On the Computation of the Norm-Euclidean Minimum Of

On the Computation of the Norm-Euclidean Minimum Of

International Journal of Algebra, Vol. 3, 2009, no. 7, 341 - 354 On the Computation of the Norm-Euclidean Minimum of Algebraic Number Fields Michele Elia Dipartimento di Elettronica Politecnico di Torino I-10129 Torino, Italy [email protected] J. Carmelo Interlando Department of Mathematics and Statistics San Diego State University, San Diego, CA, USA [email protected] Abstract Let F be an algebraic number field whose group of units has rank ≥ 1. The conjecture that the norm-Euclidean minimum M(F) is a rational number is affirmatively settled. It is proved that M(F) is lower bounded by the inverse of the smallest norm of all nonzero prime OF-ideals. Furthermore, when F/Q is a normal extension, the numer- ator and denominator of M(F) lie within finite sets of integers that can be explicitly calculated. As an application, it is proved that the known lower bounds of M(F) for 2 2 2 the cyclotomic field Q(ζ5) and the cyclic cubic fields of discriminants 103 , 109 , 117 , and 1572 are the actual values of M(F). Mathematics Subject Classification: 11R04, 13F07, 13A05, 11R27, 11Y40 Keywords: algebraic number fields, Euclidean rings, units, group actions I Introduction The problem of establishing whether an algebraic number field is a Euclidean domain has always attracted a lot of interest [6]. The Euclidean character of fields has been considered with respect to many different norms, with prevalence of the absolute value of the field norm. In this paper, F denotes an algebraic number field defined by a monic irreducible f x n Z x O F polynomial ( ) of degree in [ ]. As customary, denotes the ring of integers of . F O |N . | We say that is norm-Euclidean if its ring of integers is a Euclidean domain for ( ) , 342 M. Elia and J. Carmelo Interlando Q α, β ∈ O the absolute value of the field norm relative to . That is, given , there exist θ, ν ∈ O α θβ ν |N ν | < |N β | such that = + with ( ) ( ) . Most of the proofs concerning Euclidean domains are based on the following proposition, which is a consequence of the results contained in [5, Section 14.7, p. 212]: Proposition 1. The number field F is norm-Euclidean if and only if for every α ∈ F, β ∈ O |N α − β | < there exists such that ( ) 1. The norm-Euclidean minimum of F, denoted by M(F), is defined in [6] as M F |N ξ − ν | . ( ) = sup inf ( ) (1) ν∈Ç ξ∈ Note that F is norm-Euclidean or not according to whether M(F) < 1orM(F) > 1, respectively. If M(F) = 1, then both possibilities can occur, see [6]. In [12], van der Linden proved that M(F) is achievable if F has unit rank 1, that is, ξ F M F |N ξ0 | M F ∈ Q an element 0 in exists such that ( )= ( ) , which implies that ( ) .In [6], the Euclidean properties of number fields are surveyed and some surmises [6, p. 4] are put forth. In particular, it is conjectured that if F has unit rank ≥ 1 then: i) M(F) is achievable in F; a ii) M(F)= d where a, d ∈ Z. These conjectures are supported, or at least not contradicted, by Lemmermeyer’s extended tables [6] of number fields of given discriminant listed along with known or supposed norm- Euclidean minima. Norm-Euclidean quadratic fields have been completely classified,√ see [6, Section 4] for the succession of events that led to it. More specifically, Q( m) is norm-Euclidean if and only if m is one of the 21 integers: −11, −7, −3, −2, −1, 2, 3, 5, 6, 7, 11, 13, 17, 19, 21, 29, 33, 37, 41, 57, 73. In this paper, conjectures i) and ii) are affirmatively settled by extending an action F F/O Rn of the unit group of on to . That group action was considered by Barnes and Swinnerton-Dyer [6]. As a consequence, it is shown that M(F) is lower bounded by the O inverse of the smallest norm of all nonzero prime -ideals. Furthermore, it is shown that a in numerical evaluations of M(F)= d , the denominator d is taken from a finite set of integers which depends on the fundamental units of F. This set is explicitly defined for any F which is a normal extension of Q and has a group of units of rank greater than zero. Lastly, previously unknown values of M(F) for fields of small degree and discriminant are given in Table 1. Computation of the norm-Euclidean minimum 343 II Preliminaries Given an integral basis {ω1,...,ωn} for F,anyξ ∈ F can be represented either by a vector φ(ξ)inQn or by an n × n matrix D(ξ) with rational entries. The relation n ξ = xiωi,xi ∈ Q, i=1 n T n defines a bijective mapping φ : F → Q where for any ξ ∈ F, φ(ξ)=(x1,x2,...,xn) ∈ Q . φ F Qn φ O Zn F It follows that ( )= and ( )= . A representation of by matrices is defined by n D(ξ)= xiD(ωi),xi ∈ Q, i=1 (i) where D(ωi)=(djh) are integral matrices, see [14, pp. 1-3]. For i =1,...,n, the matrix D(ωi) is defined by the action of ωi on the row vector Ω = (ω1,...,ωn)as n (i) ωiωh = djhωj, 1 ≤ h ≤ n, j=1 that is, ωiΩ=ΩD(ωi). Note that D is a ring monomorphism from F into the ring M(n, Q) of n × n rational matrices. It follows, with some abuse of notation, that D(F) ⊆ M(n, Q) D O ⊆ M n, Z and ( ) ( ). Let σ1,...,σn be the embeddings of F in C, that is, the Q-isomorphisms of F in C. Then the eigenvalues of D(ξ) are ξ and its conjugates, that is, σ1(ξ),...,σn(ξ), since n n n ΩD(ξ)=Ω xiD(ωi)= xiΩD(ωi)= xiωiΩ=ξΩ , i=1 i=1 i=1 and σj(D(ξ)) = D(ξ) because the entries in D(ξ) are rational numbers. Using φ and D, the field product, trace, and norm are computed in F as: φ ξη D η φ ξ D ξ φ η , ξ D ξ , N ξ D ξ , ( )= ( ) ( )= ( ) ( ) Tr ( ) = tr( ( )) and ( ) = det( ( )) (2) respectively. n A complex number z of the form i=1 xiωi where the xi are rational or real numbers and at least one of the xi belongs to R \ Q will be called an irrational point of F. The representation φ(z)ofz is not unique, but for our purposes this fact is irrelevant. The irrational points of F do not belong to F, but they may be viewed as limit points of sequences of elements in F. The union of F and its irrational points will be denoted by F. Note that the representations φ and D can be both extended to F in the obvious way. Let Hn be the hypercube {(x1,...,xn) | 0 ≤ xi < 1} with vertex set Vn = {(x1,...,xn) | xi ∈ { , }} H F φ O Zn 0 1 . Note that n is the fundamental region n of the lattice ( )= . The image T φ F/O Hn n = ( ) is the set of points in with rational coordinates. 344 M. Elia and J. Carmelo Interlando ξ F ξ,O Definition 1. Given any point of , its distance dist( ) from the ring of integers O is defined as ξ,O |N ξ − ν | . dist( ) = inf ( ) ν∈Ç δ ∈ O ξ O δO Similarly, for any , the distance between and the integral -ideal is ξ,δO |N ξ − δν |. defined as dist( ) = infν∈Ç ( ) Although we will not be doing explicit calculations with irrational points, Definition 1 can be extended to irrational points z ∈ F with the assumption of referring to a specific representation φ(z). The distance of an z O irrational point to is then defined as the infimum of the distances over all possible representations of z. Definition 2. A rational or irrational point ξo ∈ F is said to be optimal if ξ , O ξ,O M F . dist( o ) = sup dist( )= ( ) n φ(ξ)∈Ê ς M F ς , O A rational or irrational optimal point o is called a 0-optimal point if ( ) = dist( o )= |N ςo | ( ) . Note that the search for the supreme can be restricted to points ξ ∈ F such that φ ξ ∈ Hn ξ,O O ς ∈ F o ( ) since dist( ) is defined modulo . A 0-optimal point can always |N · | R R+ be found since ( ) is a continuous function from into , the set of nonnegative real numbers, see [6, p. 7]. In other words, some rational or irrational optimal point ξo is always achievable. δ ∈ O Z δ O /δO δ For , let ( )= denote the residue class ring modulo [3, pp. 24-31]. The multiplicative subgroup of the elements prime with δ in Z(δ) is denoted by Z(δ)∗. The Z δ δ |Z δ | |N δ | order of ( ) is equal to the absolute value of the field norm of , that is, ( ) = ( ) . The order of Z(δ)∗ is equal to the Euler totient function ϕ(δ) counting the number of O /δO algebraic integers that are units in the ring , see [2, p. 231, Problem 4] for its ϕ δ properties. The function ( ) can be computed from the prime-power factorization of δ s βai O = i=1 i in . More specifically, s ϕ δ N δ − 1 , ( )= ( ) 1 (3) N βi i=1 ( ) see [2, p. 231, Problem 5] or [8, p. 152]. U F O Let ( ) be the group of units of .

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    14 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us