软时间窗约束下动态选址—路径优化研究

发布时间：2018-03-06 17:05

本文选题：选址-路径　切入点：软时间窗　出处：《福州大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着全球经济的飞速发展,物流企业面临着一个复杂多变、市场竞争日益激烈的动态决策环境,很多决策因素会随着时间而发生变化。因此,根据决策环境的动态变化进行合理的物流系统规划已成为一体化物流的发展趋势。在物流系统规划中,配送中心选址的确立和路径的安排之间存在相互依赖的关系,对这两个要素进行集成优化是物流系统规划的核心问题。国内外已有不少学者对其进行了研究,但主要是静态的,不符合目前的动态决策环境,导致所做出的决策属于“短视”的次优决策。同时,考虑客户需求呈现多样性发展,客户对订单的响应速度要求越来越高,使得企业处于一个基于时间竞争的市场环境。因此,研究带软时间窗的动态选址-路径优化问题不仅具有理论价值也有现实意义。本文对配送中心选址的确立和路径的安排进行集成优化,同时考虑物流系统中客户需求量、配送中心运营成本、重新打开和关闭成本等参数随时间推移而变化的动态性质,建立了动态的选址-路径优化模型。在一个工厂、多配送中心、多客户、多时间段的情况下确定了各个时间段配送中心的选址数量和位置以及配送路径的安排方案,并确定了整个计划期内的最优方案；在此基础上,考虑客户软时间窗约束,建立了带软时间窗的动态选址-路径优化模型,确定了各个时间段配送中心的选址数量和位置以及配送路径的安排方案,并确定了整个计划期内的最优方案。在模型的求解上,针对两个优化模型,本文将其分别分解成静态选址-路径优化问题和动态规划以及静态带软时间窗的选址-路径优化问题和动态规划两个子问题求解。对于静态的选址-路径优化问题和静态的带软时间窗的选址-路径优化问题分别设计了基于贪心算法的改进微粒群算法(以下简称GA-PSO算法)和罚函数的微粒群算法(以下简称PENALTY-PSO算法)进行求解,得出各个时间段的选址和路径安排方案。最后,用动态规划法求出整个规划期内的最优方案。实验结果表明,GA-PSO算法在最低总成本平均值、解的稳定性、最优解的命中率以及搜索速度上均优于P ENALTY-PSO算法。同时加入客户的软时间窗约束对最低总成本平均值也有一定的影响。算例结果说明了模型及求解算法是有效的和实用的。最后,对本文的研究成果进行简要总结,并指出需要进一步研究的方向。
[Abstract]:With the rapid development of the global economy, logistics enterprises are faced with a complex and changeable dynamic decision-making environment in which market competition is increasingly fierce. Many decision-making factors will change with time. According to the dynamic change of decision-making environment, reasonable logistics system planning has become the development trend of integrated logistics. In the logistics system planning, the establishment of distribution center location and the path arrangement are interdependent. The integration optimization of these two elements is the core problem of logistics system planning, which has been studied by many scholars at home and abroad, but it is mainly static and does not conform to the current dynamic decision-making environment. At the same time, considering the diversity of customer demand, the customer response speed is higher and higher, which makes the enterprise in a market environment based on time competition. The study of dynamic site-path optimization with soft time window has not only theoretical value but also practical significance. In this paper, the establishment of distribution center location and the path arrangement are integrated and optimized, and the customer demand in logistics system is taken into account at the same time. The dynamic properties of the operating costs, re-opening and closing costs of distribution centers over time, and a dynamic site-path optimization model is established in a factory, multi-distribution center, multi-customer, and so on. In the case of multiple time periods, the location number and location of distribution center and the arrangement scheme of distribution route in each time period are determined, and the optimal scheme in the whole plan period is determined. On this basis, the soft time window constraint of customer is considered. The dynamic site-path optimization model with soft time window is established, the location number and location of distribution center in each time period and the arrangement scheme of distribution path are determined, and the optimal scheme is determined during the whole planning period. For two optimization models, In this paper, it is decomposed into static site-path optimization problem and dynamic programming problem, as well as static site-path optimization problem with soft time window and dynamic programming. The problem and static site-path optimization problem with soft time window are solved by the improved particle swarm optimization (GA-PSO) algorithm based on greedy algorithm (hereinafter referred to as GA-PSO algorithm) and the penalty function particle swarm optimization algorithm (PENALTY-PSO algorithm), respectively. Finally, the dynamic programming method is used to solve the optimal scheme for the whole planning period. The experimental results show that the GA-PSO algorithm is stable in the mean of the lowest total cost and the stability of the solution. The hit ratio and search speed of the optimal solution are better than that of the P ENALTY-PSO algorithm. At the same time, the soft time window constraint of the customer also has a certain effect on the average minimum total cost. The results of an example show that the model and the algorithm are effective. And practical. Finally, This paper briefly summarizes the research results and points out the direction of further research.
【学位授予单位】：福州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U116.2;F252

【相似文献】