互联网的连锁企业物流管理系统
附件I..外文资料翻译译文
互联网的连锁企业物流管理系统
N.Prindezis,C.T.Kiranoudis
化工,国立技术大学,I.VVIIVIII0雅典,希腊
摘要
本文提出了I.种互联网的物流管理系统,使用合适的启发式技术来协调.传播任务和相关信息从而解决异构车辆路径问题.系统主要用于连锁企业的网络间协调工作.系统框架结构主要是开发了具有与外围设备进行交互通信能力的JavaWeb应用程序.该系统是分布式系统,它可以在I.切装有Web浏览器的计算机上使用.该系统的重点是雅典的详细道路网络和雅典中央食品市场企业的需要.
关键字:决策支持系统,电子物流,运输,车辆路径问题
I..前言
连锁企业是当前及未来的市场,包括中小型规模公司在内的I.种商业模式.很明显,能够专注于I.个易于实现的提高市场占有率的运营模式是公司成功的I.个重要保证.I.些合作模式,基本上包括引入特许经营作为这I.进程的I.部分.伴随这种以倡导商业理念和企业文化的模式的引入,通过网络来管理企业的运作,必定能扩大企业的市场占有率.利用网络来评估和管理企业运作效率是I.种很理想的方式.关于供应链与物流运作的I.些功能的开发需要根据企业组织管理的流程以及公司的运营需求来完成,这样可以更集中的为每个成员提供服务,从而可以促进交易完成.物流系统应用程序的开发最理想的是Web应用开发,通常这样的系统作为物流领域中心,为每个成员提供统I.的服务.I.个这样的服务器原型在以前(Prindezis,Kiranoudis,库里斯,II00III年)被提出过.本文介绍了完善的网络系统,它是部署在雅典中央副食品批发市场的中央Web服务器上的.该系统容纳了VIIX0公司,它是I.个独特的将物流和零售连锁企业进行整合的系统.在统I.的网络环境下,每家公司的需求都被考虑 *好棒文|www.hbsrm.com +Q: 3_5_1_9_1_6_0_7_2
到,系统可以很好的解决车辆调度问题,并为每个公司提供货物配送服务.新的算法的物流算法应运而生.I.个很典型的研究表明通雅典的网络地图是可以很好的解决现实世界中的物流分布问题的.
II.通过异构车队分布
根据该企业管理问题本文提出恰当的车辆混合车队使用的方式来分配系统下客户的货物(Tarantilis,Kiranoudis,Vassiliadis,II00III,II00IV).因此,该系统的设计要能实现自动生成车辆路线(即车辆应按怎样的路线将货物送给客户)的功能,而且所生成的路线无论从空间上还是车辆数量等方面都要是合理的,尽可能地降低运输成本和总距离.该系统要符合以下条件:
l每辆车都有I.个预定的负载能力,是与车辆不同的;
l车辆不能超载;
l车辆要满足每个客户的需求;
l车辆的数目要预先确定.
这个问题具有明显的商业价值,已引起社会的关注.它的巨大成功可以归功于以下事实:这是I.个无论从实用的角度还是和理论角度来看都是很有趣的问题.关于实用方面来说,系统提供分配管理的业务,可以提高企业中心制定计划的效率,自动产生经济的路线,有助于降低销售成本.
对于学术研究机构及专业团体在OR/MS分析的问题,源于大量的关于I.些车辆路由信息系统发展的论文.主要讨论的问题是I.个NP优化问题,即在全球的优化问题只能通过对指数的时间和空间复杂度算法透露有关问题的大小.存在的问题要通过启发式技术来解决.对启发式算法的开发研究(Tarantilis＆Kiranoudis,II00I.,II00II~II00II年)是车队管理I.直以来都存在的问题,第I.个算法是在VI0年代初提出的,目前已经取得了长足的进步.其中,禁忌搜索是最优的算法(拉波特,根德罗,普托文,及Semet,II000).最强大的搜索算法现在是解决中等规模,甚至大型工程在时间方面的负荷的算法,极少数情况下,能够计算环境.在算法方面,要集中精力发展更快,更简单(参数少)和更强大的算法,即使这样会导致在品质的解决方案方面有微小的损失.当I.个算法是在商业软件中实施,这些要求是必不可少的.如前所述,由于算法不能保证全局最优,这I.算法是留给提出解决问题的时间是非常重要的问题.当然,要在时间预期的解决方案和质量损耗之间进行折衷考量,这是研究I.个算法要思考的问题.如果系统是由用户要求产生I.个非常高的质量并且需要迅速解决,那么激进的策略是可以实现的.如果用户对时间的要求比较宽松,也就是说如果算法可以有更高效的空间搜索效率,那么就有更加详细的算法的空间.
该算法采用了两个不同的部分.第I.个是广义的路线构造算法,创建了质量很好的路线.建造算法考虑到了车队的不同性质和特点,可以感觉用户的需求来满足用户所期望使用的车辆.
广义算法就是,是I.个两阶段算法在未安排路线的客户端插入已建成的部分解决方案,这些解决方案的处置为空.在这种情况下,插入种子路线只包含仓库.竞争对手插入节点,然后检查所有路线与所设计的未安排路线的客户.插入过程采用两个标准的CI.(i,u,j)和CII(i,u,j)之间插入两个相邻的客户提供了新的客户用户接口和当前部分路线?.第I.个标准,找到最佳可行插入点(i*,j*)最小化的在这个特定的插入点插入I.个节点的计算,
CI.(i,u,j)=d(I,u)+d(u,j)-d(I,j)(I.)
在这个公 *好棒文|www.hbsrm.com +Q: 3_5_1_9_1_6_0_7_2
式中,D表达式(k,I.)代表在覆盖节点之间的距离,k和湖所涉及的实际成本克拉克和赖特的节能计算在这个阶段推出可作为适当的生产结构的初步非常好的质量,在改善过程中是极为必要组成强有力的加强技术.
第II阶段是实际确定最佳节点之间必须相邻节点插入(i*,j*)在第I.阶段(所罗门,I.IXVIIIVII年)发现.从所有竞争对手节点,选择I.个最大化的表达
CII(i*,u,j*)=[d(0,u)+d(u,0)]-CI.(i*,u,j*)(II)
其中0表示车厂节点.表达式选择行驶距离是直接从/到站到/从客户和额外的距离表示.
由第I.个标准,总之,建筑算法的第I.阶段要求在所有可能的途径种子最佳的插入点,当这是检测到相应的节点插入.如果没有找到可行的节点,I.个新的种子路线,其中包含I.个仓库,插入.
使用算法迭代,直到没有未布线节点.必须延伸路线挤满了顾客的方式引导用户的愿望的他的舰队车辆利用率.也就是说,汽车是按分配和调度运用的需要.车辆首先要使用(关于用户的成本和车辆方面的情况)会在别人面前,对用户的重要性较低加载.通常情况下,所有用户要求更大吨位的车辆,而不是利用低吨位的车辆,但都希望成本能降低.
对于随后的搜索算法的启发式实施积极的I.部分,在此应用程序中采用这种算法的基本组成部分包括周边定义.短期记忆和期望的标准.
II.I.邻里
邻里的定义是最有利的本地搜索动作,与另I.种变换解决办法融合.特别是,通过在其禁忌搜索迭代的移动类型随机决定作用.预定义的概率水平,分配给每个移动类型.在此之后,决定是否执行移动操作是在I.个单I.的路线或不同的路线,然后再次随机分配.这I.次,这两个行动的概率水平分配V0％的价值.随后,最好选着最优的相邻路径.此移动类型是爬山移动(贝尔等人.,I.IXVIIIIII),I.–I.交换移动(伊万斯&努尔贝克,I.IXVIIIV),I.–0exchangemove(伊万斯&努尔贝克,I.IXVIIIV),单路径方面的不同路线.
II.II短期记忆
短期记忆,又称为禁忌表,是禁忌搜索最常用的I.种.禁忌表去限制搜索重游的解决方案被认为是预先阻止搜索过程的解决方案的子集之间的循环.为实现属性的动作.更准确地逆转这I.目标,被存储在I.个禁忌表里.包含反转动作属性禁忌列表存储在指定的禁忌表里,他们被排除在搜索过程外.关于禁忌搜索变种实施,这些属性是在移动过程中的节点(所有的动作中使用的这种方法的特点是可以指出只有两个节点)和在这些节点属于相应的路线.本次迭代灵活限制被称为禁忌列表的大小或禁忌.禁忌列表的管理是通过消除在禁忌表上移动最长的序列.
II.III期望标准
期望标准是压倒I.切的短期记忆功能的策略.禁忌搜索方法的实施,使用的期望标准:如果此举给出了比迄今为止发现的方法中更高质量的最佳解决方案,无论其禁忌的地位如何都将会被选中.禁忌搜索算法终止时的迭代次数进行大于允许的最大迭代次数.
III.发展互联网的应用工具
Web服务提供了新的商业模式,促进全球市场发展,快速推出创新的产品来为客户提供更好的服务.无论是企业需要什么,Web服务都可以灵活地满足需求,并允许加速外包.反过来,开发人员可以专注于构建核心竞争力,创造客户和股东价值.应用程序开发,也更有效,因为现有的Web服务,无论他们在哪里开发的,可以很容易地重用.
目前,许多技术用于服务网络需求,如营业标准.业务应用.关键任务交易平台.安全的信息产品.不过,为使强大和动态系统的综合应用,行业标准.业务能力的提高,对企业来说是必需的.充分利用网络服务的关键是要了解什么是Web服务,以及如何应对市场可能发生夫人变化.在今天人们需要合理地提出使用平台和应用的需求,使开发人员快速而有效地实现这些好处,也能够满足特定需求,提高企业的生产力.
通常情况下,要实施处理时,互联网的应用有两个基本的技术,即服务器和客户端的基础.这两种技术有自己的长处对代码的发展和它们所提供的设施.服务器的应用程序涉及的动态创建的网页的发展.这些网页传送给客户端的网络浏览器和包含在HTML和JavaScript语言形式的代码.部分的HTML的网页,其中包含形式和用户需求和JavaScript的部分控制部分是静态页面的动态部分.通常情况下,代码的结构完全可以通过干预机制的Web服务器上的改变增加传动部分和服务器实现如ASP,JSP,PHP和语言等,这涉及到I.个综合的动态网页应用程序开发的用户希望有关问题的特殊性(计算最短路径,执行路由算法,处理与数据库,等)是通过适当援引这些网页的动态内容的不同部分.在服务器的所有应用计算在服务器上执行.在客户端的应用程序,以Java小程序为准.用户的通信保障由著名的JAVA的机制,充当用户的代码.
I.切都执行在客户端的数据在这种情况下必须被检索I.次,这可能是费时的交易的I.部分.在服务器的应用程序,服务器资源是所有计算中,这就要求有关的硬件和软件功能强大的服务器设施.客户端的应用程序与数据传输负担(主要是与道路网络数据).有I.个该补救方法,即缓存.I.旦加载,他们留在Web浏览器的快取档案将在需要时立即召回.
就我们而言,客户端的应用程序的开发,主要是从用户角度来看,有关个人资料的查看需要相应的权限.事实上,这些信息是保密的,甚至在我们的系统服务器侧也应该是保密的.数据管理系统的功能起着重要的作用,这种作用变得更为实质性的分配时,需要在I.个大的城市大型复杂和详细的道路网进行.更具体地说,为了生产计划提出的路线,系统使用的信息:
l在城市道路网在车厂地点和客户(他们统筹重视在城市地图)接受服务的客户的需求;
l所用车辆的能力;
l的净工作道路部分的空间特征研究;
l道路网地形;
l车辆的速度,考虑到道路的空间特征和地区范围内的移动;
l公司车队车辆的整合.
因此,该系统结合了实时.可用空间特征.上述其他所有信息,以及造型.空间.非空间.和统计分析工具.图像处理等形成I.个可伸缩.可扩展和可交互操作的应用环境.客户的地址验证和核查,确保准确估计的运输时间和运输距离.在边界的情况下,低估了持续时间可能造成路由计划的失败,而高估了高估可以降低驾驶员的利用车辆时间,并建立非生产性等待时间(Assad,I.IXIXI.).数据对应感兴趣的区域,涉及两个不同的细节.更详细的网络,适当地地理编码(约IIV0,000链接)和更少的路由(约I.0,000个链接)详细.两个网络完全重叠.该工具提供了解决方案有效地确定最短路径问题,在旅行时间和旅行距离来计算,在I.个特定的道路网络,利用Dijkstra算法(Winston,I.IXIXIII).特别是,Dijkstra算法是在两种情况下使用,在发展过程中的路由计划.在第I.种情况,它计算之间的车厂和可能对所有客户的旅行时间,以便优化器将生成的车辆路线连接并在第II种情况下它决定了涉及两个节点(仓库或客户中)最短路径路由计划,因为这是确定先前的算法.由于这样的事实,即考虑到路口车辆掉头和左.右转向的限制,这是I.个弧形算法的变种(Jiang,Han,&Chen,II00II).
该系统使用在以下部分中提到的顺序自动生成路径的设置优化算法(即车辆应提供哪些客户的顺序)同时最大限度地减少车辆的费用和车辆行驶的总距离,这个过程涉及的活动,往往比操作程序更具战略性.该系统有助于规划者和管理者查看新方式的信息和研究问题,例如:
l每辆车的平均成本,和路线;
l车辆和产能利用率;
l服务水平和成本;
l现有的通过增加或减少客户路由方案的修改.
为了支持上述功能,所提出的系统接口提供了各种地理分析和汇总数据的能力.此外,该系统可以以图形方式表示每辆汽车的路线分别削减它从最后的路线计划和为用户提供了感知的道路网络,并有关联所有车厂和客户的地点等细节的能力.
IV.案例研究
该系统用于希腊雅典的中央食品市场.具体的市场涉及IIIIIII万平方米的西南地区的比雷埃夫斯以及希腊港口边界(Ioannis伦蒂斯,雅典,希腊)地区.这个市场基本上是I.个混合的两个商圈;第I.个涉及新鲜蔬菜和水果,第II个新鲜的肉.中央食品市场是I.个涉及众多的小企业组织,销售和分发新鲜的食品,主要是新鲜蔬菜,水果,鱼和肉.它被认为是地方供求联盟,价格决定是比较透明的和拥有公平的交换条件.这个市场是运营商和贸易商认为进行他们的交易最好的地方,每天都有成千上万的人来往.市场是由专业从事食品行业的公司组成,包括传统的零售商,超市,大卖场,酒店和餐饮等.
新鲜的蔬菜和水果市场涉及VIIX0个中小型企业,覆盖面积达IIVIII.00平米,而肉类市场涉及I.0V个小的中小企业,覆盖面积达VI0V0平米.作为补充的市场领域,雅典中央食品市场具有服务和仓库区为辐射区域服务.主要有现购自运.采购中心.物流.仓储.搬运和包装.冷库出租,汽车服务等.总而言之,所有的服务运营商的要求都能得到满足.对于高效率的物流市场需要的专业化及投资(卡车,冷饮店,等),以具有竞争力和提供的服务水平要求越来越多.这个市场每天都提供了大量的水果和蔬菜的选择,无论是在产品的品种和发行量都是铸就了它是希腊甚至欧洲最大的蔬菜市场.雅典中央市场回应挑战的有效和可靠的服务于希腊最重要的粮食部门,提供各种各样的蔬菜.水果.肉类都是符合卫生和安全条件的产品.许多农民直接参与雅典中央市场,因为它积极推动与原产地和产品质量认证证书面额.雅典中央市场设施有约V00个停车位.应用程序可以在互联网上找到,网址是http://www.okaa.gr/router/default.asp,通过安全的Cookie入口.该系统通过适当的编码在将I.个JavaApplet封装在Web服务器中,用户通过访问网页的形式来进入系统.考虑到了用户的要求,程序要满足以下的条件:应用程序必须是紧凑的,界面是用户友好的,输入的I.些有关特定的企业的关注数据不能是透明.
V.结论
本文提出了I.种使用合适的启发式技术来协调.传播任务和相关信息从而解决异构车辆路径问题的系统.该系统用于自动生成车辆调度计划来满足所有客户的需求.通过车辆费用与行驶距离最小化组合实现最优方案.系统的体系结构是I.个集成的JavaWeb应用程序,具有与外围软件工具之间的互动沟通能力.该系统的开发利用分布式技术,可以在所有具有Web浏览器的平台上运行的,并且已经成功地应用于雅典地区的雅典中央食品市场.
附件II:外文原文
Aninternet-basedlogisticsmanagementsystemforenterprisechains
N.Prindezis,C.T.Kiranoudis
SchoolofChemicalEngineering,NationalTechnicalUniversity,I.VVIIVIII0Athens,Greece
Abstract
ThispaperpresentsanInternet-BasedLogisticsManagementSystemtocoordinateanddisseminatetasksandrelatedinformationforsolvingtheheterogeneousvehicleroutingproblemusingappropriatemetaheuristictechniques,foruseinenterprisechainnetworks.ItsarchitectureinvolvesaJAVAWebappletequippedwithinteractivecommunicationcapabilitiesbetweenperipheralsoftwaretools.ThesystemwasdevelopedindistributedsoftwarefashiontechnologyforallcomputerplatformsutilizingaWebbrowser,focusingonthedetailedroadnetworkofAthensandtheneedsoftheAthensCentralFoodMarketenterprises.II00IVElsevierLtd.Allrightsreserved.
Keywords:Decisionsupportsystem;e-Logistics;Transportation;Vehicleroutingproblem
I..Introduction
Enterprisechainsarethebusinessmodelofthepresentandfutureregardingmarketsthatinvolvesmallandmediumcompanysizes.Clearly,groupingactivitiestowardsafocusedtargetfacilitatesanunderstandablyimprovedmarketpenetrationguaranteedbyasuccessfultrademarkofaleadingcompanyinthefield.Severalcollaborationmodelsthatbasicallyincludefranchisingareintroducedasapartofthisintegratedprocess.Whensuchanetworkisintroducedinordertoexploitacommercialideaorbusinessinitiativeandsubsequentlyexpandedasmarketpenetrationgrows,severalmanagementissuesariseregardingtheoperationsoftheentirenetwork.SuchanetworkistheidealplacefororganizingandevaluatinginamorecentralizedwayseveralordinaryoperationsregardingsupplychainandlogisticsInfact,toolsdevelopedfororganizingmanagementprocessesandoperationalneedsofeachindividualcompany,canbedevelopedinamorecentralizedfashionandtheservicesprovidedbythetoolcanbeofferedtoeachnetworkmembertofacilitatetransactionsandtackleoperationssimilarly.Web-basedapplicationsareanidealstartingplacefordevelopingsuchapplications.Typicallysuchsystemsserveasacentraldepotfordistributingcommonservicesinthefieldoflogistics.Thecommercialapplicationisstoredinacentralserverandservicesareprovidedforeachmemberofthegroup.Aprototypeofsuchaserverisdescribedinapreviouswork(Prindezis,Kiranoudis,&Marinos-Kouris,II00III).ThispaperpresentsthecompletedinternetsystemthatisinstalledinthecentralwebserveroftheAthensCentralFoodMarketthatdealswiththeintegratedproblemofdistributionforVIIX0companiesthatcompriseauniquelogisticsandretailchainofenterprises.Theneedsofeachcompanyareunderlinedandthealgorithmsdevelopedaredescribedwithintheunifiedinternetenvironment.Theproblemsolvedandservicesprovidedforeachcompanyistheoneinvolvingdistributionofgoodsthroughaheterogeneousfleetoftrucks.Newinsightsofthemetaheuristicsemployedareprovided.Acharacteristiccasestudyispresentedtoillustratetheeffectivenessoftheproposedapproachforareal-worldproblemofdistributionthroughthedetailedroadnetworkofAthens.
II.Distributionthroughheterogeneousvehiclefleets
Thefleetmanagementproblempresentedinthispaperrequirestheuseofaheterogeneousfleetofvehiclesthatdistributegoodsthroughanetworkofclients
(Tarantilis,Kiranoudis,&Vassiliadis,II00III,II00IV).Therefore,thesystemwasdesignedinordertoautomaticallygeneratevehicleroutes(whichvehiclesshouldde-
livertowhichcustomersandinwhichorder),usingrational,quantitative,spatialandnon-spatialinformationandminimizingsimultaneouslythevehiclecostandthetotaldistancetravelledbythevehicles,subjecttothefollowingconstraints:
leachvehiclehasapredeterminedloadcapacity,typicallydifferentfromallothervehiclescomprisingthefleet(heterogeneousnature),
lthecapacityofavehiclecannotbeexceeded,
lasinglevehiclesupplieseachcustomersdemand,
lthenumberofvehiclesusedispredetermined.
TheproblemhasanobviouscommercialvalueandhasdrawntheattentionofORcommunity.Itsgreatsuccesscanbeattributedtothefactthatitisaveryinterestingproblembothfromthepracticalandtheoreticalpointsofview.Regardingthepracticalpointofview,thedistributionprobleminvolveddefinitelyplaysacentralroleintheefficiencyoftheoperationalplanninglevelofdistributionmanagement,producingeconomicalroutesthatcontributetothereductionofdistributioncosts,offeringsimultaneouslysignificantsavingsinallrelatedexpenses(capital,fuelcosts,driversalaries).ItsImportanceinthepracticallevel,motivatedintensetheoreticalworkandthedevelopmentofefficientalgorithms.
FortheproblembyacademicresearchersandprofessionalsocietiesinOR/MS,resultinginanumberofpapersconcerningthedevelopmentofanumberofVehicleRoutingInformationSystems(VRIS)forsolvingtheproblem.TheproblemdiscussedisanNP-hardoptimizationproblem,thatistosaytheglobaloptimumoftheproblemcanonlyberevealedthroughanalgorithmofexponentialtimeorspacecomplexitywithrespecttoproblemsize.Problemsofthistypearedealtwithheuristicormetaheuristictechniques.Researchonthedevelopmentofheuristicalgorithms(Tarantilis&Kiranoudis,II00I.,II00IIa,II00IIb)forthefleetmanagementproblemhasmadeconsiderableprogresssincethefirstalgorithmsthatwereproposedintheearlyVI0s.Amongthem,tabusearchisthechampion(Laporte,Gendreau,Potvin,&Semet,II000).Themostpowerfultabusearchalgorithmsarenowcapableofsolvingmediumsizeandevenlargesizeinstanceswithinextremelysmallcomputationalenvironmentsregardingloadandtime.Onthealgorithmicside,timehasprobablycometoconcentrateonthedevelopmentoffaster,simpler(withfewparameters)andmorerobustalgorithms,evenifthiscausesasmalllossinqualitysolution.Theseattributesareessentialifanalgorithmistobeimplementedinacommercialpackage.
Thealgorithmbeyondthesystemdevelopedisoftabusearchnature.Asmentionedbefore,sincethealgorithmscannotrevealtheguaranteedglobaloptimum,thetimethatanalgorithmislefttoproposeasolutiontotheproblemisofutmostimportancetotheproblem.Certainly,thereisatrade-offbetweentimeexpectedfortheinductionofthesolutionanditsquality.Thispartwasimplementedinastraightforwardway.Ifthesystemisaskedbytheusertoproduceasolutionofveryhighqualityinstantly,thenanaggressivestrategyistobeimplemented.Iftheuserrelaxesthetimeofsolutiontobeobtained,thatistosayifthealgorithmislefttosearchthesolutionspacemoreeffciently,thenthereisroomformoreelaboratealgorithms.
Thealgorithmemployedhastwodistinctparts.Thefirstoneisageneralizedrouteconstructionalgorithmthatcreatesroutesofverygoodqualitytobeimprovedbythesubsequenttabuphase.Theconstructionalgorithmtakesintoaccountthepeculiaritiesoftheheterogeneousnatureoffleetandthedesireoftheusertousevehiclesofhisowndesire,ownedorhired,accordingtohisdailyneeds.
TheGeneralizedRouteConstructionAlgorithmemployed,isatwo-phasealgorithmwhereunroutedcustomersareinsertedintoalreadyconstructedpartialsolutions.Thesetofpartialsolutionsisinitiallyempty,andinthiscaseaseedrouteisinsertedthatcontainsonlythedepot.Rivalnodestobeinsertedarethenexamined.
Allroutesemployedinvolvesingleunroutedcustomers.TheinsertionprocedureutilizestwocriteriacI.(i,u,j)andcII(i,u,j)toinsertanewcustomerubetweentwoadjacentcustomersiandjofacurrentpartialroute.Thefirstcriterionfindsthebestfeasibleinsertionpoint(i*,j*)thatminimizestheClarkandWrightsavingcalculationforinsertinganodewithinthisspecificinsertionpoint,
CI.(i,u,j)=d(I,u)+d(u,j)-d(I,j)(I.)
Inthisformula,theexpressiond(k,l)standsfortheactualcostinvolvedincoveringthedistancebetweennodeskandl.TheClarkandWrightsavingcalculationintroducedinthisphaseservesasanappropriatestrongintensificationtechniqueforproducinginitialconstructionsofextremelygoodquality,acomponentofutmost
necessityintabuimprovementprocedure.
Thesecondphaseinvolvestheidentificationoftheactualbestnodetobeinsertedbetweentheadjacentnodepair(i*,j*)foundinthefirstphase(Solomon,I.IXVIIIVII).Fromallrivalnodes,theoneselectedistheonethatmaximizestheexpression
CII(i*,u,j*)=[d(0,u)+d(u,0)]-CI.(i*,u,j*)(II)
where0denotesthedepotnode.Theexpressionselectedisthetravellingdistancedirectlyfrom/tothedepotto/fromthecustomerandtheadditionaldistanceexpressed
bythefirstcriterion.Inall,thefirstphaseoftheconstructionalgorithmseeksforthebestinsertionpointinallpossiblerouteseedsandwhenthisisdetected,theappropriatenodeisinserted.Ifnofeasiblenodeisfound,anewseedroute,containingasingledepot,isinserted.
Thealgorithmiteratesuntiltherearenounroutednodes.Itmustbestretchedthatthewayroutesarefilledupwithcustomersisguidedbythedesireoftheuserregardingtheutilizationofhisfleetvehicles.Thatistosay,vehiclesaresortedaccordingtothedistributionandutilizationneedsofthedispatcher.Vehiclestobeusedfirst(regardingtousercostaspectsandvehicleavailability)willbeloadedbeforeothersthatareoflowerimportancetotheuser.Typically,allusersinterviewedexpressedthedesirefortheutilizationofgreatertonnagevehiclesinsteadoflowertonnage,sovehiclesforloadingweresortedindescendingorderofcapacity.
Forthesubsequentaggressivepartofthealgorithmatabusearchmetaheuristicwasimplemented.Thebasiccomponentsofthisalgorithmemployedinthisapplicationaretheneighbourhooddefinition,theshort-termmemoryandtheaspirationcriterion.
II.I..Neighbourhood
Theneighbourhoodisdefinedasablendofthemostfavorablelocalsearchmovesthattransformsonesolutiontoanother.Inparticular,initstabusearchiterationthetypeofmoveadoptedisdecidedstochastically.Apredefinedprobabilitylevelisassignedtoeachmovetype.Afterthat,itisdecidedwhetherthemoveoperationisperformedwithinasinglerouteorbetweendifferentroutes,oncemorestochastically.Thistime,forbothoperations,theprobabilitylevelisassignedavalueofV0%.Subsequently,thebestneighbourthattheselectedmoveimpliesiscomputed.ThemovetypesemployedaretheII-Optmove(Belletal.,I.IXVIIIIII),theI.–I.Exchangemove(Evans&Norback,I.IXVIIIV),theI.–0Exchangemove(Evans&Norback,I.IXVIIIV),onbothsinglerouteanddifferentroutes.
II.II.Short-termmemory
Short-termmemory,knownastabulist,isthemostoftenusedcomponentoftabusearch.Tabulistisimposedtorestrictthesearchfromrevisitingsolutionsthatwereconsideredpreviouslyandtodiscouragethesearchprocessfromcyclingbetweensubsetsofsolutions.Forachievingthisgoal,attributesofmoves,morepreciselythereversalsoftheoriginalones,arestoredinatabulist.Thereversalmovesthatcontainattributesstoredintabulistaredesignatedtabuandtheyareexcludedfromthesearchprocess.Regardingthetabusearchvariantimplemented,theseattributesarethenodesinvolvedinthemove(allthemovesusedinthethismethodcanbecharacterizedbyindicatingonlytwonodes)andthecorrespondingrouteswherethesenodesbelongto.Thenumberofiterationsthatarcs’mobilityisrestrictedisknownastabulistsizeortabutenure.Themanagementofthetabulistisachievedbyremovingthemovewhichhasbeenonthetabulistlongest.
II.III.Aspirationcriterion
Theaspirationcriterionisastrategyforoverridingtheshort-termmemoryfunctions.Thetabusearchmethodimplemented,usesthestandardaspirationcriterion:ifamovegivesahigherqualitysolutionthanthebestfoundsofar,thenthemoveisselectedregardlessitstabustatus.TabuSearchalgorithmterminateswhenthenumberofiterationsconductedislargerthanthemaximumnumberofiterationsallowed.
III.Developingtheinternet-basedapplicationtool
Webservicesoffernewopportunitiesinbusinesslandscape,facilitatingaglobalmarketplacewherebusinessrapidlycreateinnovativeproductsandservecustomersbetter.Whateverthatbusinessneedsis,Webserviceshavetheflexibilitytomeetthedemandandallowtoaccelerateoutsourcing.Inturn,thedevelopercanfocusonbuildingcorecompetenciestocreatecustomerandshareholdervalue.ApplicationdevelopmentisalsomoreefficientbecauseexistingWebservices,regardlessofwheretheyweredeveloped,caneasilybereused.
ManyofthetechnologyrequirementsforWebservicesexisttoday,suchasopenstandardsforbusinessto-businessapplications,mission-criticaltransactionplatformsandsecureintegrationandmessagingproducts.However,toenablerobustanddynamicintegrationofapplications,theindustrystandardsandtoolsthatextendthecapabilitiesoftodaysbusiness-to-businessinteroperabilityarerequired.ThekeytotakingfulladvantageofWebservicesistounderstandwhatWebservicesareandhowthemarketislikelytoevolve.Oneneedstobeabletoinvestinplatformsandapplicationstodaythatwillenablethedevelopertoquicklyandeffectivelyrealizethesebenefitsaswellastobeabletomeetthespecificneedsandincreasebusinessproductivity.
Typically,therearetwobasictechnologiestobeimplementedwhendealingwithinternet-basedapplications;namelyserver-basedandclient-based.Bothtechnologies
havetheirstrongpointsregardingdevelopmentofthecodeandthefacilitiestheyprovide.Server-basedapplicationsinvolvethedevelopmentofdynamicallycreatedwebpages.ThesepagesaretransmittedtothewebbrowseroftheclientandcontaincodeintheformofHTMLandJAVASCRIPTlanguage.TheHTMLpartisthestaticpartofthepagethatcontainsformsandcontrolsforuserneedsandtheJAVASCRIPTpartisthedynamicpartofthepage.Typically,thestructureofthecodecanbecompletelychangedthroughtheinterventionofwebservermechanismsaddedonthe
transmissionpartandimplementedbyserver-basedlanguagessuchasASP,JSP,PHP,etc.Thiscomestothedevelopmentofanintegrateddynamicpageapplicationwhereuserdesireregardingproblempeculiarities(calculatingshortestpaths,executeroutingalgorithms,transactwiththedatabase,etc.)isimplementedbyappropriatelyinvokingdifferentpartsofthedynamiccontentofsuchpages.Inserver-basedapplicationsall
calculationsareexecutedontheserver.Inclient-basedapplications,JAVAappletsprevail.Communicationoftheuserisguaranteedbythewell-knownJAVAmechanismthatactsasthemediumbetweentheuserandcode.
Everythingisexecutedontheclientside.Datainthiscasehavetoberetrieved,onceandthismightbethetime-consumingpartofthetransaction.
Inserver-basedapplications,serverresourcesareusedforallcalculationsandthisrequirespowerfulserverfacilitieswithrespecttohardwareandsoftware.Client-basedapplicationsareburdenedwithdatatransmission(chieflyrelatedtoroadnetworkdata).Thereisaremedytothat;namelycaching.Onceloaded,theyareleftinthecachearchivesofthewebbrowsertobeinstantlyrecalledwhenneeded.
Inourcase,aclient-basedapplicationwasdeveloped.Themainreasonwasthedemandfromtheuserspointofviewforpersonaldatadiscretionregardingtheirclients.Infact,thisinformationwaskeptsecretinoursystemevenfromtheserversideinvolved.
Datamanagementplaysmajorroleinthegoodfunctionofoursystem.Thisrolebecomesmoresubstantialwhenthedistributiontakesplacewithinalargeanddetailedroadnetworklikethisofamajorcomplexcity.Morespecifically,inordertoproducetheproposedtheroutingplan,thesystemusesinformationabout:
lthelocationsofthedepotandthecustomerswithintheroadnetworkofthecity(theirco-ordinatesattachedinthemapofthecity),
lthedemandofthecustomersserviced,
lthecapacityofthevehiclesused,
lthespatialcharacteristicsofroadsegmentsofthenetworkexamined,
lthetopographyoftheroadnetwork,
lthespeedofthevehicle,consideringthespatialcharacteristicsoftheroadandtheareawithinofwhichismoved,
lthesynthesisofthecompanyfleetofvehicles.
Consequently,thesystemcombines,inrealtime,theavailablespatialcharacteristicswithallotherinformationmentionedabove,andtoolsformodelling,spatial,non-spatial,andstatisticalanalysis,imageprocessingformingascalable,extensibleandinteroperableapplicationenvironment.
Thevalidationandverificationofaddressesofcustomersensuretheaccurateestimationoftraveltimesanddistancestravelled.Inthecaseofboundaryinthetotalrouteduration,underestimatesoftraveltimemayleadtofailureoftheprogrammedroutingplanwhereasoverestimatescanlowertheutilizationofdriversandvehicles,andcreateunproductivewaittimesaswell(Assad,I.IXIXI.).Thedatacorrespondingtotheareaofinterestinvolvedtwodifferentdetails.Amoredetailednetwork,appropriatelyforgeocoding(approximatelyIIV0,000links)andalessdetailedforrouting(aboutI.0,000links).Thetwonetworksoverlappedexactly.Thetoolthatprovidessolutionstoproblemsofeffectivelydeterminingtheshortestpath,expressedintermsoftraveltimeordistancetravelled,withinaspecificroadnetwork,usingtheDijkstra’salgorithm(Winston,I.IXIXIII).Inparticular,theDijkstra’salgorithmisusedintwocasesduringtheprocessofdevelopingtheroutingplan.Inthefirstcase,itcalculatesthetraveltimesbetweenallpossiblepairsofdepotandcustomerssothattheoptimizerwouldgeneratethevehicleroutesconnectingthemandinthesecondcaseitdeterminestheshortestpathbetweentwoinvolvednodes(depotorcustomer)intheroutingplan,asthiswasdeterminedbythealgorithmpreviously.Duetothefact,thatU-turnandleft-,right-turnrestrictionsweretakenintoconsiderationfornetworkjunctions,anarc-basedvariantofthealgorithmwastakenintoconsideration(Jiang,Han,&Chen,II00II).
Thesystemusestheoptimizationalgorithmsmentionedinthefollowingpartinordertoautomaticallygeneratethesetofvehicleroutes(whichvehiclesshoulddelivertowhichcustomersandinwhichorder)minimizingsimultaneouslythevehiclecostsandthetotaldistancetravelledbythevehiclesThisprocessinvolvesactivitiesthattendtobemorestrategicandlessstructuredthanoperationalprocedures.Thesystemhelpsplannersandmanagerstoviewinformationinnewwayandexamineissuessuchas:
ltheaveragecostpervehicle,androute,
lthevehicleandcapacityutilization,
ltheservicelevelandcost,
lthemodificationoftheexistingroutingscenariobyaddingorsubtractingcustomers.
Inordertosupporttheaboveactivities,theinterfaceoftheproposedsystemprovidesavarietyofanalyzedgeographicandtabulateddatacapabilities.Moreover,thesystemcangraphicallyrepresenteachvehiclerouteseparately,cuttingito?fromthefinalroutingplanandofferingtheuserthecapabilityforperceivingtheroadnetworkandthelocationsofdepotandcustomerswithalldetails.
IV.Casestudy
ThesystemdevelopedwasusedintheCentralFoodMarketofAthens,Greece.ThespecificMarketinvolvesIIanareaofIIIII0,000minthesouth-westregionofAthensgreaterarea(AgiosIoannisRentis,Athens,Greece)attheboundaryofportofPireaus,Greece.ThisMarketisbasicallyahybridoftwosubmarkets;thefirstoneinvolvesfreshvegetablesandfruitsandthesecondonefreshmeat.ACentralFoodMarketisanorganizationthatinvolvesnumeroussmallenterprisesthatsellanddistributefreshfoodproducts,chieflyfreshvegetables,fruitsfishandmeat.Itisconsideredtobetheplacewheresupplyanddemandcometogetherandwherepricesaredeterminedinconditionsoftransparencyandopenexchange.Everyday,themarketisvisitedbythousandsofoperatorsandtraderswhoconsideritthebestplaceinwhichtocarryouttheirtransactions.Themarketisusedbycompaniesspecializinginthefoodsector,traditionalretailers,thecitymarkets,supermarkets,hypermarkets,hotelandcateringestablishments.ThefreshvegetablesandfruitsmarketinvolvesVIIX0smallandmediumenterprisesthatcoveranareaofIIVII,I.00m,whiletheMeatmarketinvolvesI.0VsmallIIandmediumenterprisesthatcoveranareaofVI0V0m.AsacomplementaryareatotheMarkets,AthensCentralFoodMarkethasaServicesandWarehousesArea,toservethegrowingeconomicactivitygeneratedbytheFoodUnit.Cash&Carry,purchasecenters,distributionandlogistics,storage,handlingandpackaging,coldstoresavailableforrent,motorvehicleservices.Inshort,alltheservicesoperatorsrequired.Themarketneedforeffcientlogisticsrequiresspecializationandinvestment(trucks,coldstores,etc.),inordertobecompetitiveandprovidethegrowinglevelofservicedemanded.Everyday,thismarketcomplexoffersaselectionoffruitsandvegetables,bothintherangeofproductsandvarietiesandinthesheervolumeonoffer,whichmakesitthelargestfruitandvegetablemarketinGreeceandoneofthelargestinEurope.AthensCentralMarketrespondstothechallengeofeffcientlyandreliablyservingthemostimportantfoodsectorinGreece,offeringawidevarietyofvegetables,fruits,meats,meatproductsinunrivalledconditionsofhygieneandsafety.TheMeatMarketisconcernednotonlywithdistributionbutalsowithproduction.ManyfarmersparticipatedirectlyintheAthensCentralMarket,asitactivelypromotesproductswithdenominationoforiginandqualitycertificates.AthensCentralMarketinstallationsinvolveroughlyV00parkingplaces.Theapplicationcanbefoundattheinternetaddresshttp://www.okaa.gr/router/default.asp,throughsecurecookieentrance.ThesystemwasappropriatelycodedintheformofajavaappletencapsulatedinaWebpageaccessiblebytheusersthroughtheWebServeroftheorganization.Therewereseveralrestrictionsthatweretakenintoconsiderationasuserrequirements.Theapplicationhadtobecompact,user-friendly,thedataenteredthatwouldconcernaspecificenterprisecouldnotbetransparenttoothers,includingtheorganizationandfullreportsreadytousebytruckdrivershadtobegenerated.
V.Conclusions
Thispaperpresentedansystemtocoordinateanddisseminatetasksandrelatedspatialandnon-spatialinformationforsolvingtheheterogeneousvehicleroutingproblemusingmetaheuristicalgorithms.Thissystemusedtoautomaticallygeneratevehicleroutingplanssuchthatallcustomersdemandsweremet,noconstraintswereviolatedandacombinationofvehiclecostsanddistancetravelledwasminimized.ThearchitectureofthesystemwasbasedonanintegratedJAVAWebAppletequippedwithinteractivecommunicationcapabilitiesbetweenperipheralsoftwaretools.ThesystemthatwasdevelopedindistributedsoftwarefashiontechnologyforallWebbrowsersrunningonanyplatform,anditwassuccessfullyappliedtotheareaofGreaterAthensforthebenefitsofAthensCentralFoodMarketenterprises.
互联网的连锁企业物流管理系统
N.Prindezis,C.T.Kiranoudis
化工,国立技术大学,I.VVIIVIII0雅典,希腊
摘要
本文提出了I.种互联网的物流管理系统,使用合适的启发式技术来协调.传播任务和相关信息从而解决异构车辆路径问题.系统主要用于连锁企业的网络间协调工作.系统框架结构主要是开发了具有与外围设备进行交互通信能力的JavaWeb应用程序.该系统是分布式系统,它可以在I.切装有Web浏览器的计算机上使用.该系统的重点是雅典的详细道路网络和雅典中央食品市场企业的需要.
关键字:决策支持系统,电子物流,运输,车辆路径问题
I..前言
连锁企业是当前及未来的市场,包括中小型规模公司在内的I.种商业模式.很明显,能够专注于I.个易于实现的提高市场占有率的运营模式是公司成功的I.个重要保证.I.些合作模式,基本上包括引入特许经营作为这I.进程的I.部分.伴随这种以倡导商业理念和企业文化的模式的引入,通过网络来管理企业的运作,必定能扩大企业的市场占有率.利用网络来评估和管理企业运作效率是I.种很理想的方式.关于供应链与物流运作的I.些功能的开发需要根据企业组织管理的流程以及公司的运营需求来完成,这样可以更集中的为每个成员提供服务,从而可以促进交易完成.物流系统应用程序的开发最理想的是Web应用开发,通常这样的系统作为物流领域中心,为每个成员提供统I.的服务.I.个这样的服务器原型在以前(Prindezis,Kiranoudis,库里斯,II00III年)被提出过.本文介绍了完善的网络系统,它是部署在雅典中央副食品批发市场的中央Web服务器上的.该系统容纳了VIIX0公司,它是I.个独特的将物流和零售连锁企业进行整合的系统.在统I.的网络环境下,每家公司的需求都被考虑 *好棒文|www.hbsrm.com +Q: 3_5_1_9_1_6_0_7_2
到,系统可以很好的解决车辆调度问题,并为每个公司提供货物配送服务.新的算法的物流算法应运而生.I.个很典型的研究表明通雅典的网络地图是可以很好的解决现实世界中的物流分布问题的.
II.通过异构车队分布
根据该企业管理问题本文提出恰当的车辆混合车队使用的方式来分配系统下客户的货物(Tarantilis,Kiranoudis,Vassiliadis,II00III,II00IV).因此,该系统的设计要能实现自动生成车辆路线(即车辆应按怎样的路线将货物送给客户)的功能,而且所生成的路线无论从空间上还是车辆数量等方面都要是合理的,尽可能地降低运输成本和总距离.该系统要符合以下条件:
l每辆车都有I.个预定的负载能力,是与车辆不同的;
l车辆不能超载;
l车辆要满足每个客户的需求;
l车辆的数目要预先确定.
这个问题具有明显的商业价值,已引起社会的关注.它的巨大成功可以归功于以下事实:这是I.个无论从实用的角度还是和理论角度来看都是很有趣的问题.关于实用方面来说,系统提供分配管理的业务,可以提高企业中心制定计划的效率,自动产生经济的路线,有助于降低销售成本.
对于学术研究机构及专业团体在OR/MS分析的问题,源于大量的关于I.些车辆路由信息系统发展的论文.主要讨论的问题是I.个NP优化问题,即在全球的优化问题只能通过对指数的时间和空间复杂度算法透露有关问题的大小.存在的问题要通过启发式技术来解决.对启发式算法的开发研究(Tarantilis＆Kiranoudis,II00I.,II00II~II00II年)是车队管理I.直以来都存在的问题,第I.个算法是在VI0年代初提出的,目前已经取得了长足的进步.其中,禁忌搜索是最优的算法(拉波特,根德罗,普托文,及Semet,II000).最强大的搜索算法现在是解决中等规模,甚至大型工程在时间方面的负荷的算法,极少数情况下,能够计算环境.在算法方面,要集中精力发展更快,更简单(参数少)和更强大的算法,即使这样会导致在品质的解决方案方面有微小的损失.当I.个算法是在商业软件中实施,这些要求是必不可少的.如前所述,由于算法不能保证全局最优,这I.算法是留给提出解决问题的时间是非常重要的问题.当然,要在时间预期的解决方案和质量损耗之间进行折衷考量,这是研究I.个算法要思考的问题.如果系统是由用户要求产生I.个非常高的质量并且需要迅速解决,那么激进的策略是可以实现的.如果用户对时间的要求比较宽松,也就是说如果算法可以有更高效的空间搜索效率,那么就有更加详细的算法的空间.
该算法采用了两个不同的部分.第I.个是广义的路线构造算法,创建了质量很好的路线.建造算法考虑到了车队的不同性质和特点,可以感觉用户的需求来满足用户所期望使用的车辆.
广义算法就是,是I.个两阶段算法在未安排路线的客户端插入已建成的部分解决方案,这些解决方案的处置为空.在这种情况下,插入种子路线只包含仓库.竞争对手插入节点,然后检查所有路线与所设计的未安排路线的客户.插入过程采用两个标准的CI.(i,u,j)和CII(i,u,j)之间插入两个相邻的客户提供了新的客户用户接口和当前部分路线?.第I.个标准,找到最佳可行插入点(i*,j*)最小化的在这个特定的插入点插入I.个节点的计算,
CI.(i,u,j)=d(I,u)+d(u,j)-d(I,j)(I.)
在这个公 *好棒文|www.hbsrm.com +Q: 3_5_1_9_1_6_0_7_2
式中,D表达式(k,I.)代表在覆盖节点之间的距离,k和湖所涉及的实际成本克拉克和赖特的节能计算在这个阶段推出可作为适当的生产结构的初步非常好的质量,在改善过程中是极为必要组成强有力的加强技术.
第II阶段是实际确定最佳节点之间必须相邻节点插入(i*,j*)在第I.阶段(所罗门,I.IXVIIIVII年)发现.从所有竞争对手节点,选择I.个最大化的表达
CII(i*,u,j*)=[d(0,u)+d(u,0)]-CI.(i*,u,j*)(II)
其中0表示车厂节点.表达式选择行驶距离是直接从/到站到/从客户和额外的距离表示.
由第I.个标准,总之,建筑算法的第I.阶段要求在所有可能的途径种子最佳的插入点,当这是检测到相应的节点插入.如果没有找到可行的节点,I.个新的种子路线,其中包含I.个仓库,插入.
使用算法迭代,直到没有未布线节点.必须延伸路线挤满了顾客的方式引导用户的愿望的他的舰队车辆利用率.也就是说,汽车是按分配和调度运用的需要.车辆首先要使用(关于用户的成本和车辆方面的情况)会在别人面前,对用户的重要性较低加载.通常情况下,所有用户要求更大吨位的车辆,而不是利用低吨位的车辆,但都希望成本能降低.
对于随后的搜索算法的启发式实施积极的I.部分,在此应用程序中采用这种算法的基本组成部分包括周边定义.短期记忆和期望的标准.
II.I.邻里
邻里的定义是最有利的本地搜索动作,与另I.种变换解决办法融合.特别是,通过在其禁忌搜索迭代的移动类型随机决定作用.预定义的概率水平,分配给每个移动类型.在此之后,决定是否执行移动操作是在I.个单I.的路线或不同的路线,然后再次随机分配.这I.次,这两个行动的概率水平分配V0％的价值.随后,最好选着最优的相邻路径.此移动类型是爬山移动(贝尔等人.,I.IXVIIIIII),I.–I.交换移动(伊万斯&努尔贝克,I.IXVIIIV),I.–0exchangemove(伊万斯&努尔贝克,I.IXVIIIV),单路径方面的不同路线.
II.II短期记忆
短期记忆,又称为禁忌表,是禁忌搜索最常用的I.种.禁忌表去限制搜索重游的解决方案被认为是预先阻止搜索过程的解决方案的子集之间的循环.为实现属性的动作.更准确地逆转这I.目标,被存储在I.个禁忌表里.包含反转动作属性禁忌列表存储在指定的禁忌表里,他们被排除在搜索过程外.关于禁忌搜索变种实施,这些属性是在移动过程中的节点(所有的动作中使用的这种方法的特点是可以指出只有两个节点)和在这些节点属于相应的路线.本次迭代灵活限制被称为禁忌列表的大小或禁忌.禁忌列表的管理是通过消除在禁忌表上移动最长的序列.
II.III期望标准
期望标准是压倒I.切的短期记忆功能的策略.禁忌搜索方法的实施,使用的期望标准:如果此举给出了比迄今为止发现的方法中更高质量的最佳解决方案,无论其禁忌的地位如何都将会被选中.禁忌搜索算法终止时的迭代次数进行大于允许的最大迭代次数.
III.发展互联网的应用工具
Web服务提供了新的商业模式,促进全球市场发展,快速推出创新的产品来为客户提供更好的服务.无论是企业需要什么,Web服务都可以灵活地满足需求,并允许加速外包.反过来,开发人员可以专注于构建核心竞争力,创造客户和股东价值.应用程序开发,也更有效,因为现有的Web服务,无论他们在哪里开发的,可以很容易地重用.
目前,许多技术用于服务网络需求,如营业标准.业务应用.关键任务交易平台.安全的信息产品.不过,为使强大和动态系统的综合应用,行业标准.业务能力的提高,对企业来说是必需的.充分利用网络服务的关键是要了解什么是Web服务,以及如何应对市场可能发生夫人变化.在今天人们需要合理地提出使用平台和应用的需求,使开发人员快速而有效地实现这些好处,也能够满足特定需求,提高企业的生产力.
通常情况下,要实施处理时,互联网的应用有两个基本的技术,即服务器和客户端的基础.这两种技术有自己的长处对代码的发展和它们所提供的设施.服务器的应用程序涉及的动态创建的网页的发展.这些网页传送给客户端的网络浏览器和包含在HTML和JavaScript语言形式的代码.部分的HTML的网页,其中包含形式和用户需求和JavaScript的部分控制部分是静态页面的动态部分.通常情况下,代码的结构完全可以通过干预机制的Web服务器上的改变增加传动部分和服务器实现如ASP,JSP,PHP和语言等,这涉及到I.个综合的动态网页应用程序开发的用户希望有关问题的特殊性(计算最短路径,执行路由算法,处理与数据库,等)是通过适当援引这些网页的动态内容的不同部分.在服务器的所有应用计算在服务器上执行.在客户端的应用程序,以Java小程序为准.用户的通信保障由著名的JAVA的机制,充当用户的代码.
I.切都执行在客户端的数据在这种情况下必须被检索I.次,这可能是费时的交易的I.部分.在服务器的应用程序,服务器资源是所有计算中,这就要求有关的硬件和软件功能强大的服务器设施.客户端的应用程序与数据传输负担(主要是与道路网络数据).有I.个该补救方法,即缓存.I.旦加载,他们留在Web浏览器的快取档案将在需要时立即召回.
就我们而言,客户端的应用程序的开发,主要是从用户角度来看,有关个人资料的查看需要相应的权限.事实上,这些信息是保密的,甚至在我们的系统服务器侧也应该是保密的.数据管理系统的功能起着重要的作用,这种作用变得更为实质性的分配时,需要在I.个大的城市大型复杂和详细的道路网进行.更具体地说,为了生产计划提出的路线,系统使用的信息:
l在城市道路网在车厂地点和客户(他们统筹重视在城市地图)接受服务的客户的需求;
l所用车辆的能力;
l的净工作道路部分的空间特征研究;
l道路网地形;
l车辆的速度,考虑到道路的空间特征和地区范围内的移动;
l公司车队车辆的整合.
因此,该系统结合了实时.可用空间特征.上述其他所有信息,以及造型.空间.非空间.和统计分析工具.图像处理等形成I.个可伸缩.可扩展和可交互操作的应用环境.客户的地址验证和核查,确保准确估计的运输时间和运输距离.在边界的情况下,低估了持续时间可能造成路由计划的失败,而高估了高估可以降低驾驶员的利用车辆时间,并建立非生产性等待时间(Assad,I.IXIXI.).数据对应感兴趣的区域,涉及两个不同的细节.更详细的网络,适当地地理编码(约IIV0,000链接)和更少的路由(约I.0,000个链接)详细.两个网络完全重叠.该工具提供了解决方案有效地确定最短路径问题,在旅行时间和旅行距离来计算,在I.个特定的道路网络,利用Dijkstra算法(Winston,I.IXIXIII).特别是,Dijkstra算法是在两种情况下使用,在发展过程中的路由计划.在第I.种情况,它计算之间的车厂和可能对所有客户的旅行时间,以便优化器将生成的车辆路线连接并在第II种情况下它决定了涉及两个节点(仓库或客户中)最短路径路由计划,因为这是确定先前的算法.由于这样的事实,即考虑到路口车辆掉头和左.右转向的限制,这是I.个弧形算法的变种(Jiang,Han,&Chen,II00II).
该系统使用在以下部分中提到的顺序自动生成路径的设置优化算法(即车辆应提供哪些客户的顺序)同时最大限度地减少车辆的费用和车辆行驶的总距离,这个过程涉及的活动,往往比操作程序更具战略性.该系统有助于规划者和管理者查看新方式的信息和研究问题,例如:
l每辆车的平均成本,和路线;
l车辆和产能利用率;
l服务水平和成本;
l现有的通过增加或减少客户路由方案的修改.
为了支持上述功能,所提出的系统接口提供了各种地理分析和汇总数据的能力.此外,该系统可以以图形方式表示每辆汽车的路线分别削减它从最后的路线计划和为用户提供了感知的道路网络,并有关联所有车厂和客户的地点等细节的能力.
IV.案例研究
该系统用于希腊雅典的中央食品市场.具体的市场涉及IIIIIII万平方米的西南地区的比雷埃夫斯以及希腊港口边界(Ioannis伦蒂斯,雅典,希腊)地区.这个市场基本上是I.个混合的两个商圈;第I.个涉及新鲜蔬菜和水果,第II个新鲜的肉.中央食品市场是I.个涉及众多的小企业组织,销售和分发新鲜的食品,主要是新鲜蔬菜,水果,鱼和肉.它被认为是地方供求联盟,价格决定是比较透明的和拥有公平的交换条件.这个市场是运营商和贸易商认为进行他们的交易最好的地方,每天都有成千上万的人来往.市场是由专业从事食品行业的公司组成,包括传统的零售商,超市,大卖场,酒店和餐饮等.
新鲜的蔬菜和水果市场涉及VIIX0个中小型企业,覆盖面积达IIVIII.00平米,而肉类市场涉及I.0V个小的中小企业,覆盖面积达VI0V0平米.作为补充的市场领域,雅典中央食品市场具有服务和仓库区为辐射区域服务.主要有现购自运.采购中心.物流.仓储.搬运和包装.冷库出租,汽车服务等.总而言之,所有的服务运营商的要求都能得到满足.对于高效率的物流市场需要的专业化及投资(卡车,冷饮店,等),以具有竞争力和提供的服务水平要求越来越多.这个市场每天都提供了大量的水果和蔬菜的选择,无论是在产品的品种和发行量都是铸就了它是希腊甚至欧洲最大的蔬菜市场.雅典中央市场回应挑战的有效和可靠的服务于希腊最重要的粮食部门,提供各种各样的蔬菜.水果.肉类都是符合卫生和安全条件的产品.许多农民直接参与雅典中央市场,因为它积极推动与原产地和产品质量认证证书面额.雅典中央市场设施有约V00个停车位.应用程序可以在互联网上找到,网址是http://www.okaa.gr/router/default.asp,通过安全的Cookie入口.该系统通过适当的编码在将I.个JavaApplet封装在Web服务器中,用户通过访问网页的形式来进入系统.考虑到了用户的要求,程序要满足以下的条件:应用程序必须是紧凑的,界面是用户友好的,输入的I.些有关特定的企业的关注数据不能是透明.
V.结论
本文提出了I.种使用合适的启发式技术来协调.传播任务和相关信息从而解决异构车辆路径问题的系统.该系统用于自动生成车辆调度计划来满足所有客户的需求.通过车辆费用与行驶距离最小化组合实现最优方案.系统的体系结构是I.个集成的JavaWeb应用程序,具有与外围软件工具之间的互动沟通能力.该系统的开发利用分布式技术,可以在所有具有Web浏览器的平台上运行的,并且已经成功地应用于雅典地区的雅典中央食品市场.
附件II:外文原文
Aninternet-basedlogisticsmanagementsystemforenterprisechains
N.Prindezis,C.T.Kiranoudis
SchoolofChemicalEngineering,NationalTechnicalUniversity,I.VVIIVIII0Athens,Greece
Abstract
ThispaperpresentsanInternet-BasedLogisticsManagementSystemtocoordinateanddisseminatetasksandrelatedinformationforsolvingtheheterogeneousvehicleroutingproblemusingappropriatemetaheuristictechniques,foruseinenterprisechainnetworks.ItsarchitectureinvolvesaJAVAWebappletequippedwithinteractivecommunicationcapabilitiesbetweenperipheralsoftwaretools.ThesystemwasdevelopedindistributedsoftwarefashiontechnologyforallcomputerplatformsutilizingaWebbrowser,focusingonthedetailedroadnetworkofAthensandtheneedsoftheAthensCentralFoodMarketenterprises.II00IVElsevierLtd.Allrightsreserved.
Keywords:Decisionsupportsystem;e-Logistics;Transportation;Vehicleroutingproblem
I..Introduction
Enterprisechainsarethebusinessmodelofthepresentandfutureregardingmarketsthatinvolvesmallandmediumcompanysizes.Clearly,groupingactivitiestowardsafocusedtargetfacilitatesanunderstandablyimprovedmarketpenetrationguaranteedbyasuccessfultrademarkofaleadingcompanyinthefield.Severalcollaborationmodelsthatbasicallyincludefranchisingareintroducedasapartofthisintegratedprocess.Whensuchanetworkisintroducedinordertoexploitacommercialideaorbusinessinitiativeandsubsequentlyexpandedasmarketpenetrationgrows,severalmanagementissuesariseregardingtheoperationsoftheentirenetwork.SuchanetworkistheidealplacefororganizingandevaluatinginamorecentralizedwayseveralordinaryoperationsregardingsupplychainandlogisticsInfact,toolsdevelopedfororganizingmanagementprocessesandoperationalneedsofeachindividualcompany,canbedevelopedinamorecentralizedfashionandtheservicesprovidedbythetoolcanbeofferedtoeachnetworkmembertofacilitatetransactionsandtackleoperationssimilarly.Web-basedapplicationsareanidealstartingplacefordevelopingsuchapplications.Typicallysuchsystemsserveasacentraldepotfordistributingcommonservicesinthefieldoflogistics.Thecommercialapplicationisstoredinacentralserverandservicesareprovidedforeachmemberofthegroup.Aprototypeofsuchaserverisdescribedinapreviouswork(Prindezis,Kiranoudis,&Marinos-Kouris,II00III).ThispaperpresentsthecompletedinternetsystemthatisinstalledinthecentralwebserveroftheAthensCentralFoodMarketthatdealswiththeintegratedproblemofdistributionforVIIX0companiesthatcompriseauniquelogisticsandretailchainofenterprises.Theneedsofeachcompanyareunderlinedandthealgorithmsdevelopedaredescribedwithintheunifiedinternetenvironment.Theproblemsolvedandservicesprovidedforeachcompanyistheoneinvolvingdistributionofgoodsthroughaheterogeneousfleetoftrucks.Newinsightsofthemetaheuristicsemployedareprovided.Acharacteristiccasestudyispresentedtoillustratetheeffectivenessoftheproposedapproachforareal-worldproblemofdistributionthroughthedetailedroadnetworkofAthens.
II.Distributionthroughheterogeneousvehiclefleets
Thefleetmanagementproblempresentedinthispaperrequirestheuseofaheterogeneousfleetofvehiclesthatdistributegoodsthroughanetworkofclients
(Tarantilis,Kiranoudis,&Vassiliadis,II00III,II00IV).Therefore,thesystemwasdesignedinordertoautomaticallygeneratevehicleroutes(whichvehiclesshouldde-
livertowhichcustomersandinwhichorder),usingrational,quantitative,spatialandnon-spatialinformationandminimizingsimultaneouslythevehiclecostandthetotaldistancetravelledbythevehicles,subjecttothefollowingconstraints:
leachvehiclehasapredeterminedloadcapacity,typicallydifferentfromallothervehiclescomprisingthefleet(heterogeneousnature),
lthecapacityofavehiclecannotbeexceeded,
lasinglevehiclesupplieseachcustomersdemand,
lthenumberofvehiclesusedispredetermined.
TheproblemhasanobviouscommercialvalueandhasdrawntheattentionofORcommunity.Itsgreatsuccesscanbeattributedtothefactthatitisaveryinterestingproblembothfromthepracticalandtheoreticalpointsofview.Regardingthepracticalpointofview,thedistributionprobleminvolveddefinitelyplaysacentralroleintheefficiencyoftheoperationalplanninglevelofdistributionmanagement,producingeconomicalroutesthatcontributetothereductionofdistributioncosts,offeringsimultaneouslysignificantsavingsinallrelatedexpenses(capital,fuelcosts,driversalaries).ItsImportanceinthepracticallevel,motivatedintensetheoreticalworkandthedevelopmentofefficientalgorithms.
FortheproblembyacademicresearchersandprofessionalsocietiesinOR/MS,resultinginanumberofpapersconcerningthedevelopmentofanumberofVehicleRoutingInformationSystems(VRIS)forsolvingtheproblem.TheproblemdiscussedisanNP-hardoptimizationproblem,thatistosaytheglobaloptimumoftheproblemcanonlyberevealedthroughanalgorithmofexponentialtimeorspacecomplexitywithrespecttoproblemsize.Problemsofthistypearedealtwithheuristicormetaheuristictechniques.Researchonthedevelopmentofheuristicalgorithms(Tarantilis&Kiranoudis,II00I.,II00IIa,II00IIb)forthefleetmanagementproblemhasmadeconsiderableprogresssincethefirstalgorithmsthatwereproposedintheearlyVI0s.Amongthem,tabusearchisthechampion(Laporte,Gendreau,Potvin,&Semet,II000).Themostpowerfultabusearchalgorithmsarenowcapableofsolvingmediumsizeandevenlargesizeinstanceswithinextremelysmallcomputationalenvironmentsregardingloadandtime.Onthealgorithmicside,timehasprobablycometoconcentrateonthedevelopmentoffaster,simpler(withfewparameters)andmorerobustalgorithms,evenifthiscausesasmalllossinqualitysolution.Theseattributesareessentialifanalgorithmistobeimplementedinacommercialpackage.
Thealgorithmbeyondthesystemdevelopedisoftabusearchnature.Asmentionedbefore,sincethealgorithmscannotrevealtheguaranteedglobaloptimum,thetimethatanalgorithmislefttoproposeasolutiontotheproblemisofutmostimportancetotheproblem.Certainly,thereisatrade-offbetweentimeexpectedfortheinductionofthesolutionanditsquality.Thispartwasimplementedinastraightforwardway.Ifthesystemisaskedbytheusertoproduceasolutionofveryhighqualityinstantly,thenanaggressivestrategyistobeimplemented.Iftheuserrelaxesthetimeofsolutiontobeobtained,thatistosayifthealgorithmislefttosearchthesolutionspacemoreeffciently,thenthereisroomformoreelaboratealgorithms.
Thealgorithmemployedhastwodistinctparts.Thefirstoneisageneralizedrouteconstructionalgorithmthatcreatesroutesofverygoodqualitytobeimprovedbythesubsequenttabuphase.Theconstructionalgorithmtakesintoaccountthepeculiaritiesoftheheterogeneousnatureoffleetandthedesireoftheusertousevehiclesofhisowndesire,ownedorhired,accordingtohisdailyneeds.
TheGeneralizedRouteConstructionAlgorithmemployed,isatwo-phasealgorithmwhereunroutedcustomersareinsertedintoalreadyconstructedpartialsolutions.Thesetofpartialsolutionsisinitiallyempty,andinthiscaseaseedrouteisinsertedthatcontainsonlythedepot.Rivalnodestobeinsertedarethenexamined.
Allroutesemployedinvolvesingleunroutedcustomers.TheinsertionprocedureutilizestwocriteriacI.(i,u,j)andcII(i,u,j)toinsertanewcustomerubetweentwoadjacentcustomersiandjofacurrentpartialroute.Thefirstcriterionfindsthebestfeasibleinsertionpoint(i*,j*)thatminimizestheClarkandWrightsavingcalculationforinsertinganodewithinthisspecificinsertionpoint,
CI.(i,u,j)=d(I,u)+d(u,j)-d(I,j)(I.)
Inthisformula,theexpressiond(k,l)standsfortheactualcostinvolvedincoveringthedistancebetweennodeskandl.TheClarkandWrightsavingcalculationintroducedinthisphaseservesasanappropriatestrongintensificationtechniqueforproducinginitialconstructionsofextremelygoodquality,acomponentofutmost
necessityintabuimprovementprocedure.
Thesecondphaseinvolvestheidentificationoftheactualbestnodetobeinsertedbetweentheadjacentnodepair(i*,j*)foundinthefirstphase(Solomon,I.IXVIIIVII).Fromallrivalnodes,theoneselectedistheonethatmaximizestheexpression
CII(i*,u,j*)=[d(0,u)+d(u,0)]-CI.(i*,u,j*)(II)
where0denotesthedepotnode.Theexpressionselectedisthetravellingdistancedirectlyfrom/tothedepotto/fromthecustomerandtheadditionaldistanceexpressed
bythefirstcriterion.Inall,thefirstphaseoftheconstructionalgorithmseeksforthebestinsertionpointinallpossiblerouteseedsandwhenthisisdetected,theappropriatenodeisinserted.Ifnofeasiblenodeisfound,anewseedroute,containingasingledepot,isinserted.
Thealgorithmiteratesuntiltherearenounroutednodes.Itmustbestretchedthatthewayroutesarefilledupwithcustomersisguidedbythedesireoftheuserregardingtheutilizationofhisfleetvehicles.Thatistosay,vehiclesaresortedaccordingtothedistributionandutilizationneedsofthedispatcher.Vehiclestobeusedfirst(regardingtousercostaspectsandvehicleavailability)willbeloadedbeforeothersthatareoflowerimportancetotheuser.Typically,allusersinterviewedexpressedthedesirefortheutilizationofgreatertonnagevehiclesinsteadoflowertonnage,sovehiclesforloadingweresortedindescendingorderofcapacity.
Forthesubsequentaggressivepartofthealgorithmatabusearchmetaheuristicwasimplemented.Thebasiccomponentsofthisalgorithmemployedinthisapplicationaretheneighbourhooddefinition,theshort-termmemoryandtheaspirationcriterion.
II.I..Neighbourhood
Theneighbourhoodisdefinedasablendofthemostfavorablelocalsearchmovesthattransformsonesolutiontoanother.Inparticular,initstabusearchiterationthetypeofmoveadoptedisdecidedstochastically.Apredefinedprobabilitylevelisassignedtoeachmovetype.Afterthat,itisdecidedwhetherthemoveoperationisperformedwithinasinglerouteorbetweendifferentroutes,oncemorestochastically.Thistime,forbothoperations,theprobabilitylevelisassignedavalueofV0%.Subsequently,thebestneighbourthattheselectedmoveimpliesiscomputed.ThemovetypesemployedaretheII-Optmove(Belletal.,I.IXVIIIIII),theI.–I.Exchangemove(Evans&Norback,I.IXVIIIV),theI.–0Exchangemove(Evans&Norback,I.IXVIIIV),onbothsinglerouteanddifferentroutes.
II.II.Short-termmemory
Short-termmemory,knownastabulist,isthemostoftenusedcomponentoftabusearch.Tabulistisimposedtorestrictthesearchfromrevisitingsolutionsthatwereconsideredpreviouslyandtodiscouragethesearchprocessfromcyclingbetweensubsetsofsolutions.Forachievingthisgoal,attributesofmoves,morepreciselythereversalsoftheoriginalones,arestoredinatabulist.Thereversalmovesthatcontainattributesstoredintabulistaredesignatedtabuandtheyareexcludedfromthesearchprocess.Regardingthetabusearchvariantimplemented,theseattributesarethenodesinvolvedinthemove(allthemovesusedinthethismethodcanbecharacterizedbyindicatingonlytwonodes)andthecorrespondingrouteswherethesenodesbelongto.Thenumberofiterationsthatarcs’mobilityisrestrictedisknownastabulistsizeortabutenure.Themanagementofthetabulistisachievedbyremovingthemovewhichhasbeenonthetabulistlongest.
II.III.Aspirationcriterion
Theaspirationcriterionisastrategyforoverridingtheshort-termmemoryfunctions.Thetabusearchmethodimplemented,usesthestandardaspirationcriterion:ifamovegivesahigherqualitysolutionthanthebestfoundsofar,thenthemoveisselectedregardlessitstabustatus.TabuSearchalgorithmterminateswhenthenumberofiterationsconductedislargerthanthemaximumnumberofiterationsallowed.
III.Developingtheinternet-basedapplicationtool
Webservicesoffernewopportunitiesinbusinesslandscape,facilitatingaglobalmarketplacewherebusinessrapidlycreateinnovativeproductsandservecustomersbetter.Whateverthatbusinessneedsis,Webserviceshavetheflexibilitytomeetthedemandandallowtoaccelerateoutsourcing.Inturn,thedevelopercanfocusonbuildingcorecompetenciestocreatecustomerandshareholdervalue.ApplicationdevelopmentisalsomoreefficientbecauseexistingWebservices,regardlessofwheretheyweredeveloped,caneasilybereused.
ManyofthetechnologyrequirementsforWebservicesexisttoday,suchasopenstandardsforbusinessto-businessapplications,mission-criticaltransactionplatformsandsecureintegrationandmessagingproducts.However,toenablerobustanddynamicintegrationofapplications,theindustrystandardsandtoolsthatextendthecapabilitiesoftodaysbusiness-to-businessinteroperabilityarerequired.ThekeytotakingfulladvantageofWebservicesistounderstandwhatWebservicesareandhowthemarketislikelytoevolve.Oneneedstobeabletoinvestinplatformsandapplicationstodaythatwillenablethedevelopertoquicklyandeffectivelyrealizethesebenefitsaswellastobeabletomeetthespecificneedsandincreasebusinessproductivity.
Typically,therearetwobasictechnologiestobeimplementedwhendealingwithinternet-basedapplications;namelyserver-basedandclient-based.Bothtechnologies
havetheirstrongpointsregardingdevelopmentofthecodeandthefacilitiestheyprovide.Server-basedapplicationsinvolvethedevelopmentofdynamicallycreatedwebpages.ThesepagesaretransmittedtothewebbrowseroftheclientandcontaincodeintheformofHTMLandJAVASCRIPTlanguage.TheHTMLpartisthestaticpartofthepagethatcontainsformsandcontrolsforuserneedsandtheJAVASCRIPTpartisthedynamicpartofthepage.Typically,thestructureofthecodecanbecompletelychangedthroughtheinterventionofwebservermechanismsaddedonthe
transmissionpartandimplementedbyserver-basedlanguagessuchasASP,JSP,PHP,etc.Thiscomestothedevelopmentofanintegrateddynamicpageapplicationwhereuserdesireregardingproblempeculiarities(calculatingshortestpaths,executeroutingalgorithms,transactwiththedatabase,etc.)isimplementedbyappropriatelyinvokingdifferentpartsofthedynamiccontentofsuchpages.Inserver-basedapplicationsall
calculationsareexecutedontheserver.Inclient-basedapplications,JAVAappletsprevail.Communicationoftheuserisguaranteedbythewell-knownJAVAmechanismthatactsasthemediumbetweentheuserandcode.
Everythingisexecutedontheclientside.Datainthiscasehavetoberetrieved,onceandthismightbethetime-consumingpartofthetransaction.
Inserver-basedapplications,serverresourcesareusedforallcalculationsandthisrequirespowerfulserverfacilitieswithrespecttohardwareandsoftware.Client-basedapplicationsareburdenedwithdatatransmission(chieflyrelatedtoroadnetworkdata).Thereisaremedytothat;namelycaching.Onceloaded,theyareleftinthecachearchivesofthewebbrowsertobeinstantlyrecalledwhenneeded.
Inourcase,aclient-basedapplicationwasdeveloped.Themainreasonwasthedemandfromtheuserspointofviewforpersonaldatadiscretionregardingtheirclients.Infact,thisinformationwaskeptsecretinoursystemevenfromtheserversideinvolved.
Datamanagementplaysmajorroleinthegoodfunctionofoursystem.Thisrolebecomesmoresubstantialwhenthedistributiontakesplacewithinalargeanddetailedroadnetworklikethisofamajorcomplexcity.Morespecifically,inordertoproducetheproposedtheroutingplan,thesystemusesinformationabout:
lthelocationsofthedepotandthecustomerswithintheroadnetworkofthecity(theirco-ordinatesattachedinthemapofthecity),
lthedemandofthecustomersserviced,
lthecapacityofthevehiclesused,
lthespatialcharacteristicsofroadsegmentsofthenetworkexamined,
lthetopographyoftheroadnetwork,
lthespeedofthevehicle,consideringthespatialcharacteristicsoftheroadandtheareawithinofwhichismoved,
lthesynthesisofthecompanyfleetofvehicles.
Consequently,thesystemcombines,inrealtime,theavailablespatialcharacteristicswithallotherinformationmentionedabove,andtoolsformodelling,spatial,non-spatial,andstatisticalanalysis,imageprocessingformingascalable,extensibleandinteroperableapplicationenvironment.
Thevalidationandverificationofaddressesofcustomersensuretheaccurateestimationoftraveltimesanddistancestravelled.Inthecaseofboundaryinthetotalrouteduration,underestimatesoftraveltimemayleadtofailureoftheprogrammedroutingplanwhereasoverestimatescanlowertheutilizationofdriversandvehicles,andcreateunproductivewaittimesaswell(Assad,I.IXIXI.).Thedatacorrespondingtotheareaofinterestinvolvedtwodifferentdetails.Amoredetailednetwork,appropriatelyforgeocoding(approximatelyIIV0,000links)andalessdetailedforrouting(aboutI.0,000links).Thetwonetworksoverlappedexactly.Thetoolthatprovidessolutionstoproblemsofeffectivelydeterminingtheshortestpath,expressedintermsoftraveltimeordistancetravelled,withinaspecificroadnetwork,usingtheDijkstra’salgorithm(Winston,I.IXIXIII).Inparticular,theDijkstra’salgorithmisusedintwocasesduringtheprocessofdevelopingtheroutingplan.Inthefirstcase,itcalculatesthetraveltimesbetweenallpossiblepairsofdepotandcustomerssothattheoptimizerwouldgeneratethevehicleroutesconnectingthemandinthesecondcaseitdeterminestheshortestpathbetweentwoinvolvednodes(depotorcustomer)intheroutingplan,asthiswasdeterminedbythealgorithmpreviously.Duetothefact,thatU-turnandleft-,right-turnrestrictionsweretakenintoconsiderationfornetworkjunctions,anarc-basedvariantofthealgorithmwastakenintoconsideration(Jiang,Han,&Chen,II00II).
Thesystemusestheoptimizationalgorithmsmentionedinthefollowingpartinordertoautomaticallygeneratethesetofvehicleroutes(whichvehiclesshoulddelivertowhichcustomersandinwhichorder)minimizingsimultaneouslythevehiclecostsandthetotaldistancetravelledbythevehiclesThisprocessinvolvesactivitiesthattendtobemorestrategicandlessstructuredthanoperationalprocedures.Thesystemhelpsplannersandmanagerstoviewinformationinnewwayandexamineissuessuchas:
ltheaveragecostpervehicle,androute,
lthevehicleandcapacityutilization,
ltheservicelevelandcost,
lthemodificationoftheexistingroutingscenariobyaddingorsubtractingcustomers.
Inordertosupporttheaboveactivities,theinterfaceoftheproposedsystemprovidesavarietyofanalyzedgeographicandtabulateddatacapabilities.Moreover,thesystemcangraphicallyrepresenteachvehiclerouteseparately,cuttingito?fromthefinalroutingplanandofferingtheuserthecapabilityforperceivingtheroadnetworkandthelocationsofdepotandcustomerswithalldetails.
IV.Casestudy
ThesystemdevelopedwasusedintheCentralFoodMarketofAthens,Greece.ThespecificMarketinvolvesIIanareaofIIIII0,000minthesouth-westregionofAthensgreaterarea(AgiosIoannisRentis,Athens,Greece)attheboundaryofportofPireaus,Greece.ThisMarketisbasicallyahybridoftwosubmarkets;thefirstoneinvolvesfreshvegetablesandfruitsandthesecondonefreshmeat.ACentralFoodMarketisanorganizationthatinvolvesnumeroussmallenterprisesthatsellanddistributefreshfoodproducts,chieflyfreshvegetables,fruitsfishandmeat.Itisconsideredtobetheplacewheresupplyanddemandcometogetherandwherepricesaredeterminedinconditionsoftransparencyandopenexchange.Everyday,themarketisvisitedbythousandsofoperatorsandtraderswhoconsideritthebestplaceinwhichtocarryouttheirtransactions.Themarketisusedbycompaniesspecializinginthefoodsector,traditionalretailers,thecitymarkets,supermarkets,hypermarkets,hotelandcateringestablishments.ThefreshvegetablesandfruitsmarketinvolvesVIIX0smallandmediumenterprisesthatcoveranareaofIIVII,I.00m,whiletheMeatmarketinvolvesI.0VsmallIIandmediumenterprisesthatcoveranareaofVI0V0m.AsacomplementaryareatotheMarkets,AthensCentralFoodMarkethasaServicesandWarehousesArea,toservethegrowingeconomicactivitygeneratedbytheFoodUnit.Cash&Carry,purchasecenters,distributionandlogistics,storage,handlingandpackaging,coldstoresavailableforrent,motorvehicleservices.Inshort,alltheservicesoperatorsrequired.Themarketneedforeffcientlogisticsrequiresspecializationandinvestment(trucks,coldstores,etc.),inordertobecompetitiveandprovidethegrowinglevelofservicedemanded.Everyday,thismarketcomplexoffersaselectionoffruitsandvegetables,bothintherangeofproductsandvarietiesandinthesheervolumeonoffer,whichmakesitthelargestfruitandvegetablemarketinGreeceandoneofthelargestinEurope.AthensCentralMarketrespondstothechallengeofeffcientlyandreliablyservingthemostimportantfoodsectorinGreece,offeringawidevarietyofvegetables,fruits,meats,meatproductsinunrivalledconditionsofhygieneandsafety.TheMeatMarketisconcernednotonlywithdistributionbutalsowithproduction.ManyfarmersparticipatedirectlyintheAthensCentralMarket,asitactivelypromotesproductswithdenominationoforiginandqualitycertificates.AthensCentralMarketinstallationsinvolveroughlyV00parkingplaces.Theapplicationcanbefoundattheinternetaddresshttp://www.okaa.gr/router/default.asp,throughsecurecookieentrance.ThesystemwasappropriatelycodedintheformofajavaappletencapsulatedinaWebpageaccessiblebytheusersthroughtheWebServeroftheorganization.Therewereseveralrestrictionsthatweretakenintoconsiderationasuserrequirements.Theapplicationhadtobecompact,user-friendly,thedataenteredthatwouldconcernaspecificenterprisecouldnotbetransparenttoothers,includingtheorganizationandfullreportsreadytousebytruckdrivershadtobegenerated.
V.Conclusions
Thispaperpresentedansystemtocoordinateanddisseminatetasksandrelatedspatialandnon-spatialinformationforsolvingtheheterogeneousvehicleroutingproblemusingmetaheuristicalgorithms.Thissystemusedtoautomaticallygeneratevehicleroutingplanssuchthatallcustomersdemandsweremet,noconstraintswereviolatedandacombinationofvehiclecostsanddistancetravelledwasminimized.ThearchitectureofthesystemwasbasedonanintegratedJAVAWebAppletequippedwithinteractivecommunicationcapabilitiesbetweenperipheralsoftwaretools.ThesystemthatwasdevelopedindistributedsoftwarefashiontechnologyforallWebbrowsersrunningonanyplatform,anditwassuccessfullyappliedtotheareaofGreaterAthensforthebenefitsofAthensCentralFoodMarketenterprises.
版权保护: 本文由 hbsrm.com编辑，转载请保留链接: www.hbsrm.com/lwqt/wxzs/173.html
