Service fairness and value of customer information for the stochastic container relocation problem under flexible service policy

Abstract

This paper considers the optimization of service efficiency and service fairness in a Stochastic Container Relocation Problem (SCRP) under flexible service policies at a container terminal. Under the flexible service policy, the external trucks arriving within the same time window can be served out of sequence, which may raise a concern of service fairness. We incorporate the concept of service fairness into the SCRP in two phases. In phase 1, we propose a multiple sub-time windows-based flexible service policy, under which each time window will be divided into multiple sub-time windows and the flexible service policy is only applied to each individual sub-time window. In phase 2, the SCRP is formulated as a dynamic programming model with two lexicographically ordered objectives representing both relocation efficiency and service fairness, which is solved via a hierarchical iterative approach. In addition, we investigate whether the information of trucks’ arrival probability over the time window (which represents the customer preference information) would add value to the terminal operators. Extensive computational experiments are conducted to evaluate the impacts of the number of sub-time windows and to examine the impacts and value of customer preference information in various scenarios. The results show interesting trade-offs between efficiency and fairness. As the number of sub-time windows increases, the service fairness is generally improving (but not guaranteed) while the expected number of relocations is increasing. It is found that the customer preference information can be valuable in some circumstances, especially when each truck indicates a certain arrival sub-time window.