蚁群算法与物流车辆路径优化研究外文翻译资料

The research of Ant colony algorithm with logistics vehicle routing optimization

Rizzoli E

Abstract

Logistics is gradually by enterprises as obtain and maintain competitive advantage in the market of the powerful management, caused the domestic and foreign scholars and enterprises of great importance to it. Current generally accepted definition of 'logistics' is put forward by the American warehouse and purchasing association: 'the so-called logistics is to meet customer demand for the purpose, in order to improve the raw material, and related production, from the supply to the consumption of the flow and storage of efficiency and effectiveness, and carries on the plan, execution (implementation) and control the process of. Modern Logistics (Logistics), it is to point to in order to achieve customer satisfaction, to connect the subjects of supply and demand, overcoming space and time obstacles of effective and efficient flow of goods, services, economic activity process, is based on the modern information technology, integration of transportation, packing, loading and unloading, handling, storage, distribution processing, distribution, recycling and processing and Logistics information processing and other functions to form a pattern of integrated Logistics activities. Distribution vehicle routing optimization is a key part of the logistics system optimization. Optimize the distribution vehicle route, can improve the economic benefit and realize the logistics rationalization. Distribution vehicle route optimization theory and method of the logistics system research is intensive development, build a comprehensive logistics system, set up a modern scheduling command system, development of intelligent transportation systems and to carry out the foundation of e-commerce.

Key words: vehicle routing optimization, ant colony algorithm, CVRP, VRPTW

1 Introduction

A countrys level of logistics industry development, reflect the national comprehensive national and enterprise competition ability. In the modern society, logistics and business flow, information flow, and called the three pillars of economic, systematic and rationalization of the logistics management will create enormous economic profits. Logistics field had joined cut down the consumption of resources (labor and materials), and improve labor productivity and by expanding market sales to obtain more profits after the 'third profit source'. Large amounts of data show that the economy of the developed countries or regions logistics value in a very important position in national economy. Modern logistics is a new concept of system, it includes the whole process of product life cycle of the whole physical displacement, beginning from the procurement logistics, production logistics, distribution logistics into the, to reach consumers hands, and finally returned logistics. In the whole logistics process, through the packaging, transportation, loading and unloading, handling, storage, distribution processing, distribution, information processing etc. Therefore, modern logistics covers a number of industry of the national economy, it is a product of economic development, is the pillar of economic development: is a kind of incremental economic activity, and increase the cost and affect the ecological environment of economic activity. Modern logistics as an advanced organization way and management technology, is widely regarded as enterprises in reducing material consumption, improve labor productivity outside important profit source, is a revolution of circulation industry. Vehicle scheduling is the most important part of logistics management. With the development of society and consumers for the continuous improvement of service quality, efficient vehicle scheduling, in order to improve the logistics efficiency, reduce logistics cost, improve the quality of service is of great significance to promote the development of economy healthy and stable. The vehicle routing problem, it is the vehicle and appropriate selection of the path, reasonable transportation planning formulated asked the topic. To solve this problem, can be used to speed up the speed of response to customers requirement, improve the service quality, enhance customer logistics link satisfaction, and reduce the cost of service operation. Ant colony algorithm is inspired by the real ant colony foraging behavior in the nature and put forward a kind of simulated evolutionary algorithm, in solving VRP (Vehicle Routing Problem and Vehicle Routing optimization) on combinatorial optimization problems, such as good effect is obtained. Study the main purpose of the article is to improve the existing ant colony algorithm, make it suitable for different types of vehicle routing optimization problem, get better quality.

2 Summary of vehicle routing problem

Vehicle Routing scheduling Problem is presented by the Danzig first, then summed up the deepen the Vehicle Routing Problem, or vehicles, logistics distribution route path scheduling Problem, Vehicle Routing Problem (VRP) Problem, etc., is an important content in the study of modern logistics. Vehicle routing problem, just as its name implies, the main solution is to send how many cars to transport route problem. Given the interconnection of a number of goods is the demand of the customers, number of vehicles from distribution center, the distribution of all customers order task back to the distribution center, after which go route cannot be repeated, purpose is to find the minimum cost of distribution solutions. The problem is the VRP problem. Vehicle routing problem is a combinatorial optimization in the field of one of the famous NP - hard problem, and many practical problems have similarity, such as railway transportation, bus scheduling, channel route, routing, etc., VRP research has considerable practical significance. According to the practical constraints, various types

剩余内容已隐藏，支付完成后下载完整资料

蚁群算法与物流车辆路径优化研究

Rizzoli E

摘要

物流正逐步被企业视作在市场中获取和保持竞争力优势的有力管理手段，引起了国内外学者和企业界的极大重视。目前被普遍接受的“物流”定义是美国仓储与采购协会提出的：“所谓物流是以满足客户需求为目的，为提高原料、在制品、以及相关生产，从供应到消费的流动和存储的效率和效益，并对其进行的计划、执行（实现）和控制的过程”。现代物流(Logistics)，是指为了实现顾客满意，连接供给主体和需求主体，克服空间和时间阻碍的有效、快速的商品、服务流动经济活动过程，是指以现代信息技术为基础，整合运输、包装、装卸、搬运、仓储、流通加工、配送、回收加工及物流信息处理等各种功能而形成的综合性物流活动模式。配送车辆路径优化，是物流系统优化中关键的一环。对配送车辆路线进行优化，可以提高经济效益、实现物流科学化。对配送车辆线路优化的理论与方法进行系统研究是物流集约化发展、构建综合物流系统、建立现代调度指挥系统、发展智能交通运输系统和开展电子商务的基础。

关键词：车辆路径优化，蚁群算法，CVRP，VRPTW

1 引言

一个国家的物流业发展水平，反映了这个国家综合国立和企业竞争能力。在现代社会中，物流与商流、信息流并称为经济的三大支柱，系统化、合理化的物流管理将创造巨大的经济利润。物流领域已成为继降低资源（人工和材料）消耗，提高劳动生产率及通过扩大市场销售获取更多利润之后的“第三利润源泉”。大量数据表明，经济发达国家或地区的物流产值在国民经济中处于一个十分重要的地位。现代物流是一个全新的系统概念，它包含了产品生命周期的整个物理性位移的全过程，从采购物流开始，经过生产物流，再进入销售物流，到达消费者手中，最后还有回收物流。在这整个物流过程中，要经过包装、运输、装卸搬运、仓储、流通加工、配送、信息处理等作业环节。因此，现代物流涵盖了国民经济的若干个行业，它既是经济发展的产物，又是经济发展的支柱：既是一种增值性的经济活动，又是增加成本、影响生态环境的经济活动。现代物流作为一种先进的组织方式和管理技术，被广泛认为是企业在降低物资消耗、提高劳动生产率以外的重要利润源泉，是对流通业的一场革命。车辆调度是物流管理最重要的部门。随着社会的发展以及消费者对服务质量要求的不断提高，高效的车辆调度，以提高物流效率、降低物流成本、提高服务质量对于促进经济健康稳定的发展具有重要意义。所谓的车辆路径问题，就是车辆和路径的恰当选取，运输规划的合理制定问题。解决此问题，可用加快对客户需求的响应速度，提高服务质量，增强客户对物流环节的满意度，降低服务商运作成本。蚁群算法是受自然界中真实蚁群觅食行为的启发而提出的一种模拟进化算法，在求解 VRP(Vehicle Routing Problem，车辆路径优化)等组合优化问题上取得了很好的效果。本文研究的主要目的，就在于改进已有的蚁群算法，使之适合不同类型的车辆路径优化问题，得到质量更好的解。

2车辆路径问题概述

车辆路径调度问题是由Danzig 首先提出的，后来总结了深化车辆路径问题，或者称作车辆路径调度问题、物流配送路线问题、Vehicle Routing Problem (简称 VRP)问题等，是现代物流研究中的一项重要内容。车辆路径问题，顾名思义，主要解决的是派多少辆车走什么样的路线进行运输的问题。给定了相互连通的若干有货物需求的顾客点，若干车辆从配送中心出发，完成对所有顾客点的配送任务后回到配送中心，要求所走的路线不能重复，目的是找到最小成本的方案。此问题即为VRP问题车辆路径优化问题是组合优化问题中著名的NP-hard问题之一，与众多的问题都有相似性，去铁路运输、公交调度、水道航线、路由选择等，VRP研究具有相当大的实际意义。根据实际条件约束的差异，车辆路径问题种类千变万化。所谓经典车辆路径问题，其实就是在车辆的调度中，仅仅考虑最基本的货车载重量约束的最一般化的运输问题，即有容量约束的车辆路径问题。经典的VRP是由一个服务中心向多个服务需求点进行配送服务，在已知客户和出发点的位置、顾客需求及车辆最大负荷的条件下，设计车辆的配送路径，规划设计方案，是运输成本最小化，即总代价最小。可见VRP问题实际是多目标组合优化问题，大多数情况下，我们以派出车辆最少，即运输路线最少为主要目标，行车距离最短，即总代价最小为次要目标。

3多配送中心的车辆路径问题

所谓的多配送中心的车辆路径问题，就是顾客的配送工作不再仅由一个配送中心发出的车辆完成配送，而是由分散的多个配送中心供给的一种形式，有以下几个特点：顾客点分散；配送中心分散；一个顾客点可由一个配送中心供给满足；配送中心服务人以顾客点群。衡量配送方案的优劣的标准是配送总代价的大小。需求随机的车辆路径问题即在未到达顾客点之前，此顾客点的货物需求量是不知道的。例如某畅销年货的供货问题，在车辆未到达之前，存货随时都在减少，很难确定准确的需求量。此类需求问题的一种形式。SDVRP为可切分的车辆路径问题，所谓的可切分，即顾客点货物配送不再限制只由一辆车一次配送满足，而是可以将配送货物分成几份，由不同的车辆多次配给。此类车辆路径问题增加了运输的灵活性，降低了车辆的空驶率。与经典的车辆路径问题比较，车辆除了需要完成顾客点货物配给之外，还需要将各个顾客店的某些货物回运到配送中心。此类问题分为两种，一种是先配送再收集车辆历经问题，在所有的顾客点的配送任务完成之后，在进行货物的回收，其实可以将此问题概略的分解成为两个经典车辆路径问题的叠加。另一种是收集与配送穿插结合的配送收集车辆路径问题，在配送途中同时收集货物，此类问题车辆容量的约束变得比较复杂，需要同时考虑车内货物的加减问题。此外，还有路径重复VRP以及非遍历的VRP等。

4 VRP问题的优化方法

根据各种车辆路径问题的模型，曾出现多求解方法、入系统仿真法、人机互动法以及精确解法等。尽管积几类方法都提供了解决问题的思想，但是由于他们都有各自的不足之处。而这些问题也确实影响到了他们存在的价值，并被后来出现的启发式算法逐渐代替。启发式算法不需要取得精确地最优解，而是以获得可以接受的较优解为目标，既节省了计算时间，又满足了解决问题的实际要求。启发式算法以其现实、高效的优点引起了优化研究领域的高度重视，并在近年来取得了飞速发展。遗传算法是由Holland受生物进化的启发而提出的。它将问题的求解表示成“染色体”的适者生存过程通过“染色体”间的复制、交叉和变异的操作，一代代不断进化最终收敛到“最适应环境”，禁忌搜索法是局部邻域搜索算法的推广，是一种全局逐步寻优算法。Glover在1986年提出此概念，进而形成一套完整的算法。简单TS算法的基本思想是：给定一个初始解和一个邻域，然后在当前解的邻域中缺点若干候选解，若最佳解对应的目标值优于当前最好解，则忽视其禁忌特征，用其代替当前解，并将相应的对象加入禁忌表，同时修改晋级表中个对象的任期；若不存在上述候选解，则在候选解中选择非禁忌的最佳状态为新的当前解，而无视它当前解的劣势，同时将相应的对象加入晋级表，并修改禁忌表中个对象的任期；如此重复这种迭代搜索过程，直至满足停止准则。

剩余内容已隐藏，支付完成后下载完整资料

资料编号：[32149]，资料为PDF文档或Word文档，PDF文档可免费转换为Word