As mentioned in the beginning of this paper, it was proved that many production problems could be eliminated and all could be minimized. Effective execution depends on how well this is done. In addition, other actions can help, including these:
• concurrent design engineering, with teams of engineers, planners, marketers, production people, and cost accountants developing more manufacturable product designs
• process engineers focusing on streamlining production methods, reducing setups to permit smaller batch run quantities for smoother and speedier materials flow
• better management of capacity and lead times with input/output control. Figure 12 shows this technique. The work-in-process level
(queue) in this work center is deemed too high and is to be reduced
in four weeks by producing 300 hours per week, although only 270
hours are needed. In week 9, orders totaling 275 hours of work are
input and the center puts out 305 hours—close to plan. In week 10,
265 hours are input but output drops to only 260 hours. Underqual
ified plant people would explain why the drop occurred, claim it
wouldn't happen again, and expect 270 hours input, which under-
qualified planners would release. Qualified planners, however, knew
the basic rule, "To reduce work-in-process, input less than is out
put" and released only 230 hours in week 11. Output of 280 hours
was again less than plan, so only 255 hours were input in week 12.
This week, output was up to 295 hours, essentially on plan, but cumulative deviations of 55 hours in total input and 60 hours in total output pose real problems. However, work-in-process has been cut as planned. The obvious real problem is low output; this must be solved.
Stuffing in more work makes the situation worse. Increased work-in-process and longer, more erratic lead times will aggravate on-time delivery problems. Valid schedules of released and completed orders will avoid this. Figure 13 shows how orders can be selected to get them with a conventional MRP program. Planned order release dates from it are listed by week with order quantities and work order hours. A cumulative total for week 9 of 275 hours is close enough to plan; these were the orders released as "actual input" in figure 12. The total of 288 for week 10, however, was too high; the planner held part #68 order and released the others totaling 265 hours. This resulted in smooth, controlled release, although MRP suggested an erratic total.
A deadly phenomenon known as the vicious cycle is illustrated in figure 14. Also called the lead time syndrome, this begins when a plant or work center is late on many completed orders. The cause is seen as actual loads of work in the facility exceeding its capacity. Those who don't know how to control work-in-process and lead times inevitably conclude that planned lead times are too short and must be increased.
The planning system immediately indicates that more orders should be released earlier, which increases work center loads, makes queues longer, makes lead times longer and more erratic, and causes more order due dates to be missed. Several major U.S. industries fell victims of this—machine tools, bearings, home electronic appliances, office equipment and others.
The reverse of this is the best way to reduce lead times and improve deliveries. Cut planned lead times in the planning system; this will reduce work input, allowing output to reduce work-in-process, and lead times will be shorter. Accompanied by smaller order quantities and smoother flow, this can be continued until work centers start to starve and vital throughput is lost.
Finite Capacity Scheduling (FCS)
The steps to implement this powerful technique are listed under the earlier subheading, Infinite and Finite Loading. Work standard accuracy for setup and run times will determine the validity of schedules. Figure 8 shows that transit time between operations can be significant; this can be avoided by using cells where materials move immediately between machines or work stations. Queue times should not be included; these will be calculated by FCS programs as orders arrive in centers.
Use only demonstrated capacity data discussed under the heading The Planning Phase; theoretical figures are nice to know but rarely achieved, and increasing effective capacity is the primary goal of FCS. Before running wildly ahead, learn to walk by
• using only basic modules of software programs; don't turn on the
bells and whistles
• applying FCS first to critical, bottleneck centers where capacity
changes are limited
• limiting horizons to include firm customer orders plus a very few
released plant orders
• avoiding priority criteria that will introduce too much nervousness
Much hype and humbug has characterized the claims of proponents of FCS. Citing deficiencies in MRP programs that FCS overcomes, these gems have appeared recently:
• They run too long. Fact—not if net change and weekly time peri
ods are used.
• Replanning causes earthquakes. Fact—attack the causes—unnec
• They're just order launching. Fact—they only recommend release
dates; a person should select work input using input/output control.
• They assume adequate capacity. Fact—their purpose is to permit
determining average capacity of resources needed to support the
• They use fixed order quantities. Fact—options include period or
der quantity (POQ), lot-for-lot (L4L), least total cost (LTC), and
least unit cost (LUC) techniques.
• They can't synchronize orders. Fact—"Run with codes" can link
items to be run together.
• They use only backward scheduling. Fact—they tie the provision
of components to the parents "need" dates to show when to start
parent production or procurement.
• They react only to past-due orders. Fact—this is the task of execu
tion, not planning.
• Successive runs yield no better schedules. Fact—again, this is execution's job.
Those who published these comments obviously do not understand the different roles of planning and execution. They confuse precision of calculations with accuracy of data. They don't realize production operations cannot react to reschedules as fast as computers can generate them. Used properly, FCS systems can be very useful in execution. Qualified people recognize that future plans are unlikely to be accurate, firm orders are better bases for schedules, capacity is finite, and realistic schedules are necessary for plants to be run right.
Plant operations have always been a wild environment. Murphy's Law was written in a factory. We who worked there years ago knew that if nothing bad had happened today, tomorrow would certainly be worse than normal. Our advice to peers then was, "If you don't like the heat, get out of the kitchen." In the interim, manufacturing planning and control became a profession with a unique body of knowledge, language, laws, and principles. Understanding of execution improved, causes of upsets were attacked, and most common problems were eliminated. Today, many plants produce in hours what once took months, new product designs are produced as fast as marketers can promote them, and direct labor productivity rises steadily. The plant floor has been tamed.
For balance of this article, click on the below link:
Lean Manufacturing Articles and go to Series 10