The transportation costs of low value per unit volume materials can become a major portion of the total distribution cost. A stochastic dynamic mixed-integer linear program of a transportation model containing intermediate processing stations between suppliers and demand centers has been developed which has been found to reduce total transportation costs by 39 percent over standard models. The model selects the optimal locations for the intermediate processing stations and the primary demand center from a given set of possible candidates. It includes costs of environmental constraints and includes the stochastic (e.g., seasonal) nature of the supply