This paper presents a research study which aims at determining optimal locations of regional distribution centers in a collaborative distribution network. We consider a multi-layered distribution system between a cluster of suppliers from a given region and several thousands customers spread over the whole country. The optimization problem consists of finding the locations of intermediate logistics facilities called regional distribution centers and assigning customers to these facilities according to one year of historical data. The distribution system combines full truckload (FTL) routes and less-than-truckload (LTL) shipments. The use of several rates for transportation, as well as the high impact of seasonality implies that, for each shipping date, the number of FTL routes and the cost of LTL shipments should be precisely evaluated. This problem is modeled as a mixed integer linear problem and used as a decision aiding tool on a real case study related to the distribution of horticultural products in France.