• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
An Innovative Real Time Production Management System
 

An Innovative Real Time Production Management System

on

  • 215 views

Scheduling is indeed a major issue in all manufacturing and project execution facilities world over. It is also ...

Scheduling is indeed a major issue in all manufacturing and project execution facilities world over. It is also
recognized that if scheduling is efficient and automated huge benefits could result as existing resource usage can
be maximized allowing dramatic increase in number of orders processed at the same time substantially reducing
cost of production while ensuring reliability in delivery on the committed date. No wonder scheduling is a hot
research topic and the market is flooded with scheduling systems of sorts. Still a truly efficient and automatic
scheduling system remains an elusive dream.
This white paper lists the six important reasons why a scheduling system fails in real-life situations. It then
describes how a new scheduling system called Talika PMS satisfies all the six critical requirements in detail with real
data supporting the claims from its first major installation.

Statistics

Views

Total Views
215
Views on SlideShare
215
Embed Views
0

Actions

Likes
1
Downloads
4
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    An Innovative Real Time Production Management System An Innovative Real Time Production Management System Document Transcript

    • Page 1 of 13© 2013, Laxman C MaratheAn Innovative Real Time Production Management SystemA white paperAbstractScheduling is indeed a major issue in all manufacturing and project execution facilities world over. It is alsorecognized that if scheduling is efficient and automated huge benefits could result as existing resource usage canbe maximized allowing dramatic increase in number of orders processed at the same time substantially reducingcost of production while ensuring reliability in delivery on the committed date. No wonder scheduling is a hotresearch topic and the market is flooded with scheduling systems of sorts. Still a truly efficient and automaticscheduling system remains an elusive dream.This white paper lists the six important reasons why a scheduling system fails in real-life situations. It thendescribes how a new scheduling system called Talika PMS satisfies all the six critical requirements in detail with realdata supporting the claims from its first major installation.Visit www.etalika.in for more information and free download1 IntroductionDay-to-day scheduling of any manufacturing facilityis recognized to be the most important problem tobe solved. [1] D. Ouelhadj and S. Petrovic recent[Oct 2008] study reveals that solutions based oncreation of a static schedule are impractical in real-life situations and discusses several dynamicscheduling approaches only to conclude that morework is still needed in this field of research.We wish to present here a complete dynamic real-time micro level scheduling system that is proven towork in the most complex manufacturing facilities. Itis a fully scalable, decentralized, multi-location anduser configurable system to suit any manufacturing /project environment. The core scheduling is fullyautomatic and guarantees that all currently allottedtasks in real-time can be executed with a completeand detailed schedule prediction of all activities forall orders in-hand. The system automaticallyreschedules in response to real-time events asnotified by operators’ handling current tasks on theshop floor, with an objective to maximize resourceutilization while minimizing job cycle time. It offersfull micro-level future schedule visibility of allrunning jobs to predict when each would be overgiven the current load as of NOW. The cycle ofallotting tasks, seeking task-wise feedback onallotments made from operators’ on shop floor, andre-predicting its impact in subsequent reschedulehappen every minute 24x7.Before we elucidate more on the system features wewould like to re-emphasis importance of schedulingin any manufacturing facility and why currentsolutions fail to address the problem correctly.2 Importance of schedulingThe only real differentiators to compete inestablished products and services market are Costand Reliable delivery. Quality of product / service ismostly considered a pre-condition to be in businessrather than a differentiator. Both cost and reliabledelivery of product / service are directly impacted byscheduling.2.1 Scheduling and cost of productionIt is almost axiomatic to state that a major portion ofcost of production (even exceeding 70% - 80% inmade-to-order industries) is expended incoordinating and managing production activities vis-à-vis the actual cost of value-addition involved.Most real life manufacturing involves execution ofseveral individual activities in a complex order tocreate any saleable final product or service. Thestarting point thus is in breaking down an orderrequirement into elemental activities that must becompleted in order to accomplish the final product /service deliverable: ranging from getting inputs orraw material until final packing and dispatch. Unlessthis detailing is not done, actual value additioncannot begin. Once it is known “How” the order canbe fulfilled the most difficult job of scheduling
    • Page 2 of 13© 2013, Laxman C Maratheindividual activities of orders begin. It primarilytranslates in deciding what activity must be done,where it should be done (that is using what limitedresources) and when. One can refer to them as the3W’s. Most expensive and competent personnel inany manufacturing or project execution facility areengaged fully in the process of managing productionthat involves, breaking down order execution,estimating its cost, time and wastages, decidingwhat activities to do now and next, taking feedbackon progress made, follow-up and expediting to meetdeadlines. The whole exercise is repeated all overagain by rescheduling to predict and monitorexpected completion dates for all orders in hand.Add to this already complex situation, the burden ofestimating when new orders can be delivered giventhe existing load of orders already in-hand. All this isnow possible to be completely automated resultingin a substantial reduction in the cost of production.2.2 Scheduling & reliabilityScheduling decisions taken now directly impactexpected completion times of all orders in-hand. Inreal-life situations one has to deal with severalorders, each with its own set of individualinterdependent activities requiring a certain profileof resources that are both shared and limited. It iswell impossible, even in small setups, to manuallyfigure-out impact of real-time decisions on predictedcompletion dates.Honoring delivery on committed date is moreimportant than how fast one turns around an orderin a manufacturing facility. It is only possible to doso, if one is in a position to predict impact of allscheduling decisions taken now on all orders in-handin real time as an on-going process.3 Why conventional scheduling systems fail in thereal world?3.1 Static schedulingScheduling is a widely misunderstood term. Manybelieve plotting activities to be performed ondifferent resources on a time scale (Gantt chart)make a schedule. Actually a Gantt chart is just asnap shot of what is likely to happen in the futuregiven the situation NOW. As one progresses in timethis representation will change because predictionsseldom match reality owing to unexpecteddisruptions [3] & [5].So, any scheduling system that fails to respond tochanging situation on ground by failing to rescheduleand redraw its prediction (Gantt chart) is a misfit inreal life making purported schedule optimality andefficiency claims hypothetical.3.2 No feedback mechanismA scheduling system can only be responsive to whatis happening on the shop floor if a feedbackmechanism exists. This feedback mechanism shouldbe both real-time and automatic. Peter Cowling andMarcus Johansson [2] argue in a well researchedpaper that “in many production processes real timeinformation may be obtained from process controlcomputers and other monitoring systems, but mostexisting scheduling models are unable to use thisinformation to effectively influence schedulingdecisions in real time”. This is a major disconnectmaking the schedule infeasible as it is soon out ofsynchronization with reality.We have recognized that the only authentic real-time source of feedback information from the shopfloor is the personnel (Operators’) in charge ofperforming individual activities. However, eachoperator can only give feedback on what each onedoes and that too ideally limited to the current taskin-hand. We achieve a seamless feedbackmechanism to the scheduling engine by allocatingelemental executable tasks in real-time to individualOperators, and seeking task-specific feedback foreach such allotted task. The process of taskallotment, progress feedback and subsequentreschedule to decide what to do next happens 24x7automatically.3.3 Schedule not actionableThe decision to execute an elemental task or activityof an order requires one to take into account severalaspects; availability of inputs, availability ofresources and technical feasibility of performing thetask. Most scheduling systems usually fail on thiscount. Proposed activities are either not actionableor represent a group of activities leaving the decisionof what exactly to do now to the operators. In orderto circumvent this problem, many systems offer a“drag & drop” facility to correct or manipulateproposed schedule before it is released. As [4] P.
    • Page 3 of 13© 2013, Laxman C MaratheVelaga asserts, presence of a “drag & drop” facilityindicate an inherent weakness of the schedulinglogic.3.4 Manufacturing facilities are on-going concernsGetting new orders and completing existing orders isa continual process in real facilities. Existingcommitments cannot generally be disturbed becauseof new orders. Situations can become more complexas orders could be cancelled or amended. Anyscheduling solution that considers a static order loadis therefore impractical.3.5 Working in shiftsMany manufacturing facilities work round the clockin shifts manned by a different set of personnel.Scheduling decisions impact across shifts and thebiggest challenge becomes information handoverbetween shifts. The only remedy is in having thescheduling systems work 24x7 continuously.3.6 Stability versus responsivenessMost scheduling systems provide a stable schedulefrozen for a period (usually a few days) and expectsit to hold well unless disruptions occur, whichinevitably do occur. It is reasoned that having acontinually changing schedule results in shop floornervousness. Shop floor nervousness is a mythpropagated to hide inability of doing a quickreschedule. Operators’ are only concerned with thetask in-hand. As long as the current task remainsunaltered any amendment to future task listing in noway adds to nervousness. On the contrary, impacton completion dates of all jobs in hand must beknown immediately not when the next frozenschedule is created.We propose a true scheduling system called TalikaProduction Management System (PMS) that satisfiesall the above primary requirements.4 Overview of Talika PMSThe system has a distributed architecture asindicated in Figure-1. At the center is the real-timescheduling engine working round-the-clock and isthe live heart of the system. Several different typesof consoles interact with the scheduling engine usinga proprietary protocol that is robust andasynchronous making the entire process ofcommunication absolutely safe.There are several different types of consoles eachdesigned to perform a specific function on the shopfloor. Consoles work in a standalone mode but canalso communicate with the scheduling engine, ifconnected, making the entire distributed system liveand reliable. Exhibit–1 at the end detailsfunctionality of each Console shown in Figure-1 andexplains how the automatic scheduling engine drivesother peripheral or support activities. Most ERPsystems only handle the peripheral activities sansthe driving scheduling engine at its heart, making itmore of a fancy carcass disconnected from the shopfloor.
    • Page 4 of 13© 2013, Laxman C MaratheFigure-1: Macro system schema5 How the system worksFigure-2 gives an at-a-glance view of how theentire system works.5.1 Starting pointJob Study Wizard (JSW) is the starting point. Asalready explained each sales person orconcerned agency can have a JSW of their own.Potential enquiries can be quickly converted to adetailed job definition depicted as an easy tounderstand component task (CT) diagram. CTdiagram actually represents the micro levelactivity work flow for creating one-something ofany value-added service or product. It is morelike a recipe. One can always scale it up or downto match extent of final output required keepingthe CT diagram (recipe) unchanged. It is alsopossible to create, as a one-time exercise, abank of most standard CT diagrams (standardorders) used in the facility. So, defining neworders may simply translate into picking up anappropriate or nearly matching already definedCT diagram and making minor adjustments to it.One can also create part CT diagrams forcommon work flows in the factory and savethem as sub-assemblies. Sub-assemblies arebuilding blocks one may use to quickly create anew complex job definition.Jobs are stored as proprietary files with adefault “*.tlk” extension to any media. One cansave, share and reuse stored jobs over and overagain just like a text file.5.2 Scheduling a jobIt is not necessary all defined jobs be actuallyscheduled. Jobs could be defined when wereceive an enquiry to estimate its cost andassess delivery date, but we only need toschedule the order when it matures. Whenscheduled, orders flow over to the schedulingengine and the process of executing itsconstituent tasks begin.5.3 Role of scheduling engine and shop floorinterface: Work Center Console (WCC)Scheduling engine works 24x7 and proactivelycontrols all factory work centers at a micro level.
    • Page 5 of 13© 2013, Laxman C MaratheIt decides what tasks of which orders can andshould be actually allotted for execution to theshop floor. Complete information of currentlyallotted tasks flows to the concerned WCC inreal-time. Operators’ acknowledge allottedtasks to start execution and notify interimmilestones achieved, until the allotted task isnot over. All notifications flow back in real-timeto the scheduling engine to be taken cognizanceof during the next reschedule that happensevery minute. This cycle of allotting tasks,getting progress and completion notificationfeedback, and subsequent fresh allotment oneach work center on the shop floor goes onwithout end.Figure-2: Working principle at-a-glance6 Working logic of scheduling engineScheduling engine comprises of a set of complexdaemons working round-the-clock. Like ahuman scheduler does, it always decides whattasks to execute now. The entire optimizationprinciple could be summed in one line as “ifsomething (read a task) can be done and itshould be done then it will be done”. The aboverule automatically guarantees that resourceutilization is maximized while simultaneouslyreducing job cycle time.An order is first broken-down to its elementaltasks in form of a CT diagram during definition inthe JSW itself. Only on confirmation, validorders are communicated to the schedulingengine. During order definition stage itself a lotof detailing about the job is done including de-selection of technically non-feasible workcenters to execute specific tasks of the job.User can also specify several guidelines for thescheduling engine to follow while executing theorder called “execution preferences”.Scheduling engine uses its own intelligencewhile implementing user specified guidelinesbut ensures they are honored wheneverpossible. Execution preferences are not rigid;
    • Page 6 of 13© 2013, Laxman C Marathethey can be changed even at run-time after ajob is scheduled. Execution preferences couldbe different for each task. However, user canspecify them just once with applicabilitycontrolled across parts or group of tasks or forall tasks in an order. Table-1 below lists theexecution preferences and explains what eachmeans and how the scheduling engine usesthem while making allotment decisions.Table-1: Execution Preferences and what they meanExecution preference What it means Scheduling engine usageWIP controlAttempt to minimize work-in-progress (WIP) from beingcreated too much in advanceand thus remain unused.If WIP is not on the criticalchain and has enough time leftto be produced and used thenits creation is deferred therebyminimizing WIP build-up on theshop floor.Control of taskexecution orderUser desires to change taskexecution order, if necessary, atrun-time.Tasks are allotted first by orderpriority and then by the futureburden on the task within anorder. However, user maychange this natural order ofexecution at run time.Work center choiceIf one has a choice of workcenters to perform a task thenwhich one to choose?Scheduler tries to honors userpreference with switchoversavings, if any, considered. Incase the first preferred workcenter is unavailable it tries toallot the task on the secondpreferred work center and soon.Locking OptionEnsuring a particular task isonly executed within a userspecified period.Always tries to execute the saidtask within the specified period,as far as possible.Auto-breaking optionBreaking up a task to runconcurrently on more than onework center with an intentionto reduce task execution time.If the task is on the critical chainor its execution cannot bedeferred any further schedulingengine will try to optimize andselect the most appropriatebreaking option possible.Spanning OptionStop and resume task executionafter a holiday, recess period.Commonly referred to as a non-scheduling time zone (NSTZ) inthe system.Scheduler wisely decides tospan or not to span dependingon the current situation.
    • Page 7 of 13© 2013, Laxman C MaratheExecution preference What it means Scheduling engine usageMCI optionIt may not be necessary to waitto start the next value-addingtask that uses or consumeswhat is produced by the currenttask until the current task is notover. One can overlap in timeboth tasks in order to expeditethe order. We can say thepreceding task gives a mid-course intimation (MCI) to thenext task to begin.Scheduler tries to begin thenext value adding task evenbefore the earlier one feedinginto the next one is not yetover. Time to initiate the nexttask can be user decided or leftto the scheduling engine tofigure out.Interleaving optionUser may want some tasks(orders) to be executed onlywhen there is free timeavailable. Contrast this withauto-breaking where theobjective was to expedite.Scheduler ensures the task isexecuted whenever there isnothing urgent to be done.MCF OptionEspecially in long running tasksinterim milestone reachedfeedback may be necessary tore-adjust expected taskcompletion time. We call it amid-course feedback (MCF).MCF is used constructively toadjudge the expectedcompletion time for longrunning tasks.NSTZ cut-in optionNSTZ is an acronym for non-scheduling time zones. Periodswhen the scheduling engine willnot schedule (allot) a fresh task.However, an already runningtask can either by design (orbecause it is delayed) cut-intoan impending NSTZ. Systemsupports five categories of NSTZwith varying importance anduser can define how much aparticular task can actually cutinto each of them.Scheduling engine takesappropriate decision to cut intoNSTZ whenever necessary.Working during NSTZ is anadditional cost and calculatedaccordingly.
    • Page 8 of 13© 2013, Laxman C MaratheExecution preference What it means Scheduling engine usageTask line-upIt is possible for user to specifythat in an order if some task Ais executed on a particular workcenter then preferably task Btoo should be the next onefollowed by task C and so on.Valid reason could besubstantial saving in cost andtime if done so. We call it taskcascading. This again could bea preferential cascading or aforced cascading when userinsists that the schedulingengine waits a pre-determinedperiod for the next cascadedtask to mature for execution.You can guess concept ofcascading is different fromcontrolling task executionorder. The former is applicablewithin an order whereas thelater could be across orders andtypically is a run-time userintervention.In addition to above user specified executionpreferences, the scheduling engine takes intoconsideration several other aspects as well anddoes its own run-time adjustments as listedbelow.6.1 Work center capacityChecking if it is possible for a given task to beexecuted on a work center must be done beforeeach allotment. Our system allows user todefine multi-part work centers that could eitherwork as a whole or in parts enabling one toexecute a variety of tasks each requiring it ownpart capacity profile.6.2 Activating work centerResources and work centers are conventionallythought as synonyms, but in our system aresource has a very special meaning: a workcenter to become active requires resources.What resources are required to activate a workcenter is user defined. Therefore, if a workcenter is currently not active it is necessary tocheck for resource availability. Task allotmentcan only happen if it is possible to activate awork center. This check is done automaticallyby the scheduling engine.Activating work centers could also be dependenton capacity usage. A typical case could be anindustrial oven that is uneconomical to be fired-up unless filled-up to some predefined minimumcapacity.6.3 Considering time for material movementand normalizationIn real facilities it takes a while for work-in-progress to be moved from the place it iscreated to where it is needed for further value-addition. This time too must be taken intoaccount before deciding fresh allotment. Bothfixed and variable types of material movementare considered and require separate notificationfrom a special console called MaterialMovement Console (MMC) given to the personresponsible for material movement.Additionally, certain WIP may require time toset, dry, solidify, etc. We call it time tonormalize the WIP produced before furthervalue-addition on it can begin.6.4 Deciding need for expediting or skippingtask allotmentBreaking a task over more than one workcenters for concurrent execution is onlyadvantageous if the task in question has reacheda critical stage. In our language, has sufficientlyexhausted the available leeway. Schedulerreckons how much the current leeway availableis before taking such decisions. Likewise, ifsufficient leeway is available and if the userdesires task allotment may be skipped allowing
    • Page 9 of 13© 2013, Laxman C Maratheother urgent tasks to be handled in themeanwhile.Further, if sufficient time is not available tocomplete a task as one has an impending NSTZor locked task then the scheduling engine couldeither span the task, if possible, or decide to skipallotment until later.6.5 Deselecting inappropriate work centerAt run-time work centers that were originallythought appropriate to execute a task maybecome inappropriate as they waste more thanthe reported good count of inputs actuallyavailable now. Similarly, in a multi-plant facilityif certain WIP is created in one plant and thenext value-adding work center too is available inthe same plant but not currently free then thescheduling engine may decide to wait for it tobecome free rather than send WIP to anotherplant’s work center if doing so is advantageous.6.6 Deciding to hasten-up task executionNo matter how complex a rule one may use toanticipate task’s total duration it is still anestimate. When situation demands one mayslightly expedite task execution to finish it fasterthan expected. It is a done thing in practice andthe scheduling engine too, if necessary, does thesame, of course within user permitted limits.6.7 Decision to re-purpose inputsIdentical inputs could be processed by differenttasks to produce something different.Assignment of specific task inputs is rathernotional and one can, if need be, re-purposeinputs to expedite those tasks whose otherinputs are deemed available. Human schedulersoften take such decisions and so does thescheduling engine provided user allows (ordefines) such a swapping as possible.6.8 Decision to freeze part or whole orderIn case of any reported shortfall in WIP count forany reason it makes sense to temporarily haltorder execution, make good the shortfall andthen resume executing order again. Humans dotake such decisions and so does the schedulingengine. It decides to suspend order executionwhile raising an alarm for human intervention toamend order workflow.6.9 Monitor completion is within committeddateGenerally one must keep some safety bufferbetween when actually an order will becompleted and the date of delivery committedto the customer. On each reschedule, expectedcompletion time for all orders are re-calculated.However, if for some reason order completioncrosses the cut-off date an alarm is raised by thescheduling engine.6.10 Monitoring task execution (duration,wastage, cost etc.)Expected duration, cost, wastage, time for theoutput from a task to become usable forsubsequent value-addition (normalization time)and capacity the task may partake of each workcenter it can be executed on, are all calculatedduring job definition stage itself in the JSW.User can define complex formulae and lookuptables using attribute values specific to eachtask to arrive at these figures. However, thescheduling engine also captures the actualvalues in each case. Doing so not only allowsone to control deviation task-wise at run-timebut enables periodic revision of estimation rulesin order to match them to reality as closely aspossible.For example, if any task actual execution timeexceeds its estimated duration it turn black onthe live Gantt chart allowing concernedsupervisors to only focus on late tasks. Severaluseful reports too can be generated highlightingexceptions. Actual vis-à-vis estimated data canalso be used to tailor a micro-level incentivescheme as resource capacity is translated intime terms and thus easier to assess andmonitor.6.11 Procurement and maintenance too areconsidered tasksWe consider procurement of customer inputsand raw material too as tasks performed bycustomer interaction personnel or buyers. Anydeviation in expected arrivals of inputs has abearing on the overall schedule. Likewise,
    • Page 10 of 13© 2013, Laxman C Marathemaintenance activities also keep the workcenters busy and affect the schedule and thusare treated as tasks.6.12 Actual task execution may not always besuccessfulEvery allotted task may not be completedsuccessfully. We have the following optionsavailable with the Operator for an allotted task.Operator can roll back an allotted task with arequest to reassign it later. In case, Operatorhas already started working on the task it couldstill be re-allotted: a way of telling thescheduling engine that it is not possible tocomplete the task now though it can becompleted later by me or by someone else.Operators’ can pause and resume working on atask. In the worst case, Operators’ can alsodeclare a task as terminated meaning it is nomore possible to complete the task as inputs areeither damaged or destroyed - an errorcondition requiring human intervention to makegood the shortfall. All the above impact theschedule and are considered by the schedulingengine.Then there are several more activitiesperformed by the scheduling engine like – Reassessing what is completed until now How much more time existing tasks wouldrequire The actual time, wastages and costs(including overtime cost) incurred until nowand so on.It is very easy to guess, a lot of thinking happensto ensure that each allotted task can indeed beexecuted on the shop floor and everyeventuality, even after task allotment, is takencognizance of. Technically the schedulingengine can run autonomously with inbuiltcapability to raise an alarm for humanintervention only when situation so warrants – aprecondition for realizing a true computercontrolled manufacturing facility.7 Vital statistics from the first successfulinstallationThe entire system is now mature and rigorouslytested to exacting conditions in its first full-fledged installation at a medium sizedcommercial print setup in India. It has beenworking for more than 3 years now giving us theconfidence to make it available for the benefitof the world at large.The system is user configurable and starts bydefining the manufacturing facility in detail.They include identifying:7.1 Work centersListing of individual work centers of the factory,classified by departments, and if a multi-plant(location) facility, then by plants. The firstinstallation is a multi-plant facility. Table-2 givesdetails of the work centers and theirdistribution.Table-2: First installation work center detailsNumber of individual workcenter419Number of departments 47Number of plants / locations 57.2 Tasks and what they produceTasks get executed on work centers. Tasksproduces some things recognized as“component” and may also require some thingsto add value to, again a “component”. The tasksand the components it produces actually makeup the CT diagram. User must define whatelemental value-adding tasks can be performedin the facility and what generic componentsthey produce. They are but few in type - whatchanges from order-to-order is are theattributes of generic tasks and components like,extent to be value-added, cost, wastage,duration, etc. Table-3 details the number ofgeneric tasks and components defined in thefirst typical installation.
    • Page 11 of 13© 2013, Laxman C MaratheTable-3: First installation task & componentdetailsNumber of generic tasks 27Number of genericcomponents71Actual system performance as on 19thApril2013.History: The first installation is in its fourthsuccessful year with the average mean timebetween system failure now exceeding 6months, which in the beginning was around 6minutes. That shows how reliable the systemnow is.Each time a job is scheduled it is given a runningserial number. It started from ‘1’ and now reads30900. With 706 currently active jobs, it means30194 jobs were successfully executed by or viathe system with each job having about 70elemental tasks on an average.How fast it works: The scheduling Engine workson Dell T310 Power edge server. It has 706currently active jobs with 49466 elemental tasksto schedule individually with all the complexityof decision making already described. Table-4gives an actual peek of the speed at which thesystem works on this date.Table-4: First installation Scheduling Engine loadNumber of active jobs 706Number elemental tasks toreckon with49466Time to decide what to doNOW (seconds)4Time to reschedule: predictmicro-level future schedulecompletely (seconds)25Scheduling engine work at a phenomenal speedof about 1900 tasks / second when itreschedules, that happens once every minutemaking the system live. You can guess thedecision to allot tasks now and knowing effect ofall current decisions as schedule prediction areindependent processes. Time to reschedule isdecided by the number of elemental taskspresent and varies linearly. In worst casescenario, if time to reschedule exceeds 60seconds the system automatically, for suchinstances, chooses to skip a reschedule to alignwith the next minute.8 ConclusionTalika PMS is in its infancy. It is just born. Notmany are even aware that such an inexpensive,easy to use, self-configurable, off-the-shelfproduct exist that holds the promise ofpositioning any manufacturing facility leaguesapart from its competitors in terms of cost ofproduction and reliability of service offered. It isjust a matter of time before someone makes abeginning forcing others to adopt similarsystems just to remain in business.You can know more about Talika PMS byvisiting www.etalika.in and also download afree full demo version for evaluation.
    • Page 12 of 13© 2013, Laxman C MaratheExhibit–1: Console Functionality in briefConsole Function Number & LocationJob Study Wizard(JSW)Define and estimate all aspects of anorders;Simulate or schedule orders;Monitor order progress;Manage / Change order execution;CRM functionality.JSW is a multi-use console. Salespersons, Customer Supportpersonnel, shop floor Managersand Supervisors and evencustomers all can have one for theirpersonal use.Work Center Console(WCC)It is the Operators’ console. Details ofall allotted tasks flow in real-time tothese consoles for Operators’ to notifytask progress milestones.Also gives details of all tasks alreadydone and those lined up for executionlater.Valuable machine statistics and manymore helpful features.One WCC can represent one orseveral or all work center in afacility. Users can tailor thenumber of WCCs required to coverall work centers on the shop floor.System puts no higher limit.Customer InteractionConsole (CIC)Any inputs required from customers?Track, follow-up and notify input whenthey arrive in order of requirementAs many as personnel involved inmanaging customer inputs.Maintenance Console(MTN)Preventive Maintenance as well asunexpected breakdowns engage workcenters and affect the schedule. Onecan define preventive maintenanceschedule in advance and treat it like amaintenance job that can be scheduledlike any other order. This console helpsdefine a preventive maintenanceprogram, schedule it and notify itsactivities.As many as required.Material MovementConsole (MMC)Movement and storage of work-in-progress is a critical function in thevalue-adding process and affects theschedule. We designed a console forthe person in charge of work-in-progressthat performs all functions as requiredin real-time.As many as required.InventoryManagementConsoles (IMCs)Orders may require raw material eithersourced from stores or purchased.Need for material is decided and drivenby the scheduling core and thus anentire material management system iscreated around the scheduling core.This one console takes different avatarsdepending on what specific inventoryfunctionality is required.As many as required.
    • Page 13 of 13© 2013, Laxman C MaratheConsole Function Number & LocationFactory dB Manager(FDM)System is fully user configurable and allthis information resides in one logicaldatabase called the factory dB.However, user must have a means tomodify the factory dB without affectingcurrent working system. FDM allowsone to check-out locally a copy of thefactory dB for manipulation / change,revalidate it and check-in the factory dBwhen finalized.As many as required.Money ManagementConsole (MMM)All monetary information generated bythe scheduling core is fetchedperiodically by this console for financialaccounting purposes. One can thenbuild or dove tail this information intoany existing financial ERP system.As many as required. Only inconcept stage.Human ResourceConsoles (HRC)Manpower information like past usage,current manning information being usedand future requirements too flow fromthe scheduling core. This console isdesigned to cull out or control suchinformation or feed it into any existingERP system.As many as required. Only inconcept stage.References[1] D. Ouelhadj, S. Petrovic - A survey of dynamic scheduling in manufacturing systems - Springer Science:Journal Scheduling (2009) 12: 417–431 - Published online: 28 October 2008[2] Cowling, P.; Johansson, M. - Production, Manufacturing and LogisticsUsing real time information for effective dynamic scheduling - Elsevier: European Journal of OperationalResearch 139 (2002) 230–244[3] Guilherme, E.V.; Herrmann, J.W.; Lin, E. - Rescheduling Manufacturing Systems: A framework ofstrategies, policies and methods - Journal of scheduling, Kluwer Academic Publishers, Netherlands[4] Velaga P., Ph.D. (Scheduling) President, Optisol, 3910 Stony Creek Ln, College Station, Texas 77845 -Advantages & Difficulties with Drag-and-Drop Operations – Web page link: http://www.optisol.biz/Drag-and-Drop.htm[5] Zhang L., Li, X., Gao, L., Yang, Y., Jiang , P. - Predictive/reactive scheduling with uncertain disruptions -proceedings of the 41st international Conference on Computers & Industrial Engineering P 260-265