节点文献

基于蚁群算法的集装箱港口选择与网络均衡分析

Analysis of Container Port Selection and Network Equilibrium Based on Ant Colony Algorithms

【作者】 闫博

【导师】 杨赞;

【作者基本信息】 大连海事大学 , 交通运输规划与管理, 2008, 博士

【摘要】 改革开放30年来,港口作为对外开放的窗口,对国民经济发展呈现出越来越重要的推动作用。但是随着集装箱港口的迅猛发展以及货主码头的增加,使集装箱港口之间的竞争日趋激烈,造成了当前港口生产能力的部分过剩,一方面港口经营效益下降,另一方面也导致了港口资源的巨大浪费,影响了地区经济的可持续发展。为维持自身发展,在日益激烈的竞争中取胜,港口需要获得稳定、大量的货源。了解港口使用者对港口的评价过程和结果,做好港口网络的宏观结构与微观结构研究,建立和完善港口网络发展的理论基础,对于促进我国港口持续健康发展,增强国际竞争能力,使之更好地为国民经济和社会发展服务具有重要的理论价值和现实意义,是一个极为重要的课题。本论文以集装箱港口选择与港口网络均衡分析作为理论基础,结合智能算法建立优化模型,综合考虑货主港口选择的特性与港口市场竞争的特点,对货主如何进行港口选择以及如何在港口网络中进行货流分配等问题进行深入研究。论文主要研究成果概括如下:1、本论文提出了全新的货主港口选择优化模型,其以货主港口选择几率最大化为目标,即几率越大的港口越容易被货主选择。该模型综合考虑了货主选择港口的三个主要因素,定性分析与定量计算相结合,具有良好的可操作性。由于该模型是多目标组合优化模型,结构复杂,传统方法求解困难。本论文采用了一种模拟进化算法——蚁群算法,来寻找网络中的最短路径;同时,为了提高蚁群算法的搜索效率,采用了一个新的信息素增量更新策略——ANT-weight策略。最后,通过实例对该模型和算法进行了检验,理论分析和运算结果一致,表明了模型和算法的有效性。2、货主港口选择的依据是港口网络构成的基础,港口网络建立的是否合理将会直接影响货主的利益。本论文以货主港口选择几率最大化为目标函数所建立的静态港口网络,是以被选港口吞吐量无限大、港口费率相差不大为前提条件的,但在实际运输过程中,受港口吞吐能力的限制以及港口拥挤、阻塞现象越来越严重的影响,要求货物应按照一定的规则,符合实际地分配到港口网络中的各条航线上,这就形成了动态港口网络。为了解决货流量分配问题,本论文采用了交通分配模型中的增量分配法,将OD间货流量平分成若干等分,循环分配到网络中:因为寻找最短路径是货流量分配中最基本也是最重要的计算,所以提出基于粗粒度并行蚁群算法的增量分配模型,此模型实现了动态港口网络中货流的最佳配置,使建立的动态港口网络达到均衡。最后,运用该模型对某区域集装箱港口网络进行检验,得到了集装箱港口网络规划的优化效果。3、港口网络作为运输网络的重要组成部分,服务于国家经济发展的需要,因此,分析港口网络协调发展对社会和经济带来的影响,避免紧缺社会资源的浪费,在提倡科学发展观的今天格外具有现实意义。本论文提出了基于集值统计方法的模糊综合评价模型,利用专家定性分析给出各指标的区间估计,从而形成一个集值统计序列,进一步定量计算出指标评价值,解决了具有不确定性特点的模糊评价问题,该方法用于评价港口网络协调发展具有可操作性和实用性。最后,通过实例对该模型进行了检验,结果表明可以得到更为可靠的评价。论文最后对全文进行了总结,并对有待于进一步研究的问题进行了展望。

【Abstract】 For three decades after the reform and opening-up policy,as an open window,port has been playing a more and more important role in the development of national economy.The fast development of container ports and the increase of cargo owner’s wharf,however,have made competition fiercer and resulted in surplus of port productive capacity.On the one hand,it causes the decline of operation profit;on the other hand,it brings about enormous waste on port resource and in sequence,it weakens the sustainable development of regional economy.In order to keep development and go through the fierce competition,stable supply of goods is a necessity.It’s of important theoretical value and practical significance for sustainable and healthy development of port to understand the evaluation procedures and results,to research in both macrostructure and microstructure of port network,and to establish and improve basic theory on port network development.The topic thus has become an important problem to study.Theoretically based upon the equilibrium analysis of container port selection and port network,this dissertation establishes optimization model combining intelligence algorithm,and researches into the port selection and the freight flow allocation in port network,with comprehensive consideration of the characteristics of cargo owner port selection and market competition。The main research results are summarized as below:Firstly,it puts forth a new optimization model for cargo owner port selection, aiming at the probability maximum of cargo owner port selection,which means that the greater the probability is,the easier it is to be selected by the cargo owner.This model, considering comprehensively three main factors of cargo owner port selection and combining qualitative and quantitative calculation,is easy in operation.Since it’s an optimization model of multi-targets combination with complex structure,which is difficult to be solved via traditional methods,here ant colony optimization(ACO),a new evolution algorithm,is used to search for the shortest route in network.Meanwhile, in order to improve the search efficiency of ant colony algorithm,this thesis develops a new strategy to update the increased pheromone-ANT-Weight strategy of ACO. Lastly,the model and algorithm are tested through examples,and the same results from both theoretical analysis and the test verify the effectiveness of the model and algorithm.Secondly,the basis upon which cargo owners select port is fundamental to the structure of port network.The problem whether port network is built reasonably will directly affect cargo owners’ profit.The static port network aiming at maximum probability of cargo owner port selection is set up in the precondition that the selected port is of unlimited throughput and the difference in rates of port charge is small.In practical transportation,however,the pressure from port throughput limitation,port congestion and jam is more and more severe,which requires a certain rule to distribute the cargo practically into each course.To meet this need,here the dynamic port network comes into being.In order to settle the problem of cargo flow allocation,an increment distribution method in transportation distribution model is adopted.According to this method,divide cargo flow between OD averagely into many equal parts,and allocate these parts in circulation into the network.As calculation to search for the shortest path is the most basic and crucial,a new kind of increment distribution model based on coarse-granular parallel ant colony algorithms is set forth.This model realizes the optimum distribution of cargo flow in dynamic port network,which makes the dynamic port network reach equilibrium.Lastly,after applying this model and algorithm into certain regional container port network,we get the optimized effect of container port network layout.Thirdly,as an essential part of transportation network,port network contributes to the development of national economy.It’s hence of practical significance to analyze harmonious development of port network to avoid the waste of social resource, especially today when scientific development concept is advocated.This dissertation puts forth the fuzzy aggregated evaluation model based upon set occurrence statistics methodology.This model forms a set occurrence statistics sequence by interval estimation from experts’ qualitative analysis,and further calculates the evaluated value of indicators quantitatively,which solves the problem of uncertainty in fuzzy evaluation. So this method is operative and practical in the evaluation of port network harmonious development.Lastly,this method is tested by examples and its result shows that it’s possible to have more reliable evaluation. At the end of this thesis,a sum-up is made,and problems in need of further research are put forward.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络