Research On Modeling And Optimization Of Reshuffle Problem In Container Yard

In recent years, with the increasingly mature container terminals and container throughput rapid growth, the efficiency of yard become an important bottleneck factor which restrict its development. There are a large number of reshuffle in container yard because of uncertainties such as weather variations. Also, the widely used in rail gantry and yard stacking increase aggravate the reshuffle. In conclusion, the study of effectively reducing reshuffle have become increasingly urgent.First of all, based on the allocation of the container slots in yard, which is affected by container delivery appointment system, we proposed a two-dimensional graph representation of delivery sequence and slots allocation, and then suggested a solution of expected value of reshuffles and gantry crane traveling distances. An optimization model of export container slots in container yards is established based on graphs as to minimize the crane traveling distance and the reshuffles. A tabu-search algorithm was proposed under the consideration of the characteristics of graph representation to narrow down the searching space and to optimize searching direction, thus its convergence rate can be accelerated. The results show that a satisfied solution can be gained in a reasonable time by the algorithm. Besides, in contrast with the determinative model, this model can save more than 20% operation cost.The next, considered the fact that operations tend to interference each other in container yards of mixture storage mode, we proposed an export operations sequence based method for recognizing disruption and measuring interference. An export container slots optimization disruption management model was built with respect to the related disruption influence, where the slots offset objective is for reducing the inference, the reshuffles and gantry crane traveling distance objective for saving the cost of the operation recovering. A tabu-searching based disruption recovery algorithm was developed with the local searching characteristic of model solution, where the local searching spaces are improved with searching directions and ranges to accelerate its convergence. Experiment results show that the proposed method can achieve 10% improvement than the rescheduling.Lastly, considered the problem of retrieving inbound containers in terminals, a three-phase heuristic method is proposed. The first phase generates an initial feasible movement sequence of containers. The second phase generates the virtual movement path and establishes the set of movement path conflict, which make use of virtual reshuffles. A binary integer programming model is proposed in the third phase to gain feasible movement sequences of the virtual movement paths, and then an evolutionary algorithm is developed. Numerical experiments show this heuristic can improve more 20% than the heuristic OH.