【作者】 赵伟；

【导师】 刘林忠；

【作者基本信息】 兰州交通大学 ， 管理科学与工程， 2013， 硕士

【摘要】 在当今的市场经济条件下，企业之间的竞争扩展为供应链之间的竞争。增强供应链整体的竞争能力以及提高供应链整体的服务水平，成为很多学者研究的问题。以往在供应链管理方面的研究很多是基于战略层面的，很少在运作层次上研究。此外，在实际中，由于市场因素、生产设备、工作人员、天气状况等不确定因素的存在，使得供应链的运作往往处于不确定环境中。因此，本文研究的是供应链管理运作层次上的问题，主要是在不确定环境下研究供应链运作层次上的调度问题。优化供应链调度问题，可以提高供应链的服务水平，增加客户的满意度。本文研究了两阶段的供应链调度问题。在第一阶段中处于同一个区域中的多个生产商生产货物；每个生产商在每个周期内的生产能力和准备时间不同，并且每个生产商加工货物的速度也不同，如果分配给某个生产商的货物数量太多而超过了该生产商单位周期内的生产能力，那么该生产商在下个周期内生产货物。在第二阶段中，多个车辆把各个生产商加工好的货物配送给一个分销商；每个车辆所选择的路径长度不同，车辆配送货物的速度也不一样，并且每辆车辆有各自的装载能力，如果分配给某辆车的货物太多而超过了车辆的装载能力，那么该车辆在运输完一个批次的货物后就返回到生产商那里配送下一批货物。基于上述问题描述建立了随机期望值模型，目标函数是最小化所有货物的生产和配送成本总和。根据建立模型的复杂度设计了遗传算法来进行求解。设计了两个数值算例，第一个数值算例中货物批次是30，设定每个生产商和每个运输车辆在单位时间内所需要的费用都为1；在第二个数值算例中，把货物批次扩展到50，每个生产商和每个运输车辆的在单位时间内所需要的费用并不都为1。使用遗传算法求解上述两个数值算例，分析得出的最优调度方案，计算出每个货物的生产成本和配送成本，并且计算出了每个生产商的所用生产周期和生产时间，每个运输车辆的配送批次和配送时间。最后对模型进行了参数分析，在生产商和运输车辆的单位成本为1时，供应链上生产商的数量与运输车辆的数量应该大致相同，不应该存在瓶颈，并且货物的总体成本主要是由第一阶段决定的，减少每批货物的工作量或提高生产商的加工速度，能够大幅降低产品的总成本。更多还原

【Abstract】 In current market economy conditions, the competition between the companies hasexpanded to supply chains. Enhancing supply chain’s overall competitive ability andimproving the whole service level have become crucial issue for many researchers. Numerousformer studies on supply chain management are based on strategic level, but few of them isbased on operational level. In addition, due to the existence of uncertain factors such asproduction equipment, staff member and market factor as well as the weather condition,supply chain usually operates in uncertain environment. Therefore, this article study thesupply chain management issue based on operational level, main in the scheduling problembased on supply chain operational level under uncertain environment. Optimizing supplychain scheduling can improve the service level of the supply chain and boost customersatisfaction.The study considers the two-stage supply chain scheduling problem. In the first stage,many manufacturers located in the same zone process products with different processingspeed and setup time in each period. If the quantity of the products assigned to a manufacturerexceeds its unit cycle production capacity, the manufacturer would process the left products inthe next period cycle. In the second stage, many vehicles deliver the products to a distributorwith different speed and different routes as well as loading capacity. If the quantity of theproducts assigned to a vehicle exceeds its loading capacity, the vehicle would return to themanufacture for the next batch.Based on the above-mentioned problem description, formulate the situation as thestochastic expected value model. The objective function is to minimize the expectation of allproducts production cost and delivery cost. Due to the complexity of the model, geneticalgorithm is proposed to solve this model. Moreover, two numerical examples are studied. Inthe first numerical example, the products number is30, each producer cost is1in per unittime, so as each vehicle. In the second numerical example, the quantity of the productsexpanded to50, the cost of each producer or vehicle is not always1. The optimal schedulecases can be obtained by using genetic algorithm. The production cost and delivery cost ofeach product, the number of production cycle and production time of each producer, thenumber of delivery batch and delivery time of each vehicle can be calculated by analyzing theoptimal scheduling case. In the same time, the two examples verify the feasibility andeffectiveness of the genetic algorithm. Through performing parameter analysis, when eachproducer or vehicle cost is1in per unit time, the number of the produers and the vehiclesshould be matched; the bottleneck should not exist in the supply chain. In addition, the overall cost of the products is mainly determined by the first stage. Reducing the workload of theeach product or improving each producer’s production rate can greatly lower the overall cost.更多还原