Derechos y Obligaciones de los consumidores

According to Ling et al. (2005), literature on inventory pooling can be divided into three categories: component commonality, inventory transshipments in supply chains and inventory pooling in multi-echelon supply chains. End products that use same common components open up the possibility to reduce safety stocks and maintain the service level by pooling the same inventory of the common components. In addition to product and distribution structures, ordering policies affect inventory pooling operations. These policies are a sum of the limits of the operational environment and business guidelines.

2.5.1. Partial and complete pooling

Sherbrooke (1968) introduced the METRIC model in which individual locations are supplied by with repaired items from a central warehouse. Baker et al. (1986) study a two-product system with component commonality and found out that total safety stock dropped down after pooling and while total stock of specialized parts increased. Sherbrooke (1968) designed the model lying on two echelons, as the others considered it using virtual pooling. Transshipments within an echelon create a virtual centralization of the inventory and utilize the benefits of inventory pooling of the type. According to Paterson et al. (2009) two different policies in pooling are

 Partial pooling – part of the stock of the installation is held for future demand

 Complete pooling - installation shares all of its stock

Partial pooling systems reserve items for future needs so that transshipments to satisfy all the demand are not automatically sent. Having this possibility of reserved items, the system then has additional decisions needed to make which makes it more complicated to control and optimize. Complete pooling is applicable in environments where holding costs and backordering costs are relatively high compared to transshipment costs, such as spare parts distribution systems. (Paterson et al. 2009) In literature, also term “no pooling” is used to describe the situation in which all the installations are replenishing directly from external suppliers without a shared inventory.

2.5.2. Order timing and policies

Regular orders can be timed either on continuous review or periodically. Continuous review refers to an inventory concept in which the inventory position is monitored continuously. It is based on re-order points (ROP) that are the stock levels that having fallen below will launch a batch quantity, targeting at an economic order quantity (EOQ) that is the cost formula minimizing order quantity, to be ordered. Periodic review refers to certain given points when inventory position is monitored. The need for stock replenishments is evaluated at these given points. Both concepts include a trigger for cases in which inventory position drops below a specific amount of stock. Periodic review is typically considered for items with low demand, while continuous review is used for items with high demand. According to Paterson et al. (2009), these two concepts are the most typical timing methods related to ordering. Further, policies in literature include (R, Q), (s, S) and (S-1, S). Also general types and other policies are examined.

(R, Q) and (s, S) are the most common policies in inventory control. R is an integer and stands for ROP, Q is an integer for batch quantity. In (R, Q) policy a batch quantity of Q is ordered as the inventory position drops to, or below, the re-order point R. In continuous review replenishments are ordered exactly when the inventory position

meets R or if the triggering order directly drops the inventory position below R. On periodic point of view, the inventory position may drop below R before the inspection is done at the end of the period. Because lead time, the time used from a purchase to actual delivery, (L) is not zero in practice and the inventory position will be below R as the Q is received, with continuous demand the inventory position is rarely R+Q. (Axsäter 2006) (R, Q) policy with periodic review, with period T, and continuous demand is presented on Figure 13.

Figure 13 (R, Q) policy with periodic review and continuous demand (Axsäter 2006)

(R, Q) policy is similar to KANBAN policy which is used in lean supply chains. In KANBAN there are containers with a card on their bottom and as the items in a container are used the card, which is a KANBAN, is used and a quantity of items in a container are ordered. There are always a specific amount of containers available to meet the demand. The difference between (R, Q) and KANBAN is the fact that if there are more outstanding orders than KANBANs, no more orders can be triggered, i.e. backorders are not subtracted from the inventory position. (Axsäter 2006) If this is not taken into account, on the ordering point of view they are similar even though they may differ in actual operations.

(s, S) order policy is similar to (R, Q) and if re-order point is always hit, s=R and S=R+Q. s is an integer for ROP, S is an integer for the maximum level of inventory position. In (s, S) policy meeting the re-order point will always trigger an order that at the specific moment would meet the maximum level of inventory. Multiples of a given batch size are no more ordered as the order quantity may vary. However, in (R, Q) and (s, S) if items are not consumed to a quantity below ROP, no order is triggered for that period. Because of this, some variations will always order to keep the certain inventory

position unless the consumption is zero. (Axsäter 2006) (s, S) ordering policy with periodic review and continuous demand is presented on Figure 14.

Figure 14 (s, S) policy with periodic review and continuous demand

When inspection will always trigger orders unless the consumption is zero, it is typically an (S-1, S) policy. (S-1, S) policy is equivalent to an (s, S) policy with s=S-1 and to an (R, Q) policy with R=S-1 and Q=1. With expensive and slow-moving items, low ordering costs compared to holding and backordering or lost sales costs, such as in spare parts environments, the (S-1, S) policy is very appropriate. (Olsson 2007) For single echelon systems it has been proved that the (s, S) policy is the most optimal. However, no major cost differences exist between the policies which makes (R, Q) policy popular due to its practicality and easiness of use. (Axsäter 2006)