Page 70 - Orlicky's Material Requirements Planning

CHAPTER 4 Inventory in a Manufacturing Environment 51

Conversely, demand is defined as dependent when it is directly related to or derives
from the demand for another inventory item or product. This dependency may be verti-
cal, such as when a component is needed to build a subassembly or product, or horizon-
tal, as in the case of an attachment or owner’s manual shipped with the product. This
principle was formulated originally by Joe Orlicky in 1965. In most manufacturing busi-
nesses, the bulk of the total inventory is in raw materials, component parts, and sub-
assemblies, all largely subject to dependent demand. Such demand, of course, can be cal-
culated. Dependent demand need not and should not be forecast because it can be pre-
cisely determined from the demand for the items that are its sole cause. These vertical
and horizontal dependencies can be leveraged to shorten production times dramatically.
This is further described later in this chapter.

Order-Point Characteristics
Order-point theory makes five basic assumptions:
1. Independent demand can be forecast with reasonable accuracy.
2. Such forecasts will account for all demands.
3. Safety stocks will protect against forecast errors and unexpected events.
4. Demand will be fairly uniform in the short-term future and a small fraction of
reorder quantities.
5. It is desirable to replenish inventories when they are depleted below the order-
point quantity.

Forecasts of demand for components in manufacturing are most often derived from
each item’s past usage (intrinsic forecasts), rarely from finished product or other external
demand (extrinsic forecasts). Very few people try to forecast demand separately from
each product for an item common to several products. Demand forecasting determines
only the average amount of demand expected in future time periods, not the specific tim-
ing of specific demands.
When computers and applicable software became avail able, more sophisticated
applications of the order-point approach could track actual demands and compare them
with forecasts (also updating these periodically) to indicate the probability of actual
demands exceeding the forecast. Statistical techniques then could be applied to calculate
an amount of inventory (usually called safety stock) that would ensure achieving a desired
level of “service,” meaning some minimum number of stock-outs. Again, this approach
works only if the past is an accurate indicator of the future. In today’s volatile climate,
this is a rarity.
Forecasting is inseparable from order-point techniques. All forecasting (intrinsic as
well as extrinsic) attempts to use past experience to determine the shape of the future.
Forecasting succeeds only to the extent that past performance is repeatable into the
future. In a manufacturing environment, however, future demand for a given part may
be quite unrelated to its past demand. Forecasting therefore should be the method of last