Tree based searching approaches for integrated vehicle dispatching and container allocation in a transshipment hub.

Vehicle dispatching is integrated with yard crane scheduling and storage selection.Both loading and unloading activities are considered.Three tree-based adaptive searching approaches are proposed.Our methods can support real-time planning. This study addresses the integration of Vehicle dispatching and container storage location problem with consideration of loading and unloading activities simultaneously. A MIP model is formulated to describe the interrelation between vehicle scheduling, yard crane scheduling and container storage location. A tree structure is used to represent the whole solution space. This representation has a good property as it captures the neighborhood structure and enhances the performance of local search and adaptive searching algorithms. Three variants of tree based searching approaches are developed, namely, the Nested Partitions method (NP), the Beam Search method (BS), and Stochastic Beam Search method (SBS). Extensive experiments show that these proposed methods can find a promising solution in matter of seconds for a practical problem and the Stochastic Beam Search method (SBS) method performs nearly as well as Nested Partitions method (NP) while gaining great computational efficiency. Due to this merit, SBS method is suggested to solve real time integrated vehicle dispatching problem in a relative large scale and may applied in other real time complex system scheduling.