On ``A quadratic integer program for the location of interacting hub facilities'' (Q915642)

Recently, a quadratic integer programming model for a general hub location problem in discrete solution space has been proposed by \textit{M. E. O'Kelly} [ibid. 32, 393-404 (1987; Zbl 0627.90030)]. In this note, a simplified formulation of the problem with a quadratic objective function and a potentially useful allocation rule for this problem are presented. We analyze the proximity based heuristics proposed by O'Kelly and show that they only provide upper bounds. The results presented show that the efficient heuristics for assigning demand points to hubs should take the interactions between the hubs together with the transportation costs and the distances into consideration. Our findings further clarify the differences between the location-allocation problem and the hub location problem, and their solutions.