摘 要

港口集卡调度问题是一个研究热点，在传统模式下，集卡空驶率过高，浪费了大量时间和设备资源，同时作业负荷分布不均，有的作业线集卡排队等候，而有的作业线作业机器等候集卡，导致效率低下。

为了提升码头生产效率，节约时间、设备成本，本文在基于多Agent系统的基础上，结合改进合同网协议，为集卡提供了动态调度的优化方案，研究成果对港口的集卡调度有重要现实指导意义，并具有广泛应用前景。主要研究内容如下：

1. 在研究国内外文献的基础上，确定了本文的研究目的和意义，以集卡的调度为研究对象，建立以多Agent系统为基础的集卡动态调度系统框架，利用改进的合同网协议，结合模糊控制和蚁群算法，解决集卡排队问题，任务分配问题以及路径选择和堵塞问题。
2. 利用多Agent系统理论，建立了含信息Agent，集卡Agent和集卡调度Agent的多Agent协作模型，并确定各Agent之间交互关系。
3. 改进了传统的合同网协议，作为系统的协商协议，在改进合同网协议中引入了信誉度的概念，并参考文献确定信誉度函数。对协商过程进行了说明，确定了Agent之间信息交流的格式。
4. 运用模糊控制理论实现集卡调度的决策，以任务重要性和集卡行驶距离作为模糊输入变量，调度方案作为模糊输出变量，用专家经验法确定隶属函数和规则库，面积重心法进行清晰化输出。在MATLAB软件上成功运行，并得到了输出图像。
5. 利用蚁群算法实现了集卡的路径选择，通过GPRS系统来监控现场数据，确定是否堵塞以及堵塞路段，并提出了堵塞解决方案。

研究结果表明，基于多Agent技术的集卡调度方案提高了集卡调度效率，有效缓解了拥堵。

本文特色：从全局出发，考虑到了集卡调度问题的决策、任务分配、堵塞缓解等问题，涵盖面广，具有较大的现实指导意义。

关键词：集卡调度 多Agent 模糊控制 路径选择

ABSTRACT

Rescheduling of port container track is a research hot spot.In the traditional situation,the container track’s high rate of meaningless travel wastes plenty of time and equipment resources. While the uneven distribution of operating load, and some work line track waiting in line, and some operations Line machine waiting for the track, resulting in inefficient.

In order to improve the efficiency of the terminal, save time and equipment cost. Based on the multi-agent system and the improved contract network protocol, this paper provides a dynamic scheduling optimization scheme for the container track. The research results have important reality for the port scheduling guiding significance, and has wide application prospects. The main research contents are as follows:

1. By studying the domestic and foreign literatures, determined the purpose and significance of this paper. Based on the scheduling of the container track, the framework of the dynamic scheduling system based on multi-agent system is established, and using the improved contract network Protocol, combined with fuzzy control and ant colony algorithm to solve the problem of container track’s queuing problem, task assignment and path selection and congestion.
2. Based on the theory of multi-agent system, the multi-agent cooperation model with information agent, container track agent and container track scheduling agent is established, and the interaction between agents is established.
3. To improve the traditional contract network agreement as a system of consultation agreement, in the improvement of the contract network agreement introduced the concept of credibility, and reference literature to determine the function of the credibility. The negotiation process is described, and the format of information exchange between agents is determined.
4. The fuzzy control theory is used to realize the decision of the container track dispatching. The task importance and the travel distance of the container track are taken as fuzzy input variables, and the scheduling scheme is taken as the fuzzy output variable. The expert function method is used to determine the membership function and the rule base, and output by area barycenter method. Successfully ran on MATLAB and got the output image.
5. Use the ant colony algorithm to achieve the container track path selection, through the GPRS system to monitor the field data to determine whether blocked and find the block road, and put forward a blocking solution.

The results show that the multi - agent technology based on the track scheduling program to improve the efficiency of the track scheduling, effectively alleviate the congestion.

The characteristics of this paper: from the overall situation, taking into account the track scheduling problem decision-making, task allocation, congestion mitigation problem, covering a wide range, with great practical significance.