集装箱码头外部集卡到达时间预测方法研究

发布时间：2018-04-21 13:18

本文选题：集卡到达时间 + GPS　；参考：《大连海事大学》2015年硕士论文

【摘要】：随着港口之间的竞争日益激烈,集装箱码头越来越注重提高港口的服务水平。传统的集中到达方式方便了码头布置装卸作业调度计划,但对货主和集卡车队安排运输造成不便,并引发了集装箱码头闸口排队拥堵和港区及周边交通压力过大等问题。因此,一些码头采取出口箱随机到达的策略,以提高自身竞争力和提供便利服务。在随机到达模式下,集装箱到达港口的时间和到达顺序都是未知状态,出口箱集中到达模式的传统集装箱码头的“泊位计划→配载计划→堆场计划→集装箱入港计划”顺序式计划调度方法将不再适用。由于码头无法确定特定航次出口箱每天的到达量和到达时段,无法事先预留整块箱区集中堆放,只能根据集卡实际到达情况和堆场空闲情况进行分散堆放,这无疑对集装箱码头堆场箱位分配带来巨大的难度。因此,在随机到达模式下,把握外部集卡到达规律及预测到达时间有利于提高集装箱码头集卡调度的运营效率和服务水平。本文分别建立了集卡到达时间静态和动态预测模型。静态预测模型以集卡运行的历史数据为依据,通过v-SVM模型预测集卡的行驶路径和行驶时间；动态预测模型使用v-SVM预测集卡的行驶路径和路段基线行驶时间,以集卡实际运行数据为系统输入,使用Kalman更新预测值,通过迭代得到集卡到达时间动态预测的结果。计算结果表明,静态预测模型对畅通路段的预测结果较好,长距离易拥堵的路段预测误差往往较大：动态预测模型能够根据实际运行情况对预测结果动态更新,在所有路段上的预测精度均较高,因此集卡到达时间动态预测模型具有较高的预测精确度和鲁棒性。
[Abstract]:With the increasingly fierce competition among ports, the container terminal is paying more and more attention to improving the service level of the port. The traditional centralized arrival method facilitates the scheduling of the loading and unloading operations of the wharf, but is inconvenient for the transportation of the cargo owners and the caravan fleet, and causes the congestion of the container head gate and the traffic pressure in the port area and the surrounding area. Therefore, some wharves take the strategy of random arrival of the export boxes to improve their competitiveness and provide convenience services. In the random arrival mode, the time and order of arrival of the container to the port are unknown, and the "berth plan, the loading plan, and the heap" of the traditional container terminal of the container terminal are concentrated in the mode of the container. The sequential planning and scheduling method of field plan to container entry plan will no longer apply. Because the terminal can not determine the daily arrival and arrival time of the specific voyage, it can not be reserved in the whole box area in advance. It can only be distributed according to the actual arrival situation of the container and the idle situation of the yard. This will undoubtedly be the container for the container. In the random arrival mode, it is beneficial to improve the operation efficiency and service level of the container terminal scheduling. In this paper, the static and dynamic prediction models of the arrival time of the container terminal are established. On the basis of the historical data of the line, the v-SVM model is used to predict the driving path and travel time of the collection card. The dynamic prediction model uses v-SVM to predict the driving path and the travel time of the base line, and uses the actual running data as the system input, updates the prediction value by Kalman, and obtains the dynamic prediction of the arrival time of the collection card by iteration. The calculation results show that the prediction results of the static prediction model are better for the unblocked section, and the prediction error of the long distance easily congestion is often larger: the dynamic prediction model can dynamically update the prediction results according to the actual operation conditions, and the prediction accuracy on all sections is high, so the dynamic prediction model of the arrival time of the collection card is of great importance. High accuracy and robustness.

【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U691.3

【参考文献】