Motivated by the practical import free-flow program that aims to expedite the container retrieval process, we conceptualize a new container stacking strategy, termed Smart Stacking (SS) strategy. The SS strategy aims to create relocation-free stacks (smart stacks) by utilizing customer information. The Storage Location Assignment Problem (SLAP) under the SS strategy is addressed. The problem is to determine the smart customers/containers, and the number and locations of smart stacks, when assigning a batch of import containers to a yard block at an automated container terminal to minimize the total retrieval time. Two variants of SLAP are investigated under the non-split policy and the split policy, depending on whether the containers from the same customer are allowed to be split between smart stacks and non-smart stacks. For the non-split variant, a mixed-integer programming (MIP) model is formulated first. By analyzing the properties of the optimal solution, an improved formulation with enhanced computational performance is then proposed. Based on the structure of the model, a divide-and-conquer heuristic is designed to solve the non-split variant more efficiently. For the split variant, a MIP model under the optimal partitions of the non-split model is developed. We theoretically prove that the split variant yields better results than the non-split variant. Extensive experiments are carried out to illustrate the effectiveness of smart stacking. It is found that customer information and yard utilization rate have a significant influence on the effectiveness of smart stacking.
Read full abstract