On the density of coprime \(m\)-tuples over holomorphy rings (Q2800161)

Let \(F\) an algebraic function field of genus \(g\) with full constant field \(\mathbb F_q\), \(\mathcal C\) the set of places of \(F\), and \(\mathcal S\) a nonempty proper subset of \(\mathcal S\). The holomorphy ring of \(\mathcal S\) is \(H=\bigcap_{P\in\mathcal S}{\mathcal O}_P\), where \({\mathcal O}_P\) is a valuation ring of \(P\). An \(m\)-tuple of elements \(f_1,\dots,f_m\) of \(H\) is \textit{coprime} if they generate the unit ideal in \(H\), i.e. \(\langle f_1,\dots,f_m\rangle=H\) (in analogy with \(\mathbb Z\)). For a fixed positive integer \(m\), define \(U=\{(f_1,\dots,f_m)\in H^m : \langle f_1,\dots,f_m\rangle=H\}\). The authors define a notion of \textit{density} for subsets of \(H^m\), and the main result concerns the density of \(U\). Let \(\mathcal D\) denote the set of positive divisors supported away from \(\mathcal S\) and for any \(D\in\mathcal D\), let \(\mathcal L(D)\) denote the Riemann-Roch space associated to \(D\). Then \(H=\bigcup_{D\in\mathcal D}\mathcal L(D)\). The \textit{superior density} of \(L\subset H^m\) is \(\overline{\mathbb D}(L)=\limsup_{D\in\mathcal D}\frac{|L\cap{\mathcal L(D)}^m|}{|{\mathcal L(D)}^m|}\), which is defined using Moore-Smith convergence of nets; the \textit{inferior density} \(\underline{\mathbb D}(L)\) is defined analogously. When \(\overline{\mathbb D}(L)=\underline{\mathbb D}(L)\), this value is the \textit{density} of \(L\), \(\mathbb D(L)\). The main result is Theorem 2.1: The density of the set of coprime tuples of length \(m\geq 2\) of the holomorphy ring \(H\) is \(\frac{1}{Z_H(q^{-m})}\), where \(Z_H(T)=\prod_{P\in\mathcal S}(1-T^{\deg(P)})^{-1}\) (with \(0<T<q^{-1}\)) is the zeta function of \(H\). The proof uses the Riemann-Roch Theorem and the absolute convergence of \(Z_F\), the zeta function of \(F\). Theorem 2.1 is the function field version of the classical result for \(\mathbb Z\), where the density is \(\frac{1}{\zeta(m)}\), \(\zeta\) being the Riemann zeta function. In Theorem 2.1, both \(\mathcal S\) and \(\mathcal S\backslash\mathcal C\) could be finite, and \(Z_H\) may be hard to compute. However, if \(\mathcal S\) is finite then \(Z_H\) is a finite product, and if \(\mathcal S\backslash\mathcal C\) is finite then Corollary 3.1 shows that computing the density reduces to computing the \(L\)-polynomial of \(F\). In Section 3.1, Theorem 2.1 is applied to the affine coordinate ring \(A(E)\) of an elliptic curve \(E\) over \(\mathbb F_q\) of characteristic \(\neq 2,3\). The case \(F=\mathbb F_q(x)\), \(H=\bigcap_{P\neq P_{\infty}}{\mathcal O}_P=\mathbb F_q[x]\) was studied for \(m=2\) in [\textit{H. Sugita} and \textit{S. Takanobu}, Adv. Stud. Pure Math. 49, 455--478 (2007; Zbl 1154.60006)] and for general \(m\) in [\textit{X. Guo} and \textit{G. Yang}, Linear Algebra Appl. 438, No. 6, 2675--2682 (2013; Zbl 1263.15033)]. In Section 3.2. the authors interpret this work within their general framework.