基于复杂系统理论的配送网络优化研究

【作者】 李慜；

【导师】 薛惠锋；

【作者基本信息】 西北工业大学 ， 系统工程， 2006， 硕士

【摘要】 配送网络是物流系统创造“空间价值”和“时间价值”的源泉,对企业在当前以买方市场为特征的激烈竞争环境下取得竞争优势具有关键作用,在市场中受到越来越多的重视。由于配送网络的影响因素众多,其中很多优化问题所建立的模型和算法复杂,大都具有NP-难性质,采用传统的规划方法难以得到满意方案,复杂理论的“自下向上”的思维模式无疑为配送网络这种大规模复杂问题提供了有效的解决方法。 本论文从系统工程的视角,把握现代物流配送研究的前沿动向,以理论研究为经,以实证分析为纬,全方位、多层次地研究了配送网络的重要性和复杂性以及系统中各种内、外部因素的影响,针对配送网络优化中的三大关键问题:交通网络、配送中心选址和配送路线优化,提出了基于复杂系统理论的物流配送网络优化模型。结合元胞自动机与GIS,建立了基于改进元胞自动机的求解单源点到其余各节点的最短路径算法;综合考虑配送中心的建设经济性、交通便利性和功能布局合理性,提出了基于改进遗传算法的配送中心选址模型;对遗传算子进行改进,建立了配送路线优化模型,实现了对配送车辆数目和配送路线的综合求解。 论文以西安邮政配送网络进行实证研究,分析了模型的可行性和有效性。研究结果表明,优化模型能够较好的反映客观现实,对于降低配送成本、提升配送效率、优化配送网络结构都有着重要的参考和实用价值,具有较强的理论和实践意义。更多还原

【Abstract】 As the main source bringing temporal and spatial value, distribution network of logistics system has become the focus of market competition. It’s the most important base for companies to achieve top place in buyer market. Due to the complexity of distribution network, many models and algorithms established to optimize problems are complex and have the nature of NP. It’s hard to get satisfactory scheme with them. While the bottom-up methodology of complex theory has provided a better way to deal with complex problems like distribution network.Making the viewpoint of systems engineering as stand and with proper understanding of modern logistics development, this thesis gives comprehensive and layered explanations to importance and complexity of distribution network and to the influences among internal and external factors in it by combining theoretical and empirical research. After three key problems of distribution network optimization including road network, distribution center location choice and vehicle routing are generally discussed, an optimization model of logistic distribution network based on complex system theory is put forward. This model has three folds: first, by integrating CA and GIS, an extended CA sub-model to search the shortest path from one node to all other nodes is advanced;second, by considering its building cost, traffic condition and function, a distribution center allocation sub-model is proposed by adopting an improved genetic algorithm;third, a sub-model aiming to solve vehicle routing problem is established with improved genetic operators, which can calculate the fittest number of distribution vehicles in the while of optimizing the routing order with time constraints.The whole model is applied to Xi’an Post distribution network to prove its validity and simulating the reality. The results show the optimization model could fairly reflect the fact and be good for reducing the distribution cost, improving logistic effective and optimizing the distribution network.更多还原