Triples of primes in arithmetic progressions

Let \(N\) be large and \(t\geq2\). The author proves that there are disjoint sets of primes \({\mathcal A},{\mathcal B}\subset[1,N]\), with \(| {\mathcal B}| =t\) and \(| {\mathcal A}| \geq0{\cdot}33^tN/(\log N)^{t+1}\), such that \((a+b)/2\) is prime for all \(a\in{\mathcal{A}}, b\in{\mathcal{B}}\). The proof is based on an asymptotic formula of \textit{E. Grosswald} [J. Number Theory 14, 9--31 (1982; Zbl 0478.10034)] for the number of triples of primes in arithmetic progression. The problem is then reduced to a counting argument, which in turn is based on a result due to \textit{T.~Kövári, V.~T.~Sós} and \textit{P.~Turán} [Colloq. Math. 3, 50--57 (1954; Zbl 0055.00704)] in extremal graph theory. It is possible to adapt the counting argument to deal with \(k\)-tuples of primes in arithmetic progression, provided that the initial corresponding asymptotic formula can be established. It may perhaps be appropriate to mention here that it has just been announced (by \textit{B.~Green} and \textit{T.~Tao}, see http://arxiv.org/abs/math/0404188) that the primes contain arithmetic progressions with arbitrary length, although the said asymptotic formula for the number of \(k\)-tuples is unlikely to be established in the near future.

Recommendations

Cited in

(4)

This page was built for publication: Triples of primes in arithmetic progressions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4807788)