Policy recommendations for a shipment-consolidation program

Journal of Business Logistics, 1994 by Higginson, James K, Bookbinder, James H

Management must make a number of decisions in developing a shipment-consolidation program. These decisions can be generalized as follows:

* What will be consolidated? Which customer orders will be consolidated and which shipped individually?

* When will customer orders be released? What event(s) will trigger the dispatch of a consolidated vehicle load?

* Where will the consolidation be done? Should consolidation take place at the factory, on a vehicle, at a warehouse or terminal, etc.?

* Who will consolidate? Should consolidation be performed by the manufacturer, shipper, customer, carrier, or a third party?

* How will consolidation be carried out? Which specific consolidation techniques will be used?

Excellent overviews of the shipment-consolidation problem are given by worth et al.(1) and by Newbourne.(2)

This paper deals with the "When" question: how long should customer orders be held and/or what quantity should be accumulated before a consolidated load is released? Figure 1 presents a diagram of this decision problem for a newly arrived customer order (i.e., a new request for particular merchandise). (Figure 1 omitted) We defer discussion of previous research on the "When?" question to a later section of the article.

Determining when shipments should be dispatched can be viewed as a two-stage process. First, management must select a shipment-release policy; that is, a general approach for guiding the decision. Second, the chosen shipment-release policy must be operationalized. The result of that stage is a set of shipment-release parameters.

The decision of when to dispatch a consolidated load may be based on a large variety of factors. This paper examines a special class of shipment-release policies--those based only on elapsed time and accumulated quantity. Elsewhere, we examine methods for determining shipment-release parameters for these policies.(3,4)

SHIPMENT-RELEASE POLICIES

There are three commonly used shipment-release policies. A time policy dispatches each order at a predetermined shipping date, whether or not it is consolidated. This approach sometimes is referred to as "scheduled shipping." Under a quantity policy, all orders for a particular destination are held and shipped when a minimum consolidated weight is reached. Lastly, a time and quantity policy holds all orders for a particular destination until the earliest of: (i) a predetermined shipping date, or (ii) the accumulation of a minimum weight or volume. If the latter occurs first, the orders are dispatched before the prespecified release date.

We remark that a quantity policy is probably the easiest to use of the three shipment-release policies. A quantity policy is a discrete-time policy; the "state of the system" is checked only when a new order arrives. The time and time-and-quantity policies are analogous to a continuous-time review system.

The focus of a time policy often is customer service, with the shipping date set to meet consignee requirements. Conversely, the target consolidated weight of a quantity policy usually is determined by considerations of cost. Intuitively, a time-and-quantity policy might be expected to exhibit he best characteristics of both a time policy and a quantity policy. This is not necessarily true, as will be seen later.

A time policy, as well as the time component of a time-and-quantity policy, can be approached in two ways. With a maximum holding time (or oldest order) approach, the shipping date is determined by the order that has been waiting longest. Thus, the time component dictates that a consolidated load be dispatched when the oldest order reaches a certain age. The next accumulation cycle begins upon arrival of the first order after this dispatch.

The time component of a scheduled shipping and volume approach may or may not consider the age of the oldest order. For example, the accumulation cycle may begin immediately after the dispatch of a consolidated load or when the first order of a new cycle arrives. If the quantity component is ignored, beginning a new cycle immediately upon load dispatch is similar to releasing shipments at set intervals, with the possibility of cancelling a load if the accumulated weight is too small to be economical. Powell(5) and Powell and Humblet(6) studied this case through bulk-queueing theory for passenger vehicle dispatching.

Jackson surveyed the use of these three policies by practitioners. A time policy was found to be the most frequently applied (36% of respondents), but the differences in percentage-usage between the three policies was not large.

Other authors, when investigating the effect of varying consolidation parameters, have commented on the impact of shipment-release policy. Typically, however, these studies have focused on factors other than shipment-release timing, and results are interdependent. For example, although Cooper(7,8) included both one-and four-day holding times in her simulation, the goal was comparison of direct shipping versus use of warehouses, not the testing of shipment-release policies. Jackson(9) did compare a time policy with a time-and-quantity policy. Unfortunately, he also varied the number of consolidation points and the system volume: thus his results regarding shipment-release timing are intertwined with those for other parameters.