OM Chapter 11: Capacity Management PDF
Document Details
Uploaded by ReadyOrphism6804
Tilburg University
Tags
Summary
This document discusses capacity management, a critical area of operations management. It covers the concepts of demand and supply, various ways to measure demand, and the significance of capacity plans to meet both customer satisfaction and resource efficiency in various operational contexts.
Full Transcript
Chapter 11: Capacity management 11.1 WHAT IS CAPACITY MANAGEMENT? - Capacity managementis concerned with understandingthenature of demand and supply (capacity) and attempting toreduce mismatchesbetweenthem; aims toreconcilethe competingdemands of customer...
Chapter 11: Capacity management 11.1 WHAT IS CAPACITY MANAGEMENT? - Capacity managementis concerned with understandingthenature of demand and supply (capacity) and attempting toreduce mismatchesbetweenthem; aims toreconcilethe competingdemands of customer satisfaction and resourceefficiency - Level of capacity decisions must bemade within the constraints of the operations, ability of suppliers, availability of staff,... - Short-term decisions provide important feedback for planning over longer-term time horizons > - edium-term aspectof M capacity management -> decisions madelargely within the constraints of thephysical capacityset by the operations long-term capacity strategy - Involves assessing demand forecastswith atime horizon of2-18 months-> planned output canvaried(example: by changing the numberof hours that resources are used) - In practicenot many forecasts are accurateand mostoperationsmustrespondin changes in demandon ashorttimescale->short-termcapacity management - Example: hotels and restaurants haveunexpected changesin demandfrom night to night and also experience certain days that are on average busier - Decisions taken by OM about capacity plans will affectmany aspects of performance 1) Costswill beaffected by the balancebetween demandand capacity -> capacity levels in excess of demandmeanunderutilisationof capacityandhigh unit-costs ) Revenueswill also beaffected by the balancebutin theoppositeway;capacity levels 2 equal to/higher than demandensure that demand issatisfied andno revenue is lost 3) W orking capitalwill be affected if an operation decides tobuild up finished product inventory prior to demand;demand will besatisfiedbut organization will have tofund inventory until sold 4) Qualityof services might be affected by a capacityplan that involveslarge fluctuations in capacity levels(e xample:hiring temp. staff ->new staff are adisruptionto the routine andincrease the probability of errors) 5) Dependabilityof supply also affects demand and capacity->closer demandis to operation’scapacity ceilingtheless it’s able todeal with unexpected disruption 6) Flexibilityespecially volume flexibility, will beenhancedby surplus capacity -> demand and capacity inbalance= operationwon’t respondto unexpected increase in demand - Time series of activities involvedin capacity managementare mentioned in the figure^ - 1st stepon thedemand sideis tomeasuredemandforservices and products over different time periods -> select from a range ofqualitative(panel, Delphi, scenario planning) andquantitative(time series andcausal models)toolsto create a more accurate prediction of demand - 2nd stepon thesupply side is tomeasure the capacityto deliverservices and products -> the impacts of mix, time frame, and output specifications has to be considered - 3rd stepis to considerif and how to manage demandusingdemand managementandyield management techniques - 4th stepis tomanage the supply sidebydeterminingtheright level of average capacityand either decided tokeep this constant(level capacity plan) or to adjust capacityin line with the demand patterns (chasecapacity plan) - OM mustunderstand the consequencesof different decisionon both sides 11.2 HOW IS DEMAND MEASURED? - First task of capacity management is tounderstandthe patterns of demandfor products and servicesover various time frames; however knowingif the demand is rising or falling isnot enough in itself-> we need therate of change - Example:a firm of lawyers might have todecide thepoint in which its business will take on another partner, which couldtake monthsso they must be able to forecast when they expect this to happenand startsits recruitment - Qualitativeapproaches to forecasting: 1) Panel approach - Panel is like afocus groupallowing everyone to talkopenly - Benefit ofmultiple people discussingbut also makesitdifficult to reach a consensus;possible that the loudest/ highest statuesideas might emerge -> bandwagon effect - More reliable than one person's viewbut still hastheweaknessof getting things wrong 2) Delphi method - Best knownapproach to forecasts; moreformal - Attempts toreduce influencesfrom face-to-face meetings - Steps: a) Survey of expertswhere their replies are analysedandanonymous summariesare sent back to them b) The experts asked toreconsider their original forecastswith the new arguments in mind c) This isrepeated several timesuntil aconsensus ora narrower rangeof decisions is reached - It’s possible toimprove this processbyallocatingweight to individuals and their suggestionsbased on experience, past success,people’s view of their abilities,... - Issuewith this is theconstruction of appropriatequestionnairesandselecting an appropriate panel 3) Scenario planning - Used forsituations with a lot of uncertainty - Applies tolong-range forecastingusing apanel - Members are asked todevise a range of future scenarios,that is discussed with inherent risks considered - Not concerned with reaching a consensusbut ratherlooking at arange of optionsand making plans totry to avoid the leastdesired ones+ taking actions tofollow the most desired - Quantitativeapproaches to forecasting 1) Time series analysis - Examines the pattern ofpast behaviors to forecastthe future - ooks atpatterns of time series dataand by removing underlying variations L with assignable causes,extrapolates future behaviour a) Simple moving- average forecasting - Used to estimate demand for a future period by averaging demand for the n most recent time periods - The value of n can be set at any level but usually in the range of 3-7 b) S imple exponential smoothing - The main disadvantage ofmoving averagesis thattheydon’t use data from beyond n periods in forecasting - This approachforecasts demandin the next period bytaking into account the actual current demandand the previous forecast - It uses thisformula > - Thesmoothing constantis theweight given to the last piece of information available(assumed most important) to the forecaster - Other expressions in the formula include theforecastfor the current period, which includeprevious period’s actual demand, and so on - All previous data has an effect on the next forecast c) Trend-adjusted exponential smoothing - aindisadvantageof simple exponential smoothingis that itassumes a M stable underlying average - If there is atrendin the average the exponential smoothedforecast will lag behind the changes in demand -Higher smoothing constants(>0.5) help toreduceforecast errors but there maystill be a lagif the average is changing d) Seasonality in forecasting - Most organisations experienceseasonal patterns indemand - Causes can be climatic (holidays), festive (gift purchase), financial (tax processing), social, or political - In forecasting we use the termseasonalityto describeanyregularly repeating changes in demand(e xample:quarterly, monthly, weekly,...) - Example:utility companies experience larger annualseasonality but will also face seasonal patterns over the week and across the day - Multiplicative Seasonal model-> popular techniqueof incorporating seasonality in 5 steps 1) Find the average demand for each seasonby summingthe demand for the seasons by the number of available seasons \ - Example:March had sales of 80, 75, and 100 over thelast 3 years to the average March demand is (80 + 75 + 100)/3 = 85 2) Calculate the average demand over all seasonsby dividingtotal average demand by the number of seasons - Example:total average annual demand is 1320 and thereare 12 seasons the average demand equals 1320/12= 110 3) Compute seasonal indexby 1) over 2) - Example:march seasonal index -> 85/110 = 0.773 4) Estimate the next period’s total demandusing oneor more of the qualitative or quantitativemethods described above 5) Divide this estimateby thenumber of seasonsandmultiplyby the seasonal indexto provide a seasonal forecast 2)Causal models - Causal models oftenemploy complex techniquestounderstandthe strength of the relationshipbetween the network of variables andthe impact they have on each other - Simple regression modelstry to determine thebestfit expression between two variables - Example:suppose an ice-cream company is trying toforecast its future sales; after examining previous demand it can see that the main influence is the average temperature of the previous week so they plot the demand against the temperature -> using the graph the make a reasonable prediction of demand once the average temperature is known provided that the other conditions prevailing in the market are reasonably stable - If themarket conditions are unstablethey will have to be included on a multiple regression model -> networks of many variables and relationships, each with their own assumptions and limitations - Techniquesare available to help managers undertakemore complex modellingand feed back data into the model torefine and developit further - Three key ways toassess the usefulness of a demandforecastfrom an OM perspective 1) Level of accuracy - Big help for the process, becausedemand can changeinstantaneouslybut there is usuallya lag between deciding to change capacityand change happening - We calculate theforecast errorin order to assessthe relative accuracy of a forecast 2) Indication of relative uncertainty - Mostimportantindicator becausedecisions to workextra hoursand recruit extra staffcome from forecastlevels of demand which candiffer from actual demand-> unnecessary cost, unsatisfactory customer service - Example:a forecast of demand levels in a supermarketmay showinitially slow businessthat builds up to alunch time rush, afterwhichdemandslowsandbuilds up againfor theevening rushandfall againat theend of the day; the manager can use theforecast to adjust checkout capacityhoweverno day will fit exactly these predicted patterns - It is also important toestimate how much actual demanddiffers from the average by examiningdemand statistics to build up adistributionof demand at each point of the day-> the manager will have an understandingof when they need to have reserve staff 3) Expression in terms useful for capacity management - If aforecastisexpressed only in money termsandgivesno indication of demandsplaced on an operation’s capacity it needsto beexplained in terms of realistic expectations of demandin thesame unitsas capacity(staff, machines, space,...) - Example:some retail operations use sales forecaststo allocate staff hours throughout the day, yet sales also depend on staff allocation so its better to use the number of customers as a forecast to provide enough staff 11.3 HOW IS CAPACITY MEASURED? - Capacityof an operation is themaximum level of value-addedactivityover a period of time that theprocess can achieve under normal operationconditions-> reflects the scale of capacity and its processing capabilities - Example:apharmaceutical manufacturerinvesting ina new1000l capacity reactor which gives you agood sense of the scale of capacitybutit’s a useless measure for an OM; instead they will be concerned with thelevel of outputthat can be achieved with the reactor -> batch takes one hour, planned processing capacity will be 24,000 litres per day - Measuring capacityis veryambiguousunless to operationis standardised and repetitive - Example:a theme park ride designed to process batchesof 60 people per minute; 1200 people an hour ->output capacity measureis the most appropriatebecause the operation doesn’t vary - A lot of operationsdon’t have a straightforward definitionof capacity; if there is awide range of outputsthat placesvarying demandon theprocessinput capacity measures are more useful - Almost every type of operation can use amix of bothinput and output measuresbut usually they only choose one - Anoperations ability supplyis dependent onwhatit’s being required to do - Example:a hospital may have a problem in measuringits capacity because the nature of the service v aries significantly; output depends on the mix of activities in which the hospital is engaged in and because they usually perform different types of activities output is difficult to predict input and output < capacity measures for different operations - roblems caused by variation mix can be overcome by using aggregated capacity P measures - Aggregated -> different products and services are bundled together in order to get a broad view of demand and capacity - The level of activity and output that may be achievable over short period of time is not sustainable on a regular basis - Example:a tax return processing office during theend or the financial year may be capable of processing a lot more of applications a week because they extend the working hours, discourage staff from taking vacation,... to avoid disruption -> staff o need vacations and can’t work such long hours continuously d - Design capacity: theoretical capacity of an operation that its technical designers had in mind when it was commissioned - Effective capacity planning: capacity of an operationafter planned losses are accounted for - Actual output: capacity of an operationafter bothplanned and unplanned losses are accountedfor (quality problems, machine breakdowns,absenteeism) - Two measures ofcapacity performance: - apacity leakage:reduction in capacitycaused bybothpredictable and unpredictable C losses - Overall equipment effectiveness(OEE) is a methodof assessing it: OEE = a * p * q a -> availability of process p -> performance/ speed of a process q -> quality of product/ service that the process creates - OEE works on the assumption thatsome capacity leakageoccurs that causes reduced availability - Example:availability can be lost throughtime lossessuch asset-up and changeover losses(when people are being preparedfor the next activity), breakdown failures(when the machines are being repairedor employees are being trained) or throughspeed lossessuch asequipment idling (temporarily waiting for work from another process) and when equipment is run below the optimal rate+not everything processedby an operationwill be error freeso somecapacity is lostas a result of inspection, rework, and complaint handling - or process to operate effectively high levels of performance in all dimensions has to be F achieved, however they don’t give a complete picture -> better to combine them (OEE), gives a more accurate reflection of the valuable operating time ^ OEE for a client support service team in a small software company - Operations have to also have tocope with variationin capacity - Example: especially useful ifcapacityis relativelyfixed, themarketcan be clearlysegmented,theservicecannotbestoredin any way and it issoldinadvance, and themarginalcostsof a sale arelow - Example:airlines fit this criteria because they adopta collection of methods to try to maximise profit from capacity; overbooking capacity to compensate for no-shows, but if more passengers show up than they will have more upset passengers -> theystudy past data on flight demandto balance the risks - An operation willdiscount priceswhendemand doesn’tfill capacity - Example:hotels will offer cheaper room rates outsideof holiday seasons to increase naturally lower demand, or large chains sell rooms to third parties who find customers 11.5 HOW IS SUPPLY SIDE MANAGED? - apacity managementdecisions includesetting the base capacity leveland using two C methods of managing supply –level capacity plans(nominal capacity is constant) and chase capacity plans(capacity adjusted to chase fluctuationsin demand) - Common starting point is to decide the base level based on 3 factors and then adjust it accordingly 1) Effect of performance objectives on the base level - Base level should be set toreflect performance objectivesof the operation -> if thecapacity is set too highit will result inlowlevels of utilizationof capacity, if thefixed costs are too highit can havedetrimentaleffect - If capacitybase levels are highit will create acapacity cushionthat allows for flexible outputand createsmore responsive customerservice - If theoutput can be stored, there may be atrade-offbetweenfixed and working capital - High level of base capacityneeds a lot ofinvestmentand alower base level reducesthis need…still mightrequire inventory tobe built upfor future demand ->increases working capital - Some operationscan’t afford to do thatdue toshortshelf-life (perishable food) or because theoutput cannot bestored(services) 2) The effect of perishability on the base level - Ifsupply/ demand are perishable,base capacityneedsto be set at ahigh level becauseinputs or outputs cannot be storedfor longperiods - Example:a factory producing frozen fruit needs sufficientfreezing, packing, and storing capacity to cope with the rate of the fruit harvest during the seasons; a hotel cannot store accommodation 3) The effect of demand or supply variability on the base level - Variabilityreduces the ability of an operation toprocess its inputs; thegreater the variabilityin arrival/ activity time themorethe process sufferershigh throughput times and reduced utilisation -> true for the whole operation - Long throughput times = queuesbuild up in the operation - High variability affects inventory levels-> thegreaterthe variability, themore extra capacity will be neededto compensate for thereduced utilisation - evel capacity plan->capacity is fixedthroughoutthe planning periodregardless of the L fluctuationsin forecast demand - Offersstable employment patterns,high process utilisation,andhigh productivitywithlow unit costs - Disadvantageis that it also createsconsiderableinventories of materials, customers, and info-> alsonot goodforperishableproducts, products that change rapidly and unpredictably, or forcustomizedproducts - Low utilisation effectscan makelevel capacity plansexpensive, but may be consideredif theopportunity costs of individuallost sales is high(example: high-margin retailing of jewelry and real estate agents) - Ifcapacity is set a little below the forecastpeakdemand, it willreduce the degree of underutilisation; in periods wheredemand exceeds planned capacity customer service may deteriorateand customers will have to wait longer, and be processed less sensitively - Chase capacity plan->matches demand patterns closely byvarying levels of capacity - uch more difficultbecausedifferent numbers of staff,working hours, M equipment,... may be necessary in each period - Pure chase plans are unlikelyin operations that manufacturestandard, non-perishable products or in capital intensive operations (high level of physical capacity needed) - Pure chase plan usuallyused in operations that can’tstore their output (e xample:customer processing operations, manufacturesof perishable products) - Avoids wasteful supply of excess staff that happenswith level capacity and it’s also able tosatisfy customer demandthroughout theplanned period - Ifoutput can be stored, thechase demand policy shouldbe adapted -> minimises finished goods inventory; especially iffuturedemandis unpredictable 1.6 HOW CAN OPERATIONS UNDERSTAND THE CONSEQUENCES OF THEIR CAPACITY 1 MANAGEMENT DECISIONS? - Managers attempt tobalancethe need to provide aresponsive and customer-oriented service with the need to minimise costs-> most companiesmix demand-side and supply-side capacity management strategiesto maximiseperformance - Example:an accounting firm seeks to bring forwardsome peak demand by offering discounts to selected clients…demand management plan - Capacitycan beincreased by using outsourced suppliersduring busy months… chase capacity plan - Some capacity may still be constrainedandclientstillexperience delays during high demand periods - Four methods to examine consequences of the companies decision: 1) Factoring in predictable VS unpredictable demand variation - Ifdemand is stable and predictableit’seasyto manage - Ifdemand is changeable but predictablecapacityadjustmentsmay be needed but areplanned in advance - Ifdemand has unpredictable variationthe operationhas toreact quickly otherwise the change in capacity will have little effect on the ability to deliver products 2) Using cumulative representations of demand and capacity - For anycapacity plan to meet demand as it occurs,itscumulative production linemust lieabove its cumulative demand line - Plottingdemand and capacity on a cumulative basishelps the feasibility and consequences of a plan to be assessed - Helps to create animpression of the inventory implicationsby judging the area between the cumulative production and demand curves - Chase capacity plans can also be graphed but rather than thecumulative production planhaving a constantgradient it is varieddepending on the production rate-> pure demand chase plan =cumulativeproduction and demand lines match and the gap is zero(inventoryis also zero) - There would still becosts associated with changingcapacity levels-> marginal costof making a capacity changeincreasewith the size of the change - Example:if a chocolate manufacturer wishes to increasecapacity by 5% it can be achieved by requesting its staff work overtime (simple, fast, inexpensive); if the change is 15% overtime doesn’t provide enough extra capacity and temporary staff needs to be employed (more expensive solution); if the capacity increases for more than 15% is it only possible to deal with bysubcontracting(most expensive) - Queuing theory: used when a capacity management decisionis madein an operation that cannot store output(services) -> managers acceptthat whilesome demand will be satisfied instantly,othersmight have towait… commonwhen individualdemand is difficult to predictor thetime to create service/product is uncertainso it is hard to provide adequate capacity at all times - ustomers arrive according to some probability distribution and wait to be C processed by one of the n parallel servers - ource of customers:in queue managementcustomersare not always human S (example:trucks arriving at a weighbridge, ordersarriving to be processed, machines waiting to be served,...) - Finite sourcehas aknown number of customers->example:one maintenance person serving four assembly lines knows the number of customers, i.e. 4; one line might break down and need repairing but that lowers the chance of another line breaking…number of customers arriving depends onthe number of customers already being serviced - Infinite customer sourceassumes there isa largenumber of potential customers… always possible for another customer toarrive no matter how many are being serviced -> most queuing systems that deal withoutside marketshas infinite/ close-to-infinite customer sources - Servers: a server is thefacility that processes thecustomers in the queue; usually they are configured in different ways (sometimes inparallel or in a series arrangement) - Example:self-service restaurantin which youqueueto collect a trayand cutlery, then you go to the serving area andqueueagain to order and collect the mealand do the same fordrinks, and lastlyqueueto pay… youpass four serversin aseriesarrangement - Queue systems arecomplexand there is oftenvariationin how long it takes to process each customer becausehuman servers vary intake to perform tasks -> usually described with aprobability distribution - The arrival rate: rate at whichcustomers needingto be served arrive at the servers… rarely steady and predictableand there is usuallyvariation - Thearrival ratesneed to bedescribedin terms ofprobability distributions - It isnormalthat sometimes there areno customersand sometimes there are many arriving at the same time - The queue: thewaiting listitself… when there isa limit on how many customers queue at the same time we assume that aninfinite queueis possible - Not always physical in nature->example:customerswaiting for a delivery of a customized product/ patient sitting on a waiting list for an operation - ueue discipline: set ofrules to determine the order of waitingcustomers -> usually Q first come first served - Sequencing rules from CH 10 apply - Rejecting:number of customers is at the maximumsoacustomermight berejected - Example:heavy demand on a website might block itfor some customers - Balking: customer is ahuman beingwith free willmay refuse to joining the queueand wait if they think its too long - Reneging: similar to balking but happens when acustomerqueues for some timebut leaves the queuefor a specific reason - Dilemma in managing queueing system capacity ishowmany servers to have at a time toavoid long queueing times/ low utilisation - Onlyrarely does this matchso sometimesqueues buildupor someservers become idlebecause of less customers - Even whenaverage capacity = average demandtherewill beboth queue and idle time - Too few serversmeansqueues build upandcustomersmight becomedissatisfied with the waiting time even if theutilisation levelsare high - If there aretoo many serverscustomersmightnotwait as longbut theutilisation levels will be lower - Capacity planning means atrade-off between customerwaiting time and system utilisation-> it is important to predict both ofthese factors - Queues are not something we want but theycan be managedto be more satisfactory -> they are animportantaspect becausecustomersjudge the service based on the queuing time - Management of queuing involvesmanagement of customersperceptions and expectations - Principlesthat helpevaluatingqueues 1) Unoccupied time feels longerthan occupied time 2) Pre-process waits feel longerthan in-process waits 3) Anxietymakes wait seem longer 4) Uncertain wait feels longerthan known wait 5) Unexplained wait feels longerthan explained wait 6) Unfair waits feel longerthan equitable waits 7) The more valuable the service, the longer thecustomerwill beok with waiting 8) Solo waiting feels longerthan group waiting 9) Uncomfortable waits feel longer 10)New/ infrequent users feel longer waits - These principles help with interventions and providing amore comfortable waiting experience; mitigate the negative effects - Achieved withmusic, lightning, scent, art, furnishing,colour and social elements(employee visibility, customer interaction,video games for kids) - In some circumstancesqueues have positive effects-> affectperceptionof product/ service,increase demandif they perceive a shortage, givetime for decision-making,increaselevels ofpositive anticipation - Capacity management isvery dynamicand involvesreactingto actual demand/ capacity - It is asequence of reactive decisions-> at the beginningthey consider the forecasts to understand current capacity, then they make plans for following periods, and this repeats itself - To determinesuccesswemeasure costs, revenue, workingcapital, and customer satisfaction(influence revenue) ->influencedbytheactual product/ serviceand the capacity availablein a period - Capacity management isforward-looking; it is keyto know if you plan for long or short term - Iflong-term demand is ‘good’,even‘poor’ short-termcapacity won’t make cuts in capacity-> iflong-term is ‘poor’however, therewill belarge, difficult to reverse and extra capacity Chapter 13: Inventory Management he dilemma of inventory management: in spite ofthe cost and the other disadvantages T associated with holding stocks, they do facilitate the smoothing of supply and demand. - inventories only exist because supply and demand are not exactly in harmony with each other. What is inventory Inventory:the accumulation of materials, customersor information as they flow through processes or networks. Inventories are often the result of uneven flows. If there is a difference between the timing or the rate of supply and demand at any point in a process or network then accumulations will occur. anaging these accumulations is defined as “inventory management”. Its de balance between M minimizing inventory in such a way that money is not being wasted on holding on to too much inventory,but also not too little that customers orders are not fulfilled. Customers held up in queues for too long can get irritated, angry and leave, reducing revenue. Types of inventory hysical inventory, often referred to as stock, isthe accumulation of physical materials such as P components, parts, and finished goods. Created to compensate for the differences in timing between supply and demand. Product: “In-Wait-Out” - For products, it counts with finished products and raw materials Queue of customers In services, thequeueof customers is a kind of inventory,and are an accumulation of customers, as in a queueing line or people waiting for a service at the end of phone lines. and your goal is to create a balance between those who receive the service and those who provide it. hen the rate of supply exceeds the rate of W demand, inventory increases; when the rate of demand exceeds the rate of supply, inventory decreases. So if an operation or process can match supply and demand rates, it will also succeed in reducing its inventory levels. But most organizations must cope with unequal supply and demand.
