基于粒子群算法的库存——路径问题研究(硕士论文).pdf

南京工业大学本科生毕业设计(论文)


基于粒子群算法的库存——路径问题研究
摘要

库存——路径问题(Inventory Routing Problem,IRP)是供应商库存管理
(VMI)模式下的核心问题,通过协调库存控制与运输调度,使库存和运输等
综合物流成本最低。本文深入分析 IRP 问题国内外研究现状,对 IRP 问题的分
类、各种建模方法及求解方法进行总结。在此基础上研究了多周期确定需求下
的库存路径问题,建立了混合整数规划模型。研究了粒子群算法对该模型的优
化求解;基于整数编码方法,使用随机初始化的方法产生初始解,为了提高算
法的性能,引入四种惯性权重调整策略和两种学习因子调整策略。基于 Matlab
编程进行实验仿真,使用离线性能和在线性能对算法进行评价。讨论了算法的
迭代次数、惯性权重调整策略等参数对算法性能的影响,找出了解决该类问题
的适合参数。并与遗传算法、经济订货批量法的优化结果进行了比较,结果表
明粒子群算法是求解 IRP 问题的有效算法。


关键词:库存路径问题,粒子群算法,数学建模,Matlab 编程








Ⅰ
摘要
Research on Inventory Routing Problem Based on Particle
Swarm Optimization Algorithm
ABSTRACT
Inventory routing problems (IRP) are core issues of Vendor Managed Inventory
(VMI), which aims to minimize the integrative cost of inventory and transportation
through coordinating inventory control and transportation plans. This paper firstly
surveys the current research of IRP at home and abroad, then summarizes and
classifies all kinds of IRP model and optimization algorithm .Next, a kind of
multi-period inventory routing problem with determinate need is studied, and a
mixed integer programming model based on the problem is built. Particle swarm
optimization (PSO) algorithm is proposed to optimize the model. The solution is
encoded in integer and is initialized randomly .In order to improve the performance
of PSO, four kinds of strategies which are used to adjust the parameter of inertia
weight of the PSO algorithm and two kinds of learning strategies are incorporated
into the algorithm. The algorithm is implemented in Matlab, and it is evaluated by
two criterions: on-line performance and off-line performance. The effect of different
parameter values about iteration, inertia weight and so on are discussed in the
experience,and the appropriate parameters are found. parison with the
traditional economic order quantity and ic algorithms shows that the particle
swarm optimization outperforms others and is eff