Modeling Study For Port Container Operation Planning

Since the container traffic appeared in the middle 1950s in the United States, as this mode of transport can improve the loading and unloading efficiency of break bulk cargo effectively, save wrapper, increase the security of cargo during transportation, reduce the loss and damage and be convenient for multimodal transport and so on, in just sixty years time, container traffic developed quickly. However it is also a capital-intensive industry, ship turnaround times at the terminal directly affect economic benefit of customers even the whole port. The time spent in ports largely depends on laytime, and quayside container crane is a scarce and expensive container handling machinery, its handling efficiency directly decides the profit of port. Therefore,how to make a reasonable handling order to improve the utilization ratio of quay container crane, minimize ship turnaround times, improve customer's satisfaction etc has become the important embodiment of the competitiveness between ports.Container handling sequence problem is to dispatch all the quay cranes to containers in a series of constraint conditions and to determine the handling order of the containers on cranes, achieve the purpose of minimizing the maximum handling time of cranes. The problem is very similar to the flexible job-shop scheduling problem, but due to some special factors such as location constraint of quay cranes, so it is more complex than FJSP. This paper presents a mathematical model for container handling sequence problem based on the study about scheduling problem and container handling.Container handling sequence problem focus on optimizing machine assignment and scheduling. It is a typical combinatorial optimization problem, and has been proved to be NP-Hard, the space and time complexity of solving is rather high. Combinatorial optimization algorithm mainly contains exact method and approximation method. The paper finally uses Ant colony Optimization (ACO) to solve the problem after analyzing and comparing kinds of combinatorial optimization algorithms. ACO is a new Evolutionary Algorithm, due to its feedback which is positive or negative and parallelism, it's superior to some other algorithm while solving combinatorial optimization problem. However, its development is far from mature, it also has some shortcomings such as slow convergence and premature convergence etc. To overcome the default of basic ACO, an improved ACO is introduced and discussed concretely in this paper to solve Container handling sequence problem.Finally, implement the improved ACO with C# language on computer, and confirm it by examples, taking dispatching 2 quay cranes to 5 bays for example, operation time of one quay crane is 202, the other is 304 in a traditional way in bay to schedule, however handling time of both 2 quay cranes is 256 scheduled by improved ACO, the result shows that improved ACO is much better than traditional way, to achieve the purpose of the study.