of { 1 ∗(cid:0)}. H(cid:1)ence I is also conjugate to a subgroup of { 1 ∗ } and it is 0 1 0 1 actually nontrivial because 137 ra(cid:0)mifie(cid:1)s in L (so I of order 2 since the tame inertia gro(cid:0)up o(cid:1)f any finite Galoisextensionoflocalfieldsiscyclic). Itisimmediatethatρistamelyramifiedat137asnopowerof2isdivisibleby137. Also,itisclearthat I has exactly one fixed point in P1(F ) since [ ∗ ] is the only fixed point of any nontrivial element of 16 0 { 1 ∗ }. ThisestablishestheverificationofN(ρ(cid:0))=(cid:1) 137. 0 1 (cid:0) (cid:1) 4.2 Verification oftheweight Because the weight is defined in terms of the induced local representation Gal(Q /Q ), we will try to 2 2 computesomerelevantpropertiesofthesplittingfieldL ofP overQ ,wherevisanyplaceofLabove2. v 2 Inp-adicfieldsonecanonlydocalculationswithacertainprecision,butthisdoesnotgiveanyproblems since practically all properties one needs to know can be verified rigorously using a bounded precision calculationandtheerrorboundsinthecalculationscanbekepttrackofexactly. ThepolynomialP doesnotdefineanorderwhichismaximalattheprime2.Insteadweusethepolynomial R=x17−11x16+64x15−322x14+916x13+276x12−5380x11+2748x10+6904x9−23320x8 +131500x7−140744x6−16288x5−39752x4−48840x3+102352x2+234466x−1518, whichistheminimalpolynomialof 36863+22144α+123236α2+154875α3−416913α4+436074α5+229905α6−1698406α7 (cid:0)+1857625α8−467748α9−2289954α10+2838473α11−1565993α12+605054α13−263133α14 +112104α15−22586α16 /8844, (cid:1) whereαisarootofP. We canfactorRoverQ andseethatithasonerootinQ whichhappenstobe 2 2 odd,andanEisensteinfactorofdegree16,whichwewillcallE. Thistypeofdecompositioncanberead offfromtheNewtonpolygonofRanditalsoshowsthattheorderdefinedbyRisindeedmaximalat2. Fromtheoddnessoftherootand(2)itfollowsthat. v (Disc(E))=30. (4) 2 For the action of Gal(Q /Q ) on P1(F ) the factorisation means that there is one fixed point and one 2 2 16 orbitofdegree16. Ifwe adjoina rootβ ofE to Q andfactorE overQ (β) thenwe see thatit hasan 2 2 irreduciblefactorofdegree15;in[4]onecanfindmethodsforfactorisationandirreducibilitytestingthat canbeusedtoverifythis. Thismeansthat[L :Q ]isatleast240. v 2 AsubgroupofSL (F )thatfixesapointhastobeconjugatetoasubgroupofthegroup 2 16 ∗ ∗ H := ⊂SL (F ), 2 16 0 ∗ n(cid:16) (cid:17)o 5 whichisthestabilisersubgroupof[ ∗ ]. From#H =240itfollowsthatGal(L /Q )isisomorphictoH 0 v 2 andfromnowonwewillidentifyth(cid:0)es(cid:1)etwogroupswitheachother. WecanfilterH bynormalsubgroups: H ⊃I ⊃I ⊃{e}, 2 where I is the inertia subgroup and I is the wild ramification subgroup, which is the unique Sylow 2- 2 subgroupofI. WewishtodeterminethegroupsI andI . Letk(v)betheresidueclassfieldofL . The 2 v groupH/I is isomorphicto Gal(k(v)/F ) andI/I isisomorphicto a subgroupof k(v)∗. In particular 2 2 [I :I ]|(2[H:I]−1)follows. ThegroupH hastheniceproperty 2 1 ∗ [H,H]= ∼=F , 16 (cid:26)(cid:18)0 1(cid:19)(cid:27) which is its unique Sylow 2-subgroup. As H/I is abelian, we see that [H,H] ⊂ I. We conclude that I = [H,H], since above we remarked that I is the unique Sylow 2-subgroup of I. The restriction 2 2 [I :I ]|(2[H:I]−1)leavesonlyonepossibilityforI,namelyI =I . 2 2 Let L′ be the subextension of L /Q fixed by I. Then L′ is the maximal unramified subextension as v v 2 v wellasthemaximaltamelyramifiedsubextension. ItisinfactisomorphictoQ215,theuniqueunramified extension of Q of degree 15 and the Eisenstein polynomial E from above, being irreducible over any 2 unramifiedextensionofQ2,isadefiningpolynomialfortheextensionLv/Q215.Accordingto[14,Thm.3] wecanrelatethediscriminantofL tok(ρ)asfollows: v 240· 15 =450 ifk(ρ)=2 v (Disc(L ))= 8 2 v (cid:26) 240· 19 =570 ifk(ρ)6=2 8 Itfollowsfrom(4)thatv (Disc(L /Q ))=30·15=450,soindeedk(ρ)=2. 2 v 2 4.3 Verification oftheform f NowweknowN(ρ )=137andk(ρ )=2theorem1showsthatthereisaneigenformg ∈S (Γ (137)) P P 2 1 givingrisetoρ . Using[3, Cor. 2.7]weseethatifsuchag exists, thenthereactuallyexistssuchag of P trivial Nebentypus, i.e. g ∈ S (Γ (137)) (as SL (F ) is non-solvableρ cannotbe an induced Hecke 2 0 2 16 P characterfromQ(i)). AmodularsymbolscalculationshowsthatthereexisttwoGaloisorbitsofnewformsinS (Γ (137)): the 2 0 formf weusedforourcalculationsandanotherform,gsay. Theprime2decomposesinK asaproduct g λ3µ,whereλhasinertialdegree1andµhasinertialdegree4.Soitcouldbethatgmodµgivesrisetoρ . P Wewillshownowthatfmod(2)andgmodµactuallygivethesamerepresentation. Thecompletionsof O andO attheprimes(2)andµrespectivelyarebothisomorphictoZ , theunramifiedextension Kf Kg 16 ofZ ofdegree4. AfterachoiceofembeddingsofO andO intoZ weobtaintwomodularforms 2 Kf Kg 16 f′ andg′ with coefficientsin Z andwe wish to show thata suitable choiceof embeddingsexists such 16 thattheyarecongruentmodulo2. Accordingto[19,Thm.1],itsufficestocheckthereisasuitablechoice of embeddingsthat gives a (f′) ≡ a (g′)mod2 for all n ≤ [SL (Z) : Γ (137)]/6 = 23 (in [19] this n n 2 0 theoremisformulatedformodularformswithcoefficientsintheringofintegersofanumberfield,butthe proofalsoworksforp-adicrings).Usingamodularsymbolscalculation,thiscanbeeasilyverified. 5 Acknowledgements I would like to thank Ju¨rgen Klu¨ners for proposing this computational challenge and explaining some computationalGalois theory to me. Furthermore I want to thank Bas Edixhovenfor teaching me about modularformsandthecalculationoftheircoefficients.AllthecalculationsweredonewithMAGMA(see [1]), many of them on the MEDICIS cluster (http://medicis.polytechnique.fr). For being abletomakeuseoftheclusterIwanttothankMarcGiustiandPierreLafon. 6 References [1] W.Bosma,J.J.Cannon,C.E.Playoust,ThemagmaalgebrasystemI:theuserlanguage,J.Symbolic Comput.24(1997),no.3/4,235–265. [2] J.A. Buchmann and H.W. Lenstra, Jr., Approximating rings of integers in number fields, J. The´or. NombresBordeaux6(1994),no.2,221–260. [3] K.Buzzard,Onlevel-loweringformod2representations,Math.ResearchLetters7(2000),95–110. [4] D.G. Cantor and D.M. Gordon, Factoring polynomials over p-adic fields, Proceedings of the 4th InternationalSymposiumonAlgorithmicNumberTheory,2000,185–208. [5] D. Casperson and J. McKay, Symmetric functions, m-sets, and Galois groups, Math. Comp. 63 (1994),749–757 [6] F. Diamond and J. Im, Modular forms and modular curves, Seminar on Fermat’s Last Theorem (Toronto,ON,1993-1994),CMSConf.Proc.,17,Amer.Math.Soc.,Providence,RI,1995,39–133. [7] S.J.Edixhoven,TheweightinSerre’sconjecturesonmodularforms, Invent.Math.109(1992)no. 3,563–594. [8] S.J.Edixhoven,J-M.Couveignes,R.S.deJongetal.,Onthecomputationofcoefficientsofamodular form,eprint,2006,arXivreferencemath.NT/0605244. [9] K.GeisslerandJ.Klu¨ners,Galoisgroupcomputationforrationalpolynomials,J.SymbolicComput. 30(2000),653–674. [10] M.vanHoeij,Factoringpolynomialsandtheknapsackproblem,J.NumberTheory95(2002),167– 189. [11] C. Khare and J-P. Wintenberger, Serre’s modularity conjecture: the odd conductor case (I, II), http://www.math.utah.edu/∼shekhar/papers.html [12] J.Klu¨nersandG.Malle,ExplicitGaloisrealizationoftransitivegroupsofdegreeupto15,J.Sym- bolicComput.30(2000),no.6,675–716. [13] A.K. Lenstra, H.W. Lenstra, Jr. and L. Lova´sz, Factoring polynomials with rational coefficients, Math.Ann.261(1982),no.4,515–534. [14] H. MoonandY. Taguchi,RefinementofTate’sdiscriminantboundandnon-existencetheoremsfor modpGaloisrepresentations,DocumentaMath.ExtraVolumeKato(2003),641–654. [15] J-P.Serre,Localfields,Graduatetextsinmathematics67,Springer-Verlag,NewYork,1979. [16] J-P.Serre,Surlesrepre´sentationsmodulairededegre´2deGal(Q/Q),DukeMath.J.54(1987),no. 1,179–230. [17] C.C.Sims,Computationalmethodsforpermutationgroups,in: Computationalproblemsinabstract algebra(J.Leech,ed.),Pergamon,Elmsforth,N.Y.,1970,169–184. [18] W.A.Stein,Anintroductiontocomputingmodularformsusingmodularsymbols,eprint,download- ableathttp://modular.fas.harvard.edu/papers/msri-stein-ant/ [19] J.Sturm,OntheCongruenceofModularForms,LectureNotesinMathematics1240(1987),275– 280. [20] M.Suzuki,GroupTheoryI,GrundlehrendermathematischenWissenschaften247,Springer-Verlag, NewYork,1982. 7

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.