不确定环境下铁路集装箱动态多阶段调运优化模型和算法研究

项目名称： 不确定环境下铁路集装箱动态多阶段调运优化模型和算法研究

项目编号： No.60870008

项目类型： 面上项目

立项/批准年度： 2009

项目学科： 数理科学和化学

项目作者： 李引珍

作者单位： 兰州交通大学

项目金额： 25万元

中文摘要： 本课题针对非集装箱结点车站间的空箱调运，在一个计划期内，将空箱调运视为一个具有不确定环境的动态多阶段优化问题。首先对我国铁路集装箱运输实际情况进行了深入研究与分析，特别对集装箱承运办理、集结、运输、交付等诸环节进行了研究，对集装箱运输成本的构成进行了分析；利用图与网络理论、随机理论、模糊集理论等对运输网络路径的相关问题进行了研究，为铁路集装箱空箱调运模型的建立与求解奠定了良好基础；进一步，利用多阶段指派问题、约束运输问题、多商品网络流等优化问题的优化建模思想和方法，以及随机规划、模糊规划等原理，分步分层建立了单一箱型的空箱调运模型、可代用多种箱型空箱调运模型、铁路集装箱空箱调运多商品流模型、带时间窗约束的机会损失优化模型、动态多阶段空箱调运模型、基于需求等级的铁路集装箱空箱调运优化模型；设计了求解各类模型的算法；通过大量算例验证了模型的正确性与算法的时效性。本研究从箱种代用、时间窗机会损失、一个计划期内多阶段调运、需求等级等方面拓展了空箱调运的研究内容、丰富了学科领域理论，为后续研究提供了新的方法和手段，同时，也为学科建设培养了人才。

中文关键词： 铁路集装箱调运；优化模型；进化算法

英文摘要： The research aims at empty container allocation between non-key railway stations which transact container transport. The empty container allocation within a plan round is a dynamic multi-stage optimization problem under uncertain environment. Firstly, the current situation of railway container transportation in China has been studied and discussed. The key links of accept, accumulation, carriage, delivery, and so on have been researched especially. The composition of container transportation cost has been analysed. The relevant issues of traffic network paths are studied by use of graph and network theory, stochastic theory and fuzzy set. These establish good foundation for modeling railway empty container allocation. Further, single container type model, multiple container types substituted model, multi-commodity model, opportunity loss model with time window constraint, dynamic multi-stage model, and model based on demand rank are formulated respectively with the help of the principle multi-assignment problem, constraint transportation problem, multi-commodity network flow, fuzzy programing, and stochastic programing. The algorithms for solving the models are designed. The correctness of models and timeliness of algorithms are tested and verified getting by lots of instances. The research results of container type substituted, time window opportunity loss,multi-stage allocating within a plan round, and allocation based on demand rank expand future research contents on railway empty container allocation, enrich theory of container transport subject area. The research provides new methods and means for future research, and also has trained talents for discipline construction.

英文关键词： railway container allocation;optimal model;evolutionary algorithm