A partition problem (Q1335253): Difference between revisions

Let \(M(n,k)= \{n, n+1,\dots, n+k\}\) and say that \(M(n,k)\) has an \(l\)- partition if there exists a subset \(A\) of \(M(n,k)\) such that \(l\sum_{a\in A} a= \sum_{m\in M(n,k)} m\). The problem of finding all \(n\), \(k\) for which an \(l\)-partition exists was posed in the case \(l=2\) by Z. Rubinstein, and is now fully solved in that case: the conditions for existence are (i) \(k\equiv 3\pmod 4\) or (ii) \(k\) even, \(n\equiv {1\over 2}k \pmod 2\), \(4n\leq k^ 2\). For \(l\geq 2\), let \(K(l)\) be the smallest value of \(k\) for which an \(l\)- partition exists for some \(M(n,k)\). Then a (complicated) explicit formula is obtained for \(K(l)\) when \(l\) is a prime power. In general, \(K(l)\) is related to the function \(g^* (a,b)\), defined as the smallest positive integer \(g\) such that \(hb- ga= 2n-1\), \(2n\leq gh\), hold for some \(n\geq 1\), \(h\geq 1\); this deals fully with the case where \(l\) is the product of two odd primes.