【作者】 王茜；

【导师】 符小玲；

【作者基本信息】 东南大学 ， 物流工程（专业学位）， 2020， 硕士

【摘要】 随着经济全球化和贸易自由化的日益发展,我国的保税制度和园区日渐趋于成熟,保税物流与国际贸易的关系愈加紧密。在国际贸易供应链中,保税物流的保税、退税红利为企业的现金流改善提供了空间,而企业也对保税物流的效率提高和成本控制提出了更高的要求。由于保税物流受保税政策和监管制度的绝对制约,目前国内对于保税物流的研究主要集中在宏观层面,从保税体制分析、保税区域的规划设计及功能优化等方面来讨论保税物流的发展和趋势。从微观层面对保税物流管理的深入研究亟待发掘。本文详细梳理了第三方物流企业S公司的组织结构、业务运作背景,选取保税仓二线出业务作为研究对象,总结保税仓二线出业务的业务流程和业务特性,深入分析S公司在保税仓二线出业务调度中存在的问题及产生问题的原因。在此基础上,将二线出业务调度存在的问题抽象为资源受限多项目调度问题,结合时间窗理论把惩罚成本作为衡量业务调度质量的核心指标,根据保税物流业务中的资源约束及工序约束特性,引入决策单元和资源安排顺序的概念,构建了S公司保税仓二线出业务调度模型。在该模型的求解算法上,本文根据保税物流业务中工序时序和资源使用的特性,采用多段式编码来考虑所有项目使用资源的安排。采取多策略融合选择、限制邻域变异、反向初始化种群解决选择变异后的迭代波动问题、通过扩大搜索范围解决了算法收敛容易陷入局部最优解的问题。在以上研究的基础上,选取S公司的实际业务场景及数据构建实例,通过对S公司项目组织结构的优化得到项目经理优化方案,通过资源安排计划优化、采用改进后的遗传算法求得近似最优的业务调度计划得到资源安排计划优化方案。最后通过对比实例及两个优化方案的惩罚成本金额、惩罚项目个数及具体工序环节的优化效果,验证了模型的合理性和业务调度优化的有效性,并根据优化方案对S公司的运营管理和发展战略提出了建设性意见。更多还原

【Abstract】 With the increasing development of economic globalization and trade liberalization,the bonded logistics and international trade have been increasingly connected.In the international trade supply chain,enterprises require more improvement on efficiency and cost control of the bonded logistics.Current domestic research mainly focuses on the macro level,in-depth study of bonded logistics management on the micro level needs to be explored.Analyzing the organizational structure and business background of company S,bonded warehouse domestic outbound business selected as the research object,summarizing the processes and characteristics of domestic outbound business,the problems and causes of domestic outbound business scheduling is presented.On this basis,the problems of domestic outbound business scheduling is abstracted into the resource constrained project scheduling problem,combining the theory of time window,selecting the punishment cost as the core indicator of project scheduling quality.According to the characteristics of resource constraints and process constraints of bonded logistics business,introducing the concepts of decision making units and the resource arrangement order,the scheduling model of bonded warehouse domestic outbound business is built.According to the practical situation of company S,improvement design of genetic algorithm is the crucial part.According to the process sequence and resource usage in bonded logistics business,multi-stage coding is used.In order to solve the problem of iterative fluctuation after selection and mutation,combined strategy and limited neighborhood mutation is put forward.By inversely initializing the population,the initial search range is expanded,solving the problem of falling into local optimal solution.On the basis of the above study,selecting the actual operation scenarios and data of company S,through optimization of project structure and resource arrangement,two optimization schemes are obtained,and the approximate optimal project scheduling plan was obtained through the improved genetic algorithm,the comparison of iteration curves before and after the algorithm optimization was improved also verifies the rationality of model and the algorithm optimization effectiveness.更多还原