Key Ideas
1. Inventories are held for a variety of reasons, such as customer demand for end items, smoothing production, decoupling internal operations, a hedge against stockouts and price increases, and economical purchasing.
2. The requirements for effective inventory management are:
- an accounting system to keep track of on-hand and on-order merchandise
- reliable forecasting of demand
- estimates of lead times between placing an order and receiving goods, and lead-time variability
- estimates of inventory holding costs, ordering costs and shortage (backorder) costs
- a classification system
3. The A-B-C approach can be used to classify inventory items according to some measure of importance. Very often that measure is annual dollar volume, which is the product of unit cost for an item and its annual demand or usage. After determining each item?s annual dollar volume, the resulting values are arrayed from high to low. Generally there will be a small percentage of items (10% to 15%) with relatively high annual volumes. These should be classified as A items, and given disproportionately high attention by management. At the other end of the scale will be a large percentage of items (around 50% to 60%) that have relatively low annual volume. These should be classified as C items and given disproportionately low attention by management. The middle group in terms of annual volume (25% to 40%) should be classified as B items and given moderate attention by management. This approach guides management in the allocation of resources. Once the allocation has been made, a prudent manager will review the grouping and may adjust the placement of some items on the basis of other information not reflected in annual dollar volume.
4. There are numerous examples of electronic automation in retail inventory management, such as UPC scanners, J. C. Penney scanners, and airline reservation systems.
5. The question of how much to order is often answered using some form of an EOQ model. The fundamental variants of the deterministic economic order quantity model include:
- EOQ for purchased end items or semi finished goods acquired from outside vendors
- economic run length for internal production orders in batch processing and noninstantaneous replacement
- quantity discounts
- In variants a and b the order quantity is set at a level such
that total holdingcosts = total ordering costs; unit acquisition
(purchase) cost is considered indirectly, but only insofar as it is used to
calculate holding cost.
In variant c order a quantity that minimizes the sum of three costs: acquisition, holding, and ordering costs. In order to implement a quantity discount model, it is only necessary to calculate total cost at an EOQ if it falls between two price breaks, or at a price break. A price break is a quantity at which there is a change in price.
6. An EOQ model tells you how much to order, but it does not say when to place an order. With uncertainties in delivery schedules and usage rates, there is no guarantee that there will always be material in stock to service production operations. This is where reorder point (ROP) models come into the picture. There are four variants of ROP:
(1) constant demand rate and lead time
(2) variable demand, constant lead time
(3) constant demand, variable lead time
(4) variable demand and variable lead time
The approach to (1) is simple, place the order when the stock gets to a level such that what is in stock now will all be used up when the new shipment arrives. Variants (2), (3),and (4), however, add realism by allowing the decision maker to set a safety stock as a buffer against uncertainty. The size of this safety stock depends upon whether the uncertainty is in the demand rate, the lead time, or both, and how much risk of a stockout the decision maker finds acceptable. In (2), (3), and (4) the fundamental relationship is:
ROP = Expected demand during lead time + Safety stock.
The size of the safety stock is determined by the "service level," or the probability
of no stockout. That is:
Service level = 1.00 - Stockout risk.
There are two fundamental assumptions:
(1) daily demands in lead time are independent
(2) the daily demand rates are normally distributed
In "constant demand, variable lead time," the uncertainty is with the length
of the period rather than the demand rate. The lead time (number of days) is
assumed to be normally distributed.
7. Chapter 13 includes a discussion of relationships between shortages and
service levels for reorder models with fixed lead time (LT) and variable demand.
This approach introduces the order quantity (Q) as an element that influences
safety stock decisions, and shows the interrelationship between Q, service level
(SL), safety stock (SS), annual demand (D), and the standard deviation of demand
in LT (
Table 13-3 gives the unit normal loss function, E(z,), associated with both the corresponding z-
value and the lead-time service level (LTSL). For a normally distributed lead-time
demand, E(z) is a scale factor that is multiplied by 
Since 
8. When ordering is constrained to fixed times, the fixed-order-interval model applies. With the fixed-order-interval model, the safety stock is greater than with the other models, because the ordering is done at fixed time intervals regardless of the stock level.
9. The single-period model is appropriate for cases in which one order is to be placed that is intended to satisfy demand for an entire period. The period can be a day, week, or other interval. Often, the items involved are perishable (e.g., fresh seafood), or otherwise have a limited shelf life (e.g., daily newspapers). Other examples include spare parts for equipment and inventories of rental equipment.
10. Regardless of the application, the single-period model involves computation of a theoretical service level, which is the ratio of unit shortage cost to the sum of unit shortage and excess costs:
