Читать книгу Precisely Wrong: Why Conventional Planning Systems Fail - Carol Ptak - Страница 13
ОглавлениеHow Conventional Planning Works
Chapter 1 focused on the history of conventional planning and the true objective of planning. This chapter will turn its attention to conventional planning’s attributes, requirements, and assumptions and focus on fundamentally how it works to produce a specific output. Chapter 3 will then explore problems associated with the factors and outcomes.
The Conventional Planning Schema
Conventional planning systems have two primary components—one is tactical, and the other is operational. These two elements are critically linked and must be discussed in combination in order to really understand how conventional planning systems are supposed to work.
The Master Production Schedule
The tactical element is called a master production schedule (MPS). The fourteenth edition of the APICS Dictionary defines MPS this way:
The master production schedule is a line on the master schedule grid that reflects the anticipated build schedule for those items assigned to the master scheduler. The master scheduler maintains this schedule, and in turn, it becomes a set of planning numbers that drives material requirements planning. It represents what the company plans to produce expressed in specific configurations, quantities, and dates. The master production schedule is not a sales item forecast that represents a statement of demand. The master production schedule must take into account the forecast, the production plan, and other important considerations of backlog, availability of material, availability of capacity, and management policies and goals. (p. 101)
As explained in the third edition of Orlicky’s Material Requirements Planning:1
The master production schedule expresses the overall plan of production. It is stated in terms of end items, which may be either (shippable) products or highest-level assemblies from which these products are eventually built in various configurations, according to a final assembly schedule. The span of time the master production schedule covers, termed the planning horizon, is related to the cumulative procurement and manufacturing lead time for components of the products in question. The planning horizon normally equals or exceeds this cumulative lead time.
The master production schedule serves as the main input to an MRP system, in the sense that the essential purpose of this system is to translate the schedule into individual component requirements, and other inputs merely supply reference data that are required to achieve this end. In concept, the master production schedule defines the entire manufacturing program of a plant and therefore contains not only the products the plant will produce, but also orders for components that originate from sources external to the plant, as well as forecasts for items subject to independent demand. In practice, however, such orders and forecasts are normally not incorporated into the master production schedule document, but are fed directly to the MRP system as separate inputs. (p. 100)
Simply stated, the MPS is a statement of what can and will be built by the organization. One of its primary inputs comes from the organization’s sales and operations planning (S&OP) process in the form of demand forecasts. The MPS also incorporates all known sales orders from customers as demand. The output of the MPS is the time-phased demand numbers given to MRP. The MPS can update MRP as frequently as the company chooses. This will be an important point to remember in Chapter 3.
Material Requirements Planning
The operational element for conventional planning is material requirements planning. The fourteenth edition of the APICS Dictionary defines MRP as:
A set of techniques that uses bill of material data, inventory data, and the master production schedule to calculate requirements for materials. It makes recommendations to release replenishment orders for material. Further, because it is time-phased, it makes recommendations to reschedule open orders when due dates and need dates are not in phase. Time-phased MRP begins with the items listed on the MPS and determines (1) the quantity of all components and materials required to fabricate those items and (2) the date that the components and material are required. Time-phased MRP is accomplished by exploding the bill of material, adjusting for inventory quantities on hand or on order, and offsetting the net requirements by the appropriate lead times. (p. 103)
MRP is essentially a calculation hub. The master production schedule feeds time-phased demand signals to MRP, which in turn calculates the necessary synchronized list of supply orders based on current inventory records (on hand and on order) and the product structure (bill of materials). The supply orders have date and quantity requirements that define the key elements of that synchronization plan. They are turned into transfer orders to distribution sites, purchase orders to be relayed to suppliers, and manufacturing orders to be scheduled on the shop floor. These manufacturing orders are then fed to a manufacturing execution system. It should be noted that the MPS was invented as a stabilizing filter against simply dumping forecasted demand directly into MRP—something that was proved to be extremely problematic 40 years ago.
MRP evolved because of the advent of the computer, and the age of marketing in the 1950s introduced more product variety and complexity than was managed previously. As was described in Chapter 1, industry was plagued with persistent synchronization issues. Practitioners were hopeful that the mathematical precision that was now possible with computers would solve these problems. Order point (the previous method of materials management) was clearly not up to the task.
MRP was a huge leap forward because for the first time what was required could be calculated based on what was already there compared with what was needed, with the net result time-phased even through many bill of material levels. The objective of MRP was to precisely timephase requirements and replenishments to dramatically reduce inventory from the previous order point approach where some of everything was kept around all the time. This ability to calculate dependent demand through a bill of material was a significant development. It was no longer necessary to forecast dependent demand—it could be calculated based on the expected time-phased demand for the parent part. The APICS Dictionary defines dependent demand as:
Demand that is directly related to or derived from the bill of material structure for other items or end products. Such demands are therefore calculated and need not and should not be forecast. (p. 46)
MRP systems are capacity-insensitive in that they will call for the production of items for which capacity may not, in fact, exist. This might appear to be a shortcoming of material requirements planning, but on a moment’s reflection, it can be seen that this is not really accurate. A system can be designed to answer either the question of what can be produced with a given capacity (i.e., what should the master production schedule be) or the question of what capacity is required given a master schedule, but not both simultaneously. Process industry tends to ask the capacity question first and then develop the MPS, whereas discrete manufacturing companies tend to ask the latter question first. Current MRP systems are designed to answer both questions iteratively. Advanced planning and scheduling systems were designed to provide a mathematically optimized result to both questions.
MRP Outputs
An effective MRP implementation assumes that capacity has been considered in the development of the master production schedule. An MRP system obeys the master production schedule, and the validity of MRP outputs is always relative to the validity of that schedule. Another way of stating this is to say that the master production schedule can be invalid (vis-á-vis available capacity) but that the outputs of an MRP system will be computationally correct.
The primary information outputs of an MRP system are the following:
• Order-release notices, calling for the placement of planned orders
• Rescheduling notices, calling for changes in open-order due dates
• Cancellation notices, calling for cancellation or suspension of open orders
• Planned orders scheduled for release in the future
Secondary or by-product outputs come in great variety and are being generated by the MRP system at the user’s option. These outputs include:
• Exception notices reporting errors, incongruities, and out-ofbounds situations
• Inventory-level projections (inventory forecasts)
• Purchase commitment reports
• Traces to demand sources (so-called pegged requirements reports)
• Performance reports and analytics
While the output of an MRP system is always computationally correct, it is not necessarily always realistic in terms of lead time, capacity, and availability of materials, particularly when the system plans requirements for an unrealistic master production schedule. MRP simply provides information for what you would have to be able to do in order to implement the provided schedule. The MRP calculation assumes that the demand it is given is valid and takes it from there.
This means that for most large enterprises, the MPS and MRP are critically linked. Each has specific roles and attributes. In isolation those roles and attributes have isolated effects. In combination, however, the effects become quite dramatic. This will be further discussed in Chapter 3. In this book, the term “conventional planning” is specifically about an MPS-MRP approach. Figure 2-1 depicts the conventional MPS-MRP planning schema connecting to a manufacturing execution system (MES). It focuses on the MPS-to-MRP link observed in Figure 1-1 (shown in Figure 2-1 at the left side of the graphic).
MPS and MRP have already been defined, so let’s define the other critical components of this schema.
Product Structure File
Also known as a bill of material, the product structure file is the cornerstone of MRP systems. The APICS Dictionary defines this as:
FIGURE 2-1 The conventional MPS-MRP planning schema
A listing of all subassemblies, intermediates, parts and raw materials that go into parent assembly showing the quantity of each required to make an assembly. (p. 15)
The product structure file depicts what it takes to make something. It is a hierarchal view starting with an end item and descends component level by component level. At each level there is a “parent-to-component” connection. The parent part is the upper level of the connection; the components are at the lower level. Thus, components can also be parents further down a product structure. As well, the bill of material contains component-to-parent ratios, also known as quantity per parent. These ratios define the quantity of a particular component that is required for a single immediate parent. In many manufacturing environments, component quantities increase at greater and greater ratios deeper in the bill of material. Finally, the product structure must also contain a lead time for each part number in the product structure file. Figure 2-2 is a simple example of a product structure file.
FIGURE 2-2 Product structure file example
In Figure 2-2 all the critical elements of a product structure are evident: part name, ratio, lead time, and product structure level. In this example the boxes, each with a unique item label, are connected in a hierarchal order beginning at the top box labeled “FPA.” FPA simply stands for “finished product A.” which is an end item. SAA and SAB stand for, respectively, “subassembly A.” and “subassembly B” ICB stands for “intermediate component B,” while PPB and PPC stand for, respectively, “purchased part B” and “purchased part C.” Each unique part for fit, form, or function must have a unique part name.
The numbers contained in parentheses in each box (except FPA) represent the ratio of a part to its respective parent. For example, it takes two SAAs to make an FPA. As we move down the product structure, component requirements tend to increase in many environments. For example, one FPA requires two SABs. Those two SABs will then require four ICBs (two ICBs per one SAB). Those four ICBs will then require eight PAGs (two PAGs per one ICB). Thus, producing one FPA ultimately requires eight PAGs.
Lead time for each part number is expressed in days as the white numbers in black circles at the side of each part. Most MRP systems utilize fixed lead times. For manufactured items these lead times are a fixed estimation of time (typically in days) about how long it will take to build the item when all materials and components are available. This is called a manufacturing lead time. For example, the product structure depicted in Figure 2-2 says that FPA will take two days to build if two SAAs and two SABs are available at the same time. For purchased items these lead times are a fixed estimation of time (typically in days) about how long it will take to receive the item from the supplier and be available in stock. This is called purchasing lead time. The product structure depicted in Figure 2-2 says that PPB will take five days to receive from the supplier and be available in inventory. The longest string of fixed lead times in a product structure is called the cumulative lead time of the end item. This cumulative lead time will become an important factor in Chapter 3.
In reality, we intuitively know that lead times are not fixed and will in fact vary. It does not always take two days to build FPA, and it does not always take five days to receive PPB. In both cases it could take more or less time depending on a variety of potential influencing factors including capacity load, sequencing, changes in priorities, and many other unexpected events, which we have more of in today’s more variable and complex environments. Yet in order to make a plan, you must have an estimation of the time to accomplish each task at the time of planning.
Finally, the connected boxes to the right simply indicate each level of the product structure. End-item product structures begin at level zero. FPA is at level zero of the product structure. PAG is the lowest level of the product structure at level 3. Each part number in a company environment is assigned only one level. When components are shared across multiple product structures, the lowest-level value is assigned to the part (the biggest number). This is called the part’s low-level code. The use of low-level codes is vital to how an MRP system sequences its series of calculations to avoid an infinite calculation loop.
Independent Demand Forecasts
The APICS Dictionary defines independent demand as:
The demand for an item that is unrelated to the demand of other items. Demand for finished goods, parts required for destructive testing, and service parts requirements are examples of independent demand. (p. 79)
Independent demand forecasts are simply what we think the total demand for finished goods (end items), parts for testing, or service parts will be within a typical time bucket. At the end-item level, these forecasts are commonly the output of some sort of S&OP process or a statistical forecasting technique based on past usage.
Service parts and parts for destructive testing are typically at lower levels in the product structure (bill of material). As explained earlier in the chapter, service-part orders are most often left out of the MPS input and instead put directly into MRP. This explains the dotted line moving from “Independent Demand Forecast” into the “MRP SYSTEM” in Figure 2-1. Items subject to both dependent and independent demand (such as service parts) have the independent forecast quantities simply added to the calculated dependent requirements, giving us the total gross requirements for the item from both sources of demand. Note that service-part demand is either forecast or recorded upon receipt of orders (placed by a service-part organization operating its own system), but as a rule usually not both.
External Orders for Components
These are explicit requirements from points of consumption—most commonly an actual sales or customer order. According to the APICS Dictionary, “It is often referred to as actual demand to distinguish it from forecasted demand” (p. 39). This is a customer or market interface stating specifically what is actually needed as opposed to a marketing, sales, or planning function stating what we think will be needed in a future time period.
Inventory Record File
The APICS Dictionary defines inventory record as follows:
A history of the inventory transactions of a specific material. (p. 85)
The inventory record contains the on-hand, on-hold, on-allocation, and in-transit amounts for any particular item. On-hand is the amount currently available for use. On-hold is the amount that is here but unavailable for use (e.g., quality hold). On-allocation is the amount reserved for a particular use or order. In-transit is the amount that is ordered but not yet received (also frequently called “open supply”).
Manufacturing Execution System
The APICS Dictionary defines a manufacturing execution system as:
Programs and systems that participate in shop floor control, including programmed logic controllers and process control computers for direct and supervisory control of manufacturing equipment; process information systems that gather historical performance information, then generate reports; graphical displays; and alarms that inform operations personnel what is going on in the plant currently and a very short history into the past. Quality control information is also gathered and a laboratory information management system may be part of this configuration to tie process conditions to the quality data that are generated. Thereby, cause-and-effect relationships can be determined. The quality data at times affect the control parameters that are used to meet product specifications either dynamically or off line. (p. 98)
MES is responsible for executing the plan that is generated from MRP and collecting data about that execution including the inventory transactions that are a necessary input to the MRP system. In Figure 2-1 the dotted line coming out of the MES box and going into the “Inventory Record File” box depicts this transactional activity. The balance of this chapter will focus on the rules behind MRP—how it does what it does. Chapter 3 will then focus on the problems that arise as a result of the combination of MPS and MRP. The components and the behavior of manufacturing execution systems are beyond the scope of this book because MES is not an integral part of MRP.
MRP Requirements and Assumptions
The requirements to run MRP are very simple and straightforward, and the impact of each will be explained in this chapter:
• A bill of material that exists at the time of planning (product structure file)
• A source of demand for item numbers contained in the product structure file
• Inventory records for each item
Figure 2-3 shows these three required inputs to MRP in the conventional planning schema. Each input is labeled with its number on the requirements list. Note that there are three sources of demand going into MRP. The two dotted arrows from “Independent Demand Forecasts” and “External Orders for Components” represent demand for lower-level components such as service parts that are often inserted directly into MRP (not supplied by the MPS).
FIGURE 2-3 MRP requirements
When these requirements are present as inputs, the MRP system can properly calculate. However, to expect some kind of reasonable result from the system, the following assumptions are made:
• Inputs are 100% accurate and complete. This means that demand is accurate and complete, product structure (and all its aspects) is accurate and complete, and inventory records are accurate and complete.
• Every inventory item in the product structure goes into and out of stock. This distinguishes an item as stock (a completed item) versus work-in-process inventory (an incomplete item).
• There is full allocation. No order is released unless all the components are available. An order release is simply the approval for materials and/or components and activities to commence to make (or ship) an item that has been ordered.
• Components are discrete. Every distinct part named in the product structure file can be counted and measured (no “use as required”).
• Orders are independent. Every order for an item in the product structure can be started and completed on its own (assuming full allocation).
As with any plan, when the assumptions behind the plan begin to break down, the plan becomes less realistic. We will now turn our attention to how MRP generates a plan with these requirements and assumptions.
Determining Quantity Requirements
As stated previously in this chapter, there are three primary requirements to run MRP:
• A bill of material that exists at time of planning (product structure file)
• A source of demand for item numbers contained in the product structure file
• Inventory records for each item contained in the product structure file
We will use an example environment to demonstrate how MRP uses these requirements to generate a plan. To truly appreciate the extent of that plan generation, we must use an environment with at least two product structures that have shared components between them. We will take an environment that makes two end items: FPA (finished product A) and FPB (finished product B).
Figure 2-4 depicts the product structures for FPA and FPB. The numbers in parentheses are the ratios of the component to its parent item. At the far right of the figure, the product structure level is indicated. This product structure does not yet include fixed lead times. The fixed lead times in these product structures will be revealed later.
FIGURE 2-4 Product structures for FPA and FPB
In the two product structures in Figure 2-4, we can observe that there is a common subassembly—SAA. This means that all lower-level components of SAA are also common between FPA and FPB.
When MRP is given demand for FPA and FPB, it uses the product structure file including ratios to determine quantity requirements for all components at lower levels. MRP accomplishes this through a requirements explosion. The APICS Dictionary defines requirements explosion as:
The process of calculating the demand for the components of a parent item by multiplying the parent item requirements by the component usage quantity specified in the bill of material. (p. 149)
This explosion creates dependent demand for all components throughout the product structure. The APICS Dictionary defines dependent demand as:
Demand that is directly related to or derived from the bill of material structure for other items or end products. Such demands are therefore calculated and need not and should not be forecast. A given inventory item may have both dependent and independent demand at any given time. For example, a part may simultaneously be the component of an assembly and sold as a service part. (p. 46)
Figure 2-5 shows the dependent demand requirements resulting from a demand input quantity of one each at both FPA and FPB. Dependent demand for each item is shown in the dark boxes touching each item on the right side.
FIGURE 2-5 Dependent demand requirements
Since SAA and all of its subsequent components are shared components of FPA and FPB, the gross requirement is the total amount across the two product structures. The APICS Dictionary defines gross requirements as:
The total of independent and dependent demand for a component before the netting of on-hand inventory and scheduled receipts. (p. 74)
Figure 2-6 combines the two product structures at SAA to show total gross requirements for common components. We can now see that the gross requirements for SAA and its lower-level components have been added together for dependent demand both from FPA and FPB. The example will continue with this combined product structure view.
FIGURE 2-6 Combined product structures
Determining Timing Requirements
When introduced, the MRP explosion and subsequent dependent demand generation calculations created a revolution for planning functions. The ability to calculate everything together and in relatively quick fashion promised to ensure that proper quantities of materials and components available would be synchronized to demand. But synchronization is not just about calculating gross and net quantities; another prerequisite is calculating the timing of when quantities must be available.
In order to determine this timing, now it is necessary to introduce the fixed lead times of each item into the product structure. Figure 2-7 depicts the product structure with fixed lead times for each item, shown as the white number in the dark circle next to each item. Demand for each item continues to be shown in the dark boxes touching each item on the right side.
FIGURE 2-7 Fixed lead times added to the product structure
The inclusion of these lead times will determine two critical times for each component: when the component is required for its parent and correspondingly when it should be released or ordered (for purchased items). Figure 2-8 shows the required dates and order date for each item. FPA is due at time X, and FPB is due at time Y. MRP will then back-schedule all components based on that high-level demand date and the specific lead times of all items.
If FPA has a lead time of two days, it must be started two days before time X to be ready at time X. Thus, its release date is labeled X – 2. FPB also has a lead time of two days, meaning it will need to be started two days before time Y (Y – 2). Based on the required release dates of each end item, the next level of components can be scheduled for release. If components for FPA must be ready at X – 2, then they must be released a lead time ahead of that time. ICA has a lead time of one day; thus its release date is X – 3. SAA has a lead time of four, and so its release date is X – 6.
The same logic occurs for FPB. Its due date is Y. It has a lead time of two days, and so it must be released at Y – 2. This means that the next level of components must be ready by Y – 2. If SAL has a lead time of two days, it must be released at Y – 4. SAA must be released at Y – 6. For all shared components there are required and released dates based on times X and Y.
FIGURE 2-8 Required and order dates for each component item
Each path in the product structure terminates in a purchased item and an order date for each purchased item based on its required date and respective lead time. The longest calculated path is called the cumulative lead time. The APICS Dictionary defines cumulative lead time as:
The longest planned length of time to accomplish the activity in question. It is found by reviewing the lead time for each bill of material path below the item; whichever path adds up to the greatest number defines cumulative lead time. Syn: aggregate lead time, combined lead time, composite lead time, critical path lead time, stacked lead time. (p. 38)
In our example environment the cumulative lead time for both FPA and FPB is 30 days. Cumulative lead time is the primary determinant of how far into the future forecasted demand must extend. Cumulative lead time and its impact will be discussed extensively in Chapter 3.
The view in Figure 2-8 gives only a partial picture at best. It is an oversimplified graphical depiction about how the MRP calculation is performed. Figures 2-9, 2-10, and 2-11 give a much more detailed description about how the MRP calculation determines gross requirements synchronized for specific timing and quantity to parent-level demand.
Figure 2-9 shows gross requirements for items that are unique to FPA only over the relevant time horizon. In Figure 2-8 the FPA due date is labeled simply “X.” In Figure 2-9 we have given X a value of day 32. That means that FPA is due in 32 days. Three rows are displayed for each item. The row labeled “Demand” is simply the quantity of demand for the item on a particular day. For components of FPA the demand is dependent demand derived from FPA demand. The row labeled “Planned order receipt” represents the number expected to be ready to meet the demand. The planned order receipt quantity is what is actually driving the next lower level’s required quantities. The row labeled “Planned order release” is the time and quantity in which the order and materials for it will be authorized for production. It is derived from back-scheduling from the receipt date using the fixed lead time of the item. The shaded boxes simply highlight activity across the time horizon. The view in Figure 2-9 allows the reader to see the progression of exploded demand dependent requirements down each level.
FIGURE 2-9 Gross requirements for FPA unique items
FIGURE 2-10 Gross requirements for FPB unique items
FIGURE 2-11 Gross requirements for common items
Figure 2-10 shows gross requirements for items that are unique to FPB only over the relevant time horizon. The FPB due date (referred to in Figure 2-8 as Y) is now given as day 35. That means that FPB is due in 35 days. As in Figure 2-9, this view allows the reader to see the progression of dependent demand requirements move through the product structure.
Figure 2-11 shows the gross requirements for items that are common to FPA and FPB starting with SAA. With this view we begin to gain a true appreciation of just how powerful MRP can be if the assumptions behind its plan are valid. Components with multiple end-item sources of demand can be synchronized with that demand.
One observation from seeing these tables is that for MRP to really be effective, demand must be known at least a cumulative lead time in advance. This important point will be expanded on in Chapter 3.
To this point only two of the three requirements to run MRP have been considered: the product structure file and the demand input. The final aspect of determining both quantity and timing in MRP is determining the net requirements by incorporating inventory records. The APICS Dictionary defines net requirements as:
In MRP, the net requirements for a part or an assembly are derived as a result of applying gross requirements and allocations against inventory on hand, scheduled receipts, and safety stock. Net requirements, lot-sized and offset for lead time, become planned orders. (p. 110)
The process of determining net requirements by incorporating the inventory records is called netting. Netting uses projected available balance to adjust gross requirements for each item in the future. The APICS Dictionary defines projected available balance as:
An inventory balance projected into the future. It is the running sum of on hand inventory minus requirements plus scheduled receipts and planned orders. (p. 139)
Projected available balance takes into account on-hand and scheduled receipts for each day against dependent demand required on that day. Ultimately, MRP systems are looking to net each position perfectly to zero. That means that MRP is hard-coded to take a high-level demand and determine the minimum amount of material and components required to make precisely that amount with nothing left over. At face value that sounds like exactly what any efficient-minded business would want to use and promote, but we have to remember that the MRP calculation assumes that demand is known and accurate.
Additionally, in environments where order minimums and multiples differ between item numbers in the product structure, netting to zero is all but impossible. An example of this issue will be dealt with in Chapter 3. Our FPA and FPB example assumes the same order policies across all items.
