In this paper, we present an efficient approach for solving capacitated sin
gle allocation hub location problems. We use a modified version of a previo
us mixed integer linear programming formulation developed by us for p-hub m
edian problems. This formulation requires fewer variables and constraints t
han those traditionally used in the literature. We develop good heuristic a
lgorithms for its solution based on simulated annealing (SA) and random des
cent (RDH). We use the upper bound to develop an LP-based branch and bound
solution method. The problem, as we define it, finds applications in the de
sign of postal delivery networks, particularly in the location of capacitat
ed mail sorting and distribution centres. We test our algorithms on data ob
tained from this application. To the best of our knowledge, this problem ha
s not been solved in the literature. Computational results are presented in
dicating the usefulness of our approach.