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

Storage Policies

Table 4 presents the storage policies ordered by their performance in combination with optimal routing for each pick list size. Within-aisle storage with a middle P/ D is the best storage policy for all pick lists. It is followed by the within-aisle corner PID, the perimeter and diagonal storage strategies, and the across-aisle strategies. This result is consistent with Jarvis and McDowell (1991) that within-aisle storage is the optimal storage strategy for not only one-way travel, but regardless of the aisle travel restrictions. The effect of the PID location is evident as well. Except for diagonal storage, where there is no significant difference between corner and middle P/D, the middle PID is significantly better than the corner P/D. However, this difference becomes almost nonexistent for large pick lists. Most large pick lists require a picker to traverse more of the warehouse, rendering the advantage of the middle P/D almost moot. Table 4 also shows that most of the storage policies improve their performance relative to the within-aisle strategy as the pick list grows.

Table 5 shows the storage policies ordered by their performance in combination with optimal routing for each level of demand skewness. As demand becomes more skewed, the difference between within-aisle storage with the middle P/D and the other storage policies increases. This is fairly intuitive because the distance between storage locations within an aisle is less than the distance between storage locations in different aisles. At high levels of demand skewness, the distance to retrieve items is much shorter in within-aisle and diagonal storage, where only a few aisles have to be traversed as opposed to perimeter and across-aisle storage, where the entire warehouse may have to be traversed. In addition, the middle P/D generally results in a shorter route than the corner PID. However, with low demand skewness for perimeter storage and with low and medium demand skewness for diagonal storage, the corner P/D is slightly better than the middle P/D. Perimeter storage with both corner and middle PID performs well at low demand skewness. This suggests another tentative conclusion:

2. Within-aisle storage with a middle PID is the best storage policy for almost all levels of pick list size and demand skewness.

Routing and Storage Interaction

Figure 5 shows the interaction between routing and storage policies. There are several observations that merit discussion. First, it is obvious from Figure 5 that the performance of the return and transversal strategies is dependent on the choice of a storage policy. Return works better with diagonal and across-aisle storage, whereas transversal works better with within-aisle storage. Second, the performance of the largest gap, midpoint, and optimal strategies seem to be unaffected by the choice of a storage policy. Third, the composite strategy appears to be the best routing heuristic for diagonal, within-aisle, and across-aisle storage. It was not the best general routing heuristic, because composite does not perform well with perimeter storage. However, when composite is combined with diagonal, withinaisle, and across-aisle it performs better than largest gap. In addition, Figure 5 shows the negative impact of perimeter storage on return and transversal routings. Because perimeter storage places the high volume items along the perimeter of the warehouse, both largest gap and midpoint dominate as their routes tend to follow the periphery of the warehouse.