Research On Simulation Based Optimization Approaches For Logistic Systems In Container Port

Due to the features of high transport quality, high transport efficiency and ease of multimodal transport, container transport has become the world's most advanced transportation organization. In recent years, as China's economic development and export trade's expandsion, as well as the widely application of container in transport, the container throughput in our coastal port is in substantial growth. In this case, it is indispensable for container terminals to enhance their operation efficiency with more rational and effective planning. The research on container terminals has attracted extensive attention of scholars.Current researches on container terminals often use mathetical programming approach. However, the container terminal operating system is a typically multi-link complex dynamic random system. It is difficult to describe the whole system by a simple mathematical model. Some scholars have used simulation methods in the research on container terminals. But the simulation model is just a testing and confirming platform, it is unable to provide the assistant decision-making function. Consequently, in this paper, Simulation Based Optimization (SBO) is presented to optimize the container terminal logistic systems.Compared with the traditional mathematical programming method, SBO can describe complex systems more accurately. However, there is also difficulties in its application. The main reason which leads to the restrictions on application and development of the simulation-based optimization is that the computational cost is too high.This paper focuses on the optimization problems of container terminals logistic systems. According to the steps of establishment of simulation modal, authentication of simulation modal and applications of simulation based optimization, problems on resource planning in planning level and scheduling optimization in operational level in container terminals have been studied. Specific researches include the following:(1)Research situations about simulation based optimization and container terminal logistic system are summarized. History and classification of SBO, applications and difficulties of SBO and future development of SBO are introduced. The applications and achievement of simulation method, optimization method and SBO method in container terminals are also introduced.(2) Applying object-oriented discrete event system simulation method, a simulation model is constructed to describe the whole operation system of container terminals, in which container ships, anchorages, berths, quay cranes, gantry cranes, internal container trucks, external container trucks and yard gates are all included. In the simulation modal, a number of random factors are taken into account.(3) In planning level, propose that the number of container trucks should be configured dynamically according to the volume of handling tasks in different periods. In order to solve the dynamic container truck configuration problem of container terminals, the simulation-based optimization (SBO) method is proposed. Experimental results show that the result of simulation-based optimization reduces the average waiting time of internal trucks and increases the efficiency of the internal trucks greatly.(4)Considering the excessive cost of simulation based optimization, The MPI-based parallel computing technology is introduced to the solving process. The tedious computation for SBO is well improved by the design of parallel simulation based genetic algorithm.(5) In planning level, block planning problems of container terminals under disperision stockpiling strategy are studied. The research background and problem description of block planning are introduced. The simulation based optimization method is used again. Both the parallel computing technology and case searching technology are introduced into the sloving process to improve the tedious computation of SBO. Experiment results verify the effectiveness of the two methods.(6) In operational level, quay crane scheduling problems in container terminals are studied. The problem which takes a number of random factors into account is also solved by simulation based optimization. Taking GA, PSO ,SA as its optimizer respectively, comparison of their computing performance is analysed.(7) In operational level, the pre-marshalling process to achieve high loading efficiency in container terminals is studied. Propose an integer programming mathematical model to develop a movement plan to improve the layout of containers in a bay. Design a multi-stage genetic alrorithm with binary encoding. Take the define of confuse coefficient as the object value. Introduce a heuristic operator for the evolution and another heuristic rule for the solution improvement. Experiment results show that the algorithm has better steps than that of previous research achievement. It is an effective scheduling method for pre-marshalling problem.