首页 | 本学科首页   官方微博 | 高级检索  
     

集装箱码头集卡调度模型与Q学习算法
引用本文:曾庆成,杨忠振. 集装箱码头集卡调度模型与Q学习算法[J]. 哈尔滨工程大学学报, 2008, 29(1): 1-4
作者姓名:曾庆成  杨忠振
作者单位:大连海事大学,交通工程与物流学院,辽宁,大连,116026
摘    要:研究集装箱码头装卸过程中集卡调度问题,建立了集卡调度动态模型,目的是使装卸桥等待时间最小.设计了基于Q学习算法的求解方法,获得在不同状态下的集卡调度策略.提出了应用Q学习算法求解集卡最优调度时系统状态、动作规则、学习步长与折扣因子的选择方法.结果表明,随着集卡数量的增加,Q学习算法获得的结果优于最长等待时间、最远距离、固定分配集卡等调度策略.

关 键 词:集装箱码头  强化学习  集卡调度  Q学习算法
文章编号:1006-7043(2008)01-0001-04
收稿时间:2007-04-27
修稿时间:2007-04-27

A scheduling model and Q-learning algorithm for yard trailers at container terminals
ZENG Qing-cheng,YANG Zhong-zhen. A scheduling model and Q-learning algorithm for yard trailers at container terminals[J]. Journal of Harbin Engineering University, 2008, 29(1): 1-4
Authors:ZENG Qing-cheng  YANG Zhong-zhen
Abstract:The problem of assignment and scheduling of yard trailers to quay cranes was discussed.A dynamic scheduling model for yard trailers was developed so as to minimize the waiting time of quay cranes.A Q-learning algorithm was also designed to develop decision-making policy for selecting the appropriate dispatching rules in different states.Also,a selection method was proposed whereby the Q-learning algorithm was used to choose the systematic state,action rule,learning step size,and reduction rate in yard trailer scheduling.Finally,simulation studies were performed to investigate the effect of the proposed approaches.Results show that,with an increasing number of yard trailers,the Q-learning algorithm performs better than other dispatching rules such as LW(longest waiting time),LT(longest travel),and SCO(single-crane oriented).
Keywords:container terminal  reinforcement learning  yard trailer scheduling  Q-learning algorithms
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号