An evaluation of routing and volume-based storage policies in an order picking operation

Decision Sciences, Spring 1999 by Petersen, Charles G II, Schmenner, Roger W

What the Warehouse Looks Like

The warehouse that is evaluated in this paper is shown in Figure 1, a traditional rectangular warehouse. Picking is strict-order/batch picking, where a picker departs the P/D point to retrieve items on the pick list and transports the items back to the PID point for order consolidation, packaging, and shipment. This is a manual picking environment, involving the use of picking carts to transport the picked items. The layout is consistent with the warehouse layout literature (Bassan, Roll, & Rosenblatt, 1980; Ben-Mahmud, 1987; Sims, 1991) and with observations of several order picking operations. The picking area consists of 10 aisles, with storage for 1,000 items. The distance a picker travels to complete a picking tour consists of horizontal travel along the front and back aisles and vertical travel in the 10 picking aisles. The picking aisles are wide enough to allow two-way travel, but picking can be done from both sides of the aisle. The horizontal distance within picking aisles is not considered. The location of the P/D point is usually in the front corner of the warehouse or in the middle of the front aisle. This is consistent with observations of actual order picking operations and with Gibson and Sharp (1992).

How SKUs Are Assigned to Locations and Pick Lists Generated

The items or stock-keeping units (SKUs) are assigned storage locations based on their expected demand. Therefore, high volume SKUs are placed closest to the pick-up/drop-off (P/D) point and low volume SKUs are located farther away from the P/D point. The SKUs can be demanded according to any one of three different demand patterns, referred to as high, medium, and low demand skewness. In highly skewed demand, rather few SKUs account for the vast majority of the sales volume; this is the familiar Pareto principle, or 80-20 rule.

To generate a pick list, a random number between 0 and 1 is generated from a prime modulus multiplicative linear congruential generator. This random number, given the particular demand skewness pattern, in turn determines the SKU. In this way, for a high demand skewness pattern, certain SKUs are much more likely to be selected. For low demand skewness patterns, the SKUs are more evenly distributed. Once the SKU has been generated, its storage location, determined from the particular volume storage pattern in question and dependent on the rank order volumes of SKUs demanded (i.e., the highest volume items are located closer to the P/D point), is added to the pick list. This process continues until the pick list has reached its given size (number of items). The routing heuristics and the optimal algorithm take the pick list containing the SKUs and their locations and form a picking route which, in turn, determines the route length.

ROUTING POLICIES

Six different routing strategies are evaluated in this paper: transversal, return, midpoint, largest gap, composite, and optimal. These routing strategies range from the very simple to the slightly more complex. The first four are commonly found in operating warehouses and have also been subject to other research. The performance of these routing strategies depends on the particular operating conditions of the system under study. The four commonly used routing heuristics are presented first, then a new composite routing heuristic is introduced, and then the optimal strategy is discussed. Figure 2 shows the five routing heuristics, with "p" designating a pick.