Additive functions for number systems in function fields (Q973968)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Additive functions for number systems in function fields
scientific article

    Statements

    Additive functions for number systems in function fields (English)
    0 references
    26 May 2010
    0 references
    Let \(p\) in \({\mathbb F}_q[X,Y]\) be a separable irreducible polynomial and let \({\mathcal S}={\mathbb F}_q[X,Y]/p{\mathbb F}_q[X,Y]\). A pair \((B,{\mathcal N})\) with \(B\) in \({\mathcal S}\) and \({\mathcal N}\subset {\mathbb F}_q[X]\) is a number system in \({\mathcal S}\) if every \(G\) in \({\mathcal S}\) has a unique finite representation \(G=\sum_{k=0}^{l-1} D_kB^k\) with \(D_k\) in \({\mathcal N}\) and \(D_{l-1}\neq 0\) if \(G\neq 0\). This is the \(B\)-adic representation of \(G\) of length \(L_B(G)=l\). Let \({\mathcal L}_B(m)\) denote the set of all \(G\) in \({\mathcal S}\) with \(B\)-adic length less than \(m\). A function \(f\) defined on \({\mathcal S}\) is \(B\)-additive, or strictly \(B\)-additive, if \(f(G)=\sum_{k=0}^{l-1} f(D_kB^k)\) or \(\sum_{k=0}^{l-1} f(D_k)\) respectively. Now consider \({\mathbb L}={\mathbb F}_q(X,Y)/p{\mathbb F}_q(X,Y)\) and suppose this field extension has degree \(n\) and \({\mathcal S}\) is its ring of integers. For a rational function \(\alpha=A/B\) define the valuation at infinity by \(\nu(\alpha)=\deg B - \deg A\) and let \(\omega\) be the extension of \(\nu\) to the completion \({\mathbb L}_\infty\) of \({\mathbb L}\). Set \(d(\alpha)=-\omega(\alpha)\) to extend the degree to \({\mathbb L}_\infty\). For \({\mathcal T}\subset {\mathbb L}\), set \({\mathcal T}(m)=\{\alpha\;\text{in}\;{\mathcal T}: d(\alpha)\leq m\}\). The authors investigate properties of additive functions which extend the classical results for ordinary number systems to this setting. If the \(f_i\) are \(B_i\)-additive functions on \({\mathcal S}\), the \(B_i\) are coprime and the \(M_i\) are ideals in \({\mathcal S}\), then the function values \(f_i(A)\) with \(A\) in \({\mathcal S}(n)\) are equally distributed amongst their residue classes modulo \(M_i\) and the asymptotic distribution as \(n\rightarrow \infty\) can be described by a central limit theorem. There is a version of Waring's problem in this setting. Every \(N\) in \({\mathcal S}\) with \(d(N)\) sufficiently large can be represented as a sum of \(k\)-th powers, \(N=P_1^k+\cdots+P_s^k\) with \(P_j\) in \({\mathcal S}(\lceil d(N)/k\rceil)\) and \(f_i(P_j)\equiv J_i\bmod M_i\) under a certain assumption relating to a Weyl sum which measures the digital restrictions of the system. The number of representations of \(N\) satisfies a Hardy-Littlewood type asymptotic formula. The proofs rely on properties and estimates of these Weyl sums.
    0 references
    0 references
    0 references
    0 references
    0 references
    additive functions
    0 references
    polynomial number systems
    0 references
    distribution theorems
    0 references
    Waring's problem
    0 references
    0 references