scientific article; zbMATH DE number 1303192

This paper has its genesis in the study of the set \({\mathcal T}=\{\text{trace}(\alpha)/d\}\), where \(\alpha\) is a totally positive algebraic integer of degree \(d\). It is not difficult to show that \(2\) is a limit point of \(\mathcal T\) and it is an open question whether \(2\) is the smallest limit point of \(\mathcal T\). In an earlier paper [Ann. Inst. Fourier 34, No. 3, 1-28 (1984; Zbl 0543.12001)], the author proved that with the exception of five such \(\alpha\) of degrees \(1,2,3,4\) and \(4\), \(\text{trace}(\alpha) \geq 1.7719\). The proof relied on the construction of constants \(\lambda_j>0\) and polynomials \(P_j(x)\) for which \(x - \sum_{j=1}^J\lambda_j \log|P_j(x)|\geq 1.7719\), for all \(x > 0\). In this paper, the author shows that, for any choice of \(\lambda_j\) and \(P_j(x)\), one has \(\min_{x\geq 0}\bigl(x-\sum_{j=1}^J \lambda_j \log|P_j(x)|\bigr)\leq 2-10^{-41}\). Thus it cannot be shown by this method that \(2\) is the smallest limit point of \(\mathcal T\) and this suggests that the smallest limit point is smaller than \(2\). In a remark added in proof, the author notes that J.-P.~Serre has been able to improve \(2-10^{-41}\) to \(1.8983021\) in the above inequality. For the entire collection see [Zbl 0910.00037].

Recommendations

Cited in

(8)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4249715)