On a modification of a problem of Bialostocki, Erdős, and Lefmann (Q819814)

For positive integers \(m\) and \(r\), let \(g(m,r)\) denote the least positive integer \(N\) such that for any \(r\)-colouring of \(\{1,2,\dots,N\}\) there exist \(x_1<\dots<x_m<y_1<\dots<y_m\) such that the \(x_i\) all have the same colour, the \(y_i\) all have the same colour, and \(x_m\leq\frac{1}{2}(x_1+y_m)\). The exact values of \(g(m,r)\) are determined for \(2\leq r\leq 4\), for certain infinite families of \(r\) related to the Fibonacci numbers, and estimates are given for \(r\leq 10\) and sufficiently large \(m\). The problem of Bialostocki et al.\ referred to in the title is when the condition \(x_m\leq\frac{1}{2}(x_1+y_m)\) is strengthened to \(x_m-x_1\leq y_m-y_1\) [\textit{A. Bialostocki, P. Erdős} and \textit{H. Lefmann}, Discrete Math. 137, 19--34 (1995; Zbl 0822.05046)].