El papel de la Fuerza Pública en la creación de los grupos paramilitares

Given a realistic master production schedule (MPS), the MRP system provides the answer to the questions of how much and when for all the items required supporting that

schedule. The way inventory was planned before the advent of MRP was to wait until a reorder point was reached, order the parts, and hope that the parts got there before running out. Unfortunately, the reorder point system works best for demand that is relatively continuous and stable. This type of demand is found in companies that produce parts that are high in volume and low in variety and have relatively long product life cycles.

When this order point strategy is exercised with very short lead-times and relatively low inventories this is known as kanban, or automatic replenishment from suppliers, internal or external. The supplier can quickly respond to the customer’s needs because the customer is demanding the same thing at relatively the same rate over a period of time. This type of replenishment is also known as demand pull or demand flow. Planning is still necessary in this environment to allow the supplier visibility of overall product volumes and schedules. The other place that an order point system or kanban system works relatively well is for independent demand items, those items with demand coming from the external customer. Independent demand tends to be relatively continuous and steady over a period of time. The overall demand for the end item is relatively level with some variability. When this occurs, inventory falls in a steady predictable manner, as expected. Assuming that the reorder point method for inventory replenishment is used, an order would be placed as the inventory drops to the predefined level. This predefined level is the quantity of inventory that should cover the expected demand during the lead- time. These reorder points are marked with arrows in Figure 8.2.

This type of inventory pattern is considered a traditional sawtooth pattern. Inventory declines in an orderly fashion until the order point is reached lead-time away from the need. The replenishment arrives just as the inventory reaches zero—at least that is how the theory says that it

Figure 8.3 Order point inventory with variability.

should happen. The average inventory using this theory is the order quantity divided by 2. Typically the order quantity is fixed by part number and reflects an economic quantity to produce or purchase. The reality of the situation is that the demand pattern tends to look more like Figure 8.3.

In Figure 8.3 the demand has some variability and the inventory runs out either too early for the shipment to arrive and the company experiences back orders or the demand falls after the order has been placed and there is excess inventory. These problems are exacerbated by long replenishment lead-times. Shorter lead-times allow direct replenishment to occur while minimizing the chances of running out or having excess inventory. This is the basis for supply chain systems where the supplier is directly tied to the inventory and consumption of the customer. Direct replenishment works well where the reactive lead-times are short and the product mix stays stable. As soon as product variety comes into the picture, the need for detailed MRP returns once again. Suppliers must have some knowledge of requirements by quantity, configuration, and estimated timing sufficiently in advance so that they can start their processes in time to deliver the requirements. The exact timing of the replenishment can be through direct linkages to the supplier, but the planning information has to be provided sufficiently in advance of the requirement.

Imagine if you were a supplier who has been selling blue parts to your customer on a direct replenishment contract. Your customer sends you a fax, EDI, or XML-type kanban to signal the exact timing of your shipments. Today you get a kanban and the customer now informs you that it would like to have a striped part. Your processes are incapable of producing a striped part! You have never produced a striped part! Your customer knew about its requirement for a striped part as soon as its customer ordered it, but since you are delivering in a direct replenishment pull system you did not get the visibility until today. Now what do you do? This is a formula for disaster and is the reason why MRP still has an integral role in supply chain direct replenishment systems. Thinking that kanban and direct replenishment is a replacement for good planning is a fallacy. Good planning makes for good execution. This fact is still the same no matter what technology develops.

arises from the need to build the end item. This type of demand, by definition, is lumpy and discontinuous. This phenomenon was also seen as part of the inventory strategy discussed in Chapter 4. The seeming variability in demand surprises many people until they look at it more closely. Once again lead-time and batch sizing make a significant impact here. Demand for component parts does not exist until an order is released for the end item. At that time, there is a requirement for the parent’s lot size quantity times the quantity for each item from the bill of material (BOM). If the product being built is a lamp, there is no demand for yoke subassemblies until an order is released to build additional finished lamps. If the order is for 500 lamps, there is an instant demand for 500 yokes all at once, assembly lead-time away from the finished lamp requirement. Even though the independent demand for lamps can be very level, say 25 per day, there is no demand for yokes until the order is released to manufacture completed lamps. The finished goods inventory has lamps with built-in yokes. No additional yokes are required until a manufacturing order begins. When the manufacturing orders for lamps are released, the result is simultaneous demand for all the components (BOM quantity times the order quantity) to be available when the order is due to start. Figure 8.4 examines the effect on the demand for yokes.

Figure 8.5, Figure 8.6, and Figure 8.7 show how these spikes in demand can be determined with sufficient advance notice so that little or no safety stock is required. In Figure 8.5, the customer forecast is rather level with a small amount of variability. Based on the lot size and the lead-time the

Figure 8.4 Dependent demand pattern.

Item ID: Lamp 1 2 3 4 5 6 7 8

Customer forecast 200 300 400 400 300 200 500 350 Scheduled receipts 500

Projected lamps on hand 100 300 400 500 200 500 500 150 Planned order receipts 500 500 500 500 Planned order releases 500 500 500 500

Figure 8.7 Independent vs. dependent demand and inventory.

orders to assembly are expected to be released in periods 2, 3, 5, and 6. There is no demand for assembly in periods 1, 4, 7, or 8.

In Figure 8.6, the need for lamps to be assembled starting in periods 2, 3, 5, and 6 is translated directly into gross requirements for the yokes in the same periods. Based on the inventory strategy for yokes of only ordering what is needed when it is needed, this translates to 2 orders of 500 parts—1 in period 2 and 1 in period 3. In Figure 8.7, the net result of demand and inventory is shown in one chart.

This pattern of demand bears little to no resemblance to the sawtooth pattern shown in Figure 8.2. However, by using MRP the orders can be proactively planned, providing

Figure 8.5 Independent to dependent demand planning costs.

Note: Lot size=500, Low leve 1=0, (OH =: 300, LT=1, A lloc=0, SS== 100

Figure 8.6 Dependent demand explosion.

Item ID: Yoke 1 2 3 4 5 6 7 8

Yoke gross requirements 500 500 500 500 Scheduled receipts

Projected yokes on hand 1000 500 0 0 0 0 0 0 Planned order receipts 500 500

Planned order releases 500 500

sufficient lead-time to the supplier of yokes of what to expect in quantity and timing. The other alternative given the assembly lot size strategy in this example would be to use the order point method. The demand arrives all at once, quickly depleting the available supply. This unexpected demand often causes expensive expediting as the supplying department or purchasing attempts to replenish the inventory. No visibility of declining inventory exists like in the material planning system. Unlike the predictable sawtooth shape of the independent demand graph, the graph of the demand for component items looks like unpredictable discontinuous spikes. This is because all parts are required at the beginning of the process to start the order. Even though sales for the end item may be smooth and relatively continuous, demand for component inventory is affected by the parent’s manufacturing lot size as shown in the series of Figure 8.5 to Figure 8.7. The larger the lot size, the lumpier the demand is for component inventory. Figure 8.8 demonstrates the general resulting inventory for yokes given this spiky discontinuous demand.

Using the reorder point to manage this type of inventory causes high levels of stockouts. The typical reaction is to react with safety stock,

Figure 8.9 Pull system inventory pattern.

resulting in a higher level of inventory overall. The inventory level drops all at once, triggering an order that is already too late. This is affectionately referred to as the OSWO (Oh Shoot We’re Out) theory of inventory management. On the other hand, simple logic says when the parts are not required they should not be sitting in inventory. This use of company cash and other critical resources is wasted on inactive inventory. The analysis suggested in Chapter 4 could uncover this type of incorrect planning tool usage.

One way to minimize the lumpiness of demand is to reduce the lot size and lead-time of production or procurement. This is the secret behind kanban or pull system replenishment systems. The lot sizes are so small and the replenishments occur so frequently that the lumpiness of dependent demand is not noticed. The reality looks like Figure 8.9.

The inventory quickly rises and falls given the steady demand. Both the kanban quantity and response time are very small. This allows rapid replenishment and very low overall inventory levels. However, for traditional production, the lot size selected must allow efficient utilization of all the company’s resources. Even though the inventory would be at a minimum if parts were ordered daily and built into finished goods daily, few companies have sufficient uniform demand volume to support this strategy of repetitive manufacturing. Another way is to order only what is needed, when it is needed. This is the basis and motivation for the development of MRP systems. Many people would have you believe that MRP is an old tool that is no longer necessary; there can be nothing further from the truth. MRP is an essential tool in the toolbox of a complete ERP system.

MRP plans all the items that have to be completed or purchased to support the MPS. The forward visibility allows the release of orders on time. This allows the planner to take on a proactive role in managing inventory. No longer is the inventory replenishment process reactive, waiting until a shortage occurs. Instead, requirements are projected, thus

replenishments can be planned and managed as shown in Figure 8.5 and 8.7. The on-hand inventory for yolks can fall to zero with the confidence that any new requirement will be identified with sufficient lead-time to be able to secure the part. Rather than buffering demand with expensive inventory, using MRP can be thought of as buffering demand with adequate planning. The planning is considerably less expensive than the inventory. As computer systems have evolved and become more powerful, different types of scheduling algorithms have become available. The increased downward pressure on product life cycles necessitate more planning than ever to ensure that the right parts are in the right place at the right time.

Backward Scheduling

The standard default scheduling process is backward scheduling. The master schedule date is taken as the end point and then all components are offset backwards through time by their respective lead-times to determine the required completion and start dates for each. This type of logic is used in everyday life. When determining when to get up in the morning, the thought process may be to work backwards from when you want to be at work. Assuming that you are scheduled to start at 7:30 a.m. with a 30-minute drive, you would leave your home at 7:00 a.m. Assume for a moment that traffic in your city is reliable on a daily basis and you like to eat a hearty breakfast and read the paper. That whole process takes another 30 minutes, and you know you should be ready by 6:30 a.m. If it takes 1 hour to get up and get ready for work, you set your alarm for 5:30 a.m. This kind of planning is done in many different daily activities. When you make an appointment to meet someone for lunch at 12:00 p.m., you know that you cannot leave work at noon and be at the meeting point on time.

We do not have to physically go on the trip to be able to plan leadtimes to schedule when we need to begin. Activities that we do most frequently, like the trip to work, will provide the most accurate planning, while things that we do infrequently require a larger margin of error to be planned. The same process exists in planning for the manufacturing company. Suppliers who deliver reliably over a period of time can be scheduled to a smaller margin of error than those with whom we do business infrequently can. Backward scheduling is best used when the demand is known in advance of the expected lead-time. That is, if demand is known 6 weeks in advance and the assembly lead-time is only 2 weeks, backward scheduling is the best option. This scheduling method provides the least amount of inventory by starting orders at the last possible moment.

Forward Scheduling

At times, scheduling forward rather than backward better meets the needs of the company. Just as if you get up late, say at 7 a.m., in your mind you quickly schedule the things you must do to get to work. Given the same times as used in the backward scheduling example, the projected arrival time at work is 9 a.m. Unfortunately, you have a big meeting with the boss at 8 a.m. and walking in with your coat still on is not a very good idea and could be a CLM (career limiting move). Some activities must be deleted or changed to accomplish the expedited schedule. You may forego reading the newspaper

and eat a smaller breakfast on the drive to work. Using this same technique, companies making high variety and low volume parts will typically plan when is the soonest that a part can be completed using forward scheduling. This is because the customer usually wants the part as soon as possible rather than at a defined future date. If the resulting date is not acceptable, the sequence and duration of events, both value added and queue, is reviewed and adjusted until either there is no additional time that can be taken out or the desired date has been reached.

In the example used of getting to work, sometimes things are done that seem to expedite the process, but in the long run defeat the purpose. Getting to work sooner by speeding may seem like a good idea at the time until the local friendly law enforcement officer intercedes or you lose control of the car by attempting to multitask by doing other things while driving (eating, applying makeup, shaving, etc). The time it takes to get a ticket offsets dramatically any perceived timesavings accomplished through speeding. Attempting to speed selected items through the plant by expediting can cause similar negative repercussions. The attempt may make sense at the time and sometimes you actually get away with it. When done as a routine practice, in the long run, you will always get caught and the effort is not worth the price.