Integrated strategic and operational planning of dry port container networks in a stochastic environment

Abstract

This paper proposes a dry port network design model that integrates strategic and operational decision taking into account the stochastic nature of demand. The proposed model represents the forward and backward flows of both laden and empty containers as well as the inventory levels of empty containers throughout the network. This problem is formulated as a two-stage stochastic programming model where the dynamic decisions relating to containers’ transportation and inventory are integrated with decisions that determine the number and location-allocation of dry ports in an uncertain environment. For solving this problem, a robust sample average approximation method enhanced with Benders decomposition algorithm which is accelerated by multi-cut framework, knapsack inequalities, and Pareto-optimal cut scheme is proposed. The model was tested using a hypothetical case relating to the design of a hinterland container shipping network in North Carolina, USA, that comprises a seaport and fifty manufacturing companies. The quality of solutions obtained was evaluated through extensive numerical experiments. The experimental results underline the sensitivity of network configuration and operational decisions to inventory holding costs and solution robustness. We also provide managerial insights that may lead to improvements in the network configuration, modality choice, service level, fill rate, and inventory turnover in container shipping industry.