On the maximal length of two sequences of consecutive integers with the same prime divisors (Q1264179): Difference between revisions

For each integer \(n\geq 2\) let Supp(n) denote the set of prime factors of n. Suppose that for \(1\leq i\leq k\) we have \[ (1)\quad Supp(x+i)=Supp(y+i) \] where x,y,k are positive integers. A problem of P. Erdős and A. Woods is the following: Does there exist an integer \(k\geq 2\) such that (1) implies \(x=y?\) The authors prove that (1) implies that \[ (2)\quad \log k\leq C, (\log x\quad \log \log x)^{1/2}\quad for\quad x\geq 3 \] \[ (3)\quad y-x>\exp (C_ 2 k (\log k)^ 2 (\log \log k)^{-1})\quad for\quad k\geq 3 \] and \[ (4)\quad y-x>(k\quad \log \log y)^ D\quad for\quad y\geq 27 \] where \(D=C_ 3 k (\log \log y)\) (log log log y)\({}^{-1}\). Here \(C_ 1,C_ 2,C_ 3\) are effectively computable absolute positive constants. A powerful ingredient of the proof is an inequality of A. Baker dealing with the linear forms in the logarithms of rational numbers.