Clustering and hub selection for a large scale delivery problem

In this paper, we propose a framework to decompose a large scale delivery problem into a number of smaller sub-problems to improve delivery efficiency and reduce cost. The delivery points were first classified into sectors according to Singapore Zones and Districts. Then, among a number of pre-selected candidate hubs, the decomposition problem considers how to determine the appropriate hubs among the candidates, and for each selected hub, how to determine the sectors that will get supply from the hub. The problem was formulated as a Binary Integer Programing (BIP) model and solved by IBM ILOG CPLEX Optimization Studio (ILOG OPL). Computational results show that, the prosed framework can obtain reasonably good results for clustering of the delivery points.