【作者】 杨娟；

【导师】 郭海湘；

【作者基本信息】 中国地质大学 ， 管理科学与工程， 2011， 硕士

【摘要】 随着科技的不断发展,通过降低物料消耗而获取利润(即第一利润源泉)和通过降低人工成本和提高劳动生产率而增加利润(即第二利润源泉)的潜力已经越来越小,而降低物流费用已被广泛认为是企业取得利润的第三利润源泉。从物流总费用的构成来看,运输费用在物流总费用中占有相当高的比例。配送属于支线运输,是供应链上最后一个环节,负责将实物运送至用户的手中。配送路线设计的合理与否直接关系到企业是否能以最小成本及时、按量、保质将货物运送至客户点。本文的选题来源于校企合作项目“郑煤集团公司供应物流系统工程研究”。在这个具体的应用背景下,分析了煤矿不同种类物资配送的特点：物资需求有紧急性之分,车辆可多趟运行；危险物资允许联合配送,一个客户的需求可由一辆车运行多趟来满足,或由不同车辆来进行访问；而普通物资则不允许联合配送,每辆车每趟只能访问一个客户点,而且要求车辆必须满载。在大量阅读文献的基础上,发现本研究与带时间窗的车辆路径问题、带模糊时间窗的车辆路径问题、需求可拆分的车辆路径问题这三类问题之间存在着某些共性。通过对比,前两类问题中的时间窗无法描述本研究中的需求紧急性,因为客户并没有明确指明接受服务的时间区间段,也没有一个有明确上下限的期望时间窗口,故本文提出用自然数表示的“服务优先级”来刻画客户对物资需求的紧急程度。危险物资多趟配送的特点使其优化可借鉴需求可分的车辆路径问题的研究方法来进行。确定问题的范围边界后,针对危险物资和普通物资的配送特点,分别建立相应的数学模型,设计了改进的扫描算法和改进的遗传算法对危险物资多趟联合配送模型进行求解。以郑州煤电物资供销有限公司的危险物资配送为求解实例,从最优结果的配送费用、配送里程、使用车辆数、车辆装载率、平均计算时间和保证服务优先级等指标分析两种算法的优化性能,结果表明在相同的使用车辆数和车辆装载率情况下,改进的遗传算法要优于改进的扫描算法,但需要较长的平均计算时间,而且也一定程度上破坏了服务优先级,而改进的扫描算法则较大程度保证了服务优先级。同时还与该公司实际配送情况进行了比较,进一步验证了本文算法优化的有效性。对普通物资的多趟单点配送问题也设计了相应的算法进行优化,将优化结果与实际配送情况比较分析,表明算法优化大大减少了配送的成本,缩短了配送的时间。最后,基于MATLAB的GUI平台,实现了算法的可视化。用户在输入界面输入需求和车辆等参数,即可在输出界面看到算法优化的结果,借此辅助决策。更多还原

【Abstract】 With the developing of technology, by reducing material consumption to win profits (the first source of profit), and by reducing labor cost and increasing labor productivity to increase profits (the second source of profit), the potential has been getting smaller and smaller, the lower logistics cost have been widely regarded as the third source of profit for enterprises. Among the total logistics costs, the transportation cost accounts for a quite large proportion. Distribution which is a kind branch transport and the last link in the supply chain, has responsible for deliverying goods to the users with kinds of demands. Whether the design of delivery routes is rational or not, it can directly refluences whether the enterprises delivers goods to customer sites with minimum cost, in time, right quantity and high quality.The topic of this article is from the project with the title "the supply logistics systems engineering research for Zheng Coal Company" Which is cooperated by schools and enterprises. Under this specific application background, the distribution of different kinds of materials in coal mine analyzed has some characteristics:the material demands for different customers exists the different urgency, and vehicles are allowed to run more than one times; dangerous materials are allowed to adopt the joint distribution, that means a customer’ needs can be met by a car running more times, or by different vehicles visiting it; the joint distribution of common goods is not allowed, a vehicle is allowed to visit only one customer point per trip, which must be loaded. On the basis of reading lots of literatures, we find that this research has some similarities to three types of problems the vehicle routing problem with time windows, the vehicle routing problem with fuzzy time windows, and the vehicle routing problem with the split demands. By the comparison. the time window in the first two issues can not describe the urgency of needs in this paper’s study, because clients receiving services does not specify the time interval segment, and also no one clear expected time window limit, we suggest that "priority of service" characterized the natural number is used to describe the to urgency of the customers’ material needs. The characteristic that dangerous materials may be distributed more times makes the optimization can be carried out through learning from the research methods the vehicle routing problem with spilit demands.After determining the boundary of the problem, respectively establish the corresponding mathematical models according to different distribution characteristics of dangerous materials and common materials, and design the improved sweeping algorithm and the improved genetic algorithm to solve the joint distribution model characterized multiple times of dangerous materials. Take an example for the distribution of dangerous materials for Zhengzhou Coal and Electricity materials supplying and marketing limited company, by solving the optimal distribution path results, through some indicators such as the distribution costs, delivery mileage, the number of vehicles used, vehicles’s loading rate, the average computing time and the degree of ensuring service priorities analyzes the optimal performance of two above algorithms, and that indicates that under the case of the same number of vehicles used and vehicles’loading rate, the improved genetic algorithm is superior to the improved scanning algorithm, yet sacrifices a longer average time, and also destroy the priority of service to a certain extent, but the improved sweeping algorithm can ensure the priority of service at a greater degree. And compared with the actual distribution situation of the company, further verify the validity of optimization algorithms in the paper. A simple algorithm is designed to optimize the distribution of common materials characterized multiple times and single point. By comparing and analyzing the optimal results with the actual distribution, it shows that the optimization algorithm greatly reduces the cost of delivery and shorten the delivery time.Finally, realize the visualization of the algorithm based on the GUI platform in MATLAB. The user inputs parameters related to requirements and vehicles in the input interface, and can see the optimal results of algorithms in the output interface, to aid decision-making.更多还原