A shipper is a company which outsource its transportation to carriers to deliver its customers. Typically, in the retail industry, suppliers can ship several times a week small quantities (a few pallets) to multiple retailer warehouses and superstores. In order to reduce its cost and secure its contracts with its carriers, a shipper typically resorts to integrating a mix of carriers and transportation rates (full-truckload and less-than-truckload), together with consolidation on regional hubs and collaboration with other shippers. This leads to the definition of two-echelon distribution networks. The planning of such distribution networks includes defining efficient load plans for long-haul routes, freight consolidations on regional hubs and distribution routes from hubs to customers [3]. The overall objective is mainly to take decisions related to the booking of resources such as trucks or cross-dock capacities and organize a good collaboration with other shippers. In this context, we focus on several constraints that are seldom integrated in the operations research literature : (i) customers require some regularity : a commodity has to be delivered several times a week [1] on predefined days, (ii) shippers, hubs and customers have capacity constraints that model workload balancing requirements over the days of the week. Accordingly, given a set of suppliers, hubs and customers, given an order history of several weeks, we investigate the problem of assigning a set of delivery weekdays (called delivery patterns [2]) to each customer and for each commodity, in a two-echelon distribution network, such that the sum of transportation costs are minimized and the capacity of each network node is respected.