Spelling suggestions: "subject:"groupgroup"" "subject:"progroup""
1 |
The Families with Period 1 of 2-groups of Coclass 3Smith, Duncan Alexander, Mathematics, UNSW January 2000 (has links)
The 2-groups of coclass 1 are widely known and James (in 1975) looked at the 2-groups of coclass 2. Development of the p-group generation algorithm implemented by O'Brien at ANU enabled group presentations to be provided for the 2-groups of coclass 3 by Newman and O'Brien for groups of order up to 223. Newman and O'Brien (in 1999) conjectured the number of descendants of 2n for all n. They introduced the concept of a family, with each family related to a different pro-p-group and the concept of a sporadic p-group, a p-group external to any family. They found 1782 sporadic 2-groups with order at most 214. The 70 families of 2-groups of coclass 3 can be further split according to their period, a measure of the repetitive structure of the families. Newman and O'Brien conjectured that these families had periods of 1, 2 or 4. This thesis examines the 2-groups of coclass 3 contained in families with period 1 and shows that the number of descendants conjectured by Newman and O'Brien is correct. Furthermore the presentation of all groups contained in period 1 families is provided and shown to be correct.
|
2 |
Sur quelques aspects des extensions à ramification restreinte / On some aspects of extensions with restricted ramificationRougnant, Marine 16 April 2018 (has links)
Soit p un nombre premier, soit K/k une extension galoisienne finie de corps de nombres de degré premier à p et soit S un ensemble fini de premiers de k. Le groupe de Galois G(K,S) de la pro-p extension maximale de K non ramifiée en dehors de S est l'objet central de ce mémoire.On se place dans un premier temps dans le cas modéré : on suppose que S ne contient pas les places divisant p. Les travaux combinés de Labute, Minac et Schmidt sur les pro-p groupes mild ont permis d'exhiber les premiers exemples de groupes G(K,S) de dimension cohomologique 2. En implémentant un corollaire de leur critère dans le logiciel PARI/GP, on observe un phénomène de propagation : si k=Q et si le groupe G(Q,S) est mild, un fort pourcentage des groupes G(K,S) l'est également, pour K quadratique imaginaire. En associant au groupe G(K,S) deux graphes orientés dont les arcs sont définis par la ramification dans des extensions p-élémentaires, on démontre un critère théorique pour que ce phénomène de propagation ait lieu.On considère ensuite le cas sauvage : toutes les places au-dessus de p sont contenues dans S. Le groupe de Galois Δ:=Gal(K/k) agit sur G(K,S) ; on note G le plus grand quotient de G(K,S) sur lequel Δ agit trivialement et H le sous-groupe fermé de G(K,S) correspondant. Maire a étudié la liberté du Zp[[G]]-module H^{ab}. Nous poussons plus loin ses résultats en considérant les φ-composantes de H^{ab} sous l'action de Δ. Sous de bonnes hypothèses et sous la conjecture de Leopoldt, on démontre une condition nécessaire et suffisante pour que les φ-composantes soient libres ou non. La théorie du corps de classes permet de ramener cette condition à l'étude du régulateur normalisé, et donc à la p-rationalité du corps K. Les expérimentations faites sur PARI/GP dans des familles d'extensions cubiques cycliques, diédrales et cycliques de degré 4 du corps des rationnels corroborent une conjecture de Gras selon laquelle tout corps de nombres est p-rationnel pour p suffisant grand. / Let p be a prime number, let K/k be a Galois extension of number fields and let S be a finite set of primes of K. We suppose that the degree of K/k is finite and coprime to p. We denote by G(K,S) the Galois group of the pro-p maximal extension of K unramified outside S. We focus on this thesis on two differents aspects of this pro-p group.We are first interested in the tame case : we suppose that S does not contain any place above p. The works of Labute, Minac and Schmidt about mild pro-p groups brought the first examples of groups G(K,S) of cohomological dimension two. Using a corollary of their criterium, we compute some examples with PARI/GP and we observe a propagation phenomenum : if we take K=Q and if we suppose that G(Q,S) is mild, a large part of the pro-p groups G(K,S) with K imaginary quadratic are mild too. We then associate two oriented graphs to G(K,S) and we show a theoretical criterium proving mildness of some imaginary quadratic fields.We then consider the wild case where all the places dividing p belong to S. The Galois group Δ:=Gal(K/k) acts on G(K,S). The action of Δ is trivial on some quotients of G(K,S) ; we denote by G the maximal one and by H the corresponding closed subgroup of G(K,S). Maire has studied the Zp[[G]]-freeness of the module H^{ab}. We extend his results considering the φ-component of H^{ab} under the action of Δ. In a favourable context and under Leopoldt's conjecture, we show a necessary and sufficient condition for the freeness of the φ-components. This condition is connected to p-rational fields by class field theory. We present experiments with PARI/GP in some families of cubic cyclic, dihedral and quartic cyclic extensions of Q which support the following conjecture from Gras : every number field is p-rational for sufficiently large p.
|
Page generated in 0.0364 seconds