ProblemFaced When Using Automated Guided Vehicles Systems (AGVs)
ProblemFaced When Using Automated Guided Vehicles Systems (AGVs)
Theuse of automated guided vehicle systems has become very common due tothe technological growth. As a result, most of the companies thatdeals with loading and offloading products have integrated automatedvehicle system to aid in material handling solutions, to manage theproducts, and to enhance easy movement. Additionally, the use ofautomated guided vehicle systems is common in the manufacturingplants and industrial warehouses that require minimal laborintervention. Most importantly, the automated guided vehicle systemcan be used in as a specific site conditions, installed in both oldand new plants, and products that require minimum disruption duringthe operation process.
Onthe other hand, the system uses either customized or standardvehicles that can be managed by a computer system that supervises andprocesses all operations that are either done manually or automaticthe supervisory computer framework, therefore,services all therequests to achieve maximum productivity. Moreover, the automatedguided vehicle system can pick or deliver manufactured products fromthe store, warehouse, or lines without inferring with the cranes,forklifts, or other handling equipment. Consequently, the overallsystem used shows a robust safety, guided operations, and controlledframework that can be tailored to handle particular products.
However,it is a problem to achieve proper idle-vehicle scheduling,decision-making problems, and excellent battery management.Additionally, the design used on the automated guided vehicle systemis difficult to understand. As a result, this paper analyzes theimplementation of AGVS and suggests some directions that can be takento avoid or control the problems highlighted above.
Tableof content Contents
Executive Summary 2
Table of content 3
Current Process Analysis (As-Is processes) 6
Problem Analysis 7
Technology appraisal 8
Proposed Solution 9
Project critical appraisal 10
Reference list 13
AGVsare used when handling materials that require system support tofacilitate easily loading and offloading of products in thewarehouses, manufacturing plants, terminals, and in distributioncenters. As a result, the use of computer software and hardware isalso essential when in most of the distribution centers that requirematerial handling systems. Common areas such as around the guidedvehicle’s transport loads and the area between the locations(between the loading area and offloading section) require monitoringwhich in most cases is done using computer software. When handlinggoods from the receiving lanes to storage are in the warehouses forstorage or from the storage are to the shipping lanes it is crucialto have a reliable means of transport thus the emergence of the AGVsand computer related systems that are used for monitoring the entireprocess.
However,the overall performance of the handling framework that is used tohandle the materials directly can hinder the performance of theentire system. Consequently, the performance of the automated guidedvehicle system is basically divided into the number of theguided-path designed, idle vehicles that are already positioned,estimate number of the vehicle that are required, conflict resolutionand vehicle routing, and battery management (Hu & Egbelu, 2000).All these factors are relevant when making decisions on the use ofthe AGVs. For this reason, any problem that arises during the guidepath design process is perceived as a problem right at the strategiclevel. This means that very decision that is made during the guidepath stage pose a strong effect on future decisions made at thedifferent levels of implementing the AGV systems. However, the mainproblem when using AGVs is to estimate the number of vehicles,correct positioning especially for the idle vehicles, correctscheduling of the vehicles, and managing their battery chargingscheme. Consequently, there is a need to address these problemsduring the vehicle operational and testing phase to avoid theescalation of the minor problems to big ones. As a result, during theAGV systems implementation of the design phase, it is crucial tocarry out the iterations and interactions that can be noticed betweenthe steps of implementation. With this process it can be presumedthat the number of the guided paths in the system will be influenceddirectly depending on the number of the vehicles that are requiredagain, it will influence the vehicle scheduling system and othercomplexities involved. The entire scheduling system, therefore,affects the number of the vehicles that are served by the AGV system.
Onthe other hand, in the current and modern AGV systems used today,there are differences that are noticeable on the classic AGV systemsin a number of respects (Hu & Egbelu, 2000). For example, insteadof using fixed paths, most of the AGV systems used today are of “freerange” this means that their tracks use programmed software thatare relatively easy to change especially when there are new flows orworkstations added in the program. Secondly, there is a difference inthe way the modern and the old AGV systems are controlled. This isbased on the new technologies that are used to make decisions whenusing the automated vehicles in the recent past the guided vehicleswere controlled using the central controllers. The most importantthing is that the AGV systems have resulted in the adoption ofself-learning systems that are paramount in the manufacturingindustry and warehouse loading and offloading. Additionally, AGVsystems are crucial when handling complex or huge systems thatrequire many vehicles that can result in vehicle interference.However, the problem of decision-making is still obsolete despite theuse of AGV systems. Over the past researchers, researchers have triedto investigate the traditional impact on the AGV systems when makinga decision and how new developments can be achieved to aid thedecision-making process. However, few reviews have concentrated onthe effects of the new developments that concern decision-making.Most of the researchers have only concentrated on the limited partsof the problems highlighted above.
Forexample, few researchers have concentrated on the scheduling androuting problems or only deal with limited parts of the majorproblem. Additionally, researchers have ignored crucial partsespecially battery management and the idle vehicle positioning(Ebben, 2001). Consequently, the current paper focuses on fulfillingthe identified gap through contributions on areas that requiremodification during the design and implementation phase anddecision-making process in the AGV systems. Additionally, it iscrucial to review the decision making models that are employed in theAGV systems and try to develop a new model that can offer solutionsto the idle vehicles scheduling and maybe help approve strategiesused when managing the battery. Therefore, the paper will propose newdesign and decision-making strategies that can be adopted whenimplementing the AGV systems.
CurrentProcess Analysis (As-Is processes)
Inthe as-is process, the guide path design portrays the main problem inthe current AGV system that need to be considered first. According toGademann and Van de Velde (2000), the guided path design highlydepend on the allocated shop-floor space, the overall arrangement,and the layout of the storage zones that are assigned for handlingthe working stations. The problems such as difficulties in schedulingthe idle vehicles results due to fixed shop-floor space and imposeother constraints that accelerate the problem. Additionally, the arcsin the pick-up delivery locations are used to define the pathsthrough which the vehicles will follow when moving through theassigned nodes. As a result, the directed arc that exists between thenodes shows the direction through which the vehicle is supposed toflow. However, it is not easy to assign the cost when representingthe distance between the nodes or a specified segment. This isbecause the exact time that is taken by a vehicle to travel from onenode to the other along the arc is not always equal. However, AVGsystems use a network-based framework that is able to formulate theguide path design and other problems that arise during the vehicleoperations. The guide path for the vehicle is classified using thefeatured such as flow of topology (conventional or single loop),number of lanes (single or multiple lanes), or the flow direction(can be unidirectional or bidirectional).
However,scholars such as Arifin and Egbelu (2000) tried to tackle theproblems that appear more complicated. Consequently, they developed amodel that can select a specific path and configure the pick-up pointand the destination stations at the same time. Whenever the guidepath problems are solved, other problems such as the idle vehiclepositioning and scheduling will be solved to an optimal or feasiblesolutions. Moreover, there is a need to apply the benefits of AGVsystems and capture the cost that can be fixed to set a delivery andpick-up point. The main objective should be to analyze the functionof the vehicle and not the total distance travelled (Dumas,Desrosiers, & Soumis, 1991). Additionally, the analysis of theas-is process involves determining a model that can be used todetermine the total number of vehicles that are needed to meet theexpected service range this include time that is spent in theworkstation to replenish the assigned task from the pick-up point tothe delivery point. Here, zero to one IP model can be used toapproximate the number of vehicles required in the system queuing.
Thescheduling of the idle vehicle in the dispatch area is determined bythe number of cars in the required in pick-up and delivery sections.Consequently, to solve the problem of idle vehicle scheduling, it isparamount to know the exact number of vehicles required. Here, tosolve the problem, the paper uses the dispatching rule that estimatethe required number of vehicles. Consequently, we will consider usingonly one vehicle that travel under the estimated FCFS rule. Theresults can be extended to the multiple vehicles that can beestimated using K-vehicle using one vehicle that can travel k timesfaster. The method used to solve the problem can be used toapproximate the “right” statistics that include time fractionspent by a vehicle when traveling both loaded and when empty thiswill be used to determine the number of vehicles required thusminimizing the number of the idle vehicle at a given time.
Additionally,according to Talbot (2003), there are various analytical approachesthat can be used to estimate the number of vehicles that can bescheduled correctly to lessen their idle time. Talbot (2003)suggested that statistical approach, queuing models, or multipledecision models can be used when determining the approximate numberof vehicles required to be fully operational in a station whenservicing their requests.
Withthe new technological growth it possible to have multiple loading onvehicles queuing in the as idle queue. This technology can be used toreduce the number of cars that required improving the throughput ofthe AGV systems. For instance, AGV system can be used to loadmultiple vehicles or to pick up more loads while still in the processof moving the previously assigned load. Consequently, the use of AGVsystems that use multi-load technology can be used to reduce theamount of vehicles in the idle queue, reduce the time required for asingle trip, and reduce the total distance covered thus, reducing therate at which the battery of the vehicle discharges.
Asa result, the use of multi-load technology shows effective andefficient ways of loading the vehicles as compared to the single orunit loading technology for the vehicles. Moreover, simulation can beused to experiment that multi-load technology in the AGV systemsimproves the throughput of the systems. The technology is highlyadopted in the warehouses or the manufacturing industries that facehigh transport demands. On the other hand, Berman & Edan (2002)plotted that multiple loading when using AGV systems promote system’sperformance, especially when handling multiple loads. This is becausethe loads can be picked or delivered at one location. However, thedisadvantage of using multi-loading technology in AGV systems is theuse of complex scheduling system that is a must to ensure controlledsystem.
Theproposed solution to the above problems is to use a centralizedcontrol system that can keep accurate and timely record all movementsmade by vehicles in the AGV system. Additionally, the centralizedcontrol system will make a decision on where, when, and how thevehicles will be loaded and offloaded thus creating a stable schedulewhen carrying out the tasks. Additionally, the centralized controlsystems can be used to assign specific routes to the vehicles thusavoiding problems that arise due to route congestion. Moreover, toavoid idling of the vehicle while in the queue it is important to useboth offline and online scheduling especially when giving informationabout the tasks to be accomplished. In most cases, online dispatchsystems are required to control the vehicles to avoid more time onthe idle queue (Chang & Egbelu, 1996).
Inaddition, the centralized control system proposed to solve theproblem of idle vehicle scheduling will be used to give allinformation that relate to the vehicles such as their wait time,expected pick-up and delivery time, location, vehicle status andpositions all this information should be stored in the system’sdatabase to avoid wastage of time when loading or offloading avehicle. Additionally, the introduction of controllers in the AGVsystems will allow the systems to assign loads to specific vehiclesbased on the rule used. Moreover, the centralized control systemscommunicate with the assigned vehicle and offer guidance throughoutits operations (De Koster, Le Anh, & Van der Meer, 2004). Forthis reason, based on the manner through which the transportationframework is assigned, the rules used in the dispatch can becategorized as either work initiated in the workstation or vehicleinitiated dispatch where the vehicle is given the mandate to claim atask. This process reduces the number of vehicles in the queue asidle or waiting. Lastly, the problem can be solved using time-baseddispatch rule. The time-based dispatch rule is used to dispatch thevehicles based on the tasks’ waiting time. The rule uses FCFS(first come first served) strategy and when two vehicles appear atthe same time MFCFS (modified first come first served) strategy isused (Fleischmann, Gnutzmann, & Sandvo, 2003). These rules areused to solve the problem of scheduling idle vehicles in the stationthe rules ensure that one vehicle’s request is serviced one at atime and only happen to the vehicle whose request is ‘active’.However, the proposed rule should work in such a way that no vehiclewill experience empty travel in cases there is an override in thepick-up point.
Toachieve the above-proposed solution, it is paramount to modify thedesign used in AGV systems such as the input data, the layout, andthe performance demand in order to achieve the proposed solution.Consequently, the first step should be to set the strategic levelrequired on the guide path framework. After this step, the solutionis to consider all the tactical steps that include determining theestimated number of vehicles that are required for the workstation.This will enhance effective scheduling that also enhances effectivebattery management scheme and efficient parking policy. However, forthe project to be effective, it is a must to have simultaneousdecisions during the tactical period that must be considered. This isbecause the decision-making process influences the design used on theguide path especially when using free-range AGV systems. Moreover, onthe operational level of the AGV systems it is important to designthe system that is able to guide the vehicles to their expecteddestination without route conflict (Fleischmann, Gnutzmann, &Sandvo, 2003).
Onthe other hand, to avoid conflicts on the same route the AGV systemsshould have sensor detectors installed. The sensors can be used tosense whether the vehicles are too close and stop where necessary.However, the use of sensors may not be effective for the AGV systemsthat have numerous curved guide-paths. However, paths conflicts thatcan result in idleness can be detected using Petri nets in the AGVs. According to Beamon (1998), Petri net is an approach that ispromising and is able to detect the direction of the AGV system andprevent any conflict that might arise thus, reducing the idle time ofthe vehicles.
Inconclusion, to achieve the ‘right’ scheduling for the idlevehicles in a workstation while using the AGV systems, it is crucialto determine the design and implementation of the entire system.Additionally, the proposed solution shows that factors such asguide-path design and the number of vehicles estimated in theworkstation should be considered when working with AGV systems.Additionally, it is crucial to know how to effectively manage the AGVsystems. To effectively manage the systems measures such as propervehicle scheduling, battery management and charging, and proper idlevehicle positioning and the main resolutions that should be put intoplace. Moreover, online scheduling should be preferred as compared tothe offline scheduling because of the stochastic nature of the AGVs.In addition, the online scheduling is known to work excellently, butis complex, then using systems that use simple dispatch. Lastly, itis the mandate of the scheduling system to make the decision for theAGV system on the route to follow to reach the desired destinationwithout causing collisions. Additionally, multi-load technology canbe employed to reduce the number of idle vehicles in a workstationthis can also be used to enhance proper conflict resolutions in theAGV systems.
Arifin,R. and Egbelu, P.J. 2000. Determination of vehicle requirements inautomated guided vehicle systems: a statistical approach, ProductionPlanning & Control 11 258-270.
Beamon,B.M. 1998. Performance, reliability, and performability of materialhandling systems, International Journal of Production Research 36 (2)377-393.
Berman,S. and Edan, Y. 2002. Decentralized autonomous AGV system formaterial handling, International Journal of Production Research 40(15) 3995-4006.
Chang,S. and Egbelu, P.J. 1996. Dynamic relative positioning of AGVs in aloop layout to minimize mean system response time, InternationalJournal of Production Research 34 (6) 1655-1673.
DeKoster, R., Le Anh, T. and Van der Meer, J.R. 2004 (to appear).Testing and classifying vehicle dispatching rules in three real-worldsettings, Journal of Operations Management.
Dumas,Y., Desrosiers, J. and Soumis, F. 1991. The pickup and deliveryproblem with time windows, European Journal of Operational Research54 7-22.
Ebben,M., 2001. Logistic control in automated transportation networks,Dissertation, University of Twente.
Fleischmann,B., Gnutzmann, S. and Sandvoß, E. 2003 (to appear). Dynamic vehiclerouting based on on-line traffic information, Tansportation Science.
Gademann,N. and Van de Velde, S. 2000. Positioning automated guided vehiclesin a loop layout, European Journal of Operational Research 127565-573.
Hu,C.-H. and Egbelu, P.J. 2000. A framework for the selection of idlevehicle home locations in an automated guided vehicle system,International Journal of Production Research 38 (3) 543-562.
Talbot,L., 2003. Design and performance analysis of multi-station automatedguided vehicle systems, Dissertation, Universite Catholique deLouvain.