文档介绍：华中科技大学硕士学位论文
摘要
随着三峡船闸投入运行,过坝船舶逐渐增多,三峡船闸和葛洲坝船闸的两坝联
合运行工作日益繁重。为了满足三峡—葛洲坝两个枢纽统一调度的管理需求,提高三
峡船闸和葛洲坝船闸的通过能力,逐步实现智能化调度,最大限度地满足船舶过坝要
求,实现三峡-葛洲坝联合调度工作的现代化管理,确保三峡通航“安全、有序、畅
通”,三峡通航局于 2005 年 1 月委托本研究所开发三峡—葛洲坝联合调度系统。经
过一年的努力,该系统已经成功上线运行,达到了初期的目的。
两坝联合调度中对单个船闸的计划编制是一个时间上串联的 timetable 问题,空
间上是 packing 问题。对于一个固定计划期,timetable 的时间序列可以通过很多启发
式算法来获取,例如遗传算法、模拟退火算法、禁忌搜索算法、人工神经网络。由
于每个船闸的开闸时间序列和船舶流的具体船舶面积有关,所以在制定启发式搜索
时,要嵌入二维 Packing 问题。计划编制中的 timetable 问题和 Packing 问题是个耦合
问题,Packing 算法的优劣直接影响整个计划编制算法的性能,因此本文致力于闸室
编排算法的研究。
闸室编排可以用二维 Packing 模型来描述,是一个典型 NP 完全问题,本文提出
三种不同的拟人编排算法,这三种算法在实际工程应用中取得了良好的效果,有效
地提高了闸室面积利用率。在拟人法的基础上,作者引入遗传算法来求解该问题,
实践表明基于遗传算法的闸室编排算法达到良好的效果。
目前刚投入使用的三峡-葛洲坝联合通航调度系统在计划编排算法上作出了突
破性进展。但因运行模式繁多、实际突发事件等导致算法有一定的局限性,因此作
者开发了手动计划编制与调整模块,此模块作为联合调度系统的重要组成部分,使
自动手动相结合,增强人机交互,方便用户操作,得到用户一致好评。
关键词: 三峡—葛洲坝通航调度闸室编排 Packing 问题
I
华中科技大学硕士学位论文
Abstract
With the Three Gorges Ship Lock’s establishment, the population of ships that cross
dams is gradually increased. This situation makes the cooperation between Three Gorges
Ship Lock and Gezhouba Ship Lock. In order to satisfy manage requirements of unified
schedule of Three Gorges and Gezhouba Dam, improve the navigation capacity of two
dams, meet the requirement of ships as much as possible, make the realization of modern
management, Three Gorges Navigation Bureau consigns us to develop this Three Gorges
--- Gezhouba cooperative schedule system. After one year development, this system has
essfully run in the real environment, which is the goal of this system.
Scheduling two joint dams’ project on the establishment of a single lock the timetable
is a matter of time series. Space is packing problems. Plans for a fixed period of time
series timetable can be obtained through a number of heuristic algorithms. For example,
ic algorithms simulated annealing, tabu search algorithm, artificial works.
As each ship lock gate opening time series flow and the specific area of the sh