|
 |
|||||||
We want to know how to serve you better. Through the development of additional eBooks, Video Downloads, and Direct Interactive Problem Solving Methods, we hope to do just that. But to do this we need information from you! Fill out our survey or send us an email so we can get directly involved in what you need & how best to provide answers:
PRODUCTION ACTIVITY CONTROL
OverviewPAC, Production Activity Control, is concerned with converting plans into action, reporting the results achieved, and revising plans and actions as required achieving desired results. Thus, PAC converts plans into action by providing the required direction. This requires the appropriate prior master planning of orders, work force personnel, materials, and capacity requirements.
While there are better methods available via JIT and other strategies, PAC is an essential system for managing in relation to specific orders that must be properly launched, when material and labor resources, and equipment are properly timed to be delivered or allocated for the completion of those orders. There is no such thing as a “late or past due” order. There is only correct planning and poor planning. The date of an order that cannot be completed simply has the wrong completion date. At this level, errors are due more to improper long-range and mid-range planning. The availability of resources in the short-range is only possible via proper management of the tasks prior to starting the job.
Order release, dispatching, and progress reporting are the three primary functions of PAC. Dispatching is the activation of orders per original plans. Dispatching decisions are affected by queue management, I/O control, and priority control principles and techniques that are intertwined and mutually supportive. They are useful in the management of lead-time, queue length, work center idle time, and scheduled order completion. Reports on the status of orders, materials, queues, tooling, and work center utilization are essential for control. Many report types with various information are possible. Examining a given situation will reveal which reports and information are required.
The time arrives when plans must be executed, when material requirements planning and capacity requirements planning have been completed and the detail purchasing and production schedules must be determined and released for execution. The function of production activity control (PAC)--often called shop floor control (SFC)---is to have activities performed as planned, to report on operating results, and to revise plans as required to achieve desired results. Figure 1 shows the sequence of the various planning and control activities.
The PAC system also closes the control loop, as illustrated in Figure 6, by measuring actual output and comparing it to the plan Thus, PAC is an essential component of closed-loop MRP. Although all PAC Systems perform the same basic functions, individual systems differ because each manufacturing environment is unique. Each has a specific number of products, production processes, facility layouts, and relationships of available capacity of personnel and equipment to the required capacity.
SCHEDULING IN MANUFACTURING ENVIRONMENTS
The Master Scheduling Section described different types of production environments: continuous and repetitive flow lines, batch flow lines, manufacturing cells, job shop, and project (fixed site) processes. Each environment is distinctive. Scheduling continuous and repetitive flow lines is discussed in the Section on JIT: Just-In-Time & TQC: Total Quality Control.
Scheduling for Batch Flow Lines
Batch flow lines exist in beverage companies, ice cream manufacturers, soap powder packaging facilities, and pharmaceutical plants. Typically, a group of similar items is manufactured on the batch line. A family of items may be produced in batch quantities on the same line with some changes in the setup, a cleaning of the equipment, and changes in incoming materials. (If no time is required for switching from one item in the family to another, then the different items can be mixed in the same run and a mixed model line exists.) Thus, a primary production management objective is to reduce and eventually eliminate the time required for changing between items in a group. The smaller the changeover time, the greater the scheduling flexibility and the smaller will be the scheduling problem.
The quantity of an item produced depends on that item's production rate and the length of time it is run. Deciding the item to be run next and the quantity to be run depends on the following factors:
A. The on-hand (available) quantity of each item
B. The demand rate of each item
C. The times required to change between different items
D. The production rate of each item
F. The sequence, if any, in which items should be run
When the setup (changeover) times are relatively small and independent of the sequence in which the items arc produced, the decision is relatively simple: the item with the smallest runout time is run first.
Runout time is the period existing inventory will last given forecast usage. For example, if a company uses (or sells) 20 printed circuits (Part No. 101) each day and has 80 of them in stock, the runout time of Part. No.101 is four days. Runout time (R) is calculated as follows:
R = Units in Inventory
Demand (Usage) RateIf the setup times for the items in a group are relatively short and the production lot quantities are small due to relatively low demand rates and low setup costs, there is no problem. Sufficient time usually will exist to manufacture all items on schedule.
Let's look at another example. Assume we have a table of run out times for three machined parts made on the same machine, a traditional machine tool with a larger setup time and corresponding larger production quantities.
The company has a problem. Items A, B, and C should be run immediately. Some of these items should have been manufactured last week. The purpose of this example is to point out that:
A. Manufacturing engineering should reduce the setup times and, thus, improve the production run quantities and time. A computer numerically controlled, (CNC), machine that can shift from one part to another with little or no setup time may be appropriate.
B. Proper timing of order releases is as important as the quantity decision,
In addition, the appropriateness of a model depends on the situation. Order quantity and order release decisions are more complicated when more than one group is run on the same equipment, when capacity is limited, or when items in a group must be run in a particular sequence to achieve minimum changeover times (for example first Item A, then B, C, and so on).
When sufficient inventory is available, personnel may be used for preventive maintenance, methods analysis, and setup time reduction to reduce lead time and improve quality rather than to produce unneeded parts. (These topics are discussed further in Section on JIT: Just-In-Time & TQC: Total Quality Control as part of Just-in-Time concepts.)
Job Shop Scheduling
The physical layout of a job shop usually groups equipment performing similar functions in the same area. Typically, there are many different orders being processed in the plant at the same time and relatively few have the same routing (the department-by-department path through the plant). Scheduling is the assigning of starting and completion times to orders (jobs) and frequently includes the times when orders are to arrive and leave each department. Sequencing is the assigning of the sequence in which orders are to be processed, for example, do Order C first, then B, followed by D, and so on. However, in practice and in the literature, scheduling frequently refers to both the time schedule and the sequence of orders or jobs. The selection of a scheduling system, approach, or technique depends on the objectives of the schedule and the criteria by which its results will be measured.
Management policies and objectives are the basis for scheduling decisions, However, production management may define multiple and conflicting scheduling objectives in different ways, such as: minimize average lateness of orders, minimize maximum lateness, minimize manufacturing lead time (minimum average flow time), minimize work in process, and maximize utilization of bottleneck work centers. Fortunately, many of the objectives are mutually supportive. For example, reducing manufacturing lead time reduces work in process and increases the probability of meeting due dates.
Achievement of these scheduling objectives depends on the flexibility of the manufacturing equipment and personnel The importance of achieving flexibility through methods improvement, facility layout, setup reduction, worker cross training, and the development of manufacturing cells cannot be overemphasized.
Priority Control
Many methods (sometimes called priority rules) exist for establishing the priority of orders. The priority, often expressed numerically, is used to determine the sequence in which the orders should be processed. The rules described in the following pages are probably the most common, but many variations and combinations of these methods exist. The list of Common Decision Priority Rules below, provides a good overview of the basic rules and their objectives.
To show how these rules are used to determine the priority of orders, let's consider a specific example. Say we have a table that shows data concerning four orders in a manufacturing plant in central Illinois. All orders were in the same department, which we call Department 7. The simplest priority rule to implement is earliest due date. For this example, the manufacturing sequence would be A, B, C, D. These jobs have due dates of 130, 132, 136, and 133, respectively. We now discuss the computation of slack and critical ratio rules.
Slack time (ST) is found by subtracting the present date (say, Day 125) and the total operation time remaining from the due date. That is,
ST = Due Date Present Date - Total Operation Time Remaining
For Order A,
ST = 130 - 125 - 3.0 = 2.0
The critical ratio (CR) equals the difference between the due date and the present date divided by the manufacturing lead time remaining:
CR = Due Date - Present Date
Manufacturing Lead Time RemainingCommodity Priority Decision Rules:
FCFS--First Come, First Served: Run the orders in the sequence in which they arrive at the work center. This ''fairness'' rule is especially appropriate in service organizations where most customers often either need or desire the completion of the service as soon as possible SPT, SOT--Shortest Processing (Operation) Time:
Run the orders in the inverse order of the time required to process them (smallest time first) in the department. This rule usually results in the lowest work in process, the lowest average job completion (manufacturing lead time), and average job lateness. Unless this rule is combined with a due date or slack time rule, jobs (orders) with long processing times can be extremely late.
STPT-- Shortest Total Processing Time Remaining:
Run the orders in the inverse order of the total processing time remaining. The rationale of this rule is similar to the preceding one. It accomplishes similar objectives when most jobs follow a common process.
EDD--Earliest Due Date:
Run orders with the earliest due date first. This rule works well when processing times are approximately the same.
FO -- Fewest Operations:
Run first the orders with the fewest operations remaining. The logic of this rule is that fewer operations involve less queue time and, as a result, the rule reduces average work in process, manufacturing lead time, and average lateness. However, jobs with a relatively large number of operations can take excessively long if another rule is not combined with this one.
ST -- Slack Time:
Run first the order with the smallest slack time and continue the sequence in the ascending order of their slack times. Slack time equals the due date minus the remaining processing time (setup plus run time). This rule supports the achievement of due date objectives. The slack time remaining per operation is a variation of this rule.
CR --Critical Ratio: For orders not already late (overdue), run first those orders with the lowest critical ratio. The critical ratio equals the due date minus the present date divided by the normal manufacturing lead time remaining.
For Order D,
CR =
9.0=1.44
A CR of 1.0 indicates that the order is right on schedule; a CR greater than 1.0 indicates that the order is ahead of schedule; and a CR smaller than 1.0 indicates that the order is behind schedule. The smaller the CR, the higher the priority of the order. The CR index---as most priority criteria---should be used in conjunction with one or more other criteria. For example, Order X has 2 days left to delivery and 1 day of manufacturing lead time remaining; thus, its CR is 2.0(2 + 1). Suppose Order Y has a 1.11 CR: it has 10 days left until its due date and 9 days of manufacturing lead time remaining. On the basis of CR's, Order Y has the higher priority. Both have the same slack time, one day. However, the nearer due date of Order X argues strongly for giving it first priority
In addition, the CR is not a good priority index for orders whose due date has passed. Priority indices for orders whose due dates have passed are described shortly.
Applying each of the priority rules from our list above, (except FCFS) to the four orders in our earlier example, gives the following processing sequences:
EDD (Earliest Due Date): A, B, C, D
SOT (Shortest Operation Time): B, A, C, D
STPT (Shortest Total Processing Time Remaining): A, C, B, D
P0 (Fewest Operations): D, A, C, B
ST (Stack Time Remaining): A, B, D, C
CR (Critical Ratio): B, A, C, D
Although applying priority rules to any four orders at a given time in a specific department will produce different results, the above results are not unusual. Different rules produce different sequences, but certain patterns tend to appear in most. For example, Orders A and B are scheduled first or second by most rules. One factor that also should be considered is the status of the work center to which each order goes next. There would be little point in scheduling Order A first if its next operation was in a work center overloaded with higher priority orders
An advantage of the SOT rule is that the data required to use it is readily available to the immediate supervisor, as should be the due date data. Operation and order due dates are very popular for establishing order priorities because of their simplicity, ease of understanding, and direct relationship to a primary objective of management-on time delivery. The other rules require calculations and considerably more data. Thus, they usually require a computerized shop floor control system that performs all calculations and prepares daily lists showing job priority.
Planning (determining) the priorities of orders is a prerequisite to effective production activity control. Priorities must reflect actual needs and be consistent among items going into the same assembly. Changing order priorities frequently will destroy their credibility.
Overdue Orders and Priority Indices: Overdue orders are of special interest because management is interested in minimizing the cost of late orders. Special priority indices are often used to manage overdue orders because, among other reasons, the CR technique gives confusing information when applied to overdue orders. The data in Table 1, illustrate the failure of CR in an overdue situation. Orders B and C both have a CR of 0.0 indicating identical priorities: but B is 10 days behind schedule and C is 8 days behind schedule. Clearly their priorities should not be the same. Order D has a CR of -2.5 which would indicate that it is in a poorer condition than Order E which has a CR of -1.25. This is not the case; Order F is further behind schedule than Order D.
The concept of slack time, the time ahead of or behind schedule, can be used to aid in determining priority for overdue orders. Slack time may be computed by different methods; manufacturing lead-time and processing time remaining are the two most widely used. Managers may wish to minimize the number of late orders and decide to have one job very late. In this case, jobs that can be delivered on time are not delayed to process a job that already is late.
The manufacturing lead-time remaining (MLTR) method of computing slack time computes the number of days ahead of or behind schedule by subtracting the manufacturing lead time from the actual lead time remaining. The priority is then computed based on the number of days behind or ahead of schedule. For example, Order E in Table 2 has highest priority because it is the farthest behind schedule on the basis of this method.
The processing time remaining (PTR) method of computing slack time computes the number of days behind or ahead of schedule by subtracting the processing time remaining from the actual time remaining. These computations are shown in Table 3.
The days behind schedule when computed using manufacturing lead-time remaining as in Table 2, indicate that Order B is further behind schedule than Order C and thus has a higher priority. However, a ranking based on processing time remaining rather than total manufacturing lead-time gives Order C a higher priority, as illustrated in Table 3. When queue and move time are a Large but variable portion of manufacturing lead-time, and the queue and move time can be compressed by priority sequencing, ranking is improved using days overdue plus processing time remaining rather than total lead-time remaining.
Thus, for overdue orders, two priority rules are:
1. Run those orders first that have the greatest total of days behind schedule plus manufacturing lead time remaining.
2. Run those orders first that have the greatest total of days behind schedule plus processing time remaining.
Orders for safety stock and made-to-stock items should have lower priority than items being manufactured to fill a customer order with the same due date. This is in keeping with the philosophy that the customer always comes first. In addition, safety stock and finished goods stock are manufactured to meet probable but uncertain demands, while an actual order is a certainty.
Performance Measures: Criteria for evaluating a priority control system can include the following:
1. Percentage of on time orders
a. to customers
b. to the assembly line
2. Average tardiness
3. Work in process
4. Idle time
5. Minimizing setup time
6. Energy conservation
One or two of the foregoing may be dominant over a short period. The planner must be able to recognize shifting criteria, or even different criteria in different parts of the plant, and to organize dispatch lists accordingly. A dispatch list is a document that lists the jobs in a work center and indicates the priority of each. Dispatching is discussed in detail later in this section.
QUEUE LENGTH MANAGEMENT
Queues consist of those items waiting to be processed at a work center, They usually are measured in hours of work required in the work center, that is, the length or size of the queue. The lengths of queues directly affect the value of work-in-process inventory and manufacturing lead times. In an ideal situation there arc no queues and also no idle time: an item arrives exactly at the time scheduled for its processing and the work center has just become available to perform the operation. However, ideal conditions rarely exist in job shops and queues are planned to compensate for the uneven flow of incoming work and the variations in work center processing times. The section on JIT: Just-In-Time & TQC: Total Quality Control describes how Just-in-Time concepts can reduce queues substantially. This section describes the management of queues prior to achieving the benefits of JIT.
The objective of queue length management is to control lead time and work in process and to obtain full utilization of bottleneck work centers. Material queues of only an hour's work or so may be planned in a flow line process to avoid downtime. In a job shop environment, determining the nature of queues at the critical work centers should be the first step. Meaningful queue length goals then can be established. First, we will examine queue length distributions. Then, we will investigate operation overlapping and operation splitting, two methods of managing queues and lead times.
Typical Queue Distributions
Briefly, there are four different queue situations: (1) a controlled queue, (2) an excessive queue length, (3) an uncontrolled queue, and (4) substantial idle time due to a short queue.
Consider an example that illustrates a situation where the average queue length is 30 hours, the maximum length is 55 hours, and the work center is never idle because of lack of work and is seldom overloaded. On the other hand, the data could exemplify a queue whose length is never less than 45 hours. It is obvious that the length of this queue can be reduced by 45 hours without affecting idle time. This reduction can be accomplished by releasing work to the work center (controlling the input) at a reduced rate until the queue reduces.
Queue length also can be measured statistically with planned average lengths based on the probability of a stockout, a zero length queue. This approach calculates the planned average queue length by multiplying the standard deviation of the queue length observations by the number of standard deviations required to obtain the desired coverage. It assumes a queue length distribution on the basis of historical data and counts the item being machined as part of the queue. (Zero queue length corresponds to machine downtime.)
Suppose a queue has a normal distribution with a 70-hour average length and a standard deviation of 9.7 hours. If management's objective is to have a material shortage less than 1 percent of the time (a 99 percent service level), the planned average queue length should be approximately 22.6 (2.33 x 9.7) standard hours. (The approximate number of standard deviations corresponding to 49 percent of the high side area under the normal curve is 2.33.)
The first approach to the queue in our second example, indicates that the average queue length can be reduced by up to 45 standard hours, and the statistical approach suggests that an average queue length of 22.6 standard hours (a reduction of 47.4 hours in the average queue length) will meet idle time objectives. Neither approach is exact and both should be applied with caution. Queue length distributions seldom are perfectly normal and shortening the length of a queue will, in itself, affect the distribution. In most cases, however, both approaches clearly indicate when a queue can be shortened. In most cases the change should be made gradually to minimize shop adjustment problems. A sudden decrease in queue length can cause supervisors and operators to drag out available jobs. It should be made clear to shop personnel that the backlog still exists, but it has been moved from the plant to production planning and control.
Reductions in queue length at a gateway work center, the first work center at which work is performed, are achieved through input/output control at that center. Selection of appropriate orders for processing in earlier work centers will result in the desired adjustments in downstream work centers used later in the process.
The conditions represented by substantial idle time, are typical of a work center with excess available capacity. Jobs at overloaded work centers should be moved to work centers with excess capacity when possible (when the underloaded work center is a feasible route).
Our third example illustrates a stickier situation, an uncontrolled queue. It is more likely to be found in work centers in which two or more preceding operations have been completed in other work centers. In this situation, the arrival of jobs is often very erratic. A detailed analysis revealing the major source and processing patterns of incoming loads should provide clues for possible remedies. Analysis of order sequencing alternatives also may reveal options available for reducing the unusually long queues in this type of situation. Finite scheduling techniques that have simulation capability often can be used to anticipate and avoid such situations.
Operation Overlapping (Transfer Batches)
Operation overlapping, is a technique used to reduce the total lead time of a production order by dividing the lot into two or more batches and linking at least two successive operations directly (one is performed immediately after the other). Operation overlapping is a common practice in manufacturing cells when setup is required.
Operation overlapping consists of the following:
1. A lot of parts is divided into at least two batches (transfer batches).
2. As soon as the first batch completes Operation A, it is moved to Operation B for immediate processing.
3. While Operation A is being performed on the second batch, Operation B is being performed on the first batch.
4. When Operation A has been completed on the second batch, it is moved immediately to Operation B.
If Operation B requires substantially shorter time per piece than Operation A, the first batch should be sufficiently large to avoid idle time at Operation B. Calculation of this minimum batch size is straightforward:
Q = Q1 + Q2
Q1 PB + TAB + SB > or = Q2 PA + TAB (assuming Q2 is to be at Operation B before Operation B is completed on Q1)
where Q = total lot size
Q1 = minimum size of first batch
Q2 = maximum size of second batch
SB = setup time of Operation B
PA = processing time per unit, Operation A
PB = processing time per unit, Operation B
TAB = transit time between Operations A and B
Solving the above equations for Q1, gives:
Q1 is equal to, or greater than
=
PB + PAFor example, if
Q = 100 units
Pa = 10 minutes
Pb = 5 minutes
Sb = 40 minutes
Tab = 30 minutes
then
Q1 is equal to, or greater than
=
10 + 5
15The result is checked easily. The time required to process 64 units in Operation B is 320 (64 x 5) minutes of run time plus 40 minutes for setup, a total of 360 minutes. This is exactly the time required to process the second batch of 36 units at Work Center A. Move time is the same for both. If fewer than 64 units were in the first batch, Work Center B would be idle awaiting arrival of Batch 2.
If Operation B can be set up prior to the arrival of parts, consideration of setup time drops out of the equation defining the minimum size of the first batch. For example,
Q1 is equal to, or greater than
=
10 + 5Reduction of total manufacturing lead time by the reduction of the throughput time for Operations A and B is the benefit of operation over-Lapping, as illustrated in Figure 5. The disadvantages are the added cost of increased planning and control required by doubling the number of batches and material movements, plus the requirements that the first batch be moved immediately upon completion and that capacity be available at Work Center B when the first batch arrives. Time lost by not meeting these latter two requirements decreases the savings in lead time.
The manufacturing lead time (MLT) without overlapping and no queue equals the total time for Operation A (setup and run) plus transit time plus the total time for Operation B (setup and run). Thus,
MLT = 80 + 100 x 10 + 30 + 40 + 100 x 5 = 1,650 minutes
The MLT with overlapping and prior setup of Operation B equals the time for Operation A on Batch 1 (setup and run) plus transit time from Operation A to Operation B plus the total time for Operation 13 (run only) on Batches 1 and 2, Batch 2 completes Operation A and is moved to Operation 13 while Batch I is being processed in B. Thus,
MLT = 80 + 67 x 10 + 30 + 100 x 5 = 1,280 minutes
The difference between the two conditions in lead times is 370 minutes (1,650- 1,280), approximately a 22 percent reduction. The actual savings depend on whether parts are required to set up the machine as well as the normal time an order would wait between processes. Usually the major savings from overlapping come from the elimination of queue time---frequently several times greater than total processing time-between operations.
When the processing time of Operation B is greater than that of Operation A, similar calculations can be performed to determine the batch sizes required to maximize lead time savings under the constraint of only one additional movement (dividing the lot into no more than two batches). The section on JIT, examines operation overlapping further, including multiple transfer lots and the reduction of transit time. Since operation, setup, and transit times are rarely constants, simulation of these activities, is highly advisable.
Operation Splitting
Operation splitting, reduces total lead time by reducing the run time component. A production lot is divided into two or more batches and the same operation is then performed simultaneously on each of these sub-lots. Operation splitting reduces the processing (run time) component of manufacturing lead time at the cost of an additional setup. Conditions conducive to lot splitting include a relatively high ratio of total run time to setup time, idle duplicate equipment or work force personnel, and the feasibility of an operator running more than one machine. These conditions frequently exist. For example, in the cutting of large diameter ring gears, the setup time is small in comparison with the run time of a lot of 20 or more.
Lots also may be split in a "setup offset" manner. After the first machine is set up and running, the operator sets up the second machine. For this approach to be feasible, the time required to unload one part and load the following part must be shorter than the run time per part. In addition, shop practices (and the labor contract) must allow an individual to run more than one machine. This approach reduces lead time and increases labor productivity. The appropriate mix of parts to equalize runout or to meet cycle assembly requirements is committed as a group. Both overlapping and lot splitting are normal procedures in manufacturing cells.
INPUT/OUTPUT CONTROL
Input/output (I/O) planning and control is an integrated process that includes (1) planning the acceptable input and output performance ranges per time period in each work center, (2) measuring and reporting actual inputs and outputs (feedback), and (3) correcting out-of-contro1 situations.
Input/output control is an effective technique for controlling queues, work in process, and manufacturing lead time (the time from the release of an order to its completion). This section analyses actual inputs, outputs, and work in process. Input/output control enables the planner to determine what action is necessary to achieve the desired output, work in process, and manufacturing lead time objectives. We will examine the case of a single processing center and then the more complicated case of multiple work centers and many orders with different routings.
Single Work Center Processes
Some manufacturing processes have only one work center; others have a dominant (bottleneck) work center that is the focal point for controlling input and output to the entire process. In addition, gateway work centers, continuous and repetitive batch lines, and a uniform routing through a group of work centers frequently may be treated as a single processing work center for input/output analysis purposes.
Input/output is a short-range control technique; it usually is performed using daily rather than weekly time buckets. Input/output analysis compares the scheduled order (or task) inputs to the process and the scheduled outputs to the actual inputs and the actual outputs. This information comes from production schedules and reports of actual order releases, arrivals of orders in a work center, and completions of orders in a work center. The basic concept of I/O planning and control is that ending work in process equals beginning work in process plus input minus output, as illustrated by Figure 7. Further computations can provide the cumulative input deviation, the cumulative output deviation, and the planned and actual work in process (WIP).
Lead Time
Output RateManagement must develop various measures of process performances, including an acceptable level of input and output deviation-and the acceptable level of WIP. The following examples illustrate three different situations:
(1) a process in control, (2) the use of input/output to control and reduce work in process and lead time, and (3) input/output controls under out-of-control conditions.
In the first example, illustrated in Table 4, the situation is under control. Actual input and actual output differ little from the plan; the work in process is never more than 5 hours different from the plan. Typically, management will establish an acceptable cumulative deviation, perhaps 20 hours of work in process in this case for example, as acceptable due to random events. Planned WIP is usually three to four times the standard deviation of the ending WIP, resulting in virtually no time due to lack of work. Acceptable deviation is about twice the standard deviation; beyond this, action is attempted to correct the deviation. In Table 4, the planned WIP appears to be too high.
The second situation is illustrated in Table 5. Because planned WIP is excessive, a reduction in input beginning in Day 26 and constant output are planned to reduce work in process from 32 standard hours to 20 standard hours and to reduce lead time from two days (16-hour days, i.e., each day has two shifts and utilization and efficiency equal 100 percent) to two and a half shifts. Lead time equals the work in process (the hours of work in the queue plus those being processed) divided by the production rate. At the beginning of Day 26, the lead-time equals 2.0 days (32 hours + 16 hours per day). After five days, actual results approximate the plan; the WIP is 21 hours and the lead-time is 1.31 days (21 + 16). This reduction is reasonable only if 20 hours of work in process will sustain the production through normal variations in incoming work and output. Once the desired level of work in process has been reached, the input must be returned to the output level.
Typical out-of-control situations, possible causes, and corrective actions include the following:
1. Queues exceed upper limits. Possible causes include equipment failure, inefficient processing, and excessive input. Decreasing input or increasing process output is necessary to correct the situation.
2. Output is below the lower limit. Possible causes include equipment failure, inefficient processing, inadequate input, or the wrong input at assembly work centers.
Equipment failure and inefficient processing are manufacturing engineering problems. Inadequate, excessive, or the wrong inputs are I/O problems that should be rectified by dispatching. I/O control is essential at critical (bottleneck) work centers whether they are gateway, intermediate, or the final work centers.
Table 6, illustrates an application of input/output control in an unanticipated situation. An equipment problem that began during Day 30 has decreased output and the work in process did not decrease as planned. The plan was to work 2 hours overtime in Days 31 and 32 to increase output by 25 percent to 20 hours, to bold input constant at 16 hours, and to reduce the work in process to 24 hours. However, the equipment performed erratically during Days 31 and 32 and output fell short as shown in Table 6. Solving the equipment problem is the first step in rectifying this situation. In the meantime, planned input and output should be reduced. Maintaining the present input level will only maintain the high work in process and hinder production. Planned output should be based on the actual capacity of approximately 16 standard hours per day. Excess work in process exists; so planned input for Day 33 is reduced to 12 hours to achieve planned work in process. Even if the work center performance returns to the normal 20 hours of output, sufficient work in process will be available with the planned input in Table 6.
Thus, as shown in Table 6, 12 hours of input are planned for Day 33 along with 16 hours of output, while working 2 hours overtime. If the equipment operates properly and produces 20 standard hours per day, sufficient work in process exists to prevent machine downtime due to lack of work.
Table 7, is an example of a situation in which input is insufficient to produce the planned output. This can result in late deliveries, poor customer service, poor profits in the short run, and the possibility of losing future orders. Measures should be taken to increase the actual input in Week 30; otherwise the work center will experience idle time. A work center that feeds this work center is probably causing this problem. The cause of the reduced input must be identified and corrected.
The principles of input/output control are:
1. The planned output should be realistic and should represent labor and equipment capacity.
2. A planned or actual input greater than the realistic output wilt increase WIP, hinder production, and increase manufacturing lead time.
3. All significant deviations from planned input and planned output indicate operational problems that must be identified and solved.
Multiple Work Centers
Work flow through multiple work centers is often represented schematically. Two formats are commonly used, the flow-by-order format and the rate-of-flow format. Figure 8, is a schematic representation of four possible order flow patterns in a job shop with ten work centers. Work Centers Al and M are gateway work centers. The first operation is performed in one of these two work centers. Work Centers 131, B2, B3, Cl, C2, and C3 are intermediate work centers, and Dl and D2 are the finishing or final work centers. All work centers in which processing is performed following processing in a given work center are called downstream work centers. Those work centers in which processing is performed prior to a given work center are upstream work centers. We wilt examine I/O control at each type of work center.
Gateway Work Center Control: Management of the release of orders controls the input, queues, and WIP at gateway work centers. If the work center is running smoothly, output also is controlled. The input to the gateway work centers also influences inputs to downstream work centers. There is little reason to have a long queue at a gateway work center. Keeping gateway queues at a minimum enables the dispatcher to use the latest available information when establishing order priorities. It also reduces WIP and expediting.
Downstream Work Center Control: The input and queues at downstream work centers are controlled by dispatching (order sequencing) at upstream work centers in the process flow. For example, if Work Center C3 in Figure 8, is running short of work, while there is a relatively large queue at Work Center C2, priority in Work Center 112 should be given to orders going to C3 next. This requires that order release decisions recognize the needs of downstream work centers as welt as gateway work centers. Of course, other factors such as due dates and manufacturing inventory not required must also be considered.
Final Work Center Control: The output of final work centers influences shipments, due date commitments, billings, accounts receivable, and cash flow, Final output usually is one of the dominant measures of production management performance. Controlling final work center input is necessary to achieve the desired output. This involves coordinating the flow of parts, items, and sub-assemblies required in final assemblies. Dispatching is concerned with achieving control of the volume and specific items entering the final work centers. In some complex job shops, large-scale computer simulations are used to provide completion oriented priority control that extends backward from the final work center to gateway operations (Lankford 1978).
Bottleneck Work Centers: When the capacity required exceeds the capacity available, a bottleneck exists. Often this condition either is short lived or can be solved by using the flexibility of the work force and the equipment to increase capacity. Eliminating bottlenecks with flexible capacity is one of the primary objectives of the JIT approach and is necessary to compete in world markets. Chronic bottlenecks can occur even with the best planning and, therefore, should receive the special attention of planners. A bottleneck work center limits output, and an hour lost at such a work center is an hour of output lost. Thus, the scheduling of work in bottleneck work centers is critical to achieving production objectives. As a result, measures should be taken to provide flexible and sufficient capacity to eliminate bottlenecks when designing and developing production facilities. The objective of the theory of constraints is to manage bottlenecks.
Load Order Manufacturing Control: This is an input/output control method developed at the University of Hannover and implemented successfully at more than 20 manufacturing companies in Europe. It uses statistical analysis of the time-phased relationships of order releases, manufacturing process work center requirements, and loads at downstream work centers to develop order release priority rules and guidelines for specific environments, It has had noteworthy success in reducing queues, work in process, and manufacturing lead time in an orderly, practical, and systematic manner. (Bechte 1988; Wiendahi 1987).
TRADITIONAL PAC INFORMATION SYSTEMS
Production activity control (PAC) procedures include order release, dispatching, and production reporting (see Figure 6). Queue length management, input/ output control, and priority control are interwoven and mutually supportive. Their principles and techniques are applied jointly in making order release and dispatching decisions. (JIT and theory of constraints concepts and approaches are discussed in later sections).
Flow Line Processes
In both repetitive (discrete units) production and continuous process type production, the PAC system has requirements slightly different from the job shop. The salient differences are that (1) daily run schedules are used to authorize and control production rather than job orders and (2) control is executed by counts at key points in the flow
In flow line manufacturing environments (continuous, repetitive, or manufacturing cell production), the consumption of ingredients, such as chemicals, powders, or component parts and subassemblies, may be recorded automatically when the production of the finished product is recorded. Component parts, materials, and subassemblies used in reaching a given stage in the production process are deducted from the inventory on hand by exploding the bill of material and multiplying the quantities of each required by the number of assemblies produced. This is called backflushing. For example, if 5 pounds of ammonia-nitrate are used in each 25 gallon container of a specific fertilizer, the number of such containers produced is multiplied by 5 pounds to determine the amount of ammonia-nitrate to be subtracted from the inventory of that ingredient. This occurs either at key completion stages in the process or at the final point in the process. A single-level backflush deducts only the items used in the last assembly or mixing process and is usually used when backflushing takes place more than once in a process. A superflush accounts for all items down to the lowest level in the bill of material and is appropriate when the process is relatively brief and backflushing takes place only after the final process is completed.
Backflushing reduces the amount of data capturing and processing but requires system integrity, accurate reporting of completed items, accurate measures of yield, and special reporting of unusual situations such as a batch that must be discarded (scrapped). It also results in inventory records for materials and components showing larger quantities of inventory on hand than actually is the case, for at least a short time.
Job Shops
A PAC system in a job shop must be capable of the following:
1. Releasing orders to the production department on schedule (per the order release plan), having verified materials, information (blueprints and manufacturing processes), tooling, personnel, and equipment availability
2. Informing the production department of the scheduled start and completion dates of steps (individual operations) in the production process as well as the scheduled completion date of the order
3. Informing the production department of the relative priorities of the orders released
4. Recording actual performance of steps in the production process and comparing actual performance to the schedule
5. Revising order priorities on the basis of performance and changing conditions
6. Monitoring and controlling input and output, lead times, work center queues, and work in process
7. Reporting work center efficiency, personnel attendance, operator times, and order quantity counts for planning, payroll, department efficiency, and labor distribution reports
Order Release
Order release initiates the execution phase of production; it authorizes production and/or purchasing. The planned order becomes a released (open) order. Placement of a purchase order or the initiation of manufacturing follows shortly. Order release planning may take place until the moment of order release. Authorization of order release is based fits! on the planned orders in the MRP output, the current priority, the availability of materials and tooling, and the loads specified by I/O planning. Release of an order triggers the release of the following:
1. Requisitions for material and components required by the order. If some of these items are not required immediately and have not been allocated previously, they are allocated now.
2. Production order documentation to the plant. This documentation may include a set of both engineering drawings and manufacturing specifications and a manufacturing routing sheet.
3. Requisitions for tools required in the first week or so of production. Tooling, including tapes for numerically controlled machines, required in later operations is reserved for the appropriate period. Tooling can be included in the master production schedule and the bill of material Its availability is thus coordinated with material and equipment availability.
The time required to deliver production order documentation, tooling, and materials to the first operation is included in the normal planned lead time for the order. An order is released by adding it to the dispatch list.
Dispatching
Dispatching informs first-line supervision of the released orders and their priority, that is, the sequence in which orders should be run. This information can be transmitted via a hard copy (handwritten, typed, or computer printout) or via video output on a cathode ray tube (CRT). Telephone and face-to-face conversations also can be used but do not document the decisions. In a job shop a dispatch list should be prepared for each work center with the frequency of updating depending on the typical order-processing time. If orders take a day or less to process, dispatch lists usually are prepared daily. If orders take a few days, lists may be prepared weekly with midweek revisions handled on an exception basis with on-line processing. In a flow line process environment, a single list indicating the rate of flow (or in a batch flow line, the sequence in which orders are to be started) will control work on the entire line, which may be viewed as a single work center. An example of simple dispatch list information identifies the date, the plant, and the work center; it includes the work center capacity; and it lists the orders, their quantity, their capacity requirements, and their priority. Orders usually are listed in descending priority for a specified period.
The list also may include jobs at upstream work centers to provide the supervisor with information concerning orders that will arrive shortly and an indication of their priority upon arrival. A computerized system may produce relative rankings on the basis of criteria such as critical ratio and earliest due date, as described earlier, but review by a planner is required to determine if other considerations are overriding.
The planner determines the final dispatch list ranking of orders on the basis of multiple criteria including a formal priority index such as the critical ratio or the due date, input control at downstream work centers, the availability of tooling, the status of other parts required in the same assembly, energy consumption patterns, and sequencing and assignment criteria. For example, if the next operations for Orders s-4276 and s-4518 are at work centers heavily loaded with high priority orders while the next operation for Order S-4625 is at an idle work center, Order S-4625 may be processed first in spite of its CR or due date on this operation. Such situations should not occur, but they do occasionally, even in well-managed organizations. In addition, environments in which the energy consumption costs of production are relatively high can foster scheduling rules incorporating constraints on energy consumption peaks (Baker 1979).
Dispatch List Revisions: The due dates and priorities of orders may change because of such developments as forecast revisions, cancellation of orders, and the scrapping of another lot of the same item at a later stage in the production process. For example, suppose the following events occur after the dispatch list, shown in Table 8, is released on August 1.
1. The customer has canceled his order, S-4276, for Part 9706.
2. The completion date for Order S-4609, Part M3563, has been moved back one week due to a delay in receiving other parts required in their common next assembly.
3. The due date of Order S-4625, Part H4276, has been advanced two weeks to fill requirements that were to be met by another order that was scrapped at a later operation.
The dispatcher must exercise judgment in informing shop supervision of revised priorities. If Order S-4276 is in process, there may be no point in revising its priority in Work Center M3. The priority can be changed in its next work center. Revising the priorities and listings of Orders S-4609 and S-4625 seems appropriate. However, continual revision of order priorities will destroy the credibility of dispatch lists.
The dispatch list also may include orders that are due to arrive in the department shortly, as illustrated in Table 8. This enables supervisors to include these orders in their planning.
Few dispatching decisions can be made in a programmed automatic fashion. A computer can provide valuable assistance by keeping an accurate record of order status. It also can provide an inquiry capability, responding to the requests of managers and planners concerning the status of any order. However, the dispatcher must exercise judgment in balancing operating costs and customer service when determining the final priority of orders. Often local rules, or heuristics, are developed to simplify and structure order release policies.
Organization: Dispatching may be organized in a centralized or decentralized manner. Centralized dispatching exists when decisions are made in a single location and communicated to supervisors throughout the plant. Centralization facilitates monitoring the progress of orders, coordinating the priority of orders required in the same assembly, and auditing the counts of lot quantities. Its advantage is that it can improve communications among dispatchers.
Decentralized dispatching exists when order sequencing decisions are made in the department. It has the advantage of decision making at the scene. The dispatcher may have a better grasp of the department's capabilities and efficient order sequencing. Wherever they are located, dispatchers must be aware of actual conditions in the work center and overall plant objectives and developments.
The development of computers, automatic counters, and electronic data collection devices has supported the adoption of centralized dispatching approaches. Management's desire to give more responsibility to first line supervision has supported the adoption of decentralized dispatching. Such considerations often lead to the adoption of hybrid systems. Overall order status is kept in a central location that issues sequencing recommendations, and supervisors possess the authority to alter sequences within certain limits to achieve production efficiencies.
Production Reporting
Reports describing actual production status are necessary for control. Dynamic response to changing conditions is possible only if timely, accurate, and adequate information is available. The information must enable management to take meaningful corrective action concerning production schedules.
The production environment influences the design of the production reporting system. Reporting in a line flow environment with long production runs, may take place on an exception basis with feedback occurring only when the output rate falls below an acceptable level. In a custom design and manufacturing environment, that has project management and fixed site manufacturing, emphasis is on reporting the status of activities on the critical path. All reporting systems should have an exception reporting capability to inform management whenever machine failure, material shortages, or similar events threaten planned output.
Parts fabrication in a job shop environment requires more data collection for control than continuous processes or repetitive manufacturing of discrete parts. Once a flow process is initiated, it will continue smoothly unless machine failure, employee absenteeism, scrap, a materials shortage, or production inefficiencies occur. Exception reporting usually works welt in these circumstances. Flow in a job shop is more complex, and order status estimates are less certain. Thus, the processing and movement of orders does not automatically follow their release into the production stream as do orders in a flow process. Control in a job shop usually requires information concerning the following:
1. The release of orders
2. The beginning and completion of operations
3. The movement of orders
4. The availability of processing information, tooling, and material
5. The queues in each work center
Exception reporting is frequently adequate for controlling the availability of information required for processing, tooling, and material. Reporting both the beginning and completion of operations is appropriate when the total operation times are relatively long. For example, if the estimated completion time of processing a lot of parts through a particular operation is four days, reporting initiation of the operation makes sense. On the other hand, if an operation requires only an hour and a half, reporting its completion should be sufficient.
Data Collection: On-line reporting systems directly report events as they occur, usually via a data terminal or other device capable of electronically transmitting the data to a centralized recording station. Such information is called real time since the records are updated instantaneously. Whether an organization requires real time information as provided by on-line processing or whether periodic reporting (by shift, day, or week) is sufficient for the desired control depends on the situation.
In some cases the operator reports the initiation or completion of an operation, order movement, etc., via a data terminal or by completing an operation reporting form included in the job packet. In other cases the supervisor or timekeeper is responsible for reporting this information.
Typical Reports: The status of WIP, inventory availability, and work center queues and utilization influences dispatching and order release decisions. When an on-line, real time reporting system with inquiry capability exists, management, dispatchers, and planners can obtain current status information virtually instantaneously. The response to their inquiry may be presented on a video output device and/or produced on a hard copy output. When an on-line, real time reporting system exists, daily status reports are required in most cases. In all cases, periodic summary reports are required to evaluate production performance.
The following information should be available to planners on either a real time or periodic basis.
1. Released order status (see Table 9). This report gives the status of every order that has been released physically to the plant and includes part number, description, quantity, order release date, order due date, operations completed, order location, quantity scrapped, arid quantity good.
2. Unreleased order status (see Table 10). This report lists all orders whose release is past due. It also notes the cause of the delayed release, such as long queues of higher priority orders at gateway work centers, lack of required tooling, or lack of required material or parts.
3. Dispatch list-priority scheduling report (see Table 8). This report lists in priority sequence all orders in each department plus those expected to arrive shortly perhaps in the next day. Standard hours required for processing also are listed.
4. Weekly I/O by department (see Table 4, Table 5, Table 6, and Table 7).
5. Exception reports. These should be designed to meet the needs of the organization. Possible exception reports, illustrated in Table 11, include a scrap report, a rework report, and a late orders report. A review of scrap reports will reveal if quality problems are recurrent with a particular item, operation, or operator. Scrap reports also can trigger the release of new orders or a quantity increase on unreleased orders for the same item. Rework reports also can alert management of quality problems and unplanned capacity requirements. The purpose of a late orders report is to inform management of orders that require expediting and possibly of customers who should be informed of late delivery. If the late orders list is extensive, the possibility of a capacity problem or an unrealistic MPS should be investigated. The late orders report should focus on a number of orders that can be expedited efficiently and that have high priority.
6. Performance summary report. The performance summary report should state the number and percentage of orders completed on schedule during a specific period, week or month-and the lateness of late orders. A late orders aging report, similar to an accounts receivable aging report, will reveal the magnitude of any delivery problems. Performance also should be reported in terms of volume (tons, units, feet, etc.) or dollars. The causes of late orders also should be tabulated.
The types of reports possible are many and varied. This section has included only some of them; the readings contain other examples. Too many reports diminish the value of each. Different situations require different information and different organization of that information.
PAC INFORMATION SYSTEM REQUIREMENTS
Certain data and files arc required for a PAC system, In a manufacturing firm these usually are organized in the following files:
1. Planning files:
a. Part (item) master file
b. Routing file
c. Work center file
2. Control files:
a. Production order master file
b. Production order detail file
Planning Files
The part master file is required for many activities, including material requirements planning (MRP), inventory management, cost estimating, and PAC. This file has a record for each part. Each record is identified by a part number and contains relevant data such as inventory status and standard cost. In addition, the record for each item includes the following data required for PAC:
1. Part number---the unique item number assigned to the part
2. Part description---the name of the item
3. Manufacturing lead time---the normal time required to produce the item in the typical lot quantity. This information may also be in the routing file.
4. On-hand quantity---the number of units of this part in stock
5. Allocated quantity---the number of units of this item that has been assigned to previously planned future orders
6. Available quantity---the difference between the on-hand quantity and the allocated quantity
7. On-order quantity---the total number of units due on all outstanding orders for this part
8. Lot-size quantity---the normal number of units of this item produced at one time (the order quantity)
9. Substitute items---the part numbers of items (or materials) that may be used in place of this item
The routing file and the work center file are used for 'capacity requirements planning (CRP).
Control Files
The production order master file contains a record of each active production order. The purpose of the file is to store summary data describing the nature, status, and priority of each order. It contains the following data required for PAC:
1. Production order number---the number assigned to uniquely identify each order or batch
2. Order quantity---the number of units (e.g., pounds, gallons) to be produced on this order
3. Quantity completed---the number of units (or volume) reported through the last operation and final inspection
4. Quantity scrapped---the total number of units (or volume) scrapped at any point in the production of this order. Separate records of the quantity scrapped during setup and the quantity scrapped when running the item at each work center may be kept.
5. Material disbursed---the quantity of each item of materials or component parts released from stores for the production of this order
6. Due date (original)---the initial date on which this order was scheduled for completion
7. Due date (revised)---if rescheduled, the new date on which this order is scheduled for completion
8. Priority---the value used to rank this order relative to all other orders
9. Balance due---the order (or batch) quantity minus the sum of the quantities completed and scrapped. If some units are scrapped, the material requirements system will determine if another order is necessary to meet requirements.
In a job shop environment, there is a production order detail file for each order. The file contains a record for each operation required by the production process for that order. The record for each operation typically contains the following data:
1. Operation number---the number uniquely identifying the operation
2. Description---a brief description of the operation
3. Setup time reported---the number of hours reported for setting up the equipment for this operation on the given order
4. Run time reported---the number of hours reported for performing this operation on the given order
5. Quantity reported complete---the accounted number of units meeting quality requirements on completion of this operation
6. Quantity reported scrapped-the number of units that were reported scrapped on inspection during or immediately following this operation
7. Due date (revised)-if rescheduled, the new date on which this order is scheduled for completion
Conclusions:
PAC is concerned with converting plans into action, reporting the results achieved, and revising plans and actions as required to achieve desired results. Thus, PAC converts plans into action by providing the required direction. This requires the appropriate prior master planning of orders, work force personnel, materials, and capacity requirements.
Order release, dispatching, and progress reporting are the three primary functions of PAC. Dispatching is the activation of orders per original plans. Dispatching decisions are affected by queue management, I/O control, and priority control principles and techniques which are intertwined and mutually supportive. They are useful in the management of lead time, queue length, work center idle time, and scheduled order completion. Reports on the status of orders, materials, queues, tooling, and work center utilization are essential for control. Many report types with various information are possible. Examining a given situation will real which reports and information are required.
Back...
Keeping Promises...
Overcoming Resistance...
Turning Communication Into a System...
Identifying the Heart of What Really Matters...ERP-MRP Evolution…
ERP & Hoshin Kanri…
ERP Implementations…
Profit-Ability Improvement... (¬Click here to see definitions)
Profit-Ability Management Principles... (¬Click here to see definitions)
People, Empowerment & Profit-Ability… (¬Click here to access articles)
Hoshin Kanri & Deming's Plan-Do-Check-Act... (PDCA) Cycle…
