opi IONS MANAGEMENT While all managers are involved in planning, organizing, and controlling, operations managers have the direct responsibility of “getting the job done”. They must provide the leadership that is needed to produce the product or service demanded by the customer. Operations, in this context, generally mean converting the resources (such as land, labor, money, etc.) into useful products or services that can be sold to potential customers. Some functions included in operations management a 1. Forecasting: This involves making estimates about future. The scope varies from demand forecasting to technology forecasting, from short term to long term. 2. Designing products, services and processes: It includes designing produets and services as desired by the customers so that the customers expectations are fulfilled and the firm remains competitive in terms of offering of products/services. The next step is to design processes to make those products or provide the services designed so that the production cost remains reasonable. 3. Capacity planning: This requires planning the capacity of the plant or the firm depending on the future estimates of demand so that the firm is neither left with idle capacity nor loses business due to inadequate capacity. 4, Locating production and service facilities: This means finding the best locations for production plants or service facilities so that the operating costs are ‘minimized, the resources are available, and the firm also takes care of the environment and local communities. 5. Layout planning: It entails planning the actual layout of the production or service facilities (equipment, offices etc.) within the geographical location already decided. 6. Designing jobs and standards: This includes deciding about who will do what in actual operations and making some standards as to how the work needs to be done and how much work is expected from a person. 7. Scheduling: This means making decisions about when each activity/step in ‘operations needs to be done. 8. Inventory controk: It involves deci storing or stock keeping of raw materials, finished and semi services. ing and monitoring all functions related to ished goods or Of the above functions only forecasting and inventory control will be explained in some detail. (ost ofthe material inthis section is taken from the book “Production and Operations Management” by Everett E. Adam, Jr. and Ronald J. Ebert, published by PHI)Time series analyst Yo forecasting problems, analysis of demand data plotted on a time scale to reveal patterns of demand, Demand pattern General shape ofa time series; usually constane, trend, seasonal, or some combination ofthese shapes. FOR! STING CHARACTERISTICS OF DEMAND OVER TIME ‘To systematically analyze historical data for forecasting, managers commonly use «ime series anahyis. Analysts plot demand data on atime scale, study the plots, and look: for consistent shapes or patterns. A time series of demand might have, for example, a Constant, trend, of seasonal (cyclical) pattern (Figure 3.3) or some combination of these FIGURE 3.3 Demand patterns Production demanc (unt) patterns (Figure 3.4). The pattern is the general shape of the time series. Although some individval daa points do not fall on the pattern, they all tend to cluster around it. “To describe the points clustered about a pattern, we use the term noie. Low noise ‘means all o most of the points lie very close to the pattern. High noise means many of the points lie relatively far away from the pattern. Figure 3:4 shows both high and low noise levels. If you tried to envision the data in Figure 3.4 without the solid line showing the patter, you might find it difficult to identify the pattern. Because noise in the demand pacterns can disguise the pattern, forecasting, even with computers, ean be very dfcule; the result can be forecasting etrors and, eventually, operations errors. ‘Analysts usc the term demand stability to describe the tendency ofa time series to FIGURE 3.4 Noise in demand Production demand (units)Demand stability “Tendency of atime series to retain the same fever pater over ‘Noise Dispersion of demand about a demand Independent demand Demand for an item that ‘occurs separately of, demand for any other item. Dependent demand Demand for an item that can be linked to the ‘demand foe another the same general pattern over time. The demand patterns for some products or services change aver a period of tine, and the patterns for others do not. Demands are easier to forecast when the pattern is stationary (stable) than when it is dynamic (unstable) . Figure 3.5 is taken from a study of demand for frosted microscope slides in a large medical center, an example of a dynamic demand pattern. Notice that demand shifts upward beginning at about period (week) 150. Later, these shifts become more pronounced. In the study, two forecasting models—simple exponential smoothing and. adaptive exponential smoothing—were used. These models are discussed later in this chapter; here we observe that the adaptive model responds more quickly to demand shifts than the exponential model. Actual ome Exponential model © © Adaptive model Demand (nis) uses 8 Tire (weeks) FIGURE 3.5 Frosted microscope slide demand Source: Everett E, Adams Je., Wiliam L, Berry, and D. Clay ‘Whybark, “The Hospital Administrator and Management Science,” Hospital & Health Services Administration 19, no-l (Winter 1974), 38. DEPENDENT VERSUS INDEPENDENT DEMAND [Demand for a product or service is termed independent when it occurs independently of demand for any other product or service. Conversely, when demand for one product is linked to demand for another product, the demand is termed dependent. Dependency may occur when one item demand is derived from a sccond item (vertical dependency) ‘or when one item relates in another manner to the second item (horizontal dependen- «)- Ina movie theater, for example, demand for film postage is independent of demand for popcorn. Vertical dependency might be the relationship between popcorn and theater ticket (patron) demand. Horizontal dependency might be the relationship between popcorn demand and popcorn box demand, __ Only independent demand needs forecasting; dependent demand can be derived from the independent demand to which itis linked. Our discussion in Chapter 14 on ‘material requirements planning (MRP) will furcher develop this concept.USEFUL FORECASTING MODELS FOR OPERATIONS ‘AN OVERVIEW OF SPECIFIC FORECASTING METHODS We have emphasized that forecasting is a critical part of strategic and operational planning. Rather than get too deeply into specifying types of forecasts for varying ‘we summarize: The less analytical qualitative forecasting methods are frequently used for longer-range strategic planning and facilites decisions; the more analytical, time series analysis models ae frequently used for operational planning, such asin production and inventory control. Causal forecasting techniques are used for 4 variety of planning situations but are especially helpful in intermediate-term planning. ‘Table 3.1 summarizes modern forecasting techniques. The techniques have been srouped into qualitative models, nave (time series) models, and causal models. The ‘mest frequently used techniques in operations management are the qualitative and naive (time series) models. The causal models are often moce costly to implement and do not offer the increased accuracy for short-tcrm forecasting typically needed by the production/operations manager. Even though qualitative techniques are very popular, they have definite accuracy limitations. We'll limit ourselves toa brief discussion of two qualitative methods and then proceed to some useful naive (time series) and causal models TABLE 3.1. Summary of representative forecasting techniques Model Type Description Qualitative Modes z Delphi method ‘Questions panel of expert for opinions Historical data Makes analogies tothe pas in judgmental manner Nominal group techique Group proces allowing participation with forced voting Naive (Time Series) Quancatve Models Simple average Averages past data to predict the future based on that average Exponential smoothing Weights old forecasts and most recent demand Causal Quantitative Models Regresion analysis epics functional relationship among variables Economic modeling Provides an overall forecast for variable such as ‘rose national prodvet (GNP)Delphi technique A ‘qualitative forecasting Technique in which = panel of experts working Separately and not Nominal group technique ‘A qualitative forecasting technique in which 2 panel of experts working together ina meeting, through discussion and ranking of ideas. QUALITATIVE MODELS Delphi ‘The Delphi technique is a group process intended to achieve a consensus forecast. A panel of expeits from either within or without the organization provides ‘written comments on the point in question, “The procedure works as follows: ‘A coordinator poses a question, in writing, to each expert on a pancl. Bach expert writes brief prediction. 2. The coordinator brings the written predictions together, edits them, and summarizes them. 3. On the basis ofthe summary, the coordinator rites anew set of ‘questions and gives them to the experts. These arc answered in writing. 4. Again, che coordinator edits and summarizes the answers, repeating the process until the coordinator is satisfied with the overall prediction synthesized from the experts. “The key to the Delphi technique lies in the coordinator and experts. The experts frequently have diverse backgrounds: Two physicists, a chemist, an electrical engineer, ‘and an economist might make up a panel. The coordinator must be talented enowgh to synthesize diverse and wide-ranging statements and arrive at both a structured set of questions and a forecast. "An advantage of this method is that direct interpersonal relations are avoided. Hence personalities do not conflict, nor can one strong-willed member dominate the ‘group. The Delphi method has worked successfully for technological forecasting: for Example, to forecast market penetration of solar electric energy for the year 2000.* Nominal Group Technique Like the Delphi technique, the nominal group technique involves a panel of experts. Unlike the Delphi technique, the nominal group technique affords opportunity for discussion among the expert. “The process works like this. Seven to ten experts are asked to sit around a table in full view of one another, but they are asked not to speak to onc another. A group facilitator hands out copies of the question needing a forecast. Each expert is asked to write down alist of ideas abour the question. After a few minutes, the group facilitator asks each expert in turn to share one idea from his or her list. A recorder writes each idea on a flip chart so that everyone can see it. The experts continuc to give their ideas in a round-robin manner until all the ideas have been written on the flip chart. No discussion takes place in this phase of the meeting, Usually between 15 and 25 ideas result from the round-robin. During the next phase of the meeting, the experts discuss the ideas that have been presented. The facilitator makes sure that all the ideas are discussed. Often similar ideas are combined, reducing the total number of ideas. When all discussion has ended, the experts are asked to rank the ideas, in writing, according to priority. The group concensus is the ‘mathematically derived outcome of the individual rankings. ‘The keys to che nominal group process are clearly identifying the question, allowing creativity, encouraging discussion, and ultimately, ruling for concensus. WAIVE (TIMESERIES) QUANTITATIVE MODELS Many models use historical data to calculate an average of past demand. This average is then used as a forecast. There are several ways of calculating an average; here are a few. WwSimple moving average ‘Average of demands ‘occuring in several of the most recent periods; ‘most recent periods are added and oldes ones éropped to keep —_—_—— EXAWPLE Simple Moving Average A simple moving average (MA) combines the demand data from several of the most recent periods, their average being the forecast for the next period. Once the number of past periods to be used in the calculations has been selected, itis held constant. We may use a 3:period moving average or a 20-period ‘moving average, but once we decide, we must continue to use the same number of periods. The demands for all periods are equally weighted. The average “moves” over time, in that, after each period elapses, the demand for the oldest period is discarded and the demand for the newest period is added for the next calculation, overcoming the ‘major shortcoming of the simple averaging model. A simple moving average is calculated as follows: = sum of demands for periods 3.4) MA = "chosen number of periods a 2p, +o,+ ...49, where ‘= the chosen number of periods = 1 is the oldest period in the »-period average = mis the most recent period D,= the demand in the th period Frigerware has experienced the following demand for ice coolers during the past six ‘months: Tee Coolers Month Demanded January 200 February 300 March 200 pail 400 May 500 June 600 "The plant manager has requested that you prepare a forecast using a six-period moving average to forecast July sales. It néw July 2, and we are to begin our production of ice coolers on July 6. 2. - 200 + 300 + 200 + 400 + 00 + 600 a a rae = 367 Using a six-month moving average, the July forceast is 367. Now examine the data. A three-month moving average might be a more accurate forecast. If we use three months, the forecast for July is: sua n 2= 400+ 500+ 600 js — If we use a one-month moving average, the forecasted demand for July is the actual ‘demand for June, s0 the July forecast is 600. ‘We must make some recommendation to the plant mangger for Figerwate. For now, let's recommend using a three-month moving average of 500 ice coolers for July, since that number looks more representative of the time-series pattern than does the six-month moving average, and itis based on more data than is the one-month moving averageWeighed moving sverage ‘An averaging method that allows for varying weighting of old demands —_———- EXAWPLE Weighted Moving Average Sometimes the forccaster wants to use a, moving average but does not want all periods equally weighted. A weighted moving average (WMA) allows for varying, not equal, weighting of old demands: WMA = Bach petiod’s demand times a weight, summed (3.5) cover all periods in the moving average “This model allows uneven weighting of demand. If» is three, for example, we could ‘weight the most recent period twice as heavily as the other periods by setting Cy = 25, G = 25, and C, = 50. For Frigerware, a forecast of demand for July using a three-period model with the most recent period’s demand weighted twice as heavily as cach of the previous two periods’ demand is: WMA = > G,D, = .25(400) + .25(500) + .50(600) WMA = 525 —_ ‘An advantage of this model is that it allows you to compensate for some trend oF seasonality by carefully fitting the coefficients, C, If you want to, you can weight recent ‘months most heavily and still dampen somewhat the effects of noise by placing small tings on older demands, Of course, the modeler or manager still has to choose the coefficients, and this choice is critical to model success or failure. EXPONENTIAL SMOOTHING 3 Exponential smoothing models are well known and often used in operations manage- ‘ment. The reasons for their popularity are two: They arc readily available in standard ‘computer software packages, and they require relatively little data storage’ and ‘computation, an important consideration when forecasts are needed for each of many individual items. Many computer companies have spent considerable time developing. land marketing forecasting software and educating managers in how to use it. In Addition, some major professional and trade associations, among them the American Production and Inventory Control Society (APICS), have introduced their members to these techniques. Exponential smoothing is distinguishable by the special way it weights each past demand. The pattern of weights is exponential in form. Demand for the most recent period is weighted most heavily; the weights placed on successively older periods decrease exponentially. In other words, the weights decrease in magnitude the further back in time the data are weighted; the decrease is nonlinear (exponential).First-Order Exponential Smoothing To begin, let’s examine the computational aspect of first-order exponential smoothing. The equation for ereating anew of updated forcast ses two pieces of information: actual demand forthe most recent period and ‘most recent demand forecast. As each time period expires, « new forecast is made: eal ‘demand smank forecast Forecast of next = a | for most + (1 — a)| for most period's demand recent recent period period (3.6) Fe aD + (1- af, 0S a < 1, and is the period After period ¢~ 1 ends, you know the actual demand D,., for period f~ 1. At the beginning of period ¢ — 1, you made a forecast F,, of the demand during period s ~ 1. Therefore, at the end of — 1, you have both picces of information needed for cealeulating a forecast of demand for the next period ‘Why is this model called exponential smoothing? An expansion of Equation 3.6 shows Since Fm aD, + (1- a)Fy (37) then Foy=aD.2+ (1 a): (3.8) and similarly Fey = @Dzy + (1 aE 5 69) We begin expanding by replacing F. in Equation 3.7 with its equivalent, the right side of Equation 3.8. Fm aD. + (1a) [aDpa + (1- Fs] (3:10) F,= aD, + a(1 ~ a)D.. + (I~ a): ‘We continue expanding by replacing Fz in Equation 3.10 with its equivalent, the right side of Equation 3.9: F.= aD, + a(l ~ a)D,; + (1 - a)! [a Dy + (1 ~ @)Fs) (3-11) F=aD,,+ a(t ~ a)D,; + a(1 — a) D,, + (I~ a) Fs Equation 3.11 can be rewritten as follows: F,= ali - a)*D_, + a(t — a)! Dz + all — a)? Dy + (1a)? Fy (3.12) We have expanded Equation 3.7 to obtain Equation 3.12. The expansion could be continued further, but itis not necessary for illustrating our point; Equation 3.12 shows the relative weight that is placed on each past period's demand in arriving at a new forecast. Since 0 @ $1, the terms a(1— a)°, a(1 — a}',;a(1 — a)?, and so forth are successively smaller in Equation 3.12. More specifically, these weights decrease ‘exponentially. The most recent demand, D,.., is given the most weight, while the older data are weighted less and less heavily. Suppose, for example, that we are using a = ‘Then a(1 ~ a)® = .2,a(1 — a)' = .16, a(1 ~ a)? = .128, and so forth, and these are the relative weightings being placed on D,., D.», Ds, and so forth, respectively. Remember, all of this is being accomplished automatically when you use Equation 3.7, the simple forecasting equationEXAWPLE Phocnix General Hospital has experienced irregular, and usually increasing, demand {for disposable kits throughout the hospital. The demand for a disposable plastic tubing, in pediatrics for September was 300 units and for October, 350 units. The old forecast procedure was to use last year’s average monthly demand as the forecast for each month this year. Last year's average monthly demand was 200 units. Using 200 units as the September forecast and a smoothing coefficient of .7 to weight recent demand most, hheavily, the forecast for skis month, October, would have been (r= October): F,= aD. + (1-a)k,, 7(300) + (1 ~ .7}200 10 + 60 70 ‘The forecast for November would be (¢ = November): F= aD. + (1 ak (350) + (1 — .7)270 45 + 81 = 326 Instead of last year's monthly demand for 200 units, November's forecast is 326 units. The old forecasting method, based on a simple average, provided a considerably different forecast from the exponential smoothing model.INVENTORY CONTROL Inventory: The word is used for stores of goods and stocks. Items in inventory are called stock keeping items, held at a stock (storage) point. Stock keeping items usuaily are raw materials, work-in-process (WIP), finished products and supplies. Inventory control includes all activities that maintain stock keeping items at desired levels. Since items in store require money to buy and the money remains blocked so long as the items remain in store, it is necessary to maintain the stores at some level that minimize the costs involved. ‘The operating doctrine: Operations managers must make two basic inventory policy decisions: when to reorder stock and how much stock to reorder. These decisions are referred to as the inventory control operating doctrine The time at which a reorder is placed is called the reorder point. A system signal, usually a predetermined inventory level, tells clerical or other responsible personnel when itis time to reorder stock. The amount that should be reordered is called the order quantity. INVENTORY ‘SYSTEMS inven IR Inventory System One practical way to establish an inventory sjstem is to Re ie te Seapeoue every tem ved fom inventory and place anode for more sock when foc waits ms opened inventories dwindle to a predetermined level, the reorder point. The reorder quantity, roogder. point R— toe also called the economic order quantity, is fixed in size (volume), size having been. trigger level—and an edetermined. Figures 12.5 and 12.6 illustrate owo such Q/R inventory systems. For Hearn ee eeie” the system shown in Figure 12.5, the demand for inventories, also called the usage rate, order quantity is known and constant. Replenishment inventories ae assumed to be received at the (GOC)—are fixed stock point the moment they have been ordered. Notice that at the beginning of the Units in inventory time axis (far left), an order has just arrived. As time goes by, inventory is steadily depleted until a level of R units is reached. Ris the reorder point, also called the érigger evel, an ordet for Q units is placed. ‘These units arrive at the instant they are ordered. Procurement lead time is zero. The usage pattern is then repeated, so again at level R, ‘quantity Q is ordered. In a simple ease like this, there would be no need to carry buffer instantaneous, and the demand for the inventory item is known for ‘certain, Thus R is set at zero units. In a QIR system, both the reorder quantity and the reorder point are fixed. For Figure 12.5, then R and Q are constant. " ‘The inventory situation is slightly more complex for the system shown in Figure 12.6. Usage rate is variable; we do not know in advance how rapidly inventory will be depleted. As before, R and Qare constant; however, a8 you can see, somewhat different procedures are used to determine their values. It is difficult to establish the most FIGURE 12.5 Q/R inventory system: ‘constant usage rate, Ris the reorder point and Q the order Time quantity.‘Tho inventory level that signals the lipstick and cosmetic reorder and the reorder quantity are economic decisions at the heart of the operations manager's inventory control function—whether at the factory level as shown here or in the distribution chain. Source: Superstock. economical operating doctrine when demand varies, a8 it does here, and even more difficult when lead time varies too. Since lead time is the time between placing and receiving an order, itis shown as L, and L, on the graph. When either demand or lead time varies, the time interval between orders varies—but the order quantity always remains constant. FIGURE 12.6 Q/R inventory system: variable tuaage rate. R is the reorder point, Q the order quantity, and L, the lead times. : 2 i ? ; | Tie ; EL iPeriodic inventory sgitem An operating, doctrine for which reorder points and order ‘quantities vary; stocks are replenished up 0 2 fixed base stock level “after a fixed time period has passed. Periodic Inventory System Another practical inventory control method is to count inventories at sct time intervals, periodically. With this method, the order quantity will bbe whatever is needed to bring the amount of inventory back up to some preestablished base stock level. As Figure 12.7 illustrates, the level of inventory is examined at times T, 2T, and 37 and orders are placed for quantities Q,, Q, and Q;. The base stock level and the time T between orders are set by operations management and comprise the inventory system's operating doctrine. In the periodic system, T'is constant, but 0, does not necessarily equal 0, or Q,. Although Figure 12.7 shows constant demand within any cone review period and zero lead time, these conditions could be relaxed and still allow the periodic inventory system concepts to be retained. In this book, we emphasize Q/R systems. Although we concentrate on determin- ing economic order quantities and reorder points, be aware that the procedures are similar for the periodic system. Economic order quantity in the Q/R system and base stock levels in the periodic system determine how much to order; reorder point in the ‘QIR system and time between orders in the periodic system determine when to order. i Untsin inventory FIGURE 12.7 Periodic inventory system. T is the time between orders and Q, the order quantities. INVENTORY COSTS In operating an inventory system managers should consider only those costs that vary directly with the operating doctrine in deciding when and how much to reorder; costs independent of the operating doctrine are irrelevant. Basically, there arc five types of relevant costs:Procurement costs Costs of placing an order, or setup cost if ordered items are manufaccured by the firm. Stoctout coets Coats ‘sociated with demand when stocks have been depleted; generally lose ‘sales or backorder costs. 1. Cost of the item 2. Gost of procuring the item 3. Cost of carrying the item in inventory 4, Cost associated with being out of stock when units are demanded but are unavailable (stockouts) Cost associated with data gathering and control procedures for the inventory system ‘Often these five costs are combined in one way or another, but let's discuss them separately before we consider combinations. COST OF ITEM ‘The cast, or value, of the item is usually its purchase price: the amount paid to the supplier for the item. In some instances, however, transportation, receiving, or inspection costs, for example, may be included as part of the cost ofthe item. If the cost of the item per unit is constant for all quantities ordered, the total cost of items purchased during the planning horizon is irrelevant to the operating doctrine. (See the supplement to this chapter.) If the unit cost varies with the quantity ordered, a price reduction called a quantity discount, this cost is relevant. If the facility manufactures the item, the cost of the item is its direct manufacturing cost. Again, constant unit costs mean total costs are irrelevant. PROCUREMENT COSTS Procurement costs ate the costs of placing a purchase order, or the setup costs if the item is manufactured at the facility. These costs vary directly with each purchase order placed. Procurement costs include costs of postage, telephone calls to the vendor, labor costs in purchasing and accounting, receiving costs, computer time for record keeping, and purchase order supplies. CARRYING (HOLDING) COSTS Carrying, ot holding, costs are the costs of maintaining the inventory warehouse and? protecting the inventoried items. Typical costs are insurance, security, warchouse rental, heat, ight, taxes, and losses due to pitferage, spoilage, or breakage. The cost of typing up capital in inventory is also considered a carrying cost. STOCOUT COSTS 4 Sioctout cass, associated with demand when stocks have been depleted, take the form of lost sales or backorder costs. When sales are lost because of stockouts, the firm loses, both the profit margin on unmade sales and its customers’ good will. If customers their business elsewhere, future profit margins may aso be lost. When customers to come back after inventories have been replenished, they make backorders: Backorder costs include loss of good will and money paid to reorder goods and notify ‘customers when goods arrive. As the next example shows, stockouts can and do occur it the service industries.Ina simple inventory control model, the cost of item and the stockout costs are ignored. The annual total cost will then be the sum of procurement costs (also called ordering costs) and the carrying (holding) costs. The variation of costs with order quantity is shown in Figure 12.8. ‘Ann cots a a ‘Order quantity (in nits) (2) FIGURE 12.8 Cost tradeoffs in inventory control. Q* is the optional ‘order quantity.‘The Economic Order Quantity (EOQ) formula Let us define some notations: D =Total annual demand (in units) Q= Order quantity (in units) Q* = Economic Order Quantity (EOQ) (in units) R= Reorder point (in units) L = Lead time (in time unit) S = Procurement cost (per order) C= Carrying (holding) cost (per unit per unit time) T= Total annual cost ‘The total annual cost will be: T= (Order procurement cost)(Number of orders per year) + (Cost of carrying one unit for one year)(Average number of units carried per year) T=S(DIQ)+C(Q2) This is the relation that is shown in Figure 12.8. For minimum cost the above equation can be differentiated and made equal to zero, which gives the following formula for £0Q. O=. ic This looks a very simple formula but is used often in industry and works fine in most situations, It effectively tries to equal the annual procurement costs and the annual carrying costs. The reorder point will be: R=LD