THE EFFECT OF WAREHOUSE CROSS AISLES ON ORDER …



A Preliminary Comparison of Warehouse Slotting Measures

Charles G. Petersen, Department of Operations Management and Information Systems

Northern Illinois University, DeKalb, IL 60115, (815) 753-1454, cpetersen@niu.edu

ABSTRACT

Slotting is the assignment of items or stock-keeping units (SKUs) to warehouse storage locations. This paper evaluates several slotting methods to determine which method minimizes travel distance in a warehouse. The results show that popularity and cube-per-order index result in less picker travel than other slotting methods.

Introduction

Today’s warehouses have to execute more, smaller transactions, handle and store more products, offer more product and service customization, and provide more value-added services, while having less time to process orders and with less margin for error. While many firms try to solve these challenges with more technology, a better solution may result from a careful analysis of customer orders and products in the warehouse. Frazelle [2] notes that most warehouses are spending 10 to 30 percent more per year than they should because the warehouse is improperly slotted. Slotting is the assignment of items or stock-keeping units (SKUs) to warehouse storage locations. In a typical warehouse it is estimated that less than 15 percent of the SKUs are properly slotted.

The key to proper slotting is a systematic analysis of SKU and customer order activity, commonly called warehouse activity profiling. This profiling process is designed to quickly identify root causes of material and information flow problems and to pinpoint major opportunities for process improvements. However, the practitioner literature [1,10,11] states the importance of profiling but gives no tangible information as to how it should be done. The academic literature on profiling is minimal, except for Yoon and Sharp [13,14] that only briefly discuss it.

SKU activity profiling is used to slot the warehouse. This means where to store each stock-keeping unit (SKU), how much of each SKU to store, and what storage mode to use for each SKU. This author and several others have researched the first question extensively [4,5,6,7,9,12]. However, these previous works have focused on evaluating storage assignment strategies and not on evaluating the slotting measures that can be used in determining storage assignment. The question of how much of a SKU to store includes not only the total amount to store but also whether a SKU should be stored in one location or in several different locations. Determination of the storage mode typically depends on the characteristics of the SKU. Yoon and Sharp [13,14] discuss the need to answer these questions as part of the overall design process of an order picking system. Although these two questions are also important, they are beyond the scope of this paper and will be studied in future research.

However, what slotting measure or method to use? There are several slotting measures available to a warehouse manager to use, including popularity, turnover, cube-per-order index, volume, and pick density. The popularity principle refers to the number of requests for a SKU during a given period of time. Turnover refers to the demand or number of units shipped per period. Cube-per-order index developed by Heskett [3] takes the physical size of the SKU into account as well as the daily demand for the SKU and the average order size of the SKU. Volume refers to the cubic volume of a SKU shipped per period. Pick density is the number of requests per cubic volume of a SKU. This method is sometimes used in “golden zone” picking where the SKUs with the highest pick density are assigned to the most accessible pick locations taking ease of reach and fatigue into account. This paper focuses on evaluating these five slotting measures under a variety of operating conditions.

Warehouse Simulation and Experimental Design

The warehouse for this Monte Carlo simulation is a manual bin-shelving pick area with 10 picking aisles and a front and back cross-aisle to allow access to all picking aisles (Figure 1). These aisles allow for picking from both sides of the aisles and are wide enough to permit two-way travel. The warehouse contains enough storage space to handle 1,000 SKUs. The demand for these SKUs follows the commonly observed 80-20 curve. For each order, the picker travels from the pick-up/drop-off (p/d) point to retrieve all the SKUs on the pick list and then returns to the p/d point to drop-off the SKUs before picking up a new pick list.

The factors and levels for this experiment are presented in Table 1 and results in a 5x3x2x3 design with 90 cells. The five slotting measures are popularity, turnover, volume, pick density, and cube-per-order index. The three storage assignment strategies are within-aisle, diagonal, and across-aisle. These three storage assignment strategies are shown in Figure 2 with the darkest shading indicating “A” SKUs, the medium shading for “B” SKUs, and no shading for the “C” SKUs. Petersen and Schmenner [7] evaluated these storage strategies using volume-based slotting only and found that within-aisle and diagonal reduced piker travel more than the other storage strategies. However, no one has tested these storage strategies using other slotting measures.

Figure 1 Warehouse layout

Table 1 Experimental Factors and Levels

|Factor |Levels |Notation or Values |

|Slotting measure |5 |Popularity, Turnover, Volume, |

| | |Pick Density, Cube-per-order |

| | |index |

|Storage assignment strategy|3 |Within-aisle, Diagonal, |

| | |Across-aisle |

|Routing policy |2 |Traversal, Optimal |

|Order size |3 |3, 10, 20 SKUs |

In addition to the optimal routing procedure of Ratliff and Rosenthal [8], the author chose to use traversal routing because it is commonly used in warehousing and order picking. Traversal routing requires that an order picker exit a picking aisle from the opposite end from which he or she entered. This also sometimes called serpentine or s-shaped routing.

The author chose three levels of order size corresponding to small, medium, and large orders. The literature as shown that order size (or pick list size if orders are batched) has a major effect on the performance of routing and storage policies. For each level of order size, 500 orders were randomly generated. The performance measure for this experiment is the average route length of the order picker to complete the 500 orders.

Results and Discussion

The results of the experiment are shown in Tables 2 and 3. There are several observations of note. First, the best overall slotting measures appear to be popularity and cube-per-order index, although turnover is a close third. Pick density is clearly the worst, but is performance is expected to improve when evaluated on total time to pick all orders and not just on travel distance. While the cube-per-order index is the best slotting measure when using within-aisle storage, the popularity measure is generally the best with either diagonal or across-aisle storage.

Figure 2 Storage Implementation Strategies

Within-aisle

Across-aisle

Diagonal

It is clear that within-aisle storage is clearly the best storage implementation strategy for all every factor and level except when using pick density as a slotting measure. The relative performance of the slotting measures and storage strategies does not seem to change whether optimal or traversal routing is used.

Table 2 Average Route Length with Optimal Routing (in Feet)

| |Within |Diagonal |Across |Average |

|3 SKUs | | | | |

|Popularity |99.9 |111.9 |131.9 |114.6 |

|Turnover |102.7 |118.2 |139.0 |120.0 |

|Volume |112.9 |127.8 |142.0 |127.5 |

|Density |184.7 |187.0 |184.5 |185.4 |

|Cube |100.2 |113.3 |132.2 |115.3 |

|Average |120.1 |131.7 |145.9 |132.6 |

|10 SKUs | | | | |

|Popularity |189.3 |234.7 |261.3 |228.4 |

|Turnover |195.2 |247.9 |267.6 |236.9 |

|Volume |211.9 |263.6 |282.2 |252.6 |

|Density |343.1 |354.5 |348.0 |348.5 |

|Cube |188.7 |235.9 |261.4 |228.7 |

|Average |225.6 |267.3 |284.1 |259.0 |

|20 SKUs | | | | |

|Popularity |257.0 |319.6 |341.3 |306.0 |

|Turnover |261.0 |328.0 |347.6 |312.2 |

|Volume |281.9 |352.8 |368.0 |334.2 |

|Density |458.3 |473.3 |462.5 |464.7 |

|Cube |255.0 |320.1 |340.5 |305.2 |

|Average |302.6 |358.8 |372.0 |344.5 |

|Overall | | | | |

|Popularity |182.1 |222.1 |244.8 |216.3 |

|Turnover |186.3 |231.3 |251.4 |223.0 |

|Volume |202.2 |248.1 |264.1 |238.1 |

|Density |328.7 |338.3 |331.7 |332.9 |

|Cube |181.3 |223.1 |244.7 |216.4 |

|Average |216.1 |252.6 |267.3 |245.3 |

Conclusion

The author has conducted some preliminary work on evaluating slotting measures and has found that when only considering travel distance that popularity, turnover, and cube-per-order index result in significantly less picker travel than volume and pick density slotting measures. However, this experiment needs be expanded to evaluate the total time to complete a picking tour by taking into account the picking time difference in storage location. The time to retrieve a SKU from a storage location is dependent on the height of the storage location in addition to size and weight of the SKU. Storage locations above the picker’s shoulder or below the picker’s waist require more time to retrieve. The area between the waist and shoulders is called the “golden zone” and typically SKUs with higher demand are stored there while other SKUs that are commonly ordered with the golden zone SKUs are located above or below the golden zone. Frazelle [2] suggests that golden zone picking used in conjunction with pick density slotting can be used with SKUs with a high correlation to reduce picker travel. However, no results are presented and this remains an area that needs further study.

Table 3 Average Route Length with Traversal Routing (in Feet)

| |Within |Diagonal |Across |Average |

|3 SKUs | | | | |

|Popularity |147.3 |222.8 |261.5 |210.5 |

|Turnover |150.8 |231.8 |267.2 |216.6 |

|Volume |169.0 |236.6 |263.1 |222.9 |

|Density |270.3 |269.8 |267.7 |269.3 |

|Cube |146.2 |225.4 |262.0 |211.2 |

|Average |176.7 |237.3 |264.3 |226.1 |

|10 SKUs | | | | |

|Popularity |240.3 |386.2 |464.0 |363.5 |

|Turnover |245.3 |409.0 |467.5 |373.9 |

|Volume |267.2 |415.0 |472.1 |384.8 |

|Density |473.4 |469.3 |466.9 |469.9 |

|Cube |239.4 |389.4 |464.4 |364.4 |

|Average |293.1 |413.8 |467.0 |391.3 |

|20 SKUs | | | | |

|Popularity |309.3 |493.3 |569.0 |457.2 |

|Turnover |313.3 |506.8 |575.1 |465.1 |

|Volume |337.4 |524.5 |575.3 |479.1 |

|Density |576.2 |572.9 |575.2 |574.8 |

|Cube |307.4 |491.6 |568.7 |455.9 |

|Average |368.7 |517.8 |572.6 |486.4 |

|Overall | | | | |

|Popularity |232.3 |367.4 |431.5 |343.8 |

|Turnover |236.5 |382.5 |436.6 |351.9 |

|Volume |257.9 |392.0 |436.8 |362.2 |

|Density |440.0 |437.4 |436.6 |438.0 |

|Cube |231.0 |368.8 |431.7 |343.8 |

|Average |279.5 |389.6 |434.6 |367.9 |

References available upon request from the author

-----------------------

[pic]

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download